Researchers from the University of California San Diego (UCSD) collaborated with Google to recycle “old” Pixel smartphones and give them a second life as a low-cost data center. According to Google Research, retired smartphones are part of the “embodied carbon” that is associated with manufacturing and its carbon footprint. In fact, humanity’s penchant for mobile devices and replacing them every few years is one of the biggest contributors to e-waste, so the group from UCSD planned to give these discarded devices a second life as a “general-purpose computing platform.”

The study revealed that smartphones from just three years ago still deliver a higher single-core performance compared to servers like the Asus RS720A-E11, which can be equipped with Nvidia H200 or Nvidia RTX Pro 6000 GPUs and two AMD EPYC server processors, that you frequently find in the most powerful data centers. While the latter delivers performance that a mobile device can’t even dream of, the fact that the former still scored higher in the SPEC benchmarking suite on a per-core basis meant that researchers could still use them for compute tasks with a little creativity.

The first thing they did was to strip these gadgets of non-essential components — displays, batteries, cameras, speakers, chassis, etc. Only the motherboard remains, as it plays host to the SoC needed for running compute. The Android operating system is then replaced with a general-purpose Linux distro used in data center applications, which removes unnecessary bloat found in the original consumer device and allows for the deployment of orchestration software like Kubernetes. Benchmarking results revealed that 25 to 50 old phones wereequal to the computing power of a single dual-socket server-class CPU.

UCSD determined that a 20-phone cluster can support one application that a 75+ student class requires. So, instead of hosting it on the cloud, which would entail additional costs and resource use on the data center side, it could instead run these apps on a local deployment of these used smartphones. The research team plans to use 2,000 phones to build a local data center that can support “a hundred such classes at once.” Aside from getting the advantage of running apps locally and owning the hardware needed for them, the group also says that it’s only a “fraction of the usual cost,” likely referring to building a local server made from new components. This is especially true today, with the increased pricing for memory and storage chips.

The research team says that it expects to launch the full system later this year and is looking to see how consumer parts can withstand continuous use in a data center application. But even if the experiment is successful, we don’t foresee AI hyperscalers switching to servers made from used phone parts as they would often want to work with fewer parts and the reliability delivered by specialized hardware. Still, this is a great option for universities and educational institutions, as well as smaller entities that do not have the resources to secure brand-new parts and compete against tech giants with billions of dollars to burn.

This isn’t the first time scientists have looked at giving old phones a second life — another group of researchers looked at converting old phones into “tiny data centers” last year, even using one set of four old devices for underwater monitoring. After all, even though the SoCs found in these devices are considered “outdated” by modern standards, they should still be more than capable enough for many mundane tasks. NASA even repurposed the Qualcomm 801 SoC, a mid-range chip from 2014 and found in the Ingenuity Mars helicopter, to help the Perseverance rover find its way around the Red Planet like some sort of processor for a makeshift GPS. And for smartphones that no longer work, people are finding ways to extract the gold and other resources found on their boards for recycling.

Nokia today announced its collaboration with Intel on 5G infrastructure silicon, including working together on Intel's recently announced Atom P5900 processor (formerly Snow Ridge). Nokia will also adopt Cascade Lake in its cloud data center solution for a common architecture "from cloud to edge."

Intel announced its portfolio of silicon for 5G New Radio network infrastructure early last week (it originally planned to announce this at Mobile World Congress). It introduced the Atom P5900 series as its first x86 SoCs for base stations. Nokia, ZTE and Ericsson were said to be adopters, and Intel claimed it was on track for 40% market share by 2021 already.

• Best gaming CPUs of 2020
• Intel drops 10nm Tremont Atom P5900

Intel's SoC is included in Nokia's AirScale radio access products, which are now shipping. AirScale is part of Nokia's Powered by ReefShark line-up. 

“By adopting ReefShark widely in its AirScale portfolio, Nokia is significantly boosting performance and lowering the energy footprint of 5G network rollouts," Nokia said in a statement. 

Nokia said the new ReefShark chipsets improve performance and have a reduced energy footprint. They have previously been rumored to leverage Intel Custom Foundry’s 10nm process technology, but Nokia did not specify the process technology. It is unclear how committed Intel still is to its foundry business after its 10nm delays and whether this affected adoption of its foundry business.

Additionally, Nokia said it will continue to adopt the latest, second-generation Xeon Scalable processors in its AirFrame cloud data center solution. Nokia did not clarify if this refers to the Cascade Lake refresh that Intel also announced last week, but said that this realized the benefits of a common (Intel) architecture from cloud to edge. Nokia also uses Xeons for its AirScale vRAN and 5G core solutions.

Dan Rodriguez, Corporate Vice President and General Manager of Intel’s Network Platforms Group, said: “Implementing technology innovations from the core to the edge of the network is key to unlocking the full potential of 5G. Through our collaboration with Nokia, our broad portfolio of products and ASIC capabilities, we are showing the value that can be realized with a consistent, high-performance architecture across the intelligent 5G network.”

MediaTek announced the new Dimensity 800 Series SoC for 5G smartphones at CES 2020 yesterday. The 7nm chip features an octa-core CPU, a 5G modem and an AI Processing Unit (APU) that offers 2.4 TOPS of artificial intelligence compute power.

Unlike the new X50 and X55 chipsets from Qualcomm, the Dimensity 800 is a single-chip 5G SoC. MediaTek said this allows the SoC to save on power--especially in conjunction with the reduced power draw associated with TSMC's 7nm process.

The SoC supports 5G with two carrier aggregation (2CC CA). MediaTek claims this delivers 30% wider high speed layer coverage compared to 5G without carrier aggregation. It supports both standalone and non-standalone sub-6GHz 5G networks and also has multi-mode support for 2G up to 5G connectivity and Dynamic Spectrum Sharing (DSS). Voice over New Radio (VoNR) is also supported.

On the CPU side, the octa-core design packs four big A76 cores and four low-power A55 cores, both of which are clocked at 2GHz. This is a step down from the four A77 cores in the flagship Dimensity 1000 that MediaTek announced in 2019. The GPU, meanwhile, is a Mali-G57MP4. On the AI side, the Dimensity 800 features MediaTek’s APU 3.0 with FP16 support. It offers performance of up to 2.4 TOPS--which is a notable decrease from the 4.5 TOPS performance offered by the Dimensity 1000.

On the media side, the Dimensity 800's image signal processor (ISP) supports up to 64MP image sensors as well as multi-camera setups. There is also support for 2160p at 60 frames per second video encode and decode and 90Hz Full HD+ displays.

The Dimensity 800 is expected to come to market before the end of the first half of the year.

In November, MediaTek and Intel announced a partnership to bring 5G to laptops in early 2021.

Huawei unveiled a new chip, called Kunpeng 920, aimed at the data center market. The chip is expected to help reduce China’s reliance on exports of chips from American companies amidst the ongoing trade dispute between China and the United States.

Huawei Kunpeng 920

The new chip is based on Arm instruction set architecture, which has historically not been successful in the server market. However, over the past few years, multiple companies have tried to enter this market with higher-performance Arm-based chips.

Some companies, such as Samsung and Qualcomm have given up on their server chip efforts, while other companies such as Huawei, Marvell, and some new server chip start-ups have redoubled their efforts to take on both Intel and AMD in the data center.

William Xu, Huawei’s chief strategy marketing officer, said on Monday:

“The ARM industry is seeing a new development opportunity. The [new] Kunpeng 920 CPU and TaiShan servers released by Huawei are primarily used in big data, distributed storage, and ARM native applications.

We will work with global partners in the spirit of openness, collaboration, and shared success to drive the development of the ARM ecosystem and expand the computing space, and embrace a diversified computing era.”

High-Performance Arm CPU

Earlier this year, when Huawei announced the chip for the first time, the company claimed that Kunpeng 920 is the “industry’s highest-performance Arm-based server CPU.” The processor was designed in-house by Huawei and manufactured on a 7nm process.

Huawei claimed that its design significantly improves processor performance by optimizing branch prediction algorithms, increasing the number of OP units, and improving the memory subsystem architecture. The company said that the chip outperforms the competition by 25% in performance and 30% in energy efficiency.

The Huawei Kunpeng 920 comes with 64 CPU cores with a clock speed of 2.6 GHz, supports 8-channel DDR4-3200, and the company claimed it beats the competition by 46% in terms of memory bandwidth. The chip supports PCIe 4.0 and CCIX interfaces and provides 640 Gbps total bandwidth.

Huawei Taishan X6000 V2 Server

Huawei also launched the high-density Taishan X6000 V2 server, a U2 4-node server that uses the Kunpeng 920 processors and comes in multiple versions from a 32-core 2.6 GHz version to 64-core 3GHz version.

Huawei Cloud, the company’s cloud service, has also started integrating the Taishan X6000 V2 server in its offerings, which include elastic cloud services, bare metal services, and cloud phone services.

Update: 8:20 pm PT: Qualcomm provided a couple of clarifying statements indicating that its actions are related to iPhones that use baseband processors other than the ones provided by Qualcomm affiliates, and that its limited exclusion order pertains to future devices coming into the U.S. We have added those notes in the appropriate sections below.

The legal battle between Qualcomm and Apple is heating up. Qualcomm is seeking to block the importation and sale of Apple iPhones and other products, alleging that they infringe on six Qualcomm technology patent, mostly related to extending battery life.

Qualcomm and Apple have been embroiled in a contentious legal battle over technology royalty fees for most of this year. Apple's actions have sought to make a direct impact on Qualcomm's revenues, and the chip maker's latest move would be an audacious strike back if it were to succeed. How in the world did we get here?

In January, Apple filed lawsuits in the United States and in China alleging that Qualcomm's royalty fees are too high and the license agreements too restrictive. The lawsuits seek $1 Billion and 1 Billion Yuan in damages, respectively. During an earnings call the following week, Apple CEO Tim Cook explained that he doesn’t like resorting to litigation, but he “didn’t see another way forward."

As you can imagine, Qualcomm doesn’t agree with Apple’s stance. Don Rosenberg, Qualcomm’s executive vice president and general counsel provided a statement following the second Apple legal filing.

“These filings by Apple's Chinese subsidiary are just part of Apple's efforts to find ways to pay less for Qualcomm's technology,” said Rosenberg. “Apple was offered terms consistent with terms accepted by more than one hundred other Chinese companies and refused to even consider them. These terms were consistent with our NDRC Rectification plan. [...] Qualcomm is prepared to defend its business model anywhere in the world. We are proud of our history of contributing our inventions to the development and success of the mobile communications ecosystem."

In April, Qualcomm filed a counterclaim against Apple, seeking the court's intervention in compelling Apple to adhere to the agreements between the two companies.

Later that month, Qualcomm said that Apple had withheld royalty payments from manufacturers for work done in the first quarter of 2017, which prevented the manufacturers from paying their licensing fees to Qualcomm. In May, Qualcomm revealed that four iOS device manufacturers—Foxconn, Pegatron Corporation, Wistron Corporation, and Compal Electronics—had refused to pay technology licensing fees at the behest of their client (Apple). Apple, Qualcomm said, was refusing to pay for the company's technology while the legal battle unfolds. Those companies rely heavily on Apple’s business, naturally.

Today, Qualcomm pushed further. The company filed a complaint with the United States International Trade Commission (ITC) alleging that Apple “has engaged in the unlawful importation and sale of iPhones.” Qualcomm’s complaint suggests that the Apple iPhone infringes on up to six of Qualcomm’s patents, and the company is seeking a Limited Exclusion Order (LEO), which would bar Apple from importing iPhones and other unnamed products into the Unites States. Qualcomm clarified that the company is seeking the LEO against iPhones that use cellular baseband processors other than those supplied by Qualcomm’s affiliates.

“Qualcomm’s inventions are at the heart of every iPhone and extend well beyond modem technologies or cellular standards,” said Don Rosenberg, executive vice president and general counsel for Qualcomm. “The patents we are asserting represent six important technologies, out of a portfolio of thousands, and each is vital to iPhone functions. Apple continues to use Qualcomm’s technology while refusing to pay for it. These lawsuits seek to stop Apple’s infringement of six of our patented technologies.”

Furthermore, Qualcomm is seeking a Cease and Desist Order, which would prevent Apple from selling the hardware that is already stateside and from “marketing, advertising, demonstration, and warehousing of inventory for distribution and use of those imported products . . . " A Qualcomm spokesperson said that the limited exclusion order is for future devices, so anyone who currently owns an iPhone will not be impacted.

Which Patents Is Apple Allegedly Infringing On?

Qualcomm maintains that the latest iteration of the iPhone infringes on Qualcomm’s battery life conservation technology (US Patent No. 8,698,558; US Patent No. 8,633,936; US Patent No. 9,608,675; US Patent No. 8,487,658, US Patent No. 8,838,949). The company provided an infographic (pictured partially above, with the patents described briefly below). The company also believes that the iPhone infringes on US Patent No. 9,535,490, which “enables the applications on your smartphone to get their data to and from the internet quickly and efficiently by acting as a smart ‘traffic cop’ between the apps processor and the modem.”

The battle between Apple and Qualcomm could go on for some time. Qualcomm said it expects an ITC investigation by August, with a potential trial beginning next year.

We've also reached out to Apple for comment and will add any statements from the company when we hear back.

The Federal Trade Commission (FTC) accused Qualcomm of using the popularity of its baseband processors, which are used to manage connections to wireless networks, to weaken competitors and bully manufacturers into "onerous and anticompetitive supply and licensing terms." Qualcomm said the FTC's complaint is based on "a flawed legal theory, a lack of economic support, and significant misconceptions about the mobile technology industry."

There's no question that Qualcomm dominates the baseband processor market. ABI Research said in February 2016 that Qualcomm held roughly 65% of the LTE baseband market. Samsung, its closest competitor, represented just 12% of the market after it decided to create its own basebands for its smartphones. The runners-up, Huawei and MediaTek, managed to nab just 9% of the market each. None are even close to competing with Qualcomm.

Nor could one debate the costs of doing business in the smartphone market. One report, The Smartphone Royalty Stack: Surveying Royalty Demands for the Components Within Modern Smartphones, said the hypothetical manufacturer of a $400 smartphone might expect to spend $60 on royalty demands for LTE compatibility even though the baseband processor itself costs roughly $10. Qualcomm had the highest public royalty rate in the report.

So it's clear that Qualcomm is simultaneously ubiquitous and well-paid. The FTC alleged that at least part of the company's status results from it using essential patents to negotiate unfair terms from smartphone makers. The commission said that Qualcomm asks for elevated royalties and license terms for those patents, which equates to a tax on manufacturers that use competitive baseband processors, and effectively holds critical tech for ransom.

Here's the FTC on why this is a problem:

Increased costs imposed by this tax are passed on to consumers, the complaint alleges. [...] By excluding competitors, Qualcomm impedes innovation that would offer significant consumer benefits, including those that foster the increased interconnectivity of consumer products, vehicles, buildings, and other items commonly referred to as the Internet of Things.

The FTC accused Qualcomm of violating the FTC Act. It said that it's "seeking a court order to undo and prevent Qualcomm’s unfair methods of competition" and "has asked the court to order Qualcomm to cease its anticompetitive conduct and take actions to restore competitive conditions." Qualcomm, for its part, said that the commission doesn't understand the smartphone component market well enough to make this complaint.

This is what Qualcomm general counsel Dan Rosenberg said in a statement:

In our recent discussions with the FTC, it became apparent that it still lacked basic information about the industry and was instead relying on inaccurate information and presumptions. In fact, Qualcomm was still receiving requests for information from the agency that would be necessary to an informed view of the facts when it became apparent that the FTC was driving to file a complaint before the transition to the new Administration. We have grave concerns about the two Commissioners' decision to bring this case despite a lack of evidence supporting the allegations and theories in the complaint. We look forward to defending our business in federal court, where we are confident we will prevail on the merits.

The outcome of this complaint might have a meaningful effect on the smartphone market. Qualcomm's dominance could be threatened, and if competitors offer more favorable terms as they scramble to get a bigger share of the component market, phones might even get cheaper. Perhaps more interesting devices will be introduced: the authors of The Smartphone Royalty Stack said that the cost of patent royalties and other non-hardware expenses "may be undermining industry profitability — and, in turn, diminishing incentives to invest and compete." That's not good for consumers.

Or maybe the FTC will get its first loss right as it starts to transition from the Obama administration to the incoming administration. This case is likely to have broad implications for the smartphone market either way, given the importance of both the FTC and Qualcomm to the devices in our pockets.

Imagination Technology announced a new GPU family at MWC 2016 that’s focused on low-cost smartphones, wearables, televisions, IoT devices, and other embedded applications where low-power usage is critical. The new PowerVR Series8XE is an iterative update to Series7XE that includes architectural enhancements to improve fillrate. It also adds support for the latest graphics APIs, including OpenGL ES 3.2 and Vulkan 1.0. Compared to the Series7XE, Imagination Technology said the Series8XE delivers either the same performance on a 25% smaller die, with a corresponding decrease in price, or higher performance on a chip of the same size.

Unlike high-end GPUs that focus on pixel shader performance for rich gaming, the Series8XE family focuses on pixel fillrate (texturing), which is more important for user interface rendering, smooth Web browsing, and light gaming. The new GPU family offers twice the fillrate at the same area as Series7XE, a significant but necessary gain as screen resolution increases in low-end devices.

High-end applications such as 3D gaming make extensive use of layering effects and multiple render targets. Because these operations may require multiple shader passes, 32-bit floating-point (FP32) values are necessary to retain precision. Low-end applications like those targeted by Series8XE, however, use more direct rendering methods that do not require high precision. Therefore, it makes sense to take advantage of the Rogue architecture’s dedicated FP16 ALUs to save power.

At a high level, a Series8XE USC looks just like a Series7XE or high-end Series7XT USC, with the same number of FP32 and FP16 ALU cores. However, there are architectural improvements within the ALUs to improve energy efficiency and throughput. Another feature that carries over from the Series7XT is the ability to co-issue SFU and ALU operations. Series8XE also includes a tessellation co-processor (no longer an optional component), making it surprisingly well-featured for a GPU targeting the low-end market.

Series8XE GPUs have a single USC with up to 16 pipelines. The texture unit scales with the number of pipelines in order to achieve the desired balance between shading performance and fillrate, without exceeding the area budget.

The PowerVR GE8200 and GE8300 are the first two GPUs announced in the Series8XE family and are capable of processing two and four pixels per clock, respectively. For comparison, the PowerVR GX6450 in the iPhone 6 processes eight pixels per clock, while the PowerVR GT7600 in the iPhone 6s processes 12 pixels per clock.

Series8XE also includes Imagination’s OmniShield GPU hardware virtualization, an especially important feature in the automotive space, where the same chip might be driving both the dashboard display and the infotainment system. Not only might these systems be running different operating systems (RTOS for the dashboard and Android for infotainment, for example), but the infotainment system is likely networked and needs to be isolated from the dashboard display for the sake of security.

While Series8XE is an iterative design with improved fillrate, power efficiency, and updated graphics API support, the high-performance Series8XT, which should be announced later this year, will bring more significant architectural changes.

Matt Humrick is the Mobile Editor at Tom's Hardware. Contact him at mhumrick@tomshardware.com and follow him on Twitter @digitalOut_net. Follow us @tomshardware, on Facebook and on Google+.

Lenovo announced the Vibe K5 Plus, which not too long ago would’ve been seen as an upper mid-range smartphone, for only $149. Unfortunately, it won't be available in the U.S.

The device comes with a Snapdragon 616 chip, which includes eight Cortex-A53 cores with a frequency of 1.7 GHz each, an Adreno 405 GPU, which supports OpenGL ES 3.1, and a Cat. 4 LTE modem (150/50 Mbps data speeds).

The device has a 5” Full HD display, 16 GB of internal storage with support for another 32 GB expandable storage via microSD, 2 GB of LPDDR3 RAM, as well as dual-SIM, 802.11n Wi-Fi, and Bluetooth 4.1 support. It also has a 2,750 mAh replaceable battery, which is reasonably large for a device this size and with this screen resolution.

The Lenovo Vibe K5 Plus will feature a 13MP rear camera that can shoot 1080p video, and a 5MP front-camera, which uses only fixed-focus (pictures will be blurry if you don’t keep the optimal distance).

The smartphone will run the older Android 5.1 operating system out of the box. Lenovo didn’t say whether it will receive further updates, which could be an issue for Android buyers that want either the latest version of Android or want their phones to be kept up to date and secure.

The 8.2 mm device will be available in Platinum Silver or Champagne Gold colors, and it should arrive on the market in March, this year.

Lenovo also announced its slightly lower budget brother, the Lenovo Vibe K5 (or Lenovo K5), which should be identical except that it has an HD screen and a Snapdragon 415 (also octa-core, but lower frequency). The K5 will cost $129 and should also arrive in March. Neither of the two phones will be available in the U.S., according to Lenovo.

Lucian Armasu is a Contributing Writer for Tom's Hardware. You can follow him at @lucian_armasu. 

Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Archos will unveil a higher-end smartphone than we're used to seeing from the company. Scheduled to be revealed in full at MWC later this month, the new Archos 50d Oxygen will offer higher-end spec but the same kind of affordable price tag we've come to expect from Archos.

The Archos 50d Oxygen doesn’t change too much from the previous Archos smartphones in terms of looks, nor the branding, which could become confusing to some customers. However, it strives to remain a budget device with higher hardware specifications and quality.

The phone comes with a 1.3 GHz Mediatek processor with eight Cortex-A53 cores and a Mali-T720 GPU, an integrated LTE modem, a 5” IPS display with 1920x1080 resolution, 2 GB of RAM, support for microSD, dual-SIM (dual standby), Bluetooth 4.0 Smart, 802.11n Wi-Fi and Wi-Fi Direct.  

The device also brings a 2,100 mAh battery, which is slightly below average for what you’d typically see in 5” 1080p smartphones. The Archos 50d Oxygen also has a 13MP rear camera with auto-focus on the back and a 5MP camera in the front. The phone will run Android 5.1 when it launches, even though the latest version of Android, released last November, was Android 6.0.

The Archos 50d Oxygen will only cost $150 when it launches in May, and will be showcased at the MWC this month.

Lucian Armasu is a Contributing Writer for Tom's Hardware. You can follow him at @lucian_armasu. 

Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Introduction

Huawei is not a household name in North America, but if the Chinese company's performance in other markets is any indication, this will soon change. Best known for selling telecommunications equipment, Huawei is quickly making a name for itself in smartphones, too. According to IDC, it is now the third largest smartphone vendor in the world by volume. Its success is centered in China, where its smartphone shipments increased 81 percent year-over-year in Q3 2015, but Huawei is expanding into emerging markets, such as Latin America and Africa, too. The company is also experiencing solid growth in the established European market, where smartphone shipments nearly doubled over the same period, according to Huawei.

Huawei recently pierced the North American market after being selected by Google as the manufacturing partner for the Nexus 6P. Looking to capitalize on this significant design win, Huawei is shipping another flagship phablet, the Mate 8, across the Pacific. We recently spent some time with this new device and were generally impressed (see Huawei Mate 8 Hands On). Its aluminum construction and minimal bezels give it a premium look and help hide the size of the large 6-inch screen. It comes with most of the features we've come to expect from a premium phone, including a fingerprint sensor, 3 to 4GB of RAM, 802.11ac Wi-Fi, NFC, and a decent camera. And while the battery is not removable, it's at least a generous 4000mAh.

In addition to being the first Huawei branded phone available globally, the Mate 8 has a couple of other firsts hidden inside its aluminum chassis, starting with HiSilicon's Kirin 950 system on a chip (SoC). Like MediaTek and Marvell, HiSilicon, a Huawei subsidiary, is an ARM processor licensee, which allows it to build SoCs using stock ARM cores. This approach differs from architectural licensees such as Apple and Qualcomm, who design custom CPU cores compatible with ARM's architecture.

Inside the Kirin 950 is an octa-core CPU with four ARM Cortex-A72 cores at up to 2.3GHz and four Cortex-A53 cores at up to 1.8GHz in a big.LITTLE arrangement. This makes HiSilicon only the second vendor behind MediaTek to ship a product using A72 cores, and the first to implement them on a current process node running close to its target frequency (the quad-core MT8173 is made using TSMC's older 28nm HPm node).

ARM's Mali-T880 GPU also makes its debut appearance in the Kirin 950, offering higher performance and up to a 40 percent improvement in energy efficiency over the Mali-T760 GPU. HiSilicon's implementation differs from other recent SoCs by opting for fewer, higher frequency cores. Specifically, it uses four GPU cores (MP4) that ramp up to a rather high 900MHz max frequency. In contrast, Samsung's Exynos 7420 uses the previous generation Mali-T760 GPU in an eight core configuration clocked at 772MHz. It will be interesting to see if the Mali-T880's architectural enhancements, combined with a higher clock frequency, can make up for the core count deficit.

Unlike previous Huawei devices that relied on external ISPs provided by outside vendors, the Kirin 950 benefits from Huawei's increasing vertical integration by incorporating its brand new, custom 14-bit PrimISP and IVP32 DSP into the SoC. We do not know much about these new blocks other than they enable up to 1080p60 H.264 encode/decode and 4K H.265 decode. Oddly missing is a 4K encode option.

Internally, the Kirin 950 uses ARM's CCI-400 Cache Coherent Interconnect. It's a bit odd that the newer CCI-500 interconnect is not used considering it was announced alongside the A72. This could potentially limit bandwidth between the CPUs and memory, since CCI-400 lacks a snoop filter on the interconnect itself.

Speaking of system memory, the Kirin 950 uses a hybrid LPDDR3/LPDDR4 memory controller. The Mate 8 takes advantage of LPDDR4's higher bandwidth and lower operating power, but other OEMs have the option of using less expensive LPDDR3 RAM. This feature, along with the smaller die from using only four GPU cores, should make the Kirin 950 a good fit for mid-range phones, an intriguing possibility considering the performance it should deliver.

With all of these new components, it's only fitting that the Kirin 950 is manufactured on a current generation process node, specifically TSMC's 16nm FinFET+. We've already seen FinFET's power and performance gains in our testing of Apple's A9 and Samsung's Exynos 7420 SoCs. More than a mere die shrink, FinFET is a new technology that fundamentally differs from the planar processes that came before it. It significantly reduces leakage current and power density, increasing battery life and reducing the amount of thermal throttling.

Performance Testing

This performance preview is a bit different than some of our others, because we're comparing not just the Kirin 950 SoC -- and its A72 CPU and Mali-T880 GPU cores -- to its peers, but also the Mate 8 to some other flagship phones.

We'll be comparing the Mate 8 and Kirin 950 to a range of different phones and SoCs, including the dual-core Apple A9, quad-core Intel Atom Z3580, and big.LITTLE octa-core designs using a combination of A57 and A53 cores. We're also including Qualcomm's latest Snapdragon 820 with its new, custom Kryo CPU. These results were collected from Qualcomm's smartphone MDP, which is its own internal development hardware, so they are still preliminary. Our comparison devices also contain a variety of Mali, PowerVR, and Adreno GPUs. With this range of hardware, spanning from mid-range to high-end, we should get a good feel for where the Mate 8, Kirin 950, and A72 fall.

All of these devices, other than the Qualcomm smartphone MDP, were evaluated using our standard testing procedures. Do note, however, that the Mate 8's software build is not final yet, so there could be some differences once it starts shipping.

MORE: Best Smartphones
MORE: Smartphones in the News
MORE: All Smartphone Content

CPU And System Performance

Now that we have a better understanding of the hardware, it's time to see how it performs. In this section, we'll evaluate system-level performance by running a series of synthetic and real-world workloads, along with some browser-based Web tests. If you're interested in learning more about how these benchmarks work, what versions we use, or our testing methodology, please read about how we test mobile device system performance.

The Huawei Mate 8 scores well overall in Basemark OS II, held back only by its GPU performance. In the OpenGL ES 2.0-based Graphics test, the Galaxy S6's Mali-T760MP8 easily outperforms the Mate 8's Mali-T880MP4 GPU by 54 percent, which is a larger margin than we would expect. This test does perform alpha blending and various texture operations, so it's possible having only half as many ROPs limits throughput compared to the Galaxy S6.

The Mate 8 performs much better in the CPU-centric System and Web tests. Only Apple's iPhone 6s Plus, buoyed by its Twister CPU's higher instructions per cycle (IPC), performs better. The octa-core Kirin 950 does finish ahead of Qualcomm's Snapdragon 820 by a small 7 percent margin, but keep in mind the 820's quad-core CPU is at a disadvantage in the multi-threaded portions of the test, and the 820's Kryo CPUs also run at a lower peak frequency. The fact that Snapdragon 820 finishes so close to the Kirin 950 despite these limitations seems to suggest that the Kryo CPU core has higher IPC than the A72, although we cannot pinpoint by how much from this test.

Compared to the Exynos 7420 in the Galaxy S6, the Mate 8's Kirin 950 scores about 11 percent higher in both the System and Web tests, a lower than expected result that basically mirrors the average difference in CPU clock frequency. Perhaps the A72's architectural improvements will have a bigger impact in our other tests.

While the Memory test purports to measure the speed of the internal NAND storage, it turns into more of a memory test on high-end devices due to how the OS uses a RAM cache to buffer storage access. Because this test does not work as intended, we cannot draw any definitive conclusions here.

The Mate 8 achieves the best overall score we've seen in AndEBench, narrowly defeating the Galaxy S6 and outpacing the Moto X Pure Edition by 30 percent. Its advantage comes primarily from the CoreMark-HPC CPU performance test, where its Kirin 950 outperforms the Exynos 7420 by 29 percent and Snapdragon 820 by 34 percent in a mixture of single- and multi-threaded integer and floating-point workloads. The newer revision of the Snapdragon 810 falls to the bottom of the chart, below the Snapdragon 808 that has two fewer A57 cores, because of thermal throttling. This exemplifies the A57's problems at 20nm: poor sustained performance. Unable to keep its four A57's, let alone both the A53 and A57 islands, running concurrently, the OnePlus 2 shuts down the A57's and relies only on the lower-power, lower-performing A53 cores. The combination of the relatively power-hungry A57 core and the 20nm planar process node results in a chip with a high power density and poor thermal performance. Being able to use FinFET and the power-optimized A72 CPU helps the Kirin 950 avoid this problem.

In both the streaming memory bandwidth and the memory latency tests, the Mate 8 performs similar to the Galaxy S6, not unexpected since they both use LPDDR4 RAM. The latest Snapdragon SoCs use a memory controller optimized for serial access patterns, which gives the 820 a big advantage in the memory bandwidth test (the 808 uses lower-bandwidth LPDDR3 RAM and the 810 cannot fully utilize the bandwidth of LPDDR4). This works against them, however, in the memory latency test that measures the time to complete a series of random memory operations. The memory controllers in the Kirin 950 and Exynos 7420 are more equal opportunity, favoring neither serial nor random patterns.

The Platform test mimics real-world workloads, testing CPU, memory, and storage performance. This test seems to use more random memory access patterns, which is why we see the BLU Pure XL and Asus ZenFone 2 perform quite well. The Mate 8 performs better than both of these devices and 17 percent better than the Galaxy S6. Storage performance for the Mate 8 is typical for an eMMC solution, trailing the S6's UFS 2.0 NAND, so its advantage in this test comes primarily from the A72 CPU.

Turning to the synthetic Geekbench suite, we get a clearer picture of the IPC differences between the A72, A57, and Qualcomm's Kryo CPU cores. Looking at the single-core overall scores first, the A72 in the Mate 8 performs 16 percent better in the integer tests than the A57 in the Galaxy S6, which turns into a mild 6 percent increase after accounting for the A72's 9.5 percent clock speed advantage. The A72 increases its lead to 25 percent (15 percent normalized) in floating-point workloads.

The small improvements over the A57 in these tests are not enough to match Kryo's performance, which outperforms the A72 by 12 percent (20 percent) in the integer tests and 32 percent (41 percent) in the floating-point tests, where the numbers in parentheses are the normalized differences after accounting for the A72's 7 percent clock speed advantage.

Geekbench 3 Pro Integer Results *

* Values in parentheses are the percent advantage for Kirin 950

Taking a closer look at the individual Geekbench integer workloads reveals several instances where the A72 sees no performance gain relative to the A57 (the values in the table are not adjusted for the 9.5 percent clock speed difference), including BZip2 Decompress, Sobel, and the various JPEG operations. This is not a complete surprise, because the integer execution units are similar between the A72 and A57, with a new Radix-16 integer divider (doubling bandwidth over A57) and a 1-cycle CRC unit the primary improvements. Both the Sobel and JPEG workloads rely heavily on multiplication and addition, so they see no improvement when running on the A72. The impact of the A72's improved branch predictor is also minimized when running benchmarks like Geekbench that primarily run math operations in a tight loop.

There are a few workloads, however, where the A72's architectural improvements make a noticeable difference: Lua, Dijkstra, AES, and SHA1. These tests rely heavily on lookup tables and data structures and seem to benefit from the A72's higher cache bandwidth and expanded zero-cycle forwarding.

Qualcomm's Kryo CPU clearly holds an IPC advantage over the A72 in integer operations. In our Snapdragon 820 Performance Preview, we determined that Kryo has a single integer multiply/divide unit just like the A57 and A72; however, we estimate it has only a 3-cycle latency compared to the longer 4-cycle latency of the ARM cores.

While single-core IPC is still the best indicator of smartphone application performance, multi-core throughput is becoming increasingly important, especially on Android. Comparing the multi-core integer performance of Kirin 950 to Exynos 7420 reveals an interesting pattern: the multi-core results are the exact opposite of the single-core results. Namely, tests that see a small gain when running on a single core see a larger gain when running on multiple cores and vice versa. Since the tests that see the biggest single-core gains take advantage of the A72's improved cache bandwidth, it's possible that the Kirin 950's reliance on the CCI-400 interconnect, instead of the newer CCI-500 that ARM recommends for the A72, is limiting the A72's performance. We'll need to see an example of a big.LITTLE processor using CCI-500 to know for sure, though.

Geekbench 3 Pro Floating-Point Results *

* Values in parentheses are the percent advantage for Kirin 950

In our detailed discussion of the A72 architecture we highlighted the A72's new Floating-Point/Advanced SIMD units, whose shorter pipeline lengths reduce execution latencies by up to 40 percent. There's also a new Radix-16 FP divider, which does more work per cycle, doubling bandwidth. These changes give the A72 a significant boost in floating-point performance over the A57 as shown in the table above. Other than two outliers (DGEMM and SFFT), the A72 performs anywhere from 4 percent (SGEMM) to 57 percent (Mandelbrot) better than the A57 after accounting for the difference in clock speed. 

For its Kryo CPU, Qualcomm made floating-point performance a priority, increasing IPC with a wider architecture that looks similar to Apple's Typhoon core in last year's A8 SoC. Even with latency minimized, the A72's narrower architecture cannot keep pace. It manages to almost pull even in Mandelbrot, gets less than half of Kryo's throughput executing Blur Filter, and finishes 41 percent behind overall after removing the A72's frequency advantage.

ARM's big.LITTLE philosophy of using multiple, narrower cores allows the Kirin 950 to pull ahead of the Snapdragon 820 in most of the multi-core tests. 

Unlike Geekbench, PCMark tests overall device performance by stressing the CPU, memory, and storage systems. Its lighter and more varied workloads are also affected by the device's CPU governor, just like other common apps, so it's indicative of real-world performance. 

The Mate 8 and its Kirin 950 SoC post the highest PCMark score of any device we've tested, outperforming the Moto X Pure Edition by 12 percent and the Galaxy S6 by 24 percent. Looking at the individual test scores, there's no obvious weak points for the Mate 8, which scores the highest in the Writing, Web Browsing, and Video Playback tests. It's only in the Photo Editing test, where image processing is supposed to occur on the GPU using the android.media.effect API, that it falls behind the Asus ZenFone 2 and Snapdragon 820 MDP.

The Mate 8 also does well in our JavaScript benchmarks, averaging 27 percent better than the Galaxy S6 in these three tests. Its performance is also very similar to the Snapdragon 820.

MORE: Best Smartphones
MORE: Smartphones in the News
MORE: All Smartphone Content

GPU And Gaming Performance

This section explores GPU performance with several synthetic and real-world game engine tests. To learn more about how these benchmarks work, what versions we use, or our testing methodology, please read our article about how we test mobile device GPU performance.

HiSilicon's Kirin 950 uses an ARM Mali-T880MP4 GPU. While the architectural differences between the T880 and the model it replaces, the T760, are not completely known, we do know that the T880 includes three ALU units per core versus two ALUs per core for the T760.

The table above summarizes the relevant current and previous generation GPUs from ARM and Imagination Technologies. Differences in architecture and nomenclature make it difficult to directly compare GPU hardware, so it focuses more on end performance. Qualcomm's Adreno GPUs are not included, because the company does not disclose details about its architecture.

Mali's Midgard differs from other GPU architectures in several ways. For starters, Midgard's vector SIMDs (single instruction multiple data) rely exclusively on instruction level parallelism (ILP) to keep its ALUs full of instructions. Competing GPU architectures are not vector based, and use a combination of ILP and thread level parallelism (TLP). There's pros and cons for both methods, making it difficult to say which offers better performance, but it ultimately depends on the workload.

The table highlights another difference between the Mali Midgard, PowerVR Rogue, and Qualcomm's Adreno architectures: IPC. Midgard performs fewer operations per cycle than its peers, but ramps up to a higher max frequency. The Kirin 950 in the Mate 8, for example, runs at up to 900MHz, resulting in 108 GFLOPS of FP32 throughput. This trails the A8's 115 GFLOPS (6 percent), Exynos 7420's 124 GFLOPS (15 percent), and the A9's more than 173 GFLOPS (exact GPU frequency unknown).

Here we see the Mate 8's Kirin 950 performing very much like a mid-range device rather than a flagship, its quad-core GPU besting only the PowerVR G6200 in the MediaTek Helio X10. The Mate 8 actually does pretty well in the first graphics test that focuses on vertex operations (with minimal pixel processing), essentially tying with the Galaxy S6 and performing 12 percent better than the Moto X Pure Edition's Adreno 418 GPU. It's in the second graphics test, which focuses heavily on pixel operations, where it falls behind.

The Mate 8's Physics score is surprisingly low. Since this tests CPU and memory performance, the Mate 8 should score the same as or a little better than the Galaxy S6. Instead, the S6, with its A57 CPU cores, scores 31 percent better than the Mate 8's higher clocked A72 cores. Unfortunately, we had to return our review unit before we could investigate this further. Without a closer look (and a device running final software), we do not want to make too much of this result, but it warrants further investigation.

In Basemark X, which runs on the Unity 4.2.2 game engine and uses OpenGL ES 2.0, the Galaxy S6 is 39 percent faster overall than the Mate 8. Most of its advantage comes from Dunes, a test that uses a lot of triangles (more than any of our other benchmarks), where it's 73 percent faster. HiSilicon's decision to use a quad-core GPU is definitely holding it back in this test; however, like all benchmarks, Basemark X puts more sever load on the GPU than actual games. When testing the LG G4 and its Snapdragon 808 SoC, for example, we found it played a number of modern games just fine. Considering that the Mate 8 actually performs better than the LG G4 in both offscreen and onscreen tests, and typical games do not use anywhere near as many triangles as Basemark X, the Mate 8's lower triangle performance should not be a severe limitation.

The Mate 8's 1080p display also works in its favor. By avoiding the greater rendering overhead of QHD, its onscreen Hangar results are actually better than the Galaxy S6's.

The Mali-T880MP4 GPU in the Kirin 950 really starts to struggle in the high quality test. With twice as many cores (giving it twice as much triangle and texturing throughput), the Mali-T760MP8 in the Exynos 7420 extends its lead to 70 percent overall (65 percent Dunes and 75 percent Hangar) when running offscreen.

GFXBench Manhattan runs on an OpenGL ES 3.0 based game engine that uses deferred rendering for its lighting effects. The PowerVR GT7600 in Apple's A9 and Adreno 530 in Qualcomm's Snapdragon 820 flex their ALU muscles, powering through the test's pixel operations. The Exynos 7420 in the Galaxy S6 also outperforms the Mate 8 by 35 percent. The Kirin 950 continues to outshine the Snapdragon 808 devices, though.

In the onscreen test, the Mate 8 pulls ahead of the Galaxy S6, since it only needs to render about half as many pixels.

Despite using an older OpenGL ES 2.0 based engine, we see essentially the same results in T-Rex. The Galaxy S6 is only 27 percent faster in the offscreen test than the Mate 8, and the Kirin 950 still holds a slim lead over the Snapdragon 808.

The limited number of ROPs in its quad-core GPU really holds the Kirin 950 and Mate 8 back in the Alpha Blending test, allowing the Snapdragon 808 to jump ahead of it. There's something amiss with the Galaxy S6's Alpha Blending results: The offscreen values are lower than the onscreen values, which is the opposite of what we should see. This appears to be a driver issue specific to the Galaxy S6, so we'll just ignore its results in this test.

The table at the top of this page documents the comparatively low FP32 ALU throughput of the Mali-T880MP4 GPU in Kirin 950, so it's no surprise to see it land near the bottom of the ALU performance chart. The Galaxy S6 scores 30 percent better than the Mate 8. Even the more budget friendly ZenFone 2 boasts better performance here. Qualcomm has worked steadily over the past several generations to boost the ALU performance of its Adreno GPUs, which is evident in this test. In the ALU onscreen test, nearly all of the devices are capped at the 60fps vsync limit.

While the T880 benefits from one additional ALU per core versus the T760, the Mali Midgard architecture makes no similar provision for texture units; the T880 still only has one per core. With twice as many texture units, the Galaxy S6 is 59 percent faster than the Mate 8 in the Fill test.

The Driver Overhead test measures CPU performance and driver efficiency by making a large number of draw calls. It's curious then to see the Mate 8 fall to the bottom of the chart, considering its strong CPU performance. The Mate 8 we're testing is still running prerelease software, so perhaps this is a driver issue that's left to be fixed.

MORE: Best Smartphones
MORE: Smartphones in the News
MORE: All Smartphone Content

Battery Life And Thermal Throttling

Battery life may be the most important performance metric for a mobile device. After all, it does not matter how quickly a phone or tablet can load webpages or how many frames per second the GPU can crank through once the battery runs down and the device shuts off. To learn more about how we test this critical facet of mobile computing, please read our battery testing methodology article.

The PCMark system test estimates battery life using real-world workloads and reflects how long a phone can last while continuously working on common tasks. Kirin 950's improved power efficiency, combined with the Mate 8's big 4000mAh battery, helps it last 9 hours and 23 minutes, the longest of any device we've tested, edging out Samsung's Galaxy S6 edge+ at 523 minutes and Motorola's Moto G (3rd gen) at 507 minutes. This feat is made even more impressive by the Mate 8's chart-topping performance score. Our colleagues at AnandTech examined Kirin 950's power efficiency in greater detail, which helps explain the Mate 8's outstanding results in PCMark.

The Mate 8 also does well in the GFXBench 3.0 battery test, which focuses on the GPU and is an indicator of battery life during intense gaming, lasting 3 hours and 45 minutes. We've seen a few devices last longer, but only because thermal throttling forced them to lower GPU frequency.

The Mate 8 exhibits good performance stability while gaming, staying within 94 percent of its peak performance over the first 30 minutes of T-Rex. After this we see the Mate 8 throttle back a bit, briefly reducing performance to 85 percent, but overall it's able to maintain near peak performance.

Looking at these results adds new context to HiSilicon's decision to use a quad-core GPU in the Kirin 950. By focusing on sustained performance rather than peak performance, a somewhat risky decision from a marketing standpoint, the Kirin 950 actually provides similar, if not better, performance after a few minutes of gameplay as compared to some other high-end SoCs, such as the Exynos 7420 or Snapdragon 810, at least when paired with a 1080p display like in the Mate 8.

Final Thoughts

ARM's Cortex-A72 CPU is a natural progression of the A57. At a high level, the two processors look similar, but ARM has made a number of power and performance optimizations to every stage in the pipeline. Most integer workloads see no appreciable performance gain, but there's a few specific cases, encryption in particular, that benefit from zero-cycle forwarding and the new Radix-16 integer divider. The A72's lower-latency floating-point units make a larger impact, increasing single-core Geekbench performance by about 15 percent overall relative to the A57 at the same clock frequency, with most of the individual workloads showing 30 to 60 percent gains.

Despite the A72's improvements, it's still a narrower architecture than Apple's Twister CPU or Qualcomm's new Kryo core, which limits IPC. After adjusting for differences in clock frequency, Kryo holds a 20 percent advantage in Geekbench integer and a 41 percent advantage in Geekbench floating-point performance. The A72 can, however, reach higher frequencies (we should see A72 cores running at 2.5GHz), which helps mitigate, and in some cases overcome, Kryo's greater IPC.

HiSilicon is the first to deliver the A72 on TSMC's 16nm FinFET+ process, and the first to use ARM's latest high-end GPU -- the Mali-T880 -- in its Kirin 950 SoC. This combination gives it better system performance and power efficiency than A57-based SoCs, including Qualcomm's Snapdragon 810 and Samsung's Exynos 7420. Kirin 950 also looks like it will be competitive with Snapdragon 820, at least on non-GPU related tasks, with each SoC having an edge in certain workloads.

At first glance, HiSilicon's decision to use the Mali-T880 in a quad-core configuration looks puzzling. While the T880's one additional ALU and higher max frequency helps keep it within 15 to 30 percent of the Exynos 7420's octa-core T760 GPU in shader-heavy games, having only half as many ROPs, texture units, and triangle units hurts peak performance over a wide range of gaming workloads, allowing the Exynos 7420 to extend its lead to around 60 percent. In all of our gaming benchmarks, the Kirin 950 performed more like a mid-range SoC.

A closer look, however, reveals HiSilicon's logic. Sure, the Kirin 950 is not going to wow anyone with its peak performance, but our tests show that its gaming stability is excellent, able to sustain near max performance over long periods of time. When paired with the Mate 8's 1080p display, the Kirin 950 actually performs better than the Exynos 7420 in the Galaxy S6 after a short period of time, because high temperature forces the S6 to throttle back its GPU frequency. This means the Kirin 950 and Mate 8 should not have any issues playing real-world games. We do not think the Kirin 950's Mali-T880MP4 GPU is powerful enough to drive QHD displays, however, limiting its use in some high-end flagships. The smaller GPU core and hybrid LPDDR3/LPDDR4 memory controller does make the Kirin 950 a high-performing mid-range SoC option from a cost standpoint, though.

The Kirin 950 seems a good fit for Huawei's Mate 8, which set new performance and battery life records in PCMark, our best benchmark for predicting real-world behavior. These tests corroborate our own first-hand experience: the Mate 8's UI was very fluid, Web pages loaded and scrolled quickly, and it just felt very fast overall. There's still several things we have not looked at yet, such as the display and cameras, but the Mate 8 seems like the real deal when it comes to performance and battery life.

MORE: Best Smartphones
MORE: Smartphones in the News
MORE: All Smartphone Content

Update, 1/19/16, 6:50am PT: Clarified Huawei's growth statistics in first paragraph.

Matt Humrick is a Staff Editor at Tom's Hardware, covering Smartphones and Tablets. Follow him on Twitter.

Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Samsung announced that it has begun the mass production of chips utilizing its 14nm Low-Power Plus (LPP) process node. The LPP process node is the second generation 14nm FinFET process from Samsung, which brings improvements in energy efficiency as well as performance.

Samsung's Exynos 7 Octa chip, which powers the Galaxy S6, was built on the 14nm Low-Power Early (LPE) process node. The new process node will be used to build both Samsung's own next-generation chip, the Exynos 8 Octa, as well as Qualcomm's Snapdragon 820. The chips are expected to arrive in the first half of this year.

"We are pleased to start production of our industry-leading, 2nd generation 14nm FinFET process technology that delivers the highest level of performance and power efficiency" said Charlie Bae, executive vice president of sales and marketing for System LSI Business at Samsung Electronics. "Samsung will continue to offer derivative processes of its advanced 14nm FinFET technology to maintain our technology leadership."

Samsung says that the new LPP process delivers up to 15 percent improvements in speed as well as up to 15 percent improvements in power consumption compared to the LPE process. Samsung's new process is one of the best in the chip market, surpassed only by Intel's own 14nm process.

However, Intel has a small share of the smartphone and IoT (Internet of Things) markets and it doesn't let other chips makers use its fabs. Therefore, the most cutting edge chips in the smartphone and IoT markets this year will use either Samsung's 14nm LPP process or TSMC's 16FF+ process.

Qualcomm should be the main beneficiary of Samsung's new process this year. With a new CPU core and a modern process, the Snapdragon 820 will likely avoid the overheating issues of its predecessor, the Snapdragon 810, although it will still largely depend on whether Qualcomm will push the chip past its optimum performance levels or not. Samsung's next-generation chip will also use a custom CPU core and should benefit from the same 14nm LPP improvements.

Lucian Armasu is a Contributing Writer for Tom's Hardware. You can follow him at @lucian_armasu. 

Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Introduction

ARM announced the Cortex-A72, the high-end successor to the Cortex-A57, near the beginning of 2015. For more than a year now, SoC vendors have been working on integrating the new CPU core into their products. Now that mobile devices using the A72 are imminent, it’s a good time to discuss what makes ARM’s flagship CPU tick.

With the A57, ARM looked to expand the market for its CPUs beyond mobile devices and into the low-power server market. Using a single CPU architecture for both smartphones and servers sounds unreasonable, but according to ARM’s Mike Filippo, lead architect for the A72, high-end mobile workloads put a lot of pressure on caches, branch prediction, and the translation lookaside buffer (TLB), which are also important for server workloads. Where the A57 seemed skewed towards server applications based on its power consumption, the A72 takes a more balanced approach and looks to be a better fit for mobile.

The Cortex-A72 is an evolution of the Cortex-A57; the baseline architecture is very similar. However, ARM tweaked the entire pipeline for better power and performance. Perhaps the A57’s biggest weakness was its relatively high power consumption, especially on the 20nm node, which severely limited sustained performance in mobile devices, relegating it to short, bursty workloads and forcing SoCs to use the lower-performing Cortex-A53 cores for extended use.

ARM looks to correct this issue with the A72, going back and optimizing nearly every one of the A57’s logical blocks to reduce power consumption. For example, ARM was able to realize a 35-40% reduction in dynamic power for the decoder stage, and by using an early IC tag lookup, the A72’s 3-way L1 instruction and 2-way L1 data caches also use less power, similar to what direct-mapped caches would use. According to ARM, all of the changes made to the A72 result in about a 15% reduction in energy use compared to the A57 when both cores are running the same workload at the same frequency and using the same 28nm process. The A72 sees an even more significant reduction when using a modern FinFET process, such as TSMC’s 16nm FinFET+, where an A72 core stays within a 750mW power envelope at 2.5GHz, according to ARM.

[Image Source: Hiroshige Goto PC Watch]

MORE: Best SmartphonesMORE: How We Test Smartphones & TabletsMORE: All Smartphone ContentMORE: All Tablet Content

Architecture Overview

Instruction Fetch

The A72 sees improvements to performance too, starting with a much improved branch prediction algorithm. ARM’s performance modeling group is continuously updating the simulated workloads fed to the processor, and these affect the design of the branch predictor as well as the rest of the CPU. For example, based on these workloads, ARM found that instruction offsets between branches are often close together in memory. This allowed it to make certain optimizations to its dynamic predictor, like enabling the Branch Target Buffer (BTB) to hold anywhere from 2000 large branches to 4000 small branches.

Because real-world code tends to include many branch instructions, branch prediction and speculative execution can greatly improve performance—something that’s usually not tested by synthetic benchmarks. Better branch prediction usually costs more power, however, at least in the front-end. This is at least partially offset by fewer mispredictions, saving power on the back-end by avoiding pipeline flushes and wasted clock cycles. Another way ARM is saving power is by shutting off the branch predictor in situations where it’s unnecessary. There are many blocks, for instance, where instructions between branches are greater than the 16-byte instruction window the predictor uses, so it makes sense to shut it down because the predictor will obviously not hit a branch in that section of code.

Decode / Rename

The rest of the front-end is still in-order with a 3-way decoder just like the A57. However, unlike the A57’s decoder, which decodes instructions into micro-ops, the A72 decodes instructions into macro-ops that can contain multiple micro-ops. It also supports AArch64 instruction-fusion within the decoder. These macro-ops get “late-cracked” into multiple micro-ops at the dispatch level, which is now capable of issuing up to five micro-ops, up from three on the A57. ARM quotes a 1.08 micro-ops per instruction ratio on average when executing common code. Improving the A72’s front-end bandwidth helps keep the new lower-latency execution units fed and also reduces the number of cases where the front-end bottlenecks performance.

Dispatch / Retire

As explained above, the A72 is now able to dispatch five micro-ops into the issue queues that feed the execution units. The issue queues still hold a total of 66 micro-ops like they did on the A57—eight entries for each pipeline except the branch execution unit queue that holds ten entries. While queue depth is the same, the A72 does have an improved issue-queue load-balancing algorithm that eliminates some additional cases of mismatched utilization of the FP/Advanced SIMD units.

The A72’s dispatch unit also sees significant area and power reductions by reorganizing the architectural and speculative register files. ARM also reduced the number of register read ports through port sharing. This is important because the read ports are very sensitive to timing and require larger gates to enable higher core frequencies, which causes a second-order effect to area and power by pushing other components further away. In total, the dispatch unit sees a 10% reduction in area with no adverse affect on performance.

Execution Units

Moving to the out-of-order back-end, the A72 can still issue eight micro-ops per cycle like the A57, but execution latency is significantly reduced because of the next-generation FP/Advanced SIMD units. Floating-point instructions see up to a 40% latency reduction over the A57 (5-cycle to 3-cycle FMUL), and the new Radix-16 FP divider doubles bandwidth. The changes to the integer units are less extensive, but they also make the change to a Radix-16 integer divider (doubling bandwidth over A57) and a 1-cycle CRC unit.

Reducing pipeline length improves performance directly by reducing latency and indirectly by easing pressure on the out-of-order window. So even though the instruction reorder buffer remains at 128 entries like the A57, it now has more opportunity to extract instruction-level parallelism (ILP).

Expanded zero-cycle forwarding (all integer and some floating-point instructions) further reduces execution latency. This technique allows dependent operations—where the output of one instruction is the input to the next—to directly follow each other in the pipeline without a one or more cycle wait period between them. This is an important optimization for cryptography.

Load / Store

The load/store unit also sees improvements to both performance and power. One of the big changes from A57 is a move away from separate L1 and L2 prefetchers to a more sophisticated combined prefetcher that fetches from both the L1 and L2 caches. This improves bandwidth while also reducing power. There are additional power optimizations to the L1-hit pipeline and forwarding network too.

The L2 cache sees significant optimizations for higher bandwidth workloads. Memory streaming tasks see the biggest benefit, but ARM says general floating-point and integer workloads also see an increase in performance. A new cache replacement policy increases hit rates in the L2, which again improves performance and reduces power overall.

Another optimization increases parallelism in the table-walker hardware for the memory management unit (MMU), responsible for translating virtual memory addresses to physical addresses among other things. Together with a lower-latency L2 TLB, performance improves for programs that spread data across several data pages such as Web browsers.

Final Thoughts

At a high-level, the A72 looks nearly identical to the A57, but at a lower level there are a significant number of changes throughout the entire pipeline that appear to make the A72 a decent upgrade. The most notable changes affecting performance are the improved branch prediction, increased dispatch bandwidth, lower-latency execution units, and higher bandwidth L2 cache. All of these enhancements, and many more which we did not discuss here, lead to better performance—between 16-50% across a range of synthetic benchmarks, according to ARM. Real-world performance gains will be less, of course, but A72 is definitely an improvement over the A57, especially with floating-point workloads.

The A72 is not a pure performance play, however. ARM is targeting a much higher power efficiency with this architecture than with any previous high-end CPU core. It’s clear from the lengthy list of optimizations discussed above that reducing power consumption was paramount; many of the changes are purely focused on power with no net performance gain.

Reducing power and area—the A72 achieves a 10% core area reduction overall—obviously has a positive effect on battery life and cost, but it has a secondary effect on performance too. Normally, reducing latency in the execution units puts pressure on the max attainable core frequency due to increased circuit complexity and tighter timing windows; however, the A72’s power and area optimizations elsewhere, not to mention the move to FinFET, actually allow the A72 to reach a slightly higher frequency. Reducing power also reduces thermal load, allowing for higher sustained performance, something the A57 struggles with at 20nm.

The Cortex-A72 may not be a revolutionary design that catapults it above Apple’s Twister CPU in the A9 SoC for single-core performance or undercuts the A53 in power consumption, but it’s a significant update nonetheless, addressing the A57’s issues by enabling higher peak and sustained performance while using less power.

MORE: Best SmartphonesMORE: How We Test Smartphones & TabletsMORE: All Smartphone ContentMORE: All Tablet Content

Update, 1/12/16, 10:55am PT: Clarified how the Decode/Rename/Dispatch pipeline works and added some information about the issue queues.

Matt Humrick is a Staff Editor at Tom's Hardware, covering Smartphones and Tablets. Follow him on Twitter.

Follow us on Facebook, Google+, RSS, Twitter and YouTube.

When we last heard from ImasD Technologies earlier this year, the small Spain-based company was beginning to make a little noise and looked like it was poised to be a player in the modular smartphone/tablet market.

Legal Troubles

Things have shifted for ImasD in the interim. The company is in some legal trouble, apparently because of the name of its modular mobile platform, ClickARM. (Three guesses who the other litigant is. ...it’s ARM.) ImasD’s Pedro Peláez told Tom’s Hardware that the company will go to trial with ARM. “We’ll see what happens,” he said.

The fallout from the legal issues are ongoing, but primarily, ImasD refunded ClickARM One tablet customers their money and had to discontinue using that brand name. However, Peláez said that the company can continue to sell the technology with different nomenclature, and he’s confident that ImasD will emerge the winner in the lawsuit. He posited that ImasD registered its “arm” related language in a different class than ARM, and he also reminded us of the small company’s 2011 legal victory over Apple.

The company is planning to re-launch the tablet, possibly with a different name, via a crowdfunding site with a limited run of 2,000 or so devices.

Passing On PuzzlePhone, And A Cancelled Kickstarter

ImasD Technologies was also originally part of the PuzzlePhone project -- which seems to be taking off -- but ImasD left the project, citing what amounts to creative differences. However, Peláez said that ImasD is planning to launch its own phone project "when the platform becomes ready for makers and startups."

The company was also involved with a "portable Steam machine" called Smach. After raising €160,984 on Kickstarter, the campaign was cancelled. I’m inferring here, but a likely reason for the cessation has to do with the this disclaimer on the campaign page:

“Liability disclaimer: SMACH Z project is not being developed by Valve® Corporation. SMACH Z project doesn't have any connection to Valve® or Steam™. SMACH Z is not a official Steam Machine™.”

ImasD is planning to buy the company and re-release a finished version of the device via a crowdfunding campaign.

A Better Year Ahead

At CES in January, ImasD will show off its technology inside other products, including a reference design for a smartwatch with Analog Devices and an advanced baby monitor with Cypress.

Going forward, though, Peláez is optimistic. He told me that ImasD has five modules for ClickARM based on the i.MX 6 system on module (NXP and Freescale merged in December), including single-, dual- and quad-core offerings. You can currently snap up a ClickARM core module running a Samsung Exynos 4412 chip with 2 GB memory for 45€. A complete system development kit ranges from 80-110€ depending on the selected modules.

Peláez said that ImasD is finalizing the i.MX 6 design and should have a “batch” ready for the public around Q2 2016. There should also be a new PCB hub coming in that same time frame (it’s currently available to a closed group of devs).

ImasD was already an unlikely upstart, and in light of the (somewhat self-inflicted) difficulties of the past year, the future of the company seems murky. 

Update, 12/27/15, 9:15pm PT: ImasD reached out to add some information. Specifically, the company said that it's already booked some $40 million in revenue for the next two years, starting now.

Seth Colaner is the News Director for Tom's Hardware. Follow him on Twitter @SethColaner. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Qualcomm announced that its mid-range chips for next year, the Snapdragon 618 and Snapdragon 620, will be renamed the Snapdragon 650 and Snapdragon 652, respectively, to better represent the jump in performance over the Snapdragon 615, 616 and 617.

Whereas the Snapdragon 615 came with Cortex-A53 CPU cores (the same ones in the lower-end Snapdragon 410 chip, but at a higher clock rate), the new Snapdragon 650 and 652 will bring ARM’s latest Cortex-A72 CPU core.

The new CPU core is ARM’s highest-end CPU design and a direct successor to Cortex-A15 and Cortex-A72. According to the company, the Cortex-A72 is 3.5x faster on a 16nm FinFET process than the Cortex-A15 was on a 28nm process, and 1.9x faster than the Cortex-A57 built on 20nm.

Despite this large performance improvement over the Cortex-A57, which was in this year’s highest-end chips (and even caused some overheating issues in 20nm chips), Qualcomm is still going to use it in its mid-range SoCs, such as the Snapdragon 650 and Snapdragon 652, for next year. That’s because Qualcomm must believe (or know) that its Kryo cores that will be inside its flagship chip, the Snapdragon 820, are even faster.

According to Qualcomm, the Snapdragon 650 and 652 chips have seen quite high demand from device makers so far. This could be because the two chips have, for the first time in this tier, 4k video capture and playback support, as well as strong gaming capabilities thanks to the next-generation Adreno 510 GPU. The chips also come integrated with Qualcomm’s X8 LTE generation, which brings Cat. 6 LTE speeds and carrier aggregation support.

The good news here is that chips such as Snapdragon 650 and 652 should bring to mid-range devices some features and performance that equal or surpass even this year’s high-end chips. Qualcomm’s Snapdragon 810 controversy from this year must have made the company want to do much better next year, and so far it shows, at least on paper.

It still remains to be seen how these chips will actually perform in devices. Qualcomm should certainly be much more careful about prioritizing performance over power consumption and thermal management this time, unless it wants to risk another controversy and loss of sales next year.

______________________________________________________________________

Lucian Armasu joined Tom’s Hardware in early 2014. He writes news stories on mobile, chipsets, security, privacy, and anything else that might be of interest to him from the technology world. Outside of Tom’s Hardware, he dreams of becoming an entrepreneur.

You can follow him at @lucian_armasu. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Introduction

When Apple shipped the iPhone 5s with a custom designed 64-bit CPU, it took the mobile industry by surprise. The move to 64-bit was inevitable, but nobody expected Apple to get there so quickly, including Qualcomm whose 64-bit CPU was just a dot on a long-term roadmap. Without a custom designed core of its own, Qualcomm adopted ARM’s Cortex-A53 and Cortex-A57 stock cores for its flagship Snapdragon 810 processor last year.

Working from a less than ideal position resulted in a less than ideal SoC. Even before the Snapdragon 810 made a public appearance, there were rumors about overheating and memory controller issues. Our own testing validated the overheating rumors—a product of pairing the power-hungry A57 core with TSMC's 20nm HKMG process—and we have yet to see an 810 use the full bandwidth available from LPDDR4-1600 memory, even the v2.1 revision.

Even though the 810 was a stopgap, it was not all bad. The Adreno 430 GPU improved upon the Adreno 420 in Snapdragon 805, maintaining Qualcomm’s lead in ALU performance, and a faster Category 9 X10 LTE modem moved next door to the CPU after getting booted off the SoC island in the 805.

Still, it’s hard to feel anything but disappointment when it comes to the 810. Excessive thermal throttling held back performance, forcing the A57 CPU cores to sit idle. In some scenarios, the older Snapdragon 801 and 805 SoCs, along with some mid-range all A53 designs, offered equivalent or better performance. An unenviable position for a flagship product.

Qualcomm hopes to move past these issues with the Snapdragon 820 and Kryo, its first custom-designed 64-bit CPU. Qualcomm’s goal for the 820 is not just about improving performance, however. It's also about enabling innovative user experiences by leveraging heterogeneous compute, which combines the unique abilities of each processor—CPU, GPU, DSP, and ISP—to maximize performance and minimize power use. Computer vision, advanced imaging, and virtual reality are all targeted applications.

Zeroth

Many of these new capabilities will be made possible by Zeroth, a machine learning and computer vision API that developers can use to take advantage of Snapdragon 820’s hardware. This “cognitive computing platform,” as Qualcomm calls it, should further improve the capabilities of virtual assistants on smartphones and also anything that requires more human-like intelligence. One of the ways it does this is by mimicking how humans learn through positive reinforcement. We’ve already started to see mobile devices incorporate intelligent behaviors, but these generally leverage the processing power of cloud computing. With the 820, Qualcomm believes this processing can now be done locally on the device, improving privacy as a consequence, since all of that unique user data will not have to be processed on someone else’s servers.

Qualcomm’s Scene Detect technology is the application of Zeroth to computer vision. Again taking advantage of heterogeneous computing, it uses neural networks for scene detection, object recognition, and pattern matching for both still images and video captured by the device’s camera. There are many uses for this technology, including the automatic tagging of photos for easier search retrieval and augmented reality. The video above demonstrates the basic capabilities of this system.

Smart Protect will be one of the very first “killer applications” of Zeroth. A technology that goes beyond traditional signature based antivirus protection, it will be able to identify “abnormal behavior,” such as noticing that the phone is taking pictures when the screen is locked or sending SMS messages without user interaction, by using machine learning and behavioral analysis. This feature can be used to identify either zero-day malware or “transformational malware,” which is malware created to bypass popular antivirus software.

This feature has a component running at a low-level within the Android kernel and another part running within Qualcomm's SecureMSM secure execution environment, which should make it much harder for malware to circumvent. This also puts Smart Protect in a position to effectively monitor system resources, app communication, etc.

Heterogeneous Computing Examples

Beyond Zeroth, Snapdragon 820 uses heterogeneous computing to enable a host of advanced imaging features. One example takes advantage of OpenCL 1.2 and FastCV APIs to post-process a video stream in real time, separating and blurring the background to increase privacy during a video conference. By combining the processing power of both the CPU and GPU, Qualcomm claims performance improves by more than 2x compared to using just the CPU alone and also reduces power consumption by up to 40%. This same technology is also used to improve the quality of panoramic images, removing stitching seams and cleaning up ghosting artifacts caused by moving objects. Further applications could include providing a real-time preview of video effects while recording or improving augmented reality experiences.

Qualcomm’s improveTouch feature, which is also present in the Snapdragon 810 SoC, moves functionality from an external touchscreen controller onboard the SoC. Utilizing the DSP and low-power CPU island, this improves touch latency and enables more sophisticated noise rejection algorithms. The improved processing enables sophisticated water droplet rejection, essentially making the screen usable when wet, and improves touch sensitivity while charging the device by filtering out EMI. There’s also an ultra-low power double-tap screen wakeup feature.

Tying all of the specialized processors together in an efficient manner is the job of Qualcomm’s Symphony System Manager. According to Qualcomm, “Symphony is designed to manage the entire system-on-chip in different configurations so that the most efficient and effective combination of processors and specialized cores are chosen to get the job done as quickly as possible, with the least amount of power”. That’s no easy task, so we’re eager to see what real-world battery life is like when products start shipping.

Now that we understand Qualcomm’s vision for Snapdragon 820 and its future SoCs (it’s heterogeneous computing if you have not figured it out) and some of the experiences it enables, it’s time to take a closer look at the hardware.

MORE: Best SmartphonesMORE: All Smartphone ContentMORE: All Tablet Content

Snapdragon 820 Architecture

Qualcomm has remained tight-lipped about its latest processor designs. Unlike ARM’s full-disclosure model, Qualcomm is much more Apple-like when it comes to releasing low-level information, particularly about its GPUs.

While technical details are in short supply, our curiosity about Snapdragon 820’s most interesting component—the 64-bit Kryo CPU core—remains high. The move from TSMC's 20nm HKMG planar process to Samsung’s 14nm FinFET process should mitigate the thermal issues experienced by Snapdragon 810 and allow the 820 to use less power and/or reach higher clock speeds. According to Qualcomm, with the “Kryo CPU and Snapdragon 820, you can expect up to two times the performance and up to two times the power efficiency” when compared to the A57 CPU in Snapdragon 810. This is certainly a bold claim, but based on the overheating issues we observed with the 810, coupled with the excellent performance from Samsung’s Exynos 7420 SoC, which uses Samsung’s first-generation 14nm LPE (Low Power Early) FinFET process, it might be possible. It’s not clear if the 820 will use LPE or Samsung’s second-generation 14nm LPP (Low Power Plus) FinFET process, which, according to Samsung, can achieve 10% higher frequency at lower power than LPE because of a better fin aspect ratio.

As mobile workloads continue to evolve, so do mobile SoCs. One parameter in constant flux is the optimal number of CPU cores, with designs ranging from Apple’s A9, which uses two CPU cores, to MediaTek’s Helio X20, which uses ten CPU cores in a tri-cluster, big.LITTLE arrangement. According to Tim McDonough, Qualcomm's VP of Marketing, “people don't really need more than four cores.” While this statement will likely spark some heated debate, Qualcomm’s data seems to be pointing it in that direction since the Snapdragon 820 uses four Kryo CPU cores in a dual-cluster, heterogeneous configuration. While the underlying architecture of each CPU core is the same, the clusters are optimized to operate at different frequencies and power levels, sort-of like ARM’s big.LITTLE approach. The two Kryo cores in the lower-power “Silver” cluster operate at frequencies up to 1.6GHz and share a 512KB L2 cache. The second pair of Kryo cores in the higher-performing “Gold” cluster operate at frequencies up to 2.2GHz and share a 1MB L2 cache. While L2 cache is not shared between the Gold and Silver clusters, the two L2 caches use a snooping mechanism to maintain coherency. Unlike Apple’s A9, Snapdragon 820 does not use an L3 cache. Qualcomm says it considered using an L3 cache, but ultimately decided the benefits did not outweigh the additional cost in power and die space for its design. Qualcomm is not divulging any lower-level details of its Kryo architecture, so we’ll have to see what, if anything, we can infer from our test data.

Qualcomm's Snapdragon 8xx Flagship Family

While information about Kryo is scarce, details about Snapdragon 820’s Adreno 530 GPU is nonexistent. Beyond the name, the only thing we know for sure is that it runs at 133-624MHz. When pushed for more info, Qualcomm said it made lots of small architectural changes throughout the design, which would imply that the Adreno 530 is not a drastic redesign, but an evolution of the Adreno 430. One of the changes it mentioned was making better use of data compression when moving data around within the GPU in order to reduce power consumption.

Given Qualcomm’s focus on heterogeneous computing, it’s no surprise that the GPU and CPU can both snoop into the other’s cache, enabling better sharing of data, since both processors use 64-bit virtual addresses. With a heavy focus on compute capability, we also expect the Adreno 530 to further improve ALU performance, something which Qualcomm has done for the past several generations.

The Adreno 530 supports the latest graphics API standards, including OpenGL ES 3.1 + Android Extension Pack, DirectX 12, and Vulkan (once ratified by Khronos). Like the Adreno 430, the 530 includes a dedicated fixed-function block in hardware for accelerating tessellation.

The Snapdragon 820 comes with the brand-new Kryo CPU core, a new Adreno 530 GPU, as well as a new Image Signal Processor (Spectra ISP) and a Digital Signal Processor (Hexagon 680 DSP). Each of them come with significant boosts in performance over the previous generation, but they can also work together through heterogeneous computing to finish tasks more than twice as fast than if only the CPU was being used and save up to 40 percent energy.

Testing Snapdragon 820

In usual fashion, we were invited to Qualcomm's headquarters to get a first look at Snapdragon 820’s performance. Since retail products using the 820 are not available yet (or have even been announced), we subjected Qualcomm's Mobile Development Platform (MDP) to a battery of tests.

The smartphone MDP is an oversized phone used internally by Qualcomm for testing and development. Obviously all the usual caveats apply: We’re testing pre-production hardware and software, so we could encounter some performance anomalies. Furthermore, because the MDP chassis is larger and thicker than the sleek flagship phones that will eventually house the 820, we cannot make any assessments of thermal throttling. Also, with only a limited amount of hands-on time, we’ll need to wait for shipping hardware to quantify battery life.

The data that we will be presenting then, serves as a performance preview, an early look at the 820’s potential. While the final numbers may change, our experience testing both the Snapdragon 805 and Snapdragon 810 MDPs (those were both tablets) showed good correlation with eventual shipping products.

Devices Used for Testing

Even though testing took place outside of our lab, we followed our usual testing procedures as closely as possible. We used our own benchmark files and ran each test at least twice, averaging the results, like we usually do. Because of the limited testing time, however, we did not have time to allow the MDP to cool between tests. We did not see much variation between successive runs, though, so thermal throttling did not appear to impact the results.

CPU And System Performance

Since this is our first look at Kryo, Qualcomm’s first custom-designed 64-bit CPU core, there are a number of outstanding questions. How does the dual-cluster, quad-core arrangement compare to Apple’s dual-core approach or the octa-core big.LITTLE processors common in smartphones today? Does Kryo use a narrower architecture similar to ARM’s A57/A72, or has Qualcomm gone wider and more complex like Apple’s A9? We’re also curious to see if Snapdragon 820 gets better memory performance than the 810.

Taking a look at the single- and multi-threaded CPU System test first, we see the four Kryo cores easily outpace the older Krait CPU cores and the all Cortex-A53 Helio X10 SoC. Snapdragon 820 also surpasses the hexa-core Snapdragon 808 and octa-core 810 by at least 19%. The margin of victory over the Exynos 7420, which makes better use of its four Cortex-A57 cores than the Snapdragon 810, is very small, however. Considering that the Exynos 7420 has an additional four A53 cores at its disposal, though, it appears that Kryo’s IPC (instructions per cycle) is a bit better than the A57’s. Apple’s A9 is 42% faster than the 820 here, suggesting that Kryo is a narrower architecture that has more in common with ARM’s A57 than Apple’s Twister CPU.

In the OpenGL ES 2.0 based Graphics test, the 820’s Adreno 530 GPU is well ahead of most of its competitors. The 59% advantage over the 810’s Adreno 430, using similar clock frequencies, is impressive. Only the PowerVR GT7600 in Apple’s recently released A9 gets close, pulling within 9%.

While the Memory test is meant to measure the speed of the internal NAND storage, on high-end devices it turns into a memory test due to how the OS uses a RAM cache to buffer storage access. It’s no surprise then to see the LPDDR4 based devices with higher throughput.

Snapdragon 820 manages to pull ahead overall in AndEBench, but that’s not as important as how it performs in the individual tests. In CoreMark-HPC, which measures both single- and multi-core CPU performance using a mixture of integer and floating-point workloads, Snapdragon 820’s Kryo cores perform well, but the Exynos 7420’s octa-core design gives it the edge when dealing with many threads.

The Snapdragon 810 struggles mightily in this synthetic test. Because of thermal throttling, it leaves the A57 cores idle and relies on the four A53 cores almost exclusively (click and scroll down for graph). With twice as many A53 cores, even the Helio X10 outperforms it. The 810 also falls behind the Snapdragon 808, which actually uses its two A57 cores, and only attains half the performance of the 820. While encouraging, we cannot draw any conclusions here about the 820’s heat generation for the reasons stated on the previous page.

The memory bandwidth results for Snapdragon 820 are nothing short of amazing, improving performance over the 810’s beleaguered memory controller by a factor of 2.6x. The 820 is even 62% faster than the Exynos 7420, which also uses LPDDR4 memory. We’ve noticed in previous testing that the memory controllers in Qualcomm’s latest SoCs—the Snapdragon 808 and 810—seem to be heavily optimized for serial memory access patterns like those used in the STREAM memory bandwidth tests included in AndEBench. Based on these results, it’s safe to say that the 820 extends this trend. This is why the 808, 810, and 820 do not perform as well in the memory latency test, which uses a randomly ordered data structure. In theory, this type of optimization should help improve performance when working with large, contiguous chunks of data, like high-resolution graphics textures and frame buffers or when processing pictures and video. Qualcomm is not the only company who’s gone this direction; Apple is also using serial-optimized memory controllers in its SoCs.

Once again we see the 820’s Adreno 530 GPU lead the pack in the 3D graphics test. This test is a bit more demanding than the older Basemark OS II test, but the Adreno 530 still performs 39% faster than the Adreno 430.

While we’re including the results from the Storage and Platform tests, they really are not relevant here, since we’re not concerned with NAND performance. It should be noted that the Snapdragon 820 MDP’s storage is encrypted, unlike all of the other tested devices. This could explain the MDP’s low storage performance (we were unable to determine if encryption was being accelerated in hardware), or the MDP just used slow NAND. The 820’s score in the Platform test is also affected by this, since the workload includes reading and writing files to internal storage.

Looking at single-core performance in Geekbench confirms what we’ve already seen in our other tests: Qualcomm’s Kryo CPU core is slower than Apple’s Twister CPU, but faster than ARM’s A57. After normalizing the clock frequency, Kryo’s integer performance is 27% faster than the A57. Apple’s Twister CPU ends up being 38% faster than Kryo. Based on what we know about the architecture for both Twister and A57—and after doing some back-of-the-envelope calculations—it looks like Kryo has a single integer multiply/divide unit with a 3-cycle latency. This is very similar to the A57 and A72, which also have a single integer multiply/divide unit, but with a longer 4-cycle latency.

In Geekbench’s single-core floating-point workloads, Kryo is 61% faster than the A57, which has two floating-point/NEON units, the same as A72. The A72 does see a 25-40% latency reduction for floating-point operations, among other tweaks, which should help close the gap, but we expect Kryo to maintain a slim lead over the A72 in floating-point performance. Apple’s Twister core is only 39% faster than Kryo in this test.

Once again, Snapdragon 820 shows impressive memory bandwidth in the STREAM test. Whatever memory controller issues existed in Snapdragon 810 seem to have been remedied in the 820.

Moving on to the multi-core test, we see the 820 jump ahead of the A9 in floating-point performance, thanks to its two extra CPU cores. It does fall behind the octa-core SoCs in integer performance, but there really are not many real-world apps that use eight cores at once, so this is unlikely to adversely affect the user experience.

PCMark runs several realistic workloads and is very sensitive to CPU governor behavior. For this reason, its results tend to be more device dependent. While it’s encouraging to see Snapdragon 820 top the chart in overall PCMark performance, just keep in mind that these results will likely change a bit in shipping hardware.

Looking at the individual test scores, the 820 does well with video playback, but does not really stand out in the Web Browsing or Writing tests. Where it really shines, though, is in the Photo Editing test. Most of the image processing during this test is supposed to occur on the GPU using the android.media.effect API. In our review of the Asus ZenFone 2, which also performs well in this test, we were able to confirm that it does use the GPU as intended. All of the other devices we’ve tested, however, run the Photo Editing test on the CPU. While we were unable to confirm this in the short time we had for testing, the Snapdragon 820 seems to utilize the compute power of the GPU like the ZenFone 2, serving as an example for the benefits of heterogeneous computing.

While it’s unfair to compare the scores of the A9 to the other SoCs, we’re including the results to show the effect software plays in browser benchmarks. Since Apple controls the design of its hardware and operating system, it can heavily optimize its Safari web browser, resulting in substantial performance gains.

The older version of the Opera browser we use for testing Android devices is not nearly as optimized, resulting in lower scores across the board. Still, it’s useful for making hardware comparisons among Android devices. In all three of the browser benchmarks, Snapdragon 820 eeks out a slim lead.

Based on these initial tests, Qualcomm’s new Kryo CPU performs better than ARM’s Cortex-A57, but slower than Apple’s Twister. Integer performance is a little better than A57/A72, which implies that Kryo uses the same number of integer units as ARM’s best cores, but with better latency. Kryo pulls further ahead of the A57 in floating-point performance, and while it’s more difficult to ascertain architectural differences in floating-point units based on system-level tests, our data suggests Kryo looks more like the Typhoon core in Apple’s A8 rather than A57.

GPU And Gaming Performance

In the previous section, we saw Snapdragon 820 excel in the sequential memory bandwidth tests, which should help boost overall GPU performance. With memory bandwidth a less likely bottleneck, will a weak point emerge in the Adreno 530’s architecture? Will it show an increase in ALU performance like we expect?

In 3DMark Ice Storm Unlimited, the Adreno 530 in Snapdragon 820 falls behind the PowerVR GT7600 GPU in the A9 and even the older Adreno 430 in Snapdragon 810. Curiously, the Adreno 530 trails the 430 in both pixel- and vertex-heavy workloads. As we will see, however, these results are an outlier. Because of the Snapdragon 820 MDP’s pre-production status, this could simply be a quirk of an early revision graphics driver.

The 3DMark Physics test focuses on CPU performance and, like the AndEBench Memory Latency test, uses a data structure with randomized elements. This is why we see all of Qualcomm’s most recent SoCs, including the 808, 810, and 820, along with Apple’s A9, perform poorly. Their memory controllers are optimized for orderly, sequential memory access.

Snapdragon 820 fares better in Basemark X, with a 29% advantage over the Snapdragon 810 and a slim 7% advantage over the Exynos 7420. Most of the Adreno 530’s advantage comes from the Dunes test, which is heavily skewed towards vertex processing. Historically, Qualcomm’s Adreno GPUs have always been stronger with pixel operations, with vertex processing a relative weak point compared to other GPUs. It looks like the Adreno 530 is a more balanced design.

If you’re wondering why Apple’s A9 is not included here, it’s because Basemark X will not run on iOS 9 due to some API changes.

At the high quality setting, Snapdragon 820 flexes its memory bandwidth muscles and extends its performance lead to 57% over Snapdragon 810 and 30% over the Exynos 7420. Even in the onscreen tests, where the Snapdragon 820 is processing more pixels than any other device in these charts, the Adreno 530 performs well, which makes it suitable for the high-resolution QHD screens common in smartphones today.

We ran two different versions of the GFXBench Manhattan test. The older Manhattan 3.0 version runs an OpenGL ES 3.0 based game engine, with an abundance of lighting and pixel effects. Looking at the offscreen results we see that the Adreno 530 doubles the performance of both the Adreno 430 in Snapdragon 810 and the ARM Mali-T760MP8 in Exynos 7420. It’s even 20% faster than Apple’s A9.

Manhattan 3.0 uses deferred rendering for its lighting effects, which is highly dependent on screen resolution. This is why the iPhone 6s Plus (A9) and OnePlus 2 (Snapdragon 810) move up the chart when rendering onscreen: both have native 1920x1080 displays compared to the 2560x1600 display on the Snapdragon 820 MDP.

Manhattan 3.1 is an enhanced version of the Manhattan test utilizing OpenGL ES 3.1 features. Snapdragon 820’s performance advantage in this test is quite impressive; it’s 79% faster than the Snapdragon 810 and 2.4x faster than the Exynos 7420.

While the Manhattan tests emphasize pixel shading, the OpenGL ES 2.0 based T-Rex game simulation uses a more balanced rendering pipeline. Once again the 820 offers an impressive boost over the 810, this time around 77%. The 820 even beats Apple’s best offering by 14%, which held a stranglehold on T-Rex for some time.

The GFXBench Alpha Blending test stresses rasterization and memory bandwidth. Given the 820’s excellent memory performance, it’s no surprise to see it rise to the occasion, beating Apple’s A9 by 25%. The Fill test also stresses memory bandwidth, in addition to pixel processing, both of which play to the 820’s strengths.

Qualcomm has been improving ALU performance in its Adreno GPUs for several generations now, and the Adreno 530 is no exception. Unfortunately, we cannot say for sure if the 530 includes additional ALUs or even estimate total GFLOPS, because we cannot separate the individual contributions of the ALUs and memory bandwidth from their aggregate performance. All we can say is that Snapdragon 820 delivers 43% better ALU performance than the 810 and that a good chunk of this likely comes from the increase in memory bandwidth. ALU performance is critical to Qualcomm’s heterogeneous computing vision, and it looks like the Snapdragon 820 is well equipped to execute it.

Final Thoughts

The past year has been challenging for Qualcomm. Putting four Cortex-A57 CPUs onto TSMC's 20nm HKMG planar process was not a winning combination for its stopgap Snapdragon 810 product. It was plagued by overheating issues early, likely costing Qualcomm design wins, and while improvements came in a later revision, the 810 never lived up to its performance expectations. Fortunately for Qualcomm and its OEM partners, salvation appears to be at hand courtesy of Kryo and FinFET.

Snapdragon 820 brings tangible gains to all functional blocks: CPU, GPU, ISP, DSP, and modem. We’ve been waiting awhile for the first piece of this puzzle, but Kryo, Qualcomm’s first custom-designed 64-bit CPU, is finally here. It’s architecture is clearly unique: It simultaneously resembles both ARM’s Cortex-A57 and Apple’s Typhoon CPUs. Single-core integer performance is slightly better than A57/A72, while Kryo’s floating-point performance offers bigger gains over the A57. After factoring in the A72’s floating-point enhancements, Kryo should still see an IPC advantage, ensuring it will be competitive within the Android market. Apple’s Twister CPU in the A9 retains the single-core performance crown, however, beating Kryo in both integer and floating-point workloads.

Qualcomm’s Adreno GPUs have always had strong ALU performance, giving them the edge in games that make liberal use of pixel shaders. The Adreno 530 GPU in Snapdragon 820 continues this trend, improving ALU performance even further. The 530 is no one-trick pony, though. Improvements have been made throughout the rendering pipeline, especially to vertex processing, making the Adreno 530 a more balanced GPU.

Snapdragon 820’s memory controller is also vastly improved, offering the highest memory bandwidth we’ve seen for sequential access patterns. Like the Snapdragon 808 and 810 before it, the 820’s memory system is optimized for moving around large, contiguous chunks of data, perfect for feeding high-resolution textures to the GPU or streaming pictures and video to the enhanced ISP and DSP blocks.

This preliminary look at the Snapdragon 820 did not find any weak points in its performance. While thermal and battery life assessments will have to wait until we get shipping products into our lab, the use of Samsung’s proven 14nm FinFET process helps allay fears of a Snapdragon 810 redux. Architecturally, the 820 seems to be set up well for Qualcomm’s heterogeneous computing vision.

MORE: Best SmartphonesMORE: All Smartphone ContentMORE: All Tablet Content

Matt Humrick is a Staff Editor at Tom's Hardware, covering Smartphones and Tablets. Follow him on Twitter.

Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Qualcomm hosted its Snapdragon 820 event in New York City today to discuss the new features coming in its new SoC. Although some aspects of the hardware haven't been announced yet, the company did give an abundance of information of various new hardware features in the Snapdragon 820.

Meet Kryo

Qualcomm talked some about the hardware inside of the upcoming Snapdragon 820, but most of it remains clouded in mystery. The SoC will contain four CPUs based on Qualcomm's Kryo architecture, which uses the ARMv8 instruction sets. Qualcomm claimed that the CPU will have up to double the performance and be up to twice as energy efficient as the Snapdragon 810, which used ARM's Cortex-A57 CPU architecture. This is due, at least in part, to the use of Samsung's second generation 14nm FinFet LPP.

Although big.LITTLE architectures have been quite popular in recent years, Snapdragon opted to implement only a quad-core design. "People don't really need more than four cores," said Tim McDonough, Qualcomm's VP of Marketing. Another representative from Qualcomm went on to say that it is rare for a workload to use more than four threads, and there wasn't any point to include more than four cores.

The Adreno 530, DPU and VPU

The Snapdragon 820 will also contain Qualcomm's latest Adreno 530 GPU, which replaces the Adreno 430 as Qualcomm's flagship graphics processor. Qualcomm claimed that the new GPU is capable of roughly 40 percent higher performance while simultaneously pushing power consumption down by 40 percent compared to the Adreno 430. The Adreno 530 also brings with it a number of software API improvements, extending support to OpenGL 3.1+ AEP, full OpenCL 2.0, Vulkan and DirectX 11.2.

The DPU and VPU were also upgraded and are now capable of decoding 4K 10-bit video at 60 Hz over HDMI 2.0 or Miracast 2.0. 1080p video can be decoded at up to 240 Hz, too. At the same time, Qualcomm upgraded the hardware decode abilities of the SoC to support HEVC 10-bit and VP9.

With these improved video features, Qualcomm intends to push the Snapdragon 820 for use with VR devices coming in 2016, and the company implemented several software technologies in order to enhance a VR experience -- but that's another story for another article. Unfortunately, Qualcomm didn't have any Snapdragon 820 devices in use with VR hardware on display for us to analyze its potential.

Hexagon 680 DSP And The Spectra Camera ISP

The DSP inside of the SoC was updated to the Snapdragon Hexagon 680 DSP, which is used to power several of Qualcomm's new software technologies. Qualcomm claimed that the DSP is capable of up to three times greater performance and ten times the energy efficiency of its previous generation of DSP.

Qualcomm's latest evolution of its camera ISP has been renamed Spectra. The Snapdragon 820 contains two 14-bit Spectra ISPs, which are capable of recording 28 MP video at 30 Hz, with up to 1.2 GPix/sec.

Snapdragon X12 LTE Modem

Qualcomm's Snapdragon X12 LTE modem is capable of CAT12 uplink and CAT13 downlink. The modem also features support for the upcoming 802.11ad Wi-Fi standard, which transmits on the 60 GHz band. Over LTE, the modem is capable of up to 600 Mbps of bandwidth and can reach as high as 4.6 Gbps over Wi-Fi. Upload speeds over LTE have also increased significantly, up to 150 Mbps. Qualcomm claimed that power usage on Wi-FI has decreased by up to 45 percent, too.

The last new hardware technology present is Qualcomm's Quick Charge 3.0, which the company said is capable of charging a dead smartphone to full charge in under 35 minutes.

Qualcomm claimed that the effect of all of these power savings across the various aspects of the SoC adds up to a net drop of 30 percent in power consumption.

Qualcomm stated that these devices will start to show up on the market in the first half of 2016.

______________________________________________________________________

Michael Justin Allen Sexton (or MJ) is a Contributing Writer for Tom's Hardware. As a tech enthusiast, MJ enjoys studying and writing about all areas of tech, but specializes in the study of chipsets and microprocessors. In his personal life, MJ spends most of his time gaming, practicing martial arts, studying history, and tinkering with electronics.

Follow Michael Justin Allen Sexton @EmperorSunLao. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Now that the veil of secrecy surrounding the iPhone 6s and 6s Plus has lifted, some interesting details have emerged. One of the more tantalizing discoveries was made by Chipworks, when it revealed that Apple is dual-sourcing its A9 SoC (System on a Chip) from Samsung and TSMC.

Sourcing components from multiple vendors is a common practice among OEMs, especially for memory chips such as RAM and NAND. Apple routinely dual-sources the IPS screens for its iPhones, and Samsung frequently uses entirely different SoCs, camera sensors, and baseband processors, which differ in performance and features, in its smartphones depending on which region they're sold.

What's unusual in this case, however, is that Samsung is using its 14nm LPE (Low Power Early) FinFET process, which it also uses for the Exynos 7420 SoC, while TSMC is using its own 16nm FinFET process, resulting in two different versions of the A9 SoC with different die sizes.

It's generally undesirable to manufacture a processor using two different technologies because it adds to the engineering cost. So why would Apple choose this dual-sourcing strategy for its A9 SoC? In one word: supply. Apple derives up to 70 percent of its total revenue from iPhone sales, so any component shortages for its flagship product would seriously compromise its financial performance. Because sub-20nm FinFET is a new technology for both Samsung and TSMC, it's likely neither has the capacity or yields to satisfy Apple's demand. Having a second supplier also provides a safety net in case one of the chip makers encounters a production problem and cannot meet its quota.

With two different versions of the A9 floating around and no way to tell beforehand which one comes packed inside a new iPhone, it's understandable that some people may be concerned that one chip holds a performance or battery life advantage over the other. Because both versions share the same architecture and run at the same clock frequencies, there should be no difference in peak performance. Where they could differ, however, is the voltage required to meet those frequencies. A processor using a higher voltage consumes more power, thereby reducing battery life, and produces more heat, which could degrade performance if the processor needs to throttle back clock speed to keep from overheating.

Before we explore the performance and battery life of the different A9 versions, it's important to remember that even processors using the same process technology, or even cut from the same wafer, require different voltages to meet the same clock frequency target due to natural manufacturing variability. A few extra atoms here or a thinner deposition layer there can be the difference between an efficient, lower-voltage processor and a more power-hungry, higher-voltage one.

Rather than scrapping processors that do not meet power/performance targets, manufacturers sort them into different frequency bins and price them accordingly. For example, there are three different versions of Qualcomm's Snapdragon 801 with integrated LTE baseband with maximum frequency ratings between 2.26 GHz and 2.45 GHz. Because every A9 CPU core is required to run at the same max clock frequency, Apple's tolerance range for core voltage will necessarily be tighter.

Performance And Battery Life

This topic has already sparked discussion on several Internet forums, with users posting results from various benchmarks (primarily the Geekbench battery test) showing a definitive battery life advantage for the TSMC version, prompting Apple to issue the following statement:

With the Apple-designed A9 chip in your iPhone 6s or iPhone 6s Plus, you are getting the most advanced smartphone chip in the world. Every chip we ship meets Apple's highest standards for providing incredible performance and deliver [sic] great battery life, regardless of iPhone 6s capacity, color, or model.Certain manufactured lab tests which run the processors with a continuous heavy workload until the battery depletes are not representative of real-world usage, since they spend an unrealistic amount of time at the highest CPU performance state. It's a misleading way to measure real-world battery life. Our testing and customer data show the actual battery life of the iPhone 6s and iPhone 6s Plus, even taking into account variable component differences, vary within just 2-3% of each other.

While Apple's assertion that the benchmarks being used to compare the two A9 versions do not represent real-world usage is accurate, that's not really the problem here. The benchmarks themselves are perfectly valid tools for probing power efficiency. The real issue is how the tests were conducted. Were the screens calibrated to the same brightness level for each iPhone 6s? Did each phone have the same apps installed and use the same operating system settings? What background processes were running on each phone? We've seen firsthand how changing just a few settings can affect benchmark scores by up to 15 percent, and that's before any apps such as Facebook, Instagram, or Twitter are installed that can wreak havoc with CPU benchmarks.

For this reason, Tom's Hardware has developed a rigorous testing methodology for obtaining accurate, repeatable performance results. Using these procedures, we tested two iPhone 6s Plus phones, one with a Samsung-made A9 (which identifies internally as model N66AP) and one with the TSMC A9 (model N66mAP).

As expected, there's no discernible peak performance difference between the two different A9 models. All of the CPU, GPU, and system performance scores show less than a 2 percent difference, which lies within the margin of error for these tests. The only small outlier is the Basemark OS II Memory test, but this has more to do with RAM performance and how the operating system caches disk I/O then SoC speed.

The Basemark OS II battery score accounts for both battery life and performance under CPU-intensive workloads. It's a good stress test for determining CPU efficiency and seeing if the SoC can maintain peak performance levels without thermal throttling. In this case, we see Samsung's 14nm FinFET process has a slim but noticeable advantage, lasting 10.8 percent longer (16 minutes) than the iPhone 6s Plus with the TSMC made A9.

The GFXBench battery life test is the GPU equivalent of Basemark OS II, providing a worst-case battery life based primarily on GPU, memory, and display power consumption that is similar to what you might see while playing an intense 3D game. In this scenario, both the performance and the skin temperature on the back of each iPhone are nearly identical, indicating that both A9 SoCs generate a similar amount of heat and exhibit minimal thermal throttling. The Samsung-made A9 manages to last 3.5 percent longer when pushing the GPU hard, which equates to a meager five minutes and is barely outside the margin of error in this test.

Conclusion

Based on the results of our testing, it's clear that both versions of Apple's A9 SoC deliver the same level of performance, but Samsung's 14nm FinFET process appears to offer slightly better power efficiency, extending battery life between 3.5-10.8 percent. This is a little more than the 2-3 percent quoted by Apple, but not much, and it equates to only about 5-15 minutes of runtime under the most extreme conditions.

Real-world use cases other than intense gaming do not run the CPU and GPU at 100 percent for extended periods. Instead, the CPU and GPU run at much lower voltages and frequencies the majority of the time and only ramp up to their maximum clock speeds for short bursts of activity. Because an SoC's power versus frequency relationship is nonlinear (meaning each additional 100 MHz of frequency requires larger and larger increases in core voltage), Samsung's advantage during normal use will be less than what we measured in our tests and is likely to be very close to Apple's 2-3 percent figure.

Although we believe our results are accurate, they are derived from a sample size of one. We do not know Apple's allowable core voltage range for the A9 or where our particular samples fall within this tolerance band. Our tests also fail to capture the small variability in power consumption for other components like the RAM and display. Therefore, other iPhones may show more or less variance in runtime than our samples.

Even after taking this into account, the extra few minutes of battery life you may get from a phone using the Samsung-made A9 will hardly be noticeable.

If you want a new iPhone 6s, buy it, use it, and don't worry about who made the processor.

MORE: Best SmartphonesMORE: All Smartphone Content

Google announced not one, but two new Nexus smartphones today, which is a first in the company's history of making Nexus phones. The larger 5.7" Nexus 6P made by Huawei starts at $499, and the smaller 5.2" Nexus 5X made by LG starts at $379.

The Nexus 6P is Google's "flagship" for this year and is more of a successor to the Nexus 6 from last year, while the Nexus 5X is more of a successor to the Nexus 5 from two years ago, which was also made by LG.

Nexus 6P

The Nexus 6P brings a 64-bit Snapdragon 810 processor, a 5.7" 1440p AMOLED display, 3 GB of RAM, 32 GB of storage (with 64 GB and 128 GB options), 802.11ac, LTE, and a rather large 3,450 mAh battery.

The main attraction for the Nexus 6P this time seems to be the camera performance, which in the past hasn't been that good on a Nexus device. Google seems to have taken that criticism to heart and used a higher quality 12MP camera resolution that comes with a larger 1/2.3" Sony sensor. This means the pixels are currently among the biggest on the market, at 1.55µm.

Bigger pixels means more light can be captured, and Google said this even makes Optical Image Stabilization (OIS) unnecessary. OIS is useful for stabilizing the camera and can let more light in before the picture is taken. Even if the bigger pixels can compensate for capturing more light, the lack of OIS will still be felt when doing video recording, as OIS is also useful for stabilizing videos in real-time. Google should at least use digital image stabilization, but it remains to be seen how good it will be.

In a comparison between the Nexus 6P and the iPhone 6S, the Nexus 6P seems to have won in darker environments where the phone captured more light and the image had more natural colors.

The front-camera is 8MP, and Google's David Burke said that it's a great camera for selfies. Both of the rear cameras on the Nexus 5X and Nexus 6P can capture slow-motion video at 120fps and 240fps, respectively, and come with f/2.0 lenses and laser auto-focus.

The new Nexus 6P also comes with the latest generation Gorilla Glass 4, which is supposed to be twice as hard as Gorilla Glass 3, and therefore half as likely to get scratches. It's also one of the first smartphones with USB Type-C, the new reversible USB port, which also supports fast charging.

Nexus 5X

The Nexus 5X features a slightly weaker Snapdragon 808 processor, 2 GB of RAM, 16 GB as default storage (with only a 32 GB option), a 1080p 5.2" screen, USB type-C support, 802.11ac, LTE and a 2,700 mAh battery, which is an improvement over the Nexus 5's 2,300 mAh battery.

The camera seems to be the same one as on the Nexus 6P and can also shoot 4k video, but as already mentioned, slow-motion video recording is limited to 120fps.

Both of the new Nexus devices also come with fingerprint sensors called "Nexus imprint." Unlike other fingerprint sensors, the Nexus sensor supports the new fingerprint API, which opens fingerprint recognition for the entire app ecosystem. The fingerprint registration takes only two seconds, but once trained, the sensor can recognize a fingerprint in only 600ms.

Google also introduced a feature called the Android Sensor Hub, which moves the sensor data processing off the CPU to decrease power consumption and uses the data more intelligently as well, to recognize gestures and activities.

Both devices will also come with Android M, which brings new features including finer permissions control, the "Doze" feature that will double standby time and increase overall battery life by 30 percent, and Google Now on Tap, which can scan applications that are in use and offer intelligent recommendations (which could also open up many privacy issues, as Google will now be able to see messages and emails from other providers).

Pricing

The Nexus 6P and Nexus 5X are now available for pre-order on the Google store in the U.S., UK, Ireland and Japan. More countries will be supported in the next few weeks.

In the U.S., the Nexus 5X will start at $379 for the 16 GB version and $429 for the 32 GB version. The Nexus 5X will be available in White, Black and Mint colors. The Nexus 6P will cost $499 for the 32 GB version, $549 for the 64 GB model and $649 for the 128 GB variant. The Nexus 6P will arrive in Frost White, Aluminum and Graphite. The Nexus 5X should start shipping within two to three weeks, while the Nexus 6P should ship within four to five weeks.

In Canada, the 16 GB Nexus 5X will be $499 CAD and the 32 GB model will be $559 CAD. The 32 GB Nexus 6P will start at $699 CAD, and then will cost $749 CAD for the 64 GB version and $849 CAD for the 128 GB one. The Nexus 6P should also arrive within four to five weeks in Canada, but you can order it only in Aluminum and Graphite for now. For the Nexus 5X you can only join a waiting list. 

Follow us @tomshardware, on Facebook and on Google+.

Silent Circle announced the Blackphone 2. The new model comes with a 64-bit octa-core processor, the "Spaces" feature to create up to four isolated environments, and access to the Google Play Store.

Security And Updates

Silent Circle created the Blackphone in a time when Android phones weren't getting too many security updates, or the updates came too late to be effective for a phone that's supposed to protect the privacy and security of activists or enterprise users. Things have changed a little since then, with the discovery of the Stagefright vulnerabilities, but not significantly more.

The Blackphone's primary focus is strong security and privacy through secure communications applications, but also by having an OS that better protects against hacking and can be patched up as soon as a new vulnerability is discovered and before it can affect too many users. This makes it much better suited for activists and journalists, but also for enterprise customers where data security is critical.

Specs And Price

The Blackphone 2 comes with a premium flagship price tag of $799, but its processor doesn't seem to jibe much with that cost. The device uses an octa-core Snapdragon 615, which utilizes the lower-end Cortex-A53 CPU cores (although clocked at up to 1.7 GHz). This is a strange choice considering that this is a device that's supposed to deal with so much encryption and with multiple isolated domains at the chip level and thus should have as much performance as possible.

Other features include a 5.5" 1080p display protected by Gorilla Glass 3, 32 GB of storage, microSD support for another 128 GB, 3 GB of RAM, a 13MP camera on the back, and a 5MP camera on the front. The phone also has a 3,060 mAh battery and LTE support.

Blackphone 2 vs BlackBerry Priv

The Blackphone 2 may not be the only game in town anymore when it comes to easily accessible sub-$1,000 security-focused smartphones, as BlackBerry plans to introduce its own secure phone soon, called the BlackBerry Priv. Not much is officially known about it yet, but one way in which the Priv one-ups the Blackphone 2 is by using a hardened kernel by applying the well known and highly-effective Grsecurity suite of exploit mitigations.

Blackphone 2's isolated "Spaces" is a strong security feature in its own right as well, because it can keep personal stuff and work completely separated from each other, so if you get infected in one environment, the malware can't cross into the other environment. However, not having a well protected kernel can also lead to serious vulnerabilities and exploits, and each environment could get hacked more easily. This is why we have competition, though, and both companies can learn from each other as they build their next-generation devices and operating systems.

Blackphone 2 also continues to have better support for end-to-end encrypted calls and texts by default, although now that the BlackBerry Priv is based on Android, it could also get access to either the Silent Phone app from the Play Store or the open source TextSecure and RedPhone apps made by Open Whisper Systems.

The competition for security-focused devices seems to be heating up, and this should ultimately bring more security enhancements to the mainstream Android operating system, especially if companies such as Silent Circle and BlackBerry contribute back to the community and open source their code.

Silent Circle's Blackphone 2 can now be purchased in North America from the company's web store for $799.

Follow us @tomshardware, on Facebook and on Google+.

Qualcomm announced that its upcoming flagship SoC, the Snapdragon 820, will be the first chip to support 600 Mbps downlink Cat 12 LTE, 150 Mbps uplink Cat 13 LTE, and the brand new LTE-U technology.

LTE Cat 12/13

The Snapdragon 820 looks to be an almost complete overhaul. It will get brand-new CPU cores, an Adreno 500-series GPU, the Hexagon DSP, heterogeneous computing supporting, the Sigma dual-ISPs for much improved computational photography, support for ultrasonic fingerprint sensors, a cognitive computation platform, and last but not least, a range of new wireless technologies as well.

According to Qualcomm, the Snapdragon 820 will be the first to feature the most powerful and readily available Cat 12/13 LTE technologies that bring download speeds of up to 600 Mbps and upload speeds of up to 150 Mbps.

The new X12 modem can achieve these speeds by bonding together three LTE connections for downlink using LTE Advanced carrier aggregation technology, while for uplink it can use two simultaneous LET connections. The modem also uses higher-order modulation (256-QAM) on both the downlink and uplink to achieve those speeds when carrier aggregation isn't enough.

802.11ac MU-MIMO And 802.11ad

The Snapdragon 820 will support 802.11ac 2x2 MU-MIMO Wi-Fi, which is especially useful in crowded places from airports to offices, or even at home if you're using multiple devices at the same time. The MU-MIMO technology is what helps serve multiple devices at the same time with fewer connection interruptions.

The new chip will also support 802.11ad Wi-Fi, previously known as "WiGig" before it was acquired by the Wi-Fi alliance. The new technology works on the 60 GHz spectrum, which means much shorter ranges, but it can also deliver much higher transfer rates. Qualcomm's modem will support a bandwidth of up to 4.6 Gbps. This should be enough to handle multiple UHD video streams at the same time, lossless screen mirroring to the TV, or quickly backing up your camera roll to your network drive.

The LTE modem will have to use two different chips, one for 802.11ac (QCA6174) and one for 802.11ad (QCA9500), which could increase the cost of Qualcomm's LTE modem and chip bundle, but it's likely all OEMs will want it anyway.

Smart Wi-Fi Calling

The new modem in the Snapdragon 820 can also automatically identify when it's best to use an LTE or a Wi-Fi connection, depending on signal quality, end-to-end speed, and Internet reachability. All of this happens in the background without causing your call to drop.

The chip also supports HD voice over LTE as well as native HD video. Although just like with regular calls, when you choose to use the network directly instead of an app for the call, you're trusting the content of that communication with the carrier rather than the app provider, which often uses weaker encryption or has more relaxed data access policies.

LTE-U

LTE-U stands for "LTE unlicensed," because it's LTE that works on the unlicensed 5 GHz spectrum, now occupied by Wi-Fi routers. This has also caused much controversy, because eventually carriers could crowd out Wi-Fi routers with LTE-U "routers" or cells.

Right now, it's mainly other companies, including cable companies, that offer Wi-Fi services, and the carriers focus more on their LTE services. It would be in their interest to get more people to use LTE-based technologies rather than Wi-Fi technologies. The Wi-Fi Alliance has also expressed its worries to the FCC that LTE-U could interfere with Wi-Fi and suggested that the deployment of LTE-U should be delayed until proper spectrum sharing mechanisms are put in place.

Qualcomm said that its LTE-U modem will "fairly" share the space with Wi-Fi routers, but ultimately it remains to be seen how true that is.

The company also said that when in range of an LTE-U cell, the device will use one connection in the licensed LTE spectrum and one in the unlicensed LTE-U spectrum. Then it will bond the two connections to act as a faster connection.

The Qualcomm Snapdragon 820, along with all of these wireless technologies and many other new features, is expected to arrive in the first half of 2016.

Follow us @tomshardware, on Facebook and on Google+.

At its latest event today, Apple announced the chips that will be powering its most powerful iPhones and iPads yet, the A9 and the A9X (respectively). The two chips are built using new "transistor architecture," which is likely code for FinFET transistors.

Ever since Apple began to design its own chips, it has often ended up ahead of all the other mobile competition, whether it was Samsung, Qualcomm or Nvidia's chips, or even Intel's Atom line for mobile, in both CPU and GPU performance.

GPU performance is more of a result of using Imagination's latest PowerVR GPUs, but some of the gains are also due to Apple's design of an efficient SoC that allows for more powerful GPUs. With the arrival of the Metal graphics API, Apple also allowed developers to gain more direct access to those powerful chips and extract even more performance out of them for their games.

Apple claimed that its third-generation 64-bit A9 CPU is now 70 percent faster than its one-year-old A8 CPU and has a GPU that is 90 percent faster, as well. This time, Apple hasn't increased the resolution of its displays, so all of that power should go straight to new games and apps that can take advantage of it.

The new A9 SoC also comes with a new M9 coprocessor that now includes the ability to listen to the "Hey, Siri" command without the iPhone 6S having to be plugged in or for the owner to press a button first. Such functionality was first seen in the original 2013 Moto X when the "OK, Google" command could be used directly with the phone turned off.

The chip brings along a new LTE Advanced modem as well, that is twice as fast at 300 Mbps download speeds and can support 23 LTE bands. Apple claimed this is the highest number of LTE bands being supported in a smartphone right now, making it ideally suited for travelers.

Apple also introduced a more powerful variant of the A9, called the A9X, which will power the new iPad Pro. The company claimed that its CPU is 80 percent faster than the previous A8X CPU and that the GPU is also twice as fast. When comparing them to the chip in the original iPad from 2010, the company said the CPU performance has improved over 22 times -- and 360 times for the GPU.

Apple also claimed that the new A9X is faster than "80% of portable PCs," but it didn't elaborate on what exactly it means by that or how it reached that result. With A9X likely to be faster than Intel's Atom-based Celerons and Pentiums, which now go into lower-end notebooks, as well as most if not all existing ARM chips from competitors, the A9X can certainly be considered a "PC-class" chip. However, that doesn't necessarily mean it's by any means the fastest PC-class chip out there.  

For the purpose of the iPad Pro, which is meant to be a more productive yet familiar device for those already used to iOS, the chip should be powerful enough to handle just about any productivity app developers can create for it over the next year, until an even more powerful iPad Pro appears.

Follow us @tomshardware, on Facebook and on Google+.

BLU, an American smartphone company founded in 2009, announced its next-generation flagship device called the Pure XL, which aims to compete against other large-screen smartphones on the market for a significantly lower price point of only $350.

The device brings a Mediatek Helio X10 SoC (MT6795), which contains eight Cortex-A53 cores with clock speeds up to 2.2 GHz. The chip also supports video playback for the open source VP9 codec, as well as recording and playback support for HEVC.

The BLU Pure XL has a large 6" Super AMOLED screen that comes with a 2560x1440 resolution and Gorilla Glass 3 protection. It also comes with 3 GB of RAM and 64 GB of storage by default, making it one of the few devices with a baseline 64 GB model, and it even has a microSD, too, if for some reason you need more capacity for your 4k videos and apps.

Its camera promises to be quite impressive as well, with a 24MP resolution, 1/2.3" sensor, 6-layer custom lens, sapphire lens cover, real-time HDR and of course a dual-LED flash (which most of the new smartphones being launched these days seem to have).

The Pure XL's camera even comes with phase detection autofocus (PDAF) with an autofocus speed of between 0.02s-0.2s, essentially promising to match Sony's recently announced 23MP smartphone, the Xperia Z5.

It also has a dedicated hardware button for the camera, showing just how much focus BLU has put on the photography capabilities of its flagship. In the front, the device has an 8MP camera as well, with a wide-angle lens, optimized for taking better selfies.

The BLU Pure XL brings 24-bit/192 KHz audio support, dual-stereo speakers and DTS sound, becoming part of a trend to significantly improve audio quality in smartphones.

The device comes with a 3,500 mAh battery, which is a reasonable size for a 6" smartphone. It brings a 9V/2A quick charger as well, which can give you four hours of phone call talk time with a 10 minute charge.

The Pure XL also has a fingerprint sensor that can help unlock the device in less than 10 ms, or help you lock files on your phone more quickly.

The device will be available in grey and gold color options and will retail for only $350 on Amazon starting on September 29.

Follow us @tomshardware, on Facebook and on Google+.

Introduction

Asus has long been a powerhouse in the computer industry. The company produces peripherals and components such as monitors, sound cards, and graphics cards. Asus also makes complete systems, including the Republic of Gamers desktops and laptops, ultrabooks, and all-in-one PCs. Asus even introduced what many saw as the first real contender to the iPad: the Eee Pad Transformer. With a higher screen resolution, full keyboard, and full-size USB ports, the Eee Pad Transformer promised laptop functionality in a tablet convertible form factor. The company then went on to produce one of the most well loved tablets on the market at the time — Google’s Nexus 7.

Despite not being well known as a smartphone manufacturer, Asus has had a long history in the smartphone field. Starting in 2006, the company released various phones and pocket PCs running Microsoft’s Windows Mobile OS. In 2012 the company released the Padfone, its first Android phone. Before phablets became all the rage, Asus came up with a phone and tablet dock combination that allowed the phone to be inserted into the the back of the tablet to provide a larger screen. Asus also released the Fonepad, essentially a 7-inch tablet with a cellular data connection. Then in 2014, the company released the first generation ZenFone with three different models. The ZenFone 4, ZenFone 5, and Zenfone 6, with screen sizes corresponding to the product names, had screen densities below 300 PPI and two variations of a dual-core Intel Atom processor, placing these devices in the mid-range market.

Despite its various offerings, Asus has had difficulty breaking into the US smartphone market. The company managed to get its Padfone X and Padfone X mini picked up by AT&T but neither really became blockbusters. Instead of going the carrier route, Asus has focused on selling its devices at full price online and in electronics stores. This method works well overseas, however, the average US consumer is used to purchasing a phone from a carrier and signing a two year contract to subsidize the cost. But in recent years, carriers have started moving toward device payment plans, where the full price of the device is simply split up over a set period of time and tacked on to the monthly bill. And with a growing number of phones in the $200 to $300 price range, many customers have begun to realize that paying full price for a smartphone may not be such a bad idea if it saves them money in the long run and lets them upgrade faster.

This shift in the market could be advantageous for Asus’ latest smartphone, the ZenFone 2, whose 5.5-inch HD screen, quad-core Atom CPU, and a generous 4GB of RAM—the same amount of RAM found in most Chromebooks—make it a compelling mid-range option.

Technical Specifications

We're used to seeing smartphone CPUs based on the ARMv7 or ARMv8 RISC (reduced instruction set computing) architecture from companies such as Apple, MediaTek, Qualcomm, and Samsung, but the ZenFone 2 is different. It's powered by Intel's x86 CISC (complex instruction set computing) architecture, originally designed for PC applications with relatively large power envelopes. Instead of scaling down a power-hungry PC CPU like Intel, ARM took the opposite approach, optimizing its architecture for low-power applications from the beginning, an approach that has given ARM a near monopoly on smartphone CPUs. However, Intel has continued to improve its Atom CPU to where it should now provide similar performance and stamina to the ARM-based offerings.

There are two versions of the ZenFone 2 for North America, both sporting Intel Atom 64-bit processors. It will be interesting to see how the Atom CPU compares to its ARM-based competition in our performance and battery life tests. Graphics for both versions are provided by the PowerVR G6430 GPU. This is the same unit found in the iPhone 5s, which still offers good performance.

Asus manages to keep the overall size relatively manageable, packing everything into an elegant, curved-back case just slightly bigger than the LG G4 but smaller than the Galaxy Note 4. The weight, at 170 grams, is slightly lighter than the iPhone 6 Plus but heavier than the OnePlus One.

The ZenFone 2 features a 3000mAh non-removable battery, slightly larger than the iPhone 6 Plus’ battery and the same size as the LG G4. It also comes with Asus BoostMaster technology which provides a 60% charge in 39 minutes according to Asus.

The ZenFone 2’s rear camera uses a 13 MP Toshiba sensor paired with an f/2.0 aperture lens. The Nexus 6, Moto G (3rd gen), and OnePlus 2 all use 13 MP Sony sensors with f/2.0 aperture lenses too, so the ZenFone 2 looks competitive, at least on paper. The front camera is a 5 MP wide-angle affair for higher resolution selfies.

Options

There are two different model numbers for the ZenFone 2, each with several different SKUs. The ZE550ML will be sold in markets outside of the US, while the ZE551ML is the US model we are reviewing. It’s available with 16, 32, or 64GB of internal storage, with an additional 64GB available when using a microSD card. The 16GB SKU comes with a 1.8GHz Atom Z3560 SoC and 2GB RAM, while the 32GB and 64GB versions are equipped with the 2.3GHz Atom Z3580 SoC and 4GB RAM. The 18W BoostMaster adapter, necessary to enable the quick charging feature, only comes with the 2.3GHz versions. Color choices include Glacier Gray, Glamour Red, Ceramic White, Osmium Black, and Sheer Gold.

Being the first smartphone to ship with 4GB of RAM is impressive, but perhaps the ZenFone 2’s most disruptive feature is the price. At $200 for the 16GB version and $300 for the 64GB version, it’s almost a steal. These are prices one would expect to pay for the newest flagship phone on a two year contract. For comparison, a 64GB iPhone 6 Plus costs $849 and a 64GB Galaxy S6 starts at $700. It can be purchased online from several retailers, including Newegg, Groupon, Amazon, and B&H in the US, and Newegg, Canada Computers, Memory Express, and NCIX in Canada.

Cellular

There are a total of five different SKUs of the ZenFone 2, four supporting various FDD-LTE frequency bands and one for the Chinese and Indian markets. The US SKU of model ZE551ML supports the AT&T and T-Mobile networks, along with their respective MVNOs.

The ZenFone 2 uses Intel’s XMM 7262 modem, which combines an Intel X-GOLD 726 baseband with an Intel SMARTi 4.5 transceiver. Built on a 28nm process, it features support for GSM/EDGE, UMTS (WCDMA, TD-SCDMA), and LTE (LTE-FDD, LTE-TDD), but lacks support for CDMA2000. Intel’s modem can aggregate up to 40MHz of bandwidth, delivering Category 6 LTE performance (300 Mb/s down and 50 Mb/s up). WCDMA performance tops out at 5.76 Mb/s up and 42 Mb/s down.

The ZenFone 2 comes with a feature rarely found in North American handsets: dual SIM card slots. Slot one supports 2G, 3G, and 4G standards for voice and data while slot two supports 2G voice only. The slots are Dual Active, meaning you can have two different phone numbers active at the same time (slot two utilizes the Skyworks antenna switch for the additional GSM support), useful for using a single phone for both personal and business use or when traveling abroad.

MORE: All Smartphone ContentMORE: All Tablet Content

Hardware Design

The ZenFone 2 is well-made for the price. It’s an all-plastic device, but there is very little flex or creak. The back is plastic composite with a faux brushed aluminum appearance. Plastic masquerading as other materials can sometimes seem chintzy, but Asus does a good job here; this reviewer was initially fooled into thinking it actually is aluminum. The texture is a bit rough, making it more grippy than smooth glass or metal but not as grippy as the rubberized back of the Moto X.

The design of the rear panel is very linear, with all of the elements stacked in the center. The dual-color LED flash is located towards the top. Just below this is the rear camera, which is surrounded by a raised circular ridge to help protect the lens from getting scratched.

Similar to the LG G4, the elongated volume rocker, with a texture of circular ridges, is located on the back. This arrangement takes a little getting used to, but is less prone to accidental button presses. The volume rocker is fairly clicky yet easy to press, an important distinction considering the user will have to brace the sides of the device against their fingers and thumb in order to change the volume. The rocker switch is not as easy to locate as it is on the G4, however. The ZenFone’s switch is fairly narrow, and it’s difficult to tell if you’re touching the upper or lower half without any tactile indicator at the center.

The Asus logo stands out against the brushed red finish of our review unit, while the Intel and Zenfone logos fade into the background a bit, a nice touch considering all the logos and branding emblazoned on the backs of past devices. Near the bottom edge sits the speaker grille, stretching horizontally nearly the full width of the phone. The length of the grille suggests the use of multiple speakers, however, removal of the back cover reveals only one small mono speaker on the left side. This is a little disappointing considering more recent phones like the Nexus 6, Moto X, and HTC One M9 all have stereo front-facing speakers.

Removing the back cover is a little tricky: A small indentation for inserting a fingernail or tool is situated on the lower-left side but the plastic clips that hold the cover in place are difficult to release. Underneath the cover, the microSD card slot and the two micro-SIM card slots are situated along the centerline. The 3000mAh battery can be spotted beneath the frame that holds the cards in place, however, it is non-removable. Mounted on the inside of the back cover is the black rectangle of the NFC coil.

The power button sits in the middle of the top edge, which makes it rather difficult to reach considering the size of the device, requiring some awkward finger shifting to activate one-handed. Flanking the power button are the 3.5mm headphone jack and noise canceling microphone. The microUSB 2.0 port is centrally located on the bottom edge with a microphone offset to one side. Moving the buttons to the top and back leaves the sides smooth and unmarred.

Flipping the phone over reveals a black expanse dominated by the 5.5-inch HD LCD display covered with Gorilla Glass 3. With a body to screen ratio of 72%, the bezels are thin but not quite as thin as the LG G4. Surrounding the perimeter of the front face is a slightly raised plastic trim, creating a sharp edge that feels uncomfortable when held against the ear and face during phone calls. This seemingly small detail is the one thing about the ZenFone 2’s plastic design that does feel cheap.

Centered above the screen is the earpiece and an Asus logo. To the right is the front-facing 5 MP camera and an RGB LED for notifications. To the left, just barely visible, are the proximity and ambient light sensors.

Below the screen, from left to right, are the standard Android navigation buttons for back, home, and recent apps. Asus decided to go with old-school capacitive buttons as opposed to the more modern onscreen virtual buttons. Unfortunately, the buttons are not backlit—only lined with reflective material. This becomes an issue when trying to operate the phone in a dark environment. You simply have to guess where the buttons are or become familiar enough with their location and layout that you know where to put your fingers. Beneath the sometimes-difficult-to-see row of buttons is a small plastic chin with the same circular-ridged texture as the volume rocker.

Except for the sharp edge around the front, the ZenFone 2 is pretty comfortable to hold. The back curves smoothly against the palm, reminiscent of the Nexus 6, and its all-plastic construction makes it relatively light for its size. The flat sides make gripping the phone a little easier and the rounded corners do not dig into the palm or fingers. However, the brushed finish is a bit on the slippery side so shifting one's grip can be a tricky proposition. The center of balance rests just below the Asus logo on the back so most of the time it does not feel like it’s going to flip out of your hand. As with many devices in phablet territory, it feels awkward to stretch your thumb across the screen or up near the top. Either movement requires shifting your fingers so they no longer press the device securely into your palm and shimmying your thumb into position. Thankfully, Asus includes a one-handed mode similar to Samsung’s implementation on the Note 4, but generally, the ZenFone 2 is best operated with both hands.

Display And Audio

The ZenFone 2 comes with a 5.5-inch IPS LCD display with an HD 1920x1080 resolution (401 PPI). Sharing these specs with the iPhone 6 Plus—which costs several hundred dollars more—is pretty impressive. The 5.5-inch mid-range phones are still transitioning to full HD panels and many of them are still only 720p.

Another nice feature is the four different display modes that provide some control over the panel’s output: Balance, Bluelight Filter, Vivid, and Customized. Each mode also has a slider for adjusting the display’s color temperature. The Customized mode grants additional control over hue and saturation, but it does not provide direct control over the individual RGB channels.

After testing the three primary modes at their default settings, we selected Balance mode and adjusted the color temperature slider, eventually settling on the position shown in the screenshot. The display results for this setting are labeled “Custom” in the charts and graphs below.

We’re using SpectraCal's CalMAN software and SpectraCal C6 colorimeter for display measurements. All of the charts below with a gray background were generated in CalMAN v5 Ultimate.

Display brightness is not one of the ZenFone 2’s strengths. By default, Asus caps the backlight’s maximum brightness, limiting the display to just over 300 nits. We actually received two ZenFone 2s for testing, and the second unit did slightly better, reaching a maximum of 333 nits and making outdoor visibility in sunlight passable. Installing a third-party app to control brightness unlocks the backlight’s full potential of 370 nits. Our second review unit fared even better, reaching 410 nits. While the out-of-box brightness is nothing special, using a thrid-party app boosts brightness into the acceptable range.

The ZenFone 2 appears to be using dynamic contrast to improve the black level and reduce the display’s power usage. Instead of leaving the backlight at full power when displaying a black or generally dark scene, the backlight is dimmed. This became evident during testing: After calibrating the screen to 200 nits, showing an all black screen followed by an all white screen reduced brightness to 130 nits, which gradually increased back to 200 nits in about three seconds. While this technique does save power, it can lead to screen flickering if brightness changes too quickly, in a fast-paced action movie for example. Asus tries to avoid this issue by ramping up brightness gradually, sacrificing dynamic range during the three second transition. Even though the change in brightness is noticeable when switching from an all black to an all white screen, we did not notice any ill effects when watching video.

The use of dynamic contrast gives the ZenFone 2 a low black level (the second unit measures 0.14), comparable to the iPhone 6 Plus, and a respectable contrast ratio.

Gamma is one measurement where the ZenFone 2 does not fall flat, but we do not mean that in a good way. The gamma curve increases linearly with luminance, peaking between 2.7 and 3.0 depending on which mode is selected. This behavior is undesirable because it results in a darker image overall with a loss of shadow detail and subdued highlights. Unfortunately, none of the manual controls provided affect the gamma curve, so there’s no way to improve performance here.

The ZenFone 2’s Balance and Vivid modes are a bit on the cool side, giving a slight blue tint to an all-white background. The Bluelight Filter mode is aptly named, filtering out blue light and producing a warmer screen with a noticeable green tint. Adjusting the color temperature slider gives our Custom configuration a color temperature very close to the ideal value. In all modes, the color temperature remains constant across the luminance scale, which we like to see.

The Balance and Vivid modes exhibit the same RGB balance across the full range of grayscale values, both increasingly exaggerating blue as they approach 100% white, resulting in the cooler color temperatures we saw in the previous chart. The Bluelight Filter mode does the opposite, cutting the blue channel by up to 9% and boosting green by the same amount. Our custom setting shows a much better balance between green and blue and only a small deficit in red, which should lead to better grayscale accuracy.

Both the Balance and Vivid modes show noticeable grayscale error above 50% luminance because of the increasing imbalance between blue and red. The Bluelight Filter mode fares a bit worse with a green tint that’s noticeable even at 30% luminance, gradually getting worse up to 100% white. Our “calibrated” custom configuration does very well. Grayscale error mostly remains below two, only creeping up to three at 100% white.

Since color gamut is set by the LCD backlight, we do not see any variation here between the different modes. The ZenFone 2 gets close to displaying the full sRGB color gamut, missing some shades of red and green and extending blue tones just beyond the sRGB triangle.

In Balance mode, the ZenFone 2 performs pretty well in the color saturation sweep, showing no color compression, but we do see the secondary colors cyan and magenta skew towards blue as a result of the cooler color temperature. The Bluelight Filter mode also steers clear of any color compression, but here the secondary colors are shifted slightly towards green.

Up to now, Vivid mode performed similar to Balance mode, but the saturation sweep finally reveals its purpose. There’s significant color compression above 60% saturation (the spacing between measurement points is less than 20%), resulting in colors that appear overly saturated or “vivid.” The downside to this approach is color banding, where smooth color gradients are replaced by successive constant color zones separated by sharp delineations.

The ZenFone 2 struggles a bit with color accuracy, with average ΔE2000 for the three factory modes above five and peak error above ten for the color red. For more natural looking colors, we like to see color error remain below three, but only a few tested colors meet this criteria.

Our custom setting definitely improves overall color error, with average ΔE2000 dropping to four. Almost all of the tested colors remain at or below five, and several colors measure at or below three.

Full-Size Images: [Color Palette: Balance], [Color Palette: Bluelight Filter], [Color Palette: Vivid], [Color Palette: Custom]

The color palettes above show the target color on the bottom versus the displayed color on the top and are a nice way of visualizing the color error discussed above. All of the tested modes show obvious deviations from the target colors. This is partly due to the backlight’s inability to cover the whole sRGB color gamut and partly due to the gamma curve exceeding the ideal value of 2.2. The higher gamma value makes the displayed colors appear darker, an issue that affects even our custom configuration.

Looking at the grayscale values, a slight blue tint is evident in the Balance and Vivid modes, while the Bluelight Filter mode shows a very obvious green tint.

The ZenFone 2’s display turns out to be pretty decent, at least if you’re willing to make a couple tweaks. The color temperature for each of the factory modes is a bit off, giving the screen either a blue or green tint. Thankfully, Asus provides a slider for easily adjusting the color temperature, improving the RGB balance and resulting in excellent grayscale values.

Brightness is one weak point, mostly because Asus caps max brightness at around 300 nits. Installing a 3rd-party app to control brightness helps, boosting max brightness to about 400 nits—acceptable but not stellar.

The display’s biggest issue is gamma. The gamma curve is neither flat nor close to the ideal value. This degrades color accuracy (colors look darker than they should) and washes away shadow detail (dark areas end up looking like black holes). This is not a deal breaker—the ZenFone 2’s screen still looks better than those in many mid-range phones—but the screen does appear a bit darker or more subdued than other displays.

Like most IPS screens, viewing angles are excellent, and reflections do not appear to be any worse than other glossy screens.

Audio Performance

The ZenFone 2 uses the ALC5647 audio codec from Realtek, a name instantly familiar to any PC enthusiast, which includes an on-chip headphone driver. The external speaker is powered by a Texas Instruments TPA2080D1 Class-D power amplifier that provides up to 2.2 W using a 3.6V supply into 4 Ω at 1% THD+N. It also includes an integrated Class-G boost converter that provides a 5.75V supply voltage to the Class-D amplifier when high output power is required.

The single speaker is located on the back in the lower-left corner. Putting the speaker on the back, where it directs sound away from your ears, usually compromises sound quality; however, Asus makes this arrangement work pretty well. The curved back panel leaves plenty of room between the speaker and a table, effectively using the flat surface to reflect sound back towards the listener. We also did not have any issues with our hand blocking output when holding the phone in landscape mode, its naturally cupped shape forming a suitable sound reflector instead. The speaker gets loud enough to easily hear it from across the room, but it does not get as loud as the Nexus 6 with its dual front-facing speakers.

Without any enhancement, drums sound tight but cymbals are faint, and the speaker sounds a bit dull. The equalizer does not help much, but Asus includes a separate AudioWizard app with six different sound modes: Smart, Music, Movie, Recording, Gaming, and Speech. The Gaming mode works best with the external speaker, improving the sound of the midrange (especially vocals) and treble, but adding a touch more distortion. Still, the improvement is quite dramatic and makes listening to music with the speaker reasonably pleasant.

Headphone sound quality is even better. Actually, it sounds excellent, with great clarity and only lacking a bit of punch at the low end. The enhanced bass output from the AudioWizard’s Gaming mode might work for hip-hop, but I prefer the Music mode, which boosts output overall and expands the soundstage. It effectively fills in the low end but creates a hint of sibilance at the high end.

Camera Features

Camera performance has never been a strong point for Asus devices, and that trend does not change with the ZenFone 2. The rear camera is a 13 MP PixelMaster branded shooter with a dual-color LED flash. Unlike more expensive flagship devices, there’s no optical image stabilization (OIS); however, Asus does include software based image stabilization that attempts to accomplish the same thing. There is also no phase detection autofocus (PDAF), a feature relatively new to smartphones that allows for much faster focusing. Instead, the ZenFone 2’s camera uses the more common, and less reliable, contrast detection autofocus.

The rear camera uses a Toshiba T4K37 CMOS sensor with backside illumination (BSI) and an optical format of 1/3.07. It also shares the same 1.12μm pixel size most other smartphone cameras use and shoots natively with a 4:3 aspect ratio. Most recent offerings from Apple, LG, Motorola, and Samsung use Sony sensors, making this Toshiba sensor a bit of an unknown quantity in terms of performance. The only other phone we’ve looked at that uses a Toshiba sensor is the HTC One M9, which uses a higher resolution version.

The five element Largan lens array has an f/2.0 aperture which captures 8% more light than the Moto G (3rd gen) and about the same amount of light as the iPhone 6 Plus and Nexus 6. However, it captures 25% less than the Galaxy Note 4 and 41% less than the f/1.8 aperture lens in the more expensive LG G4. The ZenFone 2 uses the same 28mm focal length (35mm equivalent) as the Galaxy S6 which is wider than many of its competitors, increasing the camera’s field of view and capturing a greater portion of the scene.

The front-facing camera uses an OmniVision OV5670 5 MP sensor and an f/2.0 aperture wide-angle lens that covers 85°. On paper this is better than what most other competitors offer, beating out the 1.2 MP of the iPhone 6 Plus and 2 MP of the Nexus 6, and inline with current generation phones like the Galaxy S6 and OnePlus 2.

Camera Software

The Auto mode camera UI is streamlined and clutter free. In landscape mode, controls for flash, front/rear camera selection, and settings line the left side, while buttons for the camera shutter, video capture, and shooting modes line the right side. The current shooting mode is displayed at the top-left and a small thumbnail at the top-right provides access to the photo gallery. Touching the gear icon in the lower-left corner opens the camera and video settings menu, which you can see in the animated image below.

While in Auto mode, Asus’ camera app will suggest which of the many shooting modes available it thinks is best for the situation. Point the camera at a dim corner of a room and a caption pops up at the bottom of the screen, “Touch to enable Low Light mode,” along with an arrow pointing at a small icon denoting the aforementioned shooting mode. Point the camera at a scene with plenty of shadows mixed with sunlight and the caption will read, “Touch to enable HDR mode.” This is a useful feature for the casual user who may not be aware of what shooting mode would produce the best picture.

Surprisingly enough, running the Manual Camera Compatibility app from the Play Store confirms that the ZenFone 2 does not support Android Lollipop’s Camera2 API. However, Asus does include a manual shooting mode that replicates some of the features found in the new API: controls for white balance, exposure value, ISO, shutter speed, and focus. There is no support for RAW image capture though. The Manual camera mode interface and all of its options are shown below.

Time Rewind Mode

Asus’ camera software provides a plethora of options for capturing images, including a whopping 18 different shooting modes. Some are genuinely useful and others seem somewhat superfluous. One of the more useful ones is Time Rewind. This mode shoots a continuous burst of pictures before and after the shutter is pressed, letting you capture those moments when your finger is simply too slow or too fast with the shutter button. Pictures are captured at a much lower 1920x1080 resolution as opposed to the native 4096x3072 (13 MP). According to Asus, Time Rewind takes a burst of 31 pictures and starts recording images two seconds before as well as one second after pressing the shutter button. We found that our review unit, running camera software version 2.0.0.150527_11, actually captures a burst of 41 pictures and records images three seconds before as well as one second after the shutter activation. After the burst is taken, the user is able to manually select the shot to keep or press the “best” button to select the picture the software deems best. This is incredibly useful when trying to capture fast moving pets or squirmy toddlers.

Smart Remove Mode

Smart Remove is another useful feature. This mode shoots a series of five pictures back to back, analyzes them for any moving objects, then stitches them together to create a picture with said objects removed. Smart Remove is quite useful when taking pictures in busy public areas, parks, attractions, etc. It works well for the most part but it is not perfect. Close inspection of the final picture often reveals details that are off around the areas where objects have been removed, for example a shadow that is suddenly cut off or a corner of playground equipment that seems to disappear into thin air. However, these pictures should work fine for sharing over social media or email.

Asus includes a few other noteworthy modes. The Super Resolution shooting mode combines multiple images to a produce a single high-resolution 51 MP (8192x6144) image. While the quality will not match a native 51 MP camera, this mode does produce photos with noticeably more detail than the standard resolution. The Low Light mode uses pixel binning (combining four adjacent pixels) together with a lower shutter speed and ISO to capture a brighter image with less noise. Since the number of effective pixels is reduced by 1/4, resolution is limited to either 2 MP or 3 MP. Night mode is similar to Low Light mode in that it also lowers the shutter speed and attempts to keep the ISO low; however, it does not use pixel binning so it can capture images at up to 13 MP. Asus’ Depth of Field works similarly to Samsung’s Selective Focus mode, taking multiple pictures at different focus points and allowing you to use a simple slider to select the desired focus depth after taking the image. This can be useful for creating a blurred background bokeh effect.

Front-Facing Camera

The front camera defaults to Beautification mode when launched (it can be turned off by selecting the Auto mode). This mode applies a variety of effects to detected faces including blush, skin softening, skin brightening, eye enhancement, and cheek thinning. All effects show a live preview except for blush. The image above shows the subset of shooting modes available to the front-facing camera. Resolution is adjustable from 2 MP (2048x1152) with 16:9 aspect ratio up to the full 5 MP (2560x1920) with 4:3 aspect ratio.

Video

Intel has surprisingly little information on the video encoding and decoding abilities of the Atom Z3580. Since the CPUs remain at their idle frequency during video playback, it appears the ZenFone 2 uses fixed-function hardware (instead of software decoding which uses the CPUs) for H.264 video decoding, saving power.

The ZenFone 2 records 1080p video at 15 Mb/s, slightly lower than the 17 Mb/s of the Galaxy Note 4, Galaxy S6, and Nexus 6 and significantly lower than the 20 Mb/s of the OnePlus One and LG G3. At 720p the ZenFone 2 records at 8 Mb/s, lower than the 12 Mb/s of both the Note 4 and Nexus 6. The settings menu offers a choice between “Performance” or “Quality” video, although this has no effect on the values listed in the table below, and its impact on visual quality is negligible.

Rear Camera Video Modes

Front Camera Video Modes

Overall, video taken with the ZenFone 2 is decent but not outstanding. Indoor video from both the front and rear cameras shows a good deal of noise in anything less than ideal lighting conditions, with the front camera being more prone to this issue. Outdoor video in natural lighting fairs much better. White balance and exposure are good enough, although the front camera tends toward washing out lighter colored areas. Video also suffers from an overall lack of sharpness. Clips shot at 1080p tend to be slightly fuzzy when looking at fine details and clips shot at 720p almost look as if they’ve been upscaled from a lower resolution. Anything shot at 480p is not worth mentioning. Curiously, the front camera seems to produce better detail at 1080p than the rear camera.

There is a software-based video stabilization option available, however, it only works at 720p and 480p resolutions. The fact that it does not work at 1080p would normally count as a negative, but the video stabilization option seems to make video quality worse not better. Any kind of panning produces a kind of high-frequency wobbling effect around objects while consistently dropping frames. It appears the performance of the ZenFone 2’s ISP is holding back its video capabilities; it struggles to keep up when applying video stabilization and there’s no 1080p@60fps option even though the Toshiba rear camera supports it.

Like most mid-range and even some flagship phones, the ZenFone 2 lacks continuous autofocus when shooting video. The camera will focus initially, but remains locked after pressing the record button for the duration of the video, leading to objects that drift in and out of focus depending on movement in the scene. It’s up to the user to constantly tap the screen to refocus.

HDR video mode is also lacking and is quite noticeable outdoors. Focusing on a mix of shade and sunlit areas causes the bright areas to be overexposed. Moving back and forth between shade and sunlit areas causes the recording to brighten and darken dramatically as the camera struggles to find the correct exposure.

The ZenFone 2 presents a simple, decent, but not outstanding video recording experience. It’s lack of continuous autofocus and HDR video mode hurts video quality. You also do not get 4K or 60fps video modes like you do with more expensive phones, and slow motion video is also MIA.

Camera Performance And Photo Quality

Asus’s website states that its PixelMaster branded camera “combines software, hardware and optical design to deliver incredible image quality.” It’s time to put that to the test by comparing the ZenFone 2 to some of the best phablet phones available, including the iPhone 6 Plus, LG G3 (the G4 was not released yet when these images were taken), and Galaxy Note 4. Of these phones only the G3 matches the ZenFone 2’s 13 MP camera resolution, while the iPhone 6 Plus and Note 4 use 8 MP and 16 MP sensors, respectively. We’ll also compare it to the HTC One M9, which uses the 20 MP version of the Toshiba sensor found in the ZenFone 2.

All images were taken using the Auto mode unless noted. Also, you can view the full-sized image for each photo by clicking the text links below the images that are within a slideshow album. The ZenFone 2, LG G3, and iPhone 6 Plus all shoot natively at 4:3 aspect ratio while the Galaxy Note 4 shoots at 16:9 and the HTC One M9 uses an unusual 10:7 aspect ratio.

Outdoors

Daylight

Full-Size Images: [ZenFone 2: daylight sign], [iPhone 6 Plus: daylight sign], [LG G3: daylight sign], [HTC One M9: daylight sign], [Note 4: daylight sign], [ZenFone 2: daylight flowers], [iPhone 6 Plus: daylight flowers], [LG G3: daylight flowers], [HTC One M9: daylight flowers], [Note 4: daylight flowers]

In the first set of images, the natural light is bright enough that all phones are able to select low ISOs and fast shutter speeds. However, the iPhone 6 Plus seems to struggle here with getting the correct exposure and ultimately chooses a lower ISO that, despite the larger pixel size, results in the most underexposed image of the bunch. [Editor’s note: The discoloration in the top-left corner of the iPhone image is my finger getting in the way.] At the opposite end of the scale, the One M9 chooses a higher ISO and ends up with an overexposed image. Flowers at the top of the bush are so washed out that very little detail is visible. The same occurs with the cloud in the upper-right corner: The other cameras capture some shadow and texture, but the M9 shows an amorphous white blob. The overall best exposures belong to the G3 and Note 4. The ZenFone 2, being a little underexposed, falls somewhere in the middle.

The ZenFone 2 ends up with the coolest color balance of the bunch, with the sky in the top-right corner tinged with a light shade of green. Color balance for the iPhone 6 Plus is unnaturally warm. The flowers are a yellowish color, while the rest of the cameras present them as a more white or light pink hue. Even the pavement of the parking lot appears a more yellowish hue compared to the rest. The One M9 does a decent job with color balance but the longer exposure washes out the colors. The Note 4 and G3 have the best overall color balance, although the G3 favors slightly cooler colors too.

As expected, the ZenFone 2 captures a fair amount of detail, more than the resolution limited iPhone but less than the sharp looking image from the Note 4. Looking at the sky, it appears that Asus has opted for little to no software-based noise reduction. This is in stark contrast to the M9, whose overly aggressive noise reduction causes the edges of the white sign to blend into the surrounding sky.

All of the cameras were once again able to select low ISOs and quick shutter speeds for the flowers in the second group of images. But this time around the M9 produces the darkest image with an almost sepia tone look, and its heavy noise reduction blurs away fine detail. The ZenFone 2, however, overexposes the image, with the sun’s bright highlights washing away color in some of the flower petals. Color balance is again slightly cool with the purple flowers at the top appearing more blue than purple. Overall the best shot probably belongs to the Note 4, which captures warmth and good color balance without sacrificing too much to the harsh sunlight.

Evening

Full-Size Images: [ZenFone 2: dusk planters], [iPhone 6 Plus: dusk planters], [LG G3: dusk planters], [HTC One M9: dusk planters], [Note 4: dusk planters], [ZenFone 2: dusk neon], [iPhone 6 Plus: dusk neon], [LG G3: dusk neon], [HTC One M9: dusk neon], [Note 4: dusk neon], [ZenFone 2: night neon], [iPhone 6 Plus: night neon], [LG G3: night neon], [HTC One M9: night neon], [Note 4: night neon]

As the sun begins to set the cameras with OIS start to come into their own. The iPhone 6 Plus, G3, and Note 4 make use of this feature to keep their shutters open longer while keeping ISO in the same range as the previous set shot in full daylight. The One M9 and ZenFone 2 lack OIS and need to drive ISO above 100 to achieve similar exposures, leading to noisier images than the rest.

In the first set of images shot just before sunset, the M9 produces the brightest image. However, the planters display a slight greenish tinge as well as noise artifacts from the software noise reduction. Also, the bricks lack some of the warmth and detail present in the rest of the shots. The ZenFone 2 ends up at the other end of the spectrum, producing the darkest image with a color balance that’s too cool, missing the yellowish-orange light of the setting sun. Details in the leaves of the front planter are a bit muddy and dark. The ZenFone 2 also has trouble getting the flowers of the rearmost planter in focus; they appear as fuzzy blobs with little detail and the flowers are completely out of focus. To be fair, however, the One M9 and iPhone 6 Plus also struggle to produce detail in the rear planter area due to aggressive noise reduction for the M9 and lower resolution for the iPhone. It is only the Note 4 and G3 that produce nicely detailed images.

The second set of images showing the building with neon lights were also taken just before sunset. Once again the devices with OIS keep their shutters open a bit longer but this time ISOs start increasing as daylight fades. The ZenFone 2’s white balance skews towards blue once again—the aqua neon lights appear blue and the yellow on the walls appears pale and muted. It produces a noisy image, but the lack of noise reduction actually keeps details looking sharp. Both the ZenFone 2 and the M9, which also has a similar Toshiba camera sensor, struggle to capture enough light, overexposing the sky while the building remains relatively dark. The M9 image has better white balance but softer edges due to noise reduction. The Note 4 captures a nice looking, bright image but loses the blue color of the sky, a detail best captured by the iPhone 6 Plus.

As night descends all cameras start to struggle: noise is present across all images, neon lights appear overexposed due to dynamic range issues, and color balance is noticeably different from shot to shot. The M9 struggles the most, producing the noisiest, blurriest image. Colors are also oversaturated, especially the blue sky and reflections on the windows. The ZenFone 2 produces the darkest image with the same cooler, muted colors seen previously on the building. However, it avoids the horrible noise artifacts and oversaturated colors that plague the M9, producing a better image overall.

The quality of the ZenFone 2’s image cannot match the other flagship phones, however. Noise and brightness are comparable to the iPhone 6 Plus’ image, but the iPhone image’s colors are more saturated and natural. The G3 uses its long exposure mode to capture the brightest image, although the bright neon lights get washed out. Taking advantage of OIS, the Note 4 holds its shutter open longer and also captures a brighter image than the ZenFone 2 with much better colors.

HDR

Full-Size Images: [ZenFone 2: park - no HDR], [ZenFone 2: park - HDR], [iPhone 6 Plus: park - no HDR], [iPhone 6 Plus: park - HDR], [LG G3: park - no HDR], [LG G3: park - HDR], [HTC One M9: park - no HDR], [HTC One M9: park - HDR], [Note 4: park - no HDR], [Note 4: park - HDR], [ZenFone 2: building - no HDR], [ZenFone 2: building - HDR], [iPhone 6 Plus: building - no HDR], [iPhone 6 Plus: building - HDR], [LG G3: building - no HDR], [LG G3: building - HDR], [HTC One M9: building - no HDR], [HTC One M9: building - HDR], [Note 4: building - no HDR], [Note 4: building - HDR]

HDR mode helps improve the dynamic range of photos and has become a standard feature across flagship phones. Most devices combine images taken with at least two different exposures in post-processing to achieve the HDR effect. Unlike Samsung’s phones, the ZenFone 2 does not provide a live preview of the final HDR image, so you will not be able to tell how the photo will look until after it’s taken. Capturing an HDR image takes about two seconds, significantly longer than current flagship devices.

Shooting a mix of sunlit and shaded areas, like the first scene with the trees above, presents a challenge for any camera. Here the ZenFone 2’s HDR mode does a good job brightening the shaded areas while preserving the brightly lit sky and clouds, which were horribly overexposed with HDR turned off. However, the skew toward cooler colors is still present and the image lacks contrast overall, making the colors look a bit washed out. By comparison, the M9’s HDR mode performs worse: It boosts the darker areas a bit but leaves the sky overexposed. On the other hand, the iPhone 6 Plus does well with the sky but does nothing to improve the shadows, and there’s some purple fringing visible in the sunlight reflecting off the cars. Samsung’s HDR mode is the best we’ve seen. The Note 4 image looks nice, although there is a hint of purple fringing in the palm trees and car reflections.

In the second set of images with the building, the ZenFone 2’s HDR mode does a decent job again, but the image shows a fair amount of noise. The tendency toward cooler colors and a lack of contrast is noticeable again in the HDR image, where the warmth of the sunlight is missing and the leaves and grass are undersaturated. There is also a ghosting effect visible around some leaves in the upper-left corner, an artifact caused by the leaves shifting position between the different exposure shots due to a light breeze. The other cameras do not exhibit this problem, probably because the image capture time between the different exposure shots is less.

Once again, the ZenFone 2 does better than the M9, whose heavy noise reduction wipes away detail (especially in the background) and creates artifacts around the edges of the leaves. The ZenFone 2 also outperforms the iPhone 6 Plus, whose HDR mode seems to have little effect. The G3 and Note 4 perform the best, however, effectively lightening the shadows while keeping the sky from getting overexposed, all with no obvious artifacts.

Indoors

The staged indoor shots below were lit by overhead LED lights, a CFL lamp from the front, and an incandescent overhead light in the background.

Bright Light

Full-Size Images: [ZenFone 2: indoor bright], [iPhone 6 Plus: indoor bright], [LG G3: indoor bright], [HTC One (M9): indoor bright], [Note 4: indoor bright]

This indoor scene is dim enough to cause most of the cameras to push ISO to 200 or above in order to achieve proper light sensitivity. The exception is the iPhone 6 Plus, which keeps noise low by holding ISO below 100. The ZenFone 2 and iPhone 6 Plus both produce images with similar exposure and color balance. However, the ZenFone 2 shows significant amounts of noise throughout the image, softening some details and making some objects appear almost out of focus. Colors look muted or undersaturated as well. The G3, One M9, and Note 4 all produce brighter images, but the G3 pushes exposure too far, producing a bright glare on the figure’s faces and heads. The Note 4 image suffers from a slight green tint.

Low Light and Flash

Full-Size Images: [ZenFone 2: indoor dark], [ZenFone 2: indoor dark low-light mode], [iPhone 6 Plus: indoor dark], [LG G3: indoor dark], [HTC One M9: indoor dark], [Note 4: indoor dark], [ZenFone 2: indoor flash], [iPhone 6 Plus: indoor flash], [LG G3: indoor flash], [HTC One M9: indoor flash], [Note 4: indoor flash]

With nothing but the incandescent light in the background lighting the scene, all of the cameras start to struggle again. Noise is present across all images and some cameras have issues focusing properly. The One M9 produces the noisiest image of the bunch; the figurine on the far left has so much noise it looks pixelated. The iPhone 6 Plus and Note 4 both produce images with decent exposure, detail, and less noise. However, the Note 4 skews toward an unnaturally warm color balance and consistently shows a green tint. The G3 image is the brightest but exhibits an ugly orange-red tinge. The amount of noise present also causes some details to be considerably softened or lost completely. The ZenFone 2 selects a shutter speed three times faster than most of the other cameras, creating the darkest image of the group. Its noise level is nearly on par with the image from the iPhone 6 Plus, no small feat considering the iPhone uses an ISO of 250 while the ZenFone 2 pushes ISO all the way to 1270 to compensate for the faster shutter speed and lack of OIS.

The ZenFone 2 also features a special low-light shooting mode that, according to Asus, uses pixel binning (combining the output of four adjacent pixels into a single pixel) and image-processing algorithms to enhance light sensitivity by up to 400%. The downside is that this mode is only able to shoot at a 3 MP (4:3) or 2 MP (16:9) resolution. The second image in the photo album above shows the result of using the 2 MP (the default selection) low-light shooting mode. It brightens the image considerably, achieving an exposure similar to the iPhone 6 Plus. However, with such a low resolution, the image is only suitable for viewing on mobile devices or the web.

Turning the flash on allows all of the cameras to capture decent images, although some look better than others. The iPhone 6 Plus and Note 4 both do quite well lighting up the scene. But where the flash from the iPhone comes across rather harsh, the Note 4’s flash lights the scene more evenly and produces more neutral colors. The One M9 does a decent job at capturing detail, although the image is slightly darker than the Note 4’s and its white balance is a little cool. At opposite ends of the spectrum are the G3, which overexposes and starts to wash out some details, and the ZenFone 2, which underexposes and casts a yellowish tinge across everything.

Front-Facing Camera

Full-Size Images: [ZenFone 2: front camera indoors 1], [ZenFone 2: front camera indoors 2], [ZenFone 2: front camera panorama]

These images were taken indoors under LED lights using Auto mode. The front-facing camera does alright in the first picture, capturing a good amount of detail, although noise creeps into the darker areas. In the second picture, noise is more apparent due to the higher ISO, and the lights are completely washed out because of the longer exposure time. The last picture uses Selfie Panorama mode, which works by taking a series of shots as the user slowly pans the camera. It’s a tricky maneuver as evidenced by the mismatched edges where the individual shots do not exactly line up.

Camera Performance

The ZenFone 2’s cameras perform as expected for a mid-range device. The rear camera’s white balance is consistently too cool (blue), and it tends to overexpose images (when not using HDR) in its attempt to capture a bright photo. Curiously, noise levels in bright light are nearly the same as low light, but post-capture noise reduction is minimal, preserving detail and avoiding ugly artifacts. In low-light conditions, the lack of OIS and some difficulty setting the correct shutter speed lead to darker images. The Toshiba sensor seems to have trouble capturing color info in lower-light conditions too, producing images with undersaturated colors.

Rear camera focus latency hovers between one to two seconds depending on lighting conditions, which is about average for contrast detection autofocus. Shot-to-shot image capture latency varies between one to three seconds, with lower-light scenes taking longer.

While not perfect, the ZenFone 2’s camera still performs better than the one in the HTC One M9, a phone that costs significantly more money. The ZenFone 2’s HDR mode also works pretty well, although the final images could use more contrast.

Software

The ZenFone 2 comes with Asus’ ZenUI running on top of Android 5.0. It’s a little heavy handed with both the design and preloaded software. It does help that performance is smooth and quick, the design generally follows Google’s Material Design guidelines, and some of the preloaded software can either be uninstalled or disabled. ZenUI also includes some handy features too such as launcher customizations, ZenMotion, one-handed mode, and various settings for both the casual and power user.

Google’s Material Design follows some general guidelines: icons are flat, colorful, and not skeuomorphic; generous use of white space to draw the eye toward certain elements; layers that create a sense of depth; and the use of animation to make the environment interactive. ZenUI follows these guidelines for the most part, however, Asus adds its own spin. By default the icons for system apps as well as app folders are all rounded squares, reminiscent of Apple’s iOS or Xiaomi’s MIUI. Pulling down the notification shade shows the usual details and pulling down again reveals quick settings. But where Google’s implementation shows a few simple options and a brightness slider, ZenUI shows a 4x4 grid of circular icons plus a brightness slider. There are enough options to bewilder the casual user; thankfully, the quick settings can be customized to the user’s liking.

The app drawer offers plenty of customizability. By default, app icons are arranged in a 4x4 grid, although there are several other options available ranging from 3x3 to 5x5. The default view mode is “Customized” which sorts all of the preloaded apps in alphabetical order first followed by downloaded apps ordered by date of installation. The “All” mode simply shows all of the installed apps in alphabetical order, while the “Downloaded” and “Frequent” modes provide options for showing only a subset of apps. You also have the ability to hide apps and group them into folders to reduce clutter. It also gives you the option to manually arrange the apps by dragging them around. Overall, the ZenUI app drawer is pretty flexible.

Asus’ custom launcher packs a plethora of options. Long pressing or swiping up on the home screen opens the “Manage Home” page which serves as a launchpad for several customization options. The included Themes app (shown above) provides an easy way for quickly changing the overall appearance of the UI, and digging through the options under “Preferences” reveals more settings for scrollable wallpaper, background transparency, folder grid size and style, icon label color, etc. These are the kind of in-depth UI tweaks usually found only in custom launchers, so it’s nice to see that Asus has included them here. ZenUI even adds a few extra buttons to the app switcher, including a shortcut to the app manager, screen pinning, and a very useful button to close all open apps.

The ZenFone 2 comes with several preinstalled apps and utilities that seem inspired by Asus’ PC heritage. For starters, in the display section of the Settings menu, Asus allows you to change the screen color mode between three different presets. There’s also a slider to manually adjust the screen color temperature and a “Customized” setting that also allows you to adjust hue and saturation. The basic file manager app grants access to local files and connects to several cloud storage providers, but it does not allow access to files below /sdcard. Auto-start Manager controls which apps are allowed to run in the background, potentially reducing RAM usage and extending battery life. And, in keeping with the PC theme, Asus includes the Clean Master and Dr. Safety apps, general utilities for managing RAM usage, cached files, and antivirus security.

Asus includes a few other options and features catering to both the casual and power user. Easy mode strips away the standard interface and replaces it with a simple 3x3 grid of app icons and enlarged text for the less tech savvy. Kid mode allows parents to restrict access to a group of select apps, block incoming calls, and set a time limit for device usage. For people looking to get work done, Asus includes a handy task manager that works with several of its other apps. After clearing items from the to-do list, you can plug in some headphones (which function as an antenna) and use the ZenFone 2 as an FM radio.

The ZenFone 2 comes with Asus’ ZenMotion, a suite of controls consisting of Touch Gesture, Motion Gesture, and a one-handed mode that streamline interactions with the large phone. Touch Gesture provides a variety of app shortcuts activated by drawing letters (see screenshot below) on the screen while in sleep mode. By default these shortcuts are assigned to system apps, however, they can be reassigned to any app the user wishes. This reviewer found it very useful to be able to draw the letter ‘Z’ and have the flashlight app quickly activate. Touch Gesture also includes double tap to wake and double tap to sleep the screen, immeasurably useful considering the awkward top-center placement of the power button. For the most part Touch Gesture works well provided you draw the letters large enough. However, occasionally the feature proved unresponsive; no amount of tapping or drawing produced any result. Only by hitting the power button to wake the screen then turning it off again would Touch Gesture be reactivated. It could be that during long periods of inactivity the CPU goes into a deep sleep mode where this feature gets deactivated.

Motion Gesture consists of a single option called “Shake Shake,” which takes a screenshot and adds it as a new task in the Do It Later task manager app. The shaking gesture is tricky to pull off; shaking too little or too much does nothing. The amount of shaking required is often enough to cause an app to switch between portrait and landscape modes, leading to possible screenshots in the wrong orientation. This feature also did not feel very responsive; there was often a gap of a second or two between the time the gesture was recognized and the time the screenshot was taken.

The ZenFone 2’s one-handed mode is quite useful for a phablet-sized device that sits between the LG G4 and Galaxy Note 4 in overall case size. This feature can be activated from the Quick Settings menu, but it’s far easier to enable using the quick trigger option which allows it to be activated or deactivated by a simple double tap of the home button. As mentioned previously, Asus’ implementation is similar to Samsung’s. Once activated, the screen shrinks and docks to one side of the display. A button at the top left restores the screen to its original size, while holding and dragging the top-right corner allows the screen to be resized as needed. The window can also be repositioned anywhere on the display. Virtual buttons for back, home, and recent apps appear at the bottom of the shrunken window. This feature works well for everything except games which still launch fullscreen. Exiting the game reverts back to the shrunken one-handed window as expected.

ZenUI performs well and looks decent. Asus adds its own visual flair and quite a few preloaded apps which might turn some people off, but many of these tweaks and apps are actually quite useful. Touch Gesture and the one-handed mode are welcome additions, especially for a phablet. It’s also nice to see so many options for customizing the UI, something missing from many OEM ROMs.

CPU And System Performance

In this section, we evaluate system-level performance by running a series of synthetic and real-world workloads, along with some browser-based Web tests. There are several facets to overall device performance, including single- and multi-threaded CPU performance, memory and storage speed, and GPU rendering, all of which will be probed by our suite of benchmarks. If you're interested in learning more about how these benchmarks work, what versions we use, or our testing methodology, please read our article about how we test mobile device system performance.

The ZenFone 2 we’re testing comes with an Intel Atom Z3580 SoC and 4GB RAM. The 64-bit, quad-core CPU idles at 500MHz and ramps up to a maximum of 2.33GHz. Because this is the first time we’ve tested the Atom in a phone, we’re going to compare it to several different SoCs (shown in the table above) to determine where it falls in terms of performance.

Looking first at the System test, which evaluates single- and multi-threaded integer and floating-point performance, the ZenFone 2 scores 1.72x better than the Moto G (3rd gen). With the same core count as the Snapdragon 410, the 1.71x difference in CPU clock frequency implies that an Atom core performs about the same as an A53 core clock for clock. The ZenFone 2’s Atom SoC scores 34% lower than the A8 in the iPhone 6 Plus and 43% lower than Samsung’s Exynos 7420, its eight CPU cores adding to its advantage.

The ZenFone 2’s PowerVR G6430 GPU is bracketed by the Adreno 405 and the Adreno 418 in the Graphics test. It’s no surprise to see the iPhone 6 Plus perform better either, since it’s using a newer, more powerful configuration of the PowerVR GPU.

In the AndEBench Pro suite, the ZenFone 2 scores only 10% lower overall than the LG G4 and 24% lower than the Galaxy S6 (currently the fastest Android phone), giving the ZenFone 2 a superior performance-per-dollar ratio.

Digging a bit deeper shows the Atom SoC in the ZenFone 2 outperforming the Snapdragon 410 in the Moto G (3rd gen) by ~2.5x in the CPU-centric CoreMark-Pro test. This is an even larger margin than we saw in Basemark OS II System, with performance scaling greater than the clock frequency difference. The ZenFone 2 even outperforms the G4 in this test by 15%. This is because the G4 only utilizes its two A57 cores (the A53 cores remain mostly idle) versus the ZenFone 2’s four CPU cores.

The Atom SoC uses LPDDR3 RAM at 800MHz, limiting maximum bandwidth to 12.8 GB/s. It’s no surprise then to see the ZenFone 2 perform similarly to the Moto G (3rd gen) and Xperia M4 Aqua in the Memory Bandwidth test. The LPDDR4-1600MHz RAM in the Galaxy S6 far outpaces it, but so does the LPDDR3-933MHz RAM (14.9 GB/s) in the G4, far more than we’d expect based on theoretical bandwidth alone. We’ve seen similar behavior from the Snapdragon 810, and we believe this is because Qualcomm’s latest memory controllers are optimized for serial access. Further evidence for this hypothesis comes from the Memory Latency test, which uses a random access memory pattern, where the G4 falls to the bottom of the chart. With its stronger performance in the Memory Latency test, the ZenFone 2’s memory subsystem should perform well during normal application use but struggle a bit when reading/writing large blocks of serial data, like photos and videos.

The Platform test, which simulates a real-world workload, shows that the ZenFone 2 does indeed perform quite well in a normal use case. It even outperforms the Galaxy S6!

The ZenFone 2’s internal storage performance is typical for a mid-range phone. Read performance is pretty decent, about on par with the Galaxy Note 4. This is important because read performance impacts the user experience more than write performance; it limits how quickly apps can launch. From these performance numbers, we can see that the ZenFone 2 launches apps nearly as quickly as the more expensive Note 4.

Looking at the Geekbench single-core results confirms what we saw in the Basemark OS II System test. Namely, a Silvermont CPU core is roughly equivalent to an A53 core at the same clock frequency. The ZenFone 2, which has a 1.71x advantage in CPU clock frequency, outperforms the A53 core in the Moto G (3rd gen) by a margin of 1.64x in the integer test and 1.93x in the floating point test, working out to 1.73x overall.

Having twice as many cores clocked at a higher frequency gives the ZenFone 2 similar performance to the iPhone 6 Plus in the multi-core test.

Theoretical performance is good for bragging rights, but how a phone performs when running real-world scenarios is what truly matters. We already saw the ZenFone 2 claim victory in the AndEBench Pro Platform test, and now we see it rise to the top once again in PCMark, beating out the Galaxy S6 by a slim 9% margin.

The Galaxy S6 falls behind mainly due to its low Video Playback score, a test the ZenFone 2 has no trouble with. The Asus phone also does well in the Web Browsing test, outpacing the G4 but trailing the S6 by 16%.

Where the ZenFone 2 really shines is the Photo Editing test, performing 52% better than the second place S6. Most of the image processing during this test is supposed to occur on the GPU using the android.media.effect API. Monitoring the CPU and GPU core frequencies shows this is true for the ZenFone 2. Both the S6 and G4, however, appear to be using their CPUs instead, resulting in lower performance. It’s possible that only the ZenFone 2’s PowerVR graphics driver supports this feature.

We use a static version of the Opera browser for Android testing to keep results comparable and avoid benchmark cheating. Unfortunately, the older version we’re using clearly does not like the x86 architecture. We’ve already seen that the Atom SoC in the ZenFone 2 performs competitively, so the abnormally low Opera browser results should be ignored.

With the Opera results compromised, we reran the tests using the stock browser app (JSBench did not work in this browser). These results show that the ZenFone 2 is more than capable of running JavaScript code. Note, however, that these results are not comparable to the other phones using Opera, since the stock browser app is newer and has a faster JavaScript engine.

So, what conclusions can we draw from our testing? The synthetic tests show that Intel’s Silvermont CPU core is roughly equivalent to an ARM A53 core at the same clock frequency, with Silvermont holding a slight advantage in floating point performance. Running its CPU cores at a higher frequency gives the Atom Z3580 a performance advantage over A53 based SoCs such as the Snapdragon 410 and 615 as well as most MediaTek SoCs, however. While it’s a strong contender in the mid-range market, Atom still cannot match the theoretical performance of Apple’s A8 or A57 based SoCs like the Exynos 7420 at the high-end of the market.

Ultimately, theoretical CPU performance is only part of the user experience equation. Running more realistic, mixed workloads that stress the entire system is when the ZenFone 2 rises to the occasion. Its combination of decent CPU performance, reasonably quick NAND reads, and a memory controller optimized for random memory accesses makes it feel like you’re using a flagship device that costs several hundred dollars more.

GPU And Gaming Performance

Mobile GPU performance is becoming increasingly important as people begin to see their phones and tablets as portable gaming machines. This section explores GPU performance with several synthetic and real-world game engine tests. To learn more about how these benchmarks work, what versions we use, or our testing methodology, please read our article about how we test mobile device GPU performance.

The PowerVR GX6450 GPU in the iPhone 6 Plus is the direct descendent of the G6430 GPU used in the ZenFone 2 (also used in the previous generation iPhone 5s). These GPUs have the same ALU performance, processing 256 FP32 FLOPs/clock. The ZenFone 2 runs its GPU at either 457MHz or 533MHz under load, and while we do not know the exact peak GPU frequency for the iPhone, it does lie within this range. This means we should see similar performance in pixel shader bound workloads. Where the GX6450 gains an advantage, is in front of and behind the unified shader cores, improving vertex processing and pixel fill rate, respectively.

In 3DMark, the ZenFone 2 keeps pace with the flagship phones. The iPhone 6 Plus leads the pack in the first graphics test that focuses on vertex processing, clearly benefitting from its front-end architectural tweaks. The ZenFone 2 is 19% slower but only 12% slower than the speedy Galaxy S6.

Qualcomm’s Adreno GPUs perform well in ALU-bound workloads, so it’s no surprise to see the Adreno 418 in the LG G4 post the highest frame rate in the second graphics test that focuses on pixel processing. The ZenFone 2 is 18% slower than the G4 and 15% slower than the iPhone 6 Plus.

The 3DMark Physics test focuses on CPU performance and uses a data structure that requires random memory accesses. As we touched on in the AndEBench memory latency test, the memory controllers in the iPhone’s A8 SoC and G4’s Snapdragon 808 SoC are optimized for serial data access patterns, which penalizes them heavily in this test despite their fast CPUs. Intel’s Atom SoC embraces random memory patterns, propelling the ZenFone 2 to the front of the pack, beating even the S6. This is one of the reasons it performs so well with real-world apps.

Basemark X is a very demanding benchmark, heavily stressing the entire graphics pipeline, especially vertex processing. At the medium quality setting, the ZenFone 2 is only 21% slower than the LG G4, but it cannot keep up with the other flagship devices, coming in 40% slower than the iPhone 6 Plus. It does manage to outperform the G4 in the onscreen test, where the latter device gets bogged down by its QHD resolution.

The order does not change at the high quality setting, although the gap between the ZenFone 2 and the flagship phones closes a bit; it’s only 11% behind the G4 and 25% behind the iPhone 6 Plus. Fed by bountiful memory bandwidth, the Mali-T760MP8 in the Galaxy S6 is a vertex processing machine, propelling it to the top of the chart.

The OpenGL ES 3.0 based Manhattan test makes heavy use of pixel shaders and other post-processing effects. As we’ve seen in the previous tests, the ZenFone 2 has no trouble outperforming the Adreno 306 and 405 found in the Moto G (3rd gen) and Xperia M4 Aqua. Although it’s not shown in our chart, the ZenFone 2 also performs better than Snapdragon 801 powered devices such as the Galaxy S5 and OnePlus One using the Adreno 330 GPU. For reference, the ZenFone 2 is 30% slower than the iPhone 6 Plus when running Manhattan.

T-Rex results mirror those from the Manhattan test, including the ~30% deficit to the iPhone 6 Plus.

We mentioned previously that the PowerVR G6430 GPU in the ZenFone 2 and the GX6450 GPU in the iPhone 6 Plus should have the same ALU performance, an observation confirmed by the GFXBench 3 ALU test. It also comes very close to matching the ALU performance of the Galaxy S6. Currently, the Adreno GPUs have the advantage in FLOPs/clock, which is why the LG G4 tops the chart in this test.

The Mali-T760MP8 GPU in the Galaxy S6, along with the two PowerVR GPUs, can all process up to eight texels per clock. However, the ZenFone 2 falls behind the iPhone 6 Plus in the Fill test by 14%. It’s likely that the iPhone’s GX6450 has larger texture caches as one of its back-end improvements. In addition to some architectural differences, the Mali GPU in the S6 runs at 700MHz versus the ZenFone 2’s 533MHz, increasing its advantage.

With its last generation GPU, the ZenFone 2 falls behind the latest flagship devices. Compared to the iPhone 6 Plus, the Asus phone is roughly 30% slower (15% to 40% based on our tests). However, it does perform better than other phones in the same price range, in some cases considerably better. Even more impressive, the less expensive version of the ZenFone 2 that comes with the Atom Z3560 SoC uses the same PowerVR G6430 GPU tested here, giving you gaming performance similar to the iPhone 5s for only $200.

Throughout this section, we kept comparing the ZenFone 2 to more expensive, flagship devices. This may seem unfair, but it’s actually a compliment. The midrange ZenFone 2 just performs more like a low-end flagship than the lower-cost phone that it is.

Battery Life And Thermal Throttling

Battery life may be the most important performance metric for a mobile device. After all, it does not matter how quickly a phone or tablet can load webpages or how many frames per second the GPU can crank through once the battery runs down and the device shuts off. To learn more about how we test this critical facet of mobile computing, please read our battery testing methodology article.

Combining a relatively large battery with a power sipping processor gives the Moto G (3rd gen) a huge advantage in battery life. The ZenFone 2 and the other phones all form a close group around the six-hour mark. If you actually want to use the phone to get work done or pass the time, however, performance is important too. By multiplying the PCMark battery life by the overall performance score, we get a composite metric that gives us an estimate for how much total work can be done on a single charge.

The Moto G’s long-lasting battery gives it the lead; you can get a lot done with the Moto G before you need to charge it again. The ZenFone 2’s excellent performance and good battery life make it a close second, beating the more expensive S6 and G4.

The score in this test accounts for both battery life and performance under CPU-intensive workloads. Both the Moto G (3rd gen) and ZenFone 2 rise to the top once again just like they did in our PCMark composite work score. This time, however, the ZenFone 2 holds a slight lead.

The GFXBench 3.0 battery test focuses on the GPU and is an indicator of battery life during intense gaming. It also effectively gauges a device’s ability to dissipate heat.

Once again, the Moto G outlasts the other phones by a significant margin, but only because its scaled-back GPU delivers less than half the performance of the other devices. The ZenFone 2 strikes a good balance between performance and longevity, both lasting longer and performing better than the Galaxy S6.

About six minutes into the GFXBench battery test, the ZenFone 2 starts throttling its GPU frequency due to the increase in temperature. GPU performance declines to 75% of the peak level by the 14 minute mark and remains there for the duration of the 60 minute test.

The skin temperature on the back of the ZenFone 2 peaks at 117 °F, which is in the same range as other high-performing devices. Looking at the thermal image, we can see a buildup of heat in the upper half of the phone. The inability to readily conduct and dissipate heat is one of the drawbacks of its all-plastic construction. The iPhone 6 Plus’ all-aluminum body functions better as a radiator, allowing its more powerful GPU to keep running at peak frequency without throttling.

The combination of a 3000mAh battery and the Atom Z3580 SoC prove to be a good combination for the ZenFone 2, striking a good balance between battery life and performance. This is most evident in PCMark’s real-world work scenarios, where it pulls ahead of several higher-priced devices. While not shown in our charts, the Galaxy Note 4 lasts longer on a charge, but the ZenFone’s higher performance gives it the edge in our composite work score and delivers a better user experience while using it — exhibiting smoother scrolling and fewer pauses overall.

Conclusion

Designing a phone for the middle of the market is tough. A flagship device gets all the bells and whistles, while a low-end device gets the bare minimum to hit its target price. In the middle, however, the OEM needs to find just the right balance between design, features, performance, and price. Fall short in just one area and the phone is a flop. For the ZenFone 2, Asus finds this balance. Performing better and costing less than its peers offsets the few minor issues we found with ergonomics and the display.

Despite its low cost and all-plastic construction, the ZenFone 2 does not feel or look cheap. The faux brushed metallic finish looks convincing and the body feels rigid and sturdy. The only rough edge, literally, is around the perimeter of the front face, which feels scratchy when held against the ear.

We do have some minor quibbles about ergonomics. Asus places the volume controls on the back and the power button on the top edge. While this gives it a streamlined look and reduces inadvertent button presses, the power button is very difficult to reach considering how tall the phone is. This issue is mostly mitigated, however, by the ability to wake or sleep the device by double tapping the screen, a feature we love and wish every phone had. The rear-mounted volume rocker is placed within easy reach, although we wish it were a bit wider and had some sort of detent in the middle to make it easier to locate. Another minor issue is that the capacitive buttons are only coated with a reflective material and are not backlit, causing us to fumble for the correct button in the dark on more than one occasion.

The ZenFone 2’s sharp looking 5.5-inch IPS LCD display sports a 1080p resolution and 401 PPI pixel density, a definite plus considering many large-screened phones in this price range only have a 720p resolution. Asus also provides a few different display modes and some manual controls for adjusting the screen’s appearance. While the factory modes are not stellar in terms of grayscale or color accuracy, a simple adjustment of the color slider improves the display accuracy significantly.

While the ZenFone 2’s display is good, there are two things holding it back from being great. The first is maximum brightness. By default, Asus caps brightness to just over 300 nits, which makes outdoor viewing more difficult. Using a third-party app to control display brightness unlocks the backlight’s full potential, boosting the display to ~400 nits, an acceptable value. The second, more serious, issue is the display’s gamma curve. Linearly increasing with luminance, the high gamma gives the screen a dark cast, washing away shadow detail and hurting color accuracy.

The 13 MP camera lacks advanced features like OIS and PDAF, which is common for phones in this price range, but it is accompanied by a dual-color LED flash. The ZenFone 2 also lacks support for Lollipop’s Camera2 API. Despite this, Asus’ streamlined camera app does have a decent manual mode and offers smart suggestions for which of the several different shooting modes to use, including a good HDR mode that not only controls dynamic range well, but also avoids the most common post-processing artifacts.

Image quality in good lighting is decent, but it tends to set the white balance a bit too cool. It also tends to overexpose images when not using HDR mode. The Toshiba rear camera sensor seems to struggle with low-light sensitivity, leading to darker images with visible noise. Asus does not appear to use any noise reduction algorithms, which makes the noise more visible, but also avoids the artifacts and loss of details that result from its overuse. The rear camera also has trouble capturing color info in lower-light conditions, making images appear a bit muted and undersaturated.

Like most mid-range phones, shooting video with the ZenFone 2 is a no-frills affair. Video quality at 1080p is ok, but the lack of continuous autofocus and an HDR video mode hurts the experience. You also do not get access to any advanced video modes, including 4K, 60fps, or slow motion.

One thing about the ZenFone 2 that did surprise us is its excellent audio quality. While it does not have front-facing or even stereo speakers, its lone rear-facing speaker actually sounds pretty good. The ZenFone 2’s headphone audio quality even rivals some of the best flagships we’ve heard.

Asus’ ZenUI was another nice surprise. Sure, Asus includes quite a few of its own apps, not all of which are useful, but it’s easy enough to sweep them out of the way. Where ZenUI shines is in its flexibility. It includes a good theme engine and the launcher includes a depth of customization usually reserved for third-party ROMs. Asus’ ZenMotion suite includes useful features like the ability to launch apps by drawing letters on the screen and a nice one-handed mode too.

At the beginning of this (novel length) review, we questioned whether the ZenFone 2’s Atom SoC could compete with its ARM-based competition in terms of performance and power given Intel’s past struggles in mobile. In terms of peak theoretical performance, Intel still has a ways to go to match the best offerings from ARM and Apple. However, when it comes to more realistic workloads, Atom does fairly well. Along with decent NAND read speeds and a memory controller that performs well with random memory accesses, the Atom-powered ZenFone 2 performs as well as, or in some cases better than, more expensive flagship phones. Pairing the Atom Z3580 SoC with a 3000mAh battery gives the ZenFone 2 pretty good battery life as well. It appears that Atom can compete in the power consumption arena too.

The ZenFone 2 turns out to be a surprisingly well-rounded device considering it only costs around $300 or less unlocked. Most OEMs offering phones in this price range make serious concessions that compromise the user experience to reduce cost. Asus avoids this pitfall, creating a phone that meets (camera) or exceeds (performance and features) our expectations for a mid-range device. This is why the Asus ZenFone 2 is Editor Recommended.

MORE: All Smartphone Content
MORE: All Tablet Content

Even though the Galaxy S6 is Samsung's latest flagship smartphone, Samsung has just launched another Galaxy S4-branded device called the Galaxy S4 Mini Plus, two years after the original Galaxy S4 Mini launched.

The Galaxy S4 Mini Plus seems to be some kind of successor to the two year old Galaxy S4 Mini, yet Samsung didn't believe it deserved a completely new name, so instead it added a "Plus" at the end.

Samsung seems to want to keep separate mid-range lines of devices that look different from each other. There's the S4 Mini line-up, then there's the S5 Mini, and there will likely be an S6 Mini, too.

However, all of this could end up confusing consumers, as they won't know what to expect from all of these "Mini" devices other than the fact that they look like their bigger brothers.

From its specs and its ~200 euro price, we expect this Galaxy S4 Mini Plus to compete against the new Moto G 2015. However, some of its specs don't seem nearly as strong.

Samsung's phone comes with an ARMv8 quad-core 1.2 GHz Cortex-A53 CPU, an Adreno 306 GPU, 1.5 GB of RAM, 8 GB of storage, LTE, Bluetooth 4.0, an 8MP rear camera with 1080p video recording, 1.9MP front-camera, and a 1,900 mAh battery.

The display is quite small by today's standards, even for people who may want smaller screens. It only has a 4.3" display with a qHD (960 x 540) resolution.

The Galaxy S4 Mini Plus seems to be for all intents and purposes a rehashed Galaxy S4 Mini that has a new processor with LTE integration (Snapdragon 410 over Snapdragon 400) and of course a new version of Android, which is also quite far from the latest, being only Android 4.4 Kit Kat.

The original S4 Mini cost significantly more than the S4 Mini Plus does now, so this seems to be Samsung's way of relaunching the same device at a much lower price point, but with some upgraded components that support a newer version of Android.

The Galaxy S4 Mini Plus, which will also be called the Galaxy S4 Mini "Value Edition," will be sold mainly in Europe, and it's already available in countries such as Germany, Austria, Romania, Bulgaria and Czech Republic for between 190 euro and 239 euro.

Follow us @tomshardware, on Facebook and on Google+.

MediaTek has been making a bit of noise in the mobile market of late, particularly as it pertains to its recently-announced deca-core Helio X20 chip. Sure, there's a sort of "core" race in the mobile market (I see your quad-core and raise you octa-core!), but putting ten cores on one chip seems almost childlike in its one-upmanship on the surface. "More cores equals a better chip, a better chip equals a faster phone, and we win," seems to be the message.

Indeed, MediaTek has felt the blowback from naysayers who see its core-happy chip design that way. Even the more midrange P10, which the company announced at Computex, and the Helio X10, are brimming with eight cores each.

But Mohit Bhushan, MediaTek's VP and GM of Marketing and Business Development, told us in an interview that those detractors don't understand the point of the design.

Think Three Clusters, Not Ten Cores

Fundamentally, Bhushan said, the X20 is not about ten cores; it's about three clusters of cores that each serve a different purpose.

"Really, you're not running 10 processors at the same time. That's a very important part. Essentially, what we did is take the concept of big.LITTLE, and we stretched it a bit. You have the performance processors, and the power-conserving processors, and you pick which breed you want to turn on at any given time."

He explained that this is a tri-frequency architecture with three frequency planes -- 2.5 GHz, 2.0 GHz, and 1.4 GHz -- which are spaced out by about half a GHz. MediaTek populated each of those planes with processors.

The 2.5 GHz plane has two ARM Cortex-A72 cores, and the 2.0 GHz and 1.4 GHz planes each have four Cortex-A53 cores. (That's 2+4+4, for 10 total cores.) "Now, could we have put two A72s, two A53s, and two A53s [clocked lower] and done six [cores]?" he asked. "Sure, we could have done that. Or could we have done 2+3+3, and done 8 [cores]? Sure, we could have done that."

But, he said, the secondary idea here is that each cluster can offer the full processing experience for different tiers. For example, quad-core SoCs are the new norm; that's why the mid- and low-tier clusters are effectively each quad-core A53s. (MediaTek would have done the same with the top cluster, but the physical space required by the A72s pushed it to stick with a dual-core option for now.)

"Now let's talk use case," Bhusan continued. "You use Facebook all the time, you touch the icon on the desktop, it launches the app, and it shows you the feed, and you start scrolling through posts. All that stuff kicks off on the middle [frequency plane]. It's neither hot nor cold. And then if you pause, and start reading/liking/commenting, it uses the lowest one, 1.4 GHz. And then you turn on a video or a game in [Facebook], it goes to the high [frequency plane]."

"In just one app, what we did is optimize Facebook to run on three frequencies. Why do we do that? Power. Power is the main reason why." He said that with the three-frequency plane versus a two-frequency configuration, "We are getting 30 percent extra power benefit for doing the same app, same process, same everything."

Coherent Cache And Fixed-Point Hardware

Bhushan continued, "Part of this scheme is that you have to keep the caches coherent. There was a lot of work that went into, 'How do you keep cache coherency across three clusters?' And then once you get the caches coherent, you have to have the ability to turn on/off the cluster as you deem fit and also share the clusters with other components of the chip, like the GPU."

"That's where CorePilot comes in," he added. "CorePilot is like a scheduler, which is looking at the underlying hardware, looking at the input queue of tasks that keep coming, because users keep touching the smartphone, kicking off the processes. Those are the things that went into designing this chip."

When we asked, Bhushan also noted that much of the work done by the X20 is accomplished by fixed-point hardware, which handles items such as audio, video decoding and more.  

"Let's say you're gaming or watching an intense 1080p video," said Bhushan. "The processor is where the app is running, but when it has to do the media decoding, it's being done in the hardware. So it's not just all CPU-driven. The mobile industry has been relying on dedicated hardware for multimedia for quite some time now. And it keeps getting better."

Efficiency Is The Key

It may seem counter-intuitive considering all the cores MediaTek crammed onto the X20, but its main goal is efficiency. Again, hearkening back to the Tri-Cluster approach, that makes some sense: If you're engaging in a low-demand activity, the chip will use the lowest-necessary cluster. If you need more oomph, you can get it with the most powerful cluster. This way, you get the maximum amount of computing power with, ostensibly at least, the most efficient cluster.

Bhushan likens this paradigm to gears on a car. It's silly to drive 10 MPH in third gear, just as it's not feasible to accelerate onto a roadway stuck in first. You use the gear that's most effective for the speed you're driving.

That all sounds well and good, but there must be an inherent inefficiency in all that switching, and Bhushan admitted that's certainly the case, but there's still a net gain that makes this paradigm practical. He said that MediaTek has profiled several apps -- Facebook, Gmail, Skype, and a few Chinese apps -- and compared the performance between chips using the Tri-Cluster setup and those using a dual approach. At an overall system level, said Bhushan, "We are finding -- from kickoff to usage to going on and doing something else -- we're finding 20-25 percent power savings, despite the switching cost."

We asked where MediaTek saw itself penetrating deeper into the smartphone space -- on high-end flagships, low-end new-market devices, or somewhere in the middle -- but Bhushan turned the question around, pointing away from that high/middle/low conversation and aiming at what MediaTek is really concerned with, which is the user experience.

(We infer that MediaTek believes its three gears can satisfy users at all performance levels.)

Bhushan talked first about battery issues. "The battery has to last longer. You simply cannot afford to have your battery run out in the middle of the day. Batteries need to now last at least two or three days," he said.

"We had to innovate on how to conserve power, with Tri-Cluster as an example. There are more things being done in future chips, like obviously going to the next node, which is always the right thing to do. We're also looking at interesting technologies on components that suck less power, circuit technologies right on the board, which really improve line losses and how you design boards. So there's a strong R&D effort underway," he added.

MediaTek expects the Helio X20 to ship in consumer devices by the end of the year. When that happens, we'll look forward to putting the company's claims to the test.

Seth Colaner is the News Director at Tom's Hardware. Follow him on Twitter @SethColaner. Follow us @tomshardware, on Facebook and on Google+.

Mediatek has seen rapid adoption in low-end and mid-range phones in the past few years, mainly in China, but its chips are quickly spreading to the western markets, as well. One of its more popular chips was the octa-core Cortex-A53-based MT6752, which is now about to get replaced by the Helio P10 SoC.

The Helio P10, announced today, also has an eight-core CPU split into two clusters. Mediatek didn't specify the frequency of the lower-end cluster (presumably no lower than 1.2 GHz), but the higher-end one goes up to 2 GHz. That's the same frequency as the "mid-range" cluster inside the Helio X20 will have when it comes out by the end of the year.

As the CPU is based on the 64-bit ARMv8-A architecture, it should be future proof from a platform standpoint, and it also comes with some useful features such as hardware acceleration for AES encryption. That should enable encryption out of the box with future Android versions without experiencing significant performance slowdowns. Users could also simply enable it manually from the Security settings.

The Helio P10 comes with a next-generation Mali-T860 GPU that's focused on higher power efficiency compared to previous generations. The GPU only has two "cores," but it still manages to be 20 percent higher performance than the Mali-T760 inside the MT6752 while consuming 30 percent less power.

The GPU also brings support for the new Adaptive Scalable Texture Compression (ASTC) technology and the HEVC video codec. However, it won't support the VP9 codec, which is now required to decode HTML5 YouTube videos efficiently in hardware.

The SoC comes with its own integrated Cat. 6 (300 Mbps downlink, 50 Mbps uplink) LTE modem. The modem supports 2x20 carrier aggregation and the CDMA2000 band, which could make it useful for the U.S. market.

Mediatek's new SoC will also support cameras up to 21MP resolution, or dual-cameras that come with a 16MP+8MP setup. It can play videos of up to 1080p resolution at 60fps.

The Helio P10 will be manufactured on TSMC's 28HPC process node, which should bring slightly more power efficiency to the chip compared to the previous 28HPM process node. However, it should still be behind TSMC's 20 nm node as well as its 16 nm FinFET process node, which should be adopted by higher-end chips from other companies later this year.

Mediatek has shown some aggressive chip designs this year, both at the mid-range and the high-end, but its slow adoption of next-generation process nodes should keep its chips away from more premium devices.

Mediatek's Helio P10 will sample in Q3 this year and will enter mass production by the end of the year, which means we probably won't see it in too many devices until early 2016.

Follow us @tomshardware, on Facebook and on Google+.

Introduction And Specifications

Samsung remains the most prolific Android OEM, partly because they aren't afraid to experiment with new device classes, covering everything from Galaxy Gear wearables to the 12” Galaxy Note Pro tablet. While some of these trials miss the mark and leave us scratching our heads, many of its devices are among the most popular in their category. Probably Samsung's most successful gamble is the Galaxy Note line of smartphones.

Samsung's original Galaxy Note, while not the first phablet (remember the Dell Streak?), generated interest in larger phones, and through its utility, won a loyal following. With each generation of the Note series, more people embraced the concept and the Note's popularity grew. While the size of phablets aren't for everyone, the influx of large phones by other manufacturers indicates there’s a growing market for phones blurring the line with tablets.

The most noticeable feature of the Galaxy Note 4 is clearly the size of its screen, which has grown each generation from 5.3- to 5.5- to 5.7-inches in the Galaxy Note 3, a dimension which remains unchanged in the Note 4. Screen resolution, however, continues to climb, up from 1920x1080 in the Note 3 to 2560x1440. Samsung has also improved the quality of what's displayed on the screen by calibrating it with better precision.

The big screen isn't the Note line's only distinguishing feature. The S Pen is what defines the Note brand, and along with Samsung's supporting software, it's a true differentiator from other competing devices and can change the way you use your phone when fully embraced. Sure, you can write notes with it and use it as a pointer, but it can also be used to initiate a phone call, map an address, or markup images and PDFs. Paired with the big screen, the S Pen is a useful navigation aid and helps get real work done.

For the 4th generation Note, Samsung has also improved the design and construction. Like the Galaxy Alpha that preceded it, the Note 4 has a machined aluminum frame that surrounds a magnesium and plastic chassis. This aluminum frame looks great and the magnesium chassis adds rigidity.

Like the Galaxy S5, the Note 4 also uses Synaptics’ Natural ID capacitive fingerprint reader integrated into the home button. It works exactly the same as on the S5, requiring you to swipe a finger across it to authenticate. This method is a little frustrating to use when compared to the iPhone’s Touch ID (and the new sensor in the Galaxy S6) that simply requires you to touch the sensor. The Note 4’s physical size makes the swipe sensor even more challenging to use. At least Samsung has improved the fingerprint reader since it debuted on the S5, making it less fussy about swipe speed and partial and off-angle finger swipes.

The fingerprint scanner can be used to unlock the phone, log into your Samsung account, and gain access to items that have been protected by Samsung’s "Private Mode" feature. You can also use it with a select number of 3rd party apps such as PayPal and LastPass. Being FIDO-compliant, it can also be used to log into websites when using the Samsung browser, but not with Chrome. Business users who use Samsung’s KNOX can also enable two-factor authentication when logging into their work profile using the scanner.

Another new focus for Samsung with the Galaxy Note 4 is health and fitness. The Note 4 has hardware sensors capable of counting the number of steps you take each day, as well as your heart rate, blood oxygen level, stress level, and even the UV index. These sensors work together with the S Health app, which is fairly comprehensive in its own right, including support for Samsung’s line of wearable fitness products.

Samsung Galaxy Note 4 Tech Specs

It's only taken a few years for the large-screen smartphone category to become rather crowded. Every OEM from Apple to ZTE offers a phone with a 5.5-inch or larger display, creating fierce competition. Fortunately, the Note 4 isn't all about screen area. Its aluminum frame surrounds some very nice hardware. The Snapdragon 805 SoC still offers great performance even though it has been surpassed by the newer Cortex-A57 based Snapdragon 808/810 and Exynos 7420 chips. The 3GB of RAM is a generous amount and still the standard for current flagship phones. There's also a reasonable 32GB of internal storage, which can be expanded with removable SD cards.

Compared to the previous generation Note 3, there's a few big changes. Screen resolution has been increased from HD to QHD for the Note 4, reducing the graininess that can result from using a PenTile pixel structure. The Note 4 also includes an upgraded Snapdragon SoC and new front and rear camera sensors. Overall dimensions are very similar, with the Note 4 being ever so slightly taller and thicker than the Note 3. The aluminum frame increases the Note 4's rigidity, reducing flex and giving it a premium feel, but at the cost of 8g of additional weight.

Cellular

There are 22 different model numbers for the Galaxy Note 4, each supporting different regional or carrier specific frequency bands. Some versions substitute a Samsung Exynos 5433 [ARM Cortex-A57 (4x @ 1.9GHz) + ARM Cortex-A53 (4x @ 1.3GHz), Mali T760MP6 GPU] for the Snapdragon 805 SoC. There's also differences in audio codecs, voice processors, and baseband processors, with some using an Ericsson M7450 and others using Cat 6 basebands such as the Intel XMM7260 and Samsung's M303. To keep things from getting too complicated, the table above focuses on just the North American models.

Qualcomm's MDM9x25M baseband is a third-generation Category 4 LTE modem built on a 28nm HPm process offering 150 Mb/s down and 50 Mb/s up with carrier aggregation. It also supports HSPA+ Release 10 for 84 Mb/s down using dual-carrier HSDPA. The MDM9625M incorporates all major radio modes, including GSM/EDGE, UMTS (WCDMA, TD-SCDMA), LTE (LTE-FDD, LTE-TDD) and CDMA2000.

The Note 4 uses Qualcomm's WTR1625L transceiver paired with another WTR1625, rather than the more common WFR1620 receive-only transceiver, in order to bond two 10MHz channels and reach full Category 4 LTE speeds, a requirement when carriers don't have 20MHz LTE channels.

The RF front-end uses Qualcomm's QFE1100 envelope tracker, which dynamically adjusts voltage to the power amplifiers, wasting less energy and reducing heat generation, but none of the other components in Qualcomm's RF360 suite.

Options

The North American versions of the Note 4 all come with 32GB of onboard storage, and color options are limited to Charcoal Black or Frost White.

As evidenced by the impressive hardware specifications and premium pricing, the Galaxy Note 4 is targeted squarely at the high-end market. For consumers willing to spend top dollar on the ultimate smartphone, the question becomes one of comparative value in regards to other devices in the same category. A question answered by the design of the Galaxy Note 4 and the value added by hardware and software features.

Hardware Design

If you’ve owned or used any of the Samsung Galaxy S phones over the last several years, the Galaxy Note 4 will be familiar, though there are some fit and polish changes. The aluminum frame, modeled after the Galaxy Alpha with slightly indented and flattened sides that make it easy to grip, has a painted matte finish that matches the color of the phone, either white or charcoal grey. This finish is very durable, since over the months that we have been testing the Note 4 it hasn’t worn or scratched off. The edges of the metal sides are beveled and polished, giving them an attractive chrome finish that accents the Note 4’s design. The edges of the metal buttons on either side, the home button, and the rear camera components also receive the bevel and polish treatment.

In spite of the large display the phone is relatively lightweight, and feels incredibly thin. Because of the height and width of the phone you’ll often find yourself adjusting your grip in order to reach various edges when operating the device with one hand. We tested a white AT&T model and white and charcoal Telus models. Both colors have a leather-like texture on the back which helps maintain a hold on the phone while changing your grip. However, the white model’s texture has a slightly different, smoother feel when compared to the charcoal. This is likely to make the white finish more impervious to dirt and grime.

The top of the Galaxy Note 4 features a 3.5mm headphone jack, located toward the left of the device as it’s facing you. A slight bump on the back of the top edge is a result of the width of the headphone jack. The right side houses an Infrared LED, which allows the device to function as a remote control in conjunction with an app. A microphone used for the speakerphone and noise cancellation is also located on the top of the device.

The face of the Galaxy Note 4 is dominated by the 5.7-inch QHD Super AMOLED display which spans nearly the entire width of the device. There's an RGB LED in the top-left corner, cleverly hidden when not in use, which provides visible feedback of pending notifications and charging status when the screen is off. Centered above the screen is the phone earpiece and a chrome Samsung logo. The 3.7MP front-facing, wide-angle camera is located in the top-right corner above the display.

Note: The charcoal model has a screen protector applied.
Centered below the screen is the home button/fingerprint sensor, the only physical button on the face of the device, with Samsung's classic arrangement of capacitive buttons to either side—recent apps on the left and back on the right. The bezel to either side of the display is minimal in order to keep the device width manageable, while the bezels above and below are large enough to accommodate the navigation buttons and sensors, as well as providing something to grasp when using the device in landscape mode. The bezel has a pattern of horizontal lines below the glass, adding a bit of a textured look to the front of the phone.

The metal volume rocker is located toward the top of the left edge and either end of the rocker is raised slightly to facilitate controlling the volume by feel. The rocker gives an audible and tactile click when the button is fully pressed.

The metal power button is near the top of the right edge and provides the same tactile click as the volume rocker. It's slightly offset from the volume rocker on the opposite side, minimizing the chances of accidentally squeezing the volume rocker with a finger when clicking the power button with the thumb. Lefties will have a bit more trouble with this, since the left thumb naturally falls on the volume rocker.

Around back, the plastic rear cover retains the faux leather look seen on the Note 3, but ditches the over-the-top stitching around the perimeter. The 16 MP camera is housed within a square hump, with a rounded rectangle containing the LED flash and tricorder-like sensors for measuring your heart rate and blood oxygen levels as well as a sensor for measuring the UV level in sunlight. The lone speaker sits in the lower-left corner, which is not our preferred location. Other markings on the back are device specific and may include the Samsung logo, carrier logo, or Galaxy Note 4 branding.

The Note 4 has a removable rear cover which exposes the 3220 mAh battery, a microSD slot which accepts up to 128GB cards, and a slot for a micro-SIM card. This feature should be especially appealing for people looking to upgrade from an older Galaxy S phone who do not like the redesigned Galaxy S6. Unlike the S6, however, the Note 4 does not natively support wireless charging, although the Qi standard can be added with optional Samsung rear covers and cases.

Gone is the Note 3's USB 3.0 port, replaced by the more common but slower microUSB 2.0 variety centered on the bottom edge. The smaller port is bracketed by two microphones. The S Pen docks in the holder located on the right side of the bottom edge.

The S Pen stylus that comes with the Note 4 uses Wacom technology, so does not require any power source. It is pressure sensitive, which makes it a great tool to use with drawing applications that support this feature, such as the included Samsung S Note app and Autodesk’s SketchBook Pro. There is a button on it too, which is used to bring up the radial "Air Command" S Pen action menu and perform other actions with the pen such as selecting text. Compared to the Note 3’s S Pen, the Note 4’s is slightly larger and has double the pressure sensitivity—2048 pressure levels versus 1024.

Accessories

The Galaxy Note 4 ships with a USB charger and a microUSB cable. The included charger uses Samsung’s Adaptive Fast Charging technology (compatible with Qualcomm's Quick Charge 2.0), that gives you 50% battery capacity in only 30 minutes and 100% capacity in just under 2 hours.

Samsung’s S-View Flip Cover cases have a window which allows viewing and interacting with a portion of the display when closed. Some of these actions include the ability to respond to incoming phone calls, take pictures, or use the heart rate sensor. Thanks to the SAMOLED display, only the visible portion of the screen is powered on when the cover is closed saving power. S-View cases for the Galaxy Note 4 retail for $49.99 and come in Charcoal Black, Frost White, Bronze Gold, and Plum Red.

Another unique case for the Note 4 is Samsung’s LED cover. At first glance, it looks like a traditional windowless flip cover, but it actually has embedded low-power LEDs in the front that light up. They are used to show you various information such as the time, music info, and notifications about messages and missed calls. The LED cover comes in Charcoal Black, Bronze Gold, and Plum Red and retails for $69.99.

If you want to add Qi wireless charging to the Note 4, you can buy replacement back covers that add this feature for $34.99, or a version of the S-View Flip Cover with Qi charging for $69.99. You’ll need a charging pad to use wireless charging of course, and Samsung sells one for $49.99.

The other accessory what Samsung is making available for the Galaxy Note 4 is the Samsung Gear VR headset, which is powered by Oculus. The Note 4 is inserted into the front of the Gear VR headset, using the QHD display to present a virtual reality gaming experience using a combination of head movement and touch controls. The Samsung Gear VR headset is available for $200.

Display And Audio Performance

With each generation, Samsung improves its SAMOLED display technology, achieving higher brightness levels and more accurate colors. The Galaxy Note 4 is no exception. In addition to these improvements, screen resolution has been increased from HD to QHD for the Note 4, which helps compensate for the PenTile pixel layout having fewer red and blue subpixels as compared to an RGB stripe LCD display. Even after accounting for the discrepancy in subpixels, the Note 4 has a higher pixel density than a 1080p LCD.

In an effort to improve our mobile reviews, we are now using SpectraCal's CalMAN software and SpectraCal C6 colorimeter for display measurements. All of the charts below with a gray background were generated in CalMAN v5 Ultimate.

Full Brightness chart including values for APL=50 and auto-boost.

We are now reporting two different brightness levels for AMOLED displays: APL=50% and APL=100% (APL stands for Average Picture Level). If you are unfamiliar with APL, here is a good article explaining what it means. Basically, the brightness of an AMOLED display changes depending on what content is actually being displayed. The APL values we chose to measure provide a good upper and lower bound for what's practically achievable.

At an APL of 100%, which serves as a worst case condition, the Note 4 manages to exceed 300 nits, matching the brightness of the display in the newer Galaxy S6. This is also significantly higher than the AMOLED display in the Nexus 6, although it's still less than what LCDs can achieve. Looking at the full brightness chart shows only a modest brightness increase for an APL of 50%, again tying the display in the GS6 but a little shy of the Nexus 6. With a max brightness between 333 to 363 nits, the Note 4 is sufficiently bright for any indoor scenario, but not bright enough for outdoor viewing in sunlight.

Fortunately, Samsung provides a little trick for just this situation. If the Auto brightness mode is activated and the ambient light exceeds a certain threshold, the display brightness gets a significant boost. In this "boost" mode, max brightness shoots up to between 500 and 630 nits, equaling or even exceeding the best LCD displays. While no OEM will ever be able to make a display that can compete with the sun's brightness, this at least makes the screen visible. Of course a boosted display uses more power, and it can only maintain this brightness level for a short period of time before the screen overheats and the brightness falls back to normal levels.

The Galaxy Note 4 has four different screen modes with different gamma curves, color gamuts, and color saturation levels: Basic, Cinema, Photo, and Adaptive display, which automatically adjusts—or "optimizes", according to Samsung—these screen parameters based on the app that's being used. This mode is rather limited, however, only affecting six specific apps, making it difficult to test. Instead, we'll focus on the other three modes. For testing purposes, the "Auto adjust screen tone" setting, which automatically adjusts the screen luminance based on the content being displayed, is turned off.

The average gamma for the Note 4's Cinema mode comes the closest to the ideal value of 2.2; however, its gamma curve tells a different story. Below a grayscale level of about 55%, gamma spikes to 2.56 leading to darker shadows and a loss of highlights. Above 55%, gamma drops to a minimum of 1.34 resulting in a significant loss of shadow detail.

In contrast, gamma remains very consistent across a full grayscale sweep in both the Basic and Photo modes, coming in just above the ideal value.

Color temperature is spot on in both the Basic and Photo modes, with essentially no variation across a full grayscale sweep (values close to 0% are not accurate). Cinema mode eschews absolute accuracy and instead opts for a cooler color temperature similar to what we see in most mobile displays.

Looking at the RGB balance we can see why the Basic and Photo modes land so close to the ideal color temperature: no single primary color varies by more than ±4% for any grayscale level. Cinema mode on the other hand clearly emphasizes blue over red, leading to its cooler color temperature.

The grayscale accuracy for the Note 4 in both the Basic and Photo modes is excellent, nearly matching the exceptional accuracy of the Galaxy S6. Average ΔE2000 is below two and error at any single grayscale level remains below three. Grayscale error goes up dramatically when switching to Cinema mode, climbing steadily towards white due to the blue shift we saw in the RGB balance chart and producing visible errors similar to what we saw with the Nexus 6.

The Note 4's Basic mode does a good job of covering the sRGB color space, only missing on the blue corner of the triangle. Photo mode is similar, but extends green tones beyond sRGB. Cinema is a true wide-gamut mode covering about 130% of the sRGB color gamut. This leads to over saturated, neon-like colors that do not look natural, just like we saw with the Nexus 6.

In the color saturation sweep, we see the Basic mode perform very well once again. Each saturation level hits the target and there's no hint of color compression. Photo mode shows a slight red shift in magenta tones, but still no significant color compression. This trend does not hold when activating Cinema mode, where we see significant color compression for all colors. Each box/dot pair is a 20% step in saturation, so, for example, anything above 80% saturation for green essentially looks the same.

These graphs also show the effects of using a wide-gamut screen to view sRGB content. In Cinema mode, for example, a picture showing green grass at a 60% saturation would be displayed at 100% saturation relative to the sRGB gamut, effectively making the grass look "too green" or vivid.

Color accuracy for the Note 4's Basic mode is very good, with nearly every tested color showing an error of less than three. Things get progressively worse for the Photo and Cinema modes, however. Cinema mode is nearly as bad as the Nexus 6; most tested colors have an error above five and a maximum error of almost 13.

Full Size Images: [Color Palette: Basic], [Color Palette: Cinema], [Color Palette: Photo]

The color palettes above show the target color on the bottom versus the displayed color on the top and are a less abstract way of looking at color accuracy. Since our content system automatically applies additional compression to JPEG images, we've included links to the originals devoid of the heavy compression artifacts.

The display in the Galaxy Note 4 is one of the best currently on the market. Max brightness for AMOLED screens may still trail LCDs, but Samsung's auto-boost feature makes up the difference for extreme visibility cases at least.

Irregardless of your display preferences, you should find one of the Note 4's display modes to your liking. Basic mode is a well-calibrated, proper sRGB mode that delivers accurate colors and grayscale values that purists will enjoy. For those less concerned about accuracy and who like their colors brighter with more "pop," there's Cinema mode. Finally, Photo mode provides a reasonable compromise between these two extremes.

Audio Performance

Naming a product "Note" is a clear indication that its primary purpose is not for playing media, a decision exemplified by the mono external speaker. Located around back, it's mostly an afterthought. A single-channel class D amplifier from Maxim Integrated helps it get reasonably loud without succumbing to distortion, but obviously bass notes sound anemic.

Having the speaker firing in the opposite direction of your ears hurts both perceived loudness and the sound that eventually reaches you. Since your basically listening to reflected sound, the treble is attenuated and everything sounds a bit hollow. Cupping and orienting your hand behind the phone just right improves the volume and sound dramatically, but this is hardly an ideal way to hold your phone and the sound still falls short of other devices. Also, when watching a video on the Note 4 in landscape, our fingers would fall directly on top of the speaker, partially blocking the sound. A similar problem occurs when setting the phone down on a table. The raised ridge in the middle of the speaker ensures a small air gap between it and the surface it's resting on, but it still sounds muffled. This applies to voices on conference calls too. Raising the phone off the table surface by wedging a pen or something underneath helps voices sound much clearer.

It's probably best to just leave the external speaker for ringer duties and use some nice headphones for everything else. Thanks to Qualcomm's WCD9330 audio codec, headphone output is excellent. Based on our subjective listening tests, the Note 4 rivals the Sony Z3 (which uses the previous generation WCD9320 codec) and iPhone 6 in quality. Also, twisting the plug while seated in the headphone jack did not produce any static like it does on the iPad Air 2.

Camera: Hardware And Software

The Note series has always been a platform for Samsung’s premium mobile imaging experience (if you don’t count the Galaxy Camera), and the Note 4 is no different. Each version of the Note has added a stand-out feature: The Note 3 added 4K video capture, while the Note 4 adds optical image stabilization (OIS).

Other than the addition of OIS, the Note 4’s camera specs appear similar to the Galaxy S5 and the Note 3. The Note 4 has a 16 MP sensor with phase detection autofocus (PDAF) like the S5 but uses a Sony Exmor RS IMX240 sensor instead of Samsung's own ISOCELL unit. Its 31mm f/2.2 optical stack is also similar to the Note 3's.

Compared to its phablet competitors, the Note 4's 16 MP sensor is physically the largest at 1/2.6", but maintains the same 1.12μm pixel size as the 13 MP sensors in the Note 3, Nexus 6, G Flex 2, and OnePlus One. The iPhone 6 Plus, with only an 8 MP resolution, trades pixel quantity for pixel size, which should give it an advantage in low-light performance. It's interesting that all of these phones are using some kind of Sony sensor, a testament to Sony's strength in this market.

As for the optics, the Note 4's f/2.2 31mm lens is comparable to its competition. It's only this year that we're seeing flagship phones such as the LG G4 and Galaxy S6 move to faster sub f/2.0 optics for the rear camera. All of the current premium phablets even have OIS, apart from the OnePlus One.

The front-facing camera on the Note 4 uses a Samsung S5K6D1YX sensor, which we haven’t seen in any other phone. Images have a 16:9 aspect ratio and a 3.7 MP (2560x1440) resolution. Paired with a 90-degree f/1.9 lens, it seems to be a better front camera package, on paper at least, than what's offered by its competitors, except for the OnePlus One's 5 MP shooter.

The Sony Exmor RS IMX240 sensor in the Note 4—the same sensor being used in the new Galaxy S6—is a bit of an unknown quantity. We can’t find any official documentation for it, but it seems to be a custom design for Samsung and is not found in any other vendor's phones. While it may be a custom chip, it does seem to be largely similar feature-wise (apart from resolution) to the upcoming Stacked CMOS 21MP IMX230 sensor that just started shipping in April of this year.

As now seems to be standard on Samsung’s phones, the IMX240 captures images natively at a 16:9 aspect ratio. It also supports phase detection AF (PDAF), allowing the Note 4 to focus on its subject very quickly, greatly improving the overall camera experience. However, unlike the advanced HDR features of the IMX230, the IMX240 is not capable of using HDR when shooting video in 4K, or anything beyond 1080p at 30fps.

The optical image stabilization that Samsung has combined with this sensor is what it's calling Smart OIS, which is a combination of both optical and digital stabilization. It provides 2-axis stabilization, but not 3-axis like the OIS found on the LG G3 and G4. Despite that, the Smart OIS system should improve the Note 4’s camera performance in lower-light conditions when shooting hand-held by minimizing camera shake.

Note 4 Camera Software

The camera interface is virtually unchanged in the Lollipop update. Some of the UI icons are slightly smaller, but the control layout and settings have remained the same, except the options for fast and slow motion video are now specific modes, accessible by tapping the "MODE" button.

It's disappointing that the Note 4's Lollipop update does not support Google’s new Camera2 API. This API, introduced with Lollipop, allows for full manual control of the phone’s camera, including focus and shutter speed. It also adds support for RAW image capture. Unfortunately, this limitation extends beyond the stock camera app, affecting 3rd-party apps such as Camera FV-5. Perhaps Samsung wanted to reserve these features for the GS6, since there's no technical reason why the Note 4 can’t support this API.

The main camera UI screen puts all the basic camera controls, except for flash, within easy reach. Touching the gear icon on the bottom-left brings up a quick settings menu for toggling the shot timer and flash. The top button brings up the full settings menu. You can also quickly change the camera resolution from the native 16 MP (5312x2988) down to 2.4 MP (2048x1152). The middle button allows you to apply live image filters to your pictures. There are twelve filters pre-installed, and you can download more from the Samsung App store. These filters cannot be applied if HDR is on.

The screenshots above show some of the options available in the settings menu. ISO can be manually adjusted up to a maximum value of 800. One thing to note is that when HDR is turned on you cannot adjust exposure, ISO, or metering modes.

Selective Focus

The "MODE" menu lets you choose between nine different camera modes. One of the more interesting ones is Selective Focus, which tries to replicate the bokeh effect you’d get if you were taking a close-up shot with a lens set to a much lower aperture, such as that on a DSLR camera. It does this by taking multiple pictures of a subject that is less than 20 inches from the camera and combining them in software, allowing you to shift the focus point and blur the background. After taking the shot, you can choose between Near focus, Far focus, and Pan focus before saving the final version of the image.

Full Sized Image: [Note 4: selective focus]

The picture takes a few seconds to take making this mode unsuitable for moving subjects. Also, in order for the effect to work properly, there has to be a clear separation between the subject and its background, otherwise the software gets confused and applies the soft focus to the wrong part of the image.

“Shot & More” Mode

Shot & More mode consists of a number of different effects that can be applied to images taken of subjects in motion. In the slideshow below are a series of images showing the process of using each of these Shot & More effects (apart from Best face).

Full Size Images: [Note 4: Shot & More Best Photo], [Note 4: Shot & More Drama Shot], [Note 4: Shot & More Eraser], [Note 4: Shot & More Panning Shot]

The Best photo effect allows you to scroll through a series of shots of a subject in motion and choose the best one. Drama shot combines multiple images of a subject in motion, so it repeats throughout the shot. Eraser lets you remove moving objects from a picture. Lastly, panning shot applies adjustable motion blur to the background.

There's several other special camera modes too, including Rear-cam selfie and Panorama. Virtual tour stitches together still images taken while moving through an interior space to create an animated GIF, while Dual camera mode that lets you use the front-facing camera to create a picture-in-picture shot superimposed on the rear camera image.

Front-Facing Camera

The Note 4 has several modes and options for the front-facing camera too. Beauty mode, which applies a softening filter to detected faces, is on by default, but can be turned off. Resolution is adjustable from 0.3 MP (640x480) with a 4:3 aspect ratio to the full 3.7 MP (2560x1440) with a 16:9 aspect ratio. The same twelve filters available when using the rear camera are also available, and you can manually adjust exposure levels. The exclusive mode for the front-facing camera is "Wide selfie," which stitches together multiple images to create a wider image, perfect for fitting in more people in a group selfie. For video, you can shoot up to QHD (2560x1440) resolution.

Video

The Snapdragon 805 in the Note 4 is capable of encoding and decoding up to 4K UHD H.264 video in hardware, which brings better performance and lower power consumption. While it can also decode 4K UHD H.265 video in hardware, encoding must still be done in software. Therefore, all video on the Note 4 is encoded in H.264.

Rear Camera Video Modes

Front Camera Video Modes

The Note 4 offers a good selection of video modes, including 60fps 1080p, slow motion, and fast motion (time-lapse) video. At 4k, the Note 4 records at a 48 Mb/s bit rate, higher than the 42 Mb/s used by the Nexus 6 but lower than the Galaxy S5's and OnePlus One's 57 Mb/s average bit rate. The Note 4 records 1080p video at the same 17 Mb/s bit rate as the Galaxy S5, which is just shy of the 20 Mb/s rate used by some older phones in this size range such as the LG G3 or OnePlus One.

Video quality from both the front and rear cameras is very good overall, with good white balance and exposure. There's very little noise from the rear camera in most conditions, although some noise artifacts do appear in video shot with the front camera. The rear camera does a good job maintaining focus even during transitions, only getting flustered in some high dynamic range or lower-light conditions. The iPhone 6's PDAF autofocus is still the best we've seen, but the Note 4 is not too far behind.

When shooting 30fps video at any resolution, the video gets pretty jerky while panning. This is especially noticeable for any objects with straight edges. To be fair, all phone cameras exhibit this behavior when shooting at 30fps, however, on the Note 4 it seems to be just a bit more pronounced. Switching to the 1080p 60fps mode eliminates this issue, producing buttery smooth pans.

There are a few restrictions when shooting at higher speed or resolution though. Video stabilization and HDR are unavailable in the 1080p 60fps mode. Both the UHD and QHD modes have a five minute time limit for any single clip. Also, choosing either of these modes disables dual camera mode, HDR, video effects, and video stabilization. Turning on HDR in any of the HD modes also disables video stabilization.

There's three different slow-motion speed choices: 1/2, 1/4, and 1/8. To capture slow-motion video, you first tap the "MODE" button and then select the "Slow motion" tile. The resulting H.264 Baseline video is captured in 720p @ 120fps, with a rather low bit rate of 6 Mb/s. Next, click the video preview square to open the slow-motion video editor and use the slider controls to trim the length of the video and select the region you want to appear in slow motion. After using a toggle to select one of the three speeds, you can export the final video.

All of the exported videos playback at 15fps. The slowest 1/8 speed is just the original 120fps video played back at 15fps. At 1/4 speed, every other frame is dropped and at 1/2 speed, only one in four frames are kept, resulting in choppy looking video. The low bit rate further degrades the video quality, showing significant compression artifacts and loss of detail. The Note 4 also does not record audio in slow-motion mode.

Camera: Performance And Photo Quality

Camera hardware specs and software features mean nothing if the photos produced are of poor quality. It's time to see how the Note 4 compares to several other large-screened phones, including the iPhone 6 Plus, Nexus 6, LG G3 (same camera as the G Flex 2), and HTC Desire Eye (13MP Sony IMX214 sensor and f/2.0 lens). Unfortunately, we did not have the Note 3 and OnePlus One available for this test.

All images were taken using the Auto mode unless noted. Also, you can view the full-sized image for each photo by clicking the text links below the images that are within a slideshow album. Both the Note 4 and Desire Eye shoot natively at a 16:9 aspect ratio, while the other phones shoot in 4:3.

Outdoors

Daylight

Full Size Images: [Note 4: outdoor car], [iPhone 6 Plus: outdoor car], [Nexus 6: outdoor car], [LG G3: outdoor car], [HTC Desire Eye: outdoor car], [Note 4: outdoor boxcar], [iPhone 6 Plus: outdoor boxcar], [Nexus 6: outdoor boxcar], [LG G3: outdoor boxcar], [HTC Desire Eye: outdoor boxcar]

Despite being taken in the late afternoon, the light when the car was photographed was enough that all the phones apart from the Desire Eye were able to select very low ISOs. The shutter speed selected was also high enough that there are no focus issues from camera shake. However, the iPhone 6 Plus was able to select the highest speed because the bigger pixels of its 8MP sensor were able to capture more light.

When it comes to detail in these conditions, the 16MP camera in the Note 4 trumps its competitors, but the 13MP sensors in the Nexus 6, G3, and Desire Eye are not far behind. The color accuracy and saturation of the Note 4’s image is pretty good. However, the auto white balance is perhaps a little too cool, as is the Nexus 6’s and Desire Eye's. The iPhone 6 Plus produces the most accurate color in this case. Also, while the Note 4’s dynamic range in this picture is good, the iPhone 6 Plus’s shot does look better in this department, successfully capturing the blue sky.

The image of the boxcar was shot in brighter light and allowed all the phones to select a low ISO and a very fast shutter speed. All five cameras produce a decent image in these conditions with good dynamic range. The yellow in the Note 4's image skews slightly towards green, but when zoomed in close to the rusty patina on the boxcar, the Note 4’s 16MP sensor has captured the most detail, especially when compared to the iPhone’s 8MP sensor.

Overall, the Note 4 performs very well in daylight, but to be honest, so does nearly every other current smartphone. We'll need more challenging conditions to see how good its camera really is.

Night

Full Size Images: [Note 4: outdoor night], [iPhone 6 Plus: outdoor night], [Nexus 6: outdoor night], [LG G3: outdoor night], [HTC Desire Eye: outdoor night], [Note 4: outdoor Gremlin], [iPhone 6 Plus: outdoor Gremlin], [Nexus 6: outdoor Gremlin], [LG G3: outdoor Gremlin]

When the light drops below a certain threshold, the Note 4 activates a night shooting mode that seems to take and combine multiple images to produce a single, brighter shot. This means shot-to-shot shooting time is slower and ghosting can be an issue for moving objects. There's also no ISO or shutter speed info recorded while in this mode. The only way to turn this mode off is to set the ISO manually. However, 800 is the highest you can set this way and it may not be enough in very low light.

The first Note 4 image in the slideshow is one shot in this mode. Zoomed out it looks ok and is brighter than that of the iPhone 6 Plus and Nexus 6, but when zoomed in you can see that the detail is soft due to the aggressive noise reduction. You can also see some haloing around some of the bright areas. Of the other images of the candy apple red 30’s Ford sedan, the iPhone 6 Plus’ is clearly the winner. Its sensor’s large pixels and OIS allowed it to take a crisp, in-focus image with a longer exposure time, while reducing noise with an ISO of 200. Still, the Note 4’s picture looks to be the runner-up, since the others either have too much noise from being taken at very high ISOs, are too dark, or in the case of the LG G3, a less effective long exposure mode.

In the second image of the white Gremlin, the Note 4 does not use its special night mode, capturing a clear image at ISO 250 and a shutter speed of 1/12 seconds thanks to its OIS. Again the iPhone 6 Plus trumps the competition, being able to produce a decent image at even lower ISO and less noise thanks to its larger pixels. However, its 8MP shot at ISO 64 is not that much more detailed than the Note 4's 16MP shot at ISO 250.

The Note 4’s low light image quality is the best of the current crop of large-screen Android phones, and certainly a step above that of its predecessors, the Galaxy S5 and Galaxy Note 3. However, the iPhone 6 Plus’ larger pixels and effective OIS mean that its low-light pictures look the best, but only by a slim margin.

HDR

All of the cameras being compared have the now obligatory HDR modes, though their implementations are slightly different. The Note 4 leverages the sensor and the SoC’s ISP to preview the HDR effect in real time, so you can see how it looks when composing the shot. The other phones, apart from the Desire Eye, apply HDR as a post-processing effect only after the photo is taken, so you will not know how the picture will look before taking it. The Desire Eye does show you a preview of the HDR effect when composing your shot, but this preview does not always reflect what the final HDR image looks like.

Full Size Images: [Note 4: no HDR], [Note 4: HDR], [iPhone 6 Plus: no HDR], [iPhone 6 Plus: HDR], [Nexus 6: no HDR], [Nexus 6: HDR], [LG G3: no HDR], [LG G3: HDR], [HTC Desire Eye: no HDR], [HTC Desire Eye: HDR]

In general, these comparison images show that all of the phones here have issues with dynamic range when shooting in conditions like this. In the non-HDR images the sky is overpowering the subject of the shot, so using HDR is the only way to get a good image in this light. The challenge, though, for a good HDR mode, is to brighten the areas in shadow but keep from overexposing the already bright areas.

The Note 4, G3, and Desire Eye all do a great job, though we’d have to say the Note 4 and G3 HDR images look the best. The Desire Eye goes a little too far and creates an image that looks unnatural. As for the iPhone 6 Plus and Nexus 6, their image processing is not aggressive enough and the HDR effect is minimal at best.

Additional HDR Images

Full Size Images: [Note 4: no HDR 1], [Note 4: HDR 1], [Note 4: no HDR 2], [Note 4: HDR 2], [Note 4: no HDR 3], [Note 4: HDR 3], [Note 4: no HDR 4], [Note 4: HDR 4], [Note 4: no HDR 5], [Note 4: HDR 5], [Note 4: no HDR 6], [Note 4: HDR 6], [Note 4: no HDR 7], [Note 4: HDR 7]

The Note 4's HDR mode produces some dramatic results, whether it's a street scene shot at dusk or an interior shot with a bright window in the background. Samsung definitely has the best HDR implementation right now, from the live HDR preview to the quality of the final shot. It's so good in fact, that we recommend just leaving it on all the time.

Indoors

The staged indoor shots below were lit by overhead LED lights, a CFL lamp from the front, and an incandescent overhead light in the background.

Bright Light

Full Size Images: [Note 4: indoor bright light], [iPhone 6 Plus: indoor bright light], [Nexus 6: indoor bright light], [LG G3: indoor bright light], [HTC Desire Eye: indoor bright light]

Even with all of the lights turned on, the scene is not exceptionally bright, forcing all of the Android phones to raise their ISO value. The Note 4 and Nexus 6 hold ISO a bit lower, reducing noise and producing reasonable images. Both the G3 and Desire Eye exhibit a noticeable amount of noise, even under these lighting conditions. The iPhone 6 Plus is the only camera that produces an image with lower than 100 ISO, though its detail level is noticeably less when viewed closely due to its lower resolution. As for color accuracy, the Desire Eye's and iPhone 6 Plus' white balance looks the best, while the other images are a little too cool.

There's no clear winner for this scene, but we'd say it’s a toss-up between the Note 4 and iPhone 6 Plus.

Low Light

Full Size Images: [Note 4: indoor low light], [iPhone 6 Plus: indoor low light], [Nexus 6: indoor low light], [LG G3: indoor low light], [HTC Desire Eye: indoor low light]

This scene was only lit by the incandescent light in the background, creating a very challenging scenario. The Note 4 engages its special night mode and creates an image with roughly the same brightness as the iPhone. The Note 4’s image also retains a lot of detail and is not too noisy. However, its white balance is off, giving the whole picture a green tint. The iPhone 6 Plus does well with white balance and effectively uses its OIS to hold its shutter open longer and ISO at only 250, reducing noise; however, its noise reduction algorithm leaves some odd artifacts. The rest of the phones fall down in this test. The Nexus 6’s shot is much too dark and noisy, taken at ISO 1196. The G3’s night shot is the brightest, but unnaturally so, and the color balance skews heavily towards red. Lastly, the Desire Eye’s picture is a dark, unusable mess shot at ISO 2500.

The iPhone 6 Plus has a slight edge over the Note 4 in lower-light situations, but the Note 4 is pretty close and clearly the best of the rest.

Flash

Full Size Images: [Note 4: indoor with flash], [iPhone 6 Plus: indoor with flash], [Nexus 6: indoor with flash], [LG G3: indoor with flash], [HTC Desire Eye: indoor with flash]

Both the Note 4 and iPhone 6 Plus take very nice images with the flash turned on. The Note 4 has the edge in detail, but the iPhone 6 Plus' True Tone flash produces more natural looking colors. While the flash on the Nexus 6 and G3 is capable of adequately lighting the scene, the color is too cool and the images are overexposed.

Front-Facing Camera

Full Size Images: [Note 4: Front-facing camera sample 1], [Note 4: Front-facing camera sample 2], [Note 4: Front-facing camera sample 3]

The samples from the front-facing camera were taken with the unnatural "Beauty mode" turned off. The last image is an example of the Note 4’s "Wide selfie" mode.

Additional Sample Images

The slideshow below contains a selection of images taken with the Note 4 in a variety of locations and lighting conditions, and should give you a good idea of the kind of photographs it's capable of taking.

Full Size Images: [Note 4: sample 19], [Note 4: sample 20], [Note 4: sample 21], [Note 4: sample 18], [Note 4: sample 8], [Note 4: sample 13], [Note 4: sample 14], [Note 4: sample 11], [Note 4: sample 12], [Note 4: sample 9], [Note 4: sample 15], [Note 4: sample 3], [Note 4: sample 2], [Note 4: sample 10], [Note 4: sample 1], [Note 4: sample 16], [Note 4: sample 17], [Note 4: sample 5], [Note 4: sample 6], [Note 4: sample 4], [Note 4: sample 7], [Note 4: sample 22]

Overall Camera Performance

The Note 4's camera system, from its lens to the Sony IMX240 sensor to the Snapdragon 805’s ISP (image signal processor) to Samsung's software, work in harmony to make it one of the best cameras of any smartphone. It's capable of producing excellent high-resolution images that are better than all its large-screened Android peers. Only in lower light do we see the iPhone 6 Plus surpass the Note 4, and even then not by a significant amount. The Note 4 also has a number of useful camera modes, such as "Selective focus" and "Shot & more" that can produce some great-looking, unique images without having to take them into an image editor.

With PDAF, the Note 4 can focus on its subject in approximately 300ms, which is faster than the Galaxy S5's PDAF system and the G3's laser autofocus system, along with most other current smartphones. Only the iPhone 6 Plus' PDAF can focus faster. Combining this with the Note 4’s shot-to-shot performance, which is also very good and one of the quickest we’ve seen, means you won't miss capturing that fleeting moment.

Software

Samsung has taken some heat in years past for its TouchWiz UI, both for its cluttered design and feature bloat. For this iteration, Samsung cleans things up a bit by making some of its less-used software optional rather than installed by default. Few would call TouchWiz uncluttered, but it does add several features that are particularly useful for the Note 4, such as Multi Window, S Pen integration, and special modes for one-handed use, that are accessible via the Settings menu, widgets, or popup menus.

The Samsung Galaxy Note 4 shipped with Android KitKat 4.4.4 when it launched last fall. It took quite awhile to update its flagship devices to Lollipop: The AT&T Note 4 we tested didn’t get updated until March 2015, and the Canadian Telus unit we tested didn’t get the update until April 2015. With the update to Lollipop, one would expect to see quite a few changes. However, Samsung decided to keep using the same version of their TouchWiz UI as was found in the KitKat build, rather than using the redesigned version included with the recently released Galaxy S6. In the slideshow below you can see some comparison screenshots, with KitKat on the left and Lollipop on the right.

These screenshots show very little difference between KitKat and Lollipop on the Note 4. The most noticeable change is the implementation of Lollipop’s new notification system. On the TouchWiz lock screen, you can only have two Lollipop notifications, with additional ones shown as icons in an overflow section below.

The homescreen, icons, quick settings buttons, task switcher, and menus have not changed much. However, Samsung has tweaked the appearance of some system fonts, controls, and apps, such as Phone and Messages, to be more inline with Google's Material Design, giving TouchWiz a cleaner, less crowded look, with nicer color choices. The TouchWiz launcher did not not get the same streamlining treatment the Galaxy S6 received, however.

One-Handed Features

One-handed use of the Galaxy Note 4 can be problematic due to the dimensions of the screen, but the Note 4—like the Note 3—offers a couple of software features which can greatly increase usability with a single hand. The "Reduce screen size" option is aptly named, shrinking the size of the entire screen into something more reasonable for one handed navigation. Once enabled, you can swipe your thumb quickly from one edge of the screen to the middle and back to the edge to shrink the screen. Repeating this gesture, or tapping the button in the top-right corner of the screen, restores the screen to its original size. This miniature view can be resized by dragging the top corner or repositioned by holding the top bar and dragging. There's even soft-key versions of the Android nav buttons and volume controls located below the window, putting them within easy reach. App compatibility in this mode appears to be a non-issue, as the entire display is simply scaled down.

Another one-handed feature shrinks some onscreen controls, including the dialing keypad and in-call buttons in Phone, the Samsung keyboard, the Calculator, and the unlock pattern on the lock screen. The shrunken controls can be moved to one side of the screen or the other for ambidextrous use.

The side key panel is a small fly-out menu that holds up to four shortcuts chosen from the following list: recent apps, Home, Back, more options menu, reduce screen size, and app drawer. The panel can be repositioned and snaps to either the left or right screen edge. It autohides when not in use and appears as a small tab on the edge of the display. Swiping over it exposes or hides it from view.

Multitasking with Multi Window

Samsung's Multi Window feature makes good use of the Note 4's large display for multitasking. A Multi Window compatible app can be docked to one half of the screen with another app docked to the opposite side, much like Windows 8 apps, in either portrait or landscape orientation. The divider between the two apps is adjustable to show more of one app and less of the other. You can also open compatible apps in a pop-up view, where the active app appears in a resizable and movable window that hovers above apps in the background. The pop-up views can be minimized to a small, circular icon that remains visible at all times.

Apps can be launched in Multi Window view from the recent apps menu by clicking the stacked rectangle button next to the close button. You can also enable the Muti Window tray, which is activated by pressing and holding the Back button. Dragging and dropping an app from the tray to the main window opens it in split-screen view. Tapping the app icon will open it as a pop-up window.

Multi Window mode works reasonably well, although there are some app compatibility issues. Expect a learning curve with the Multi Window features, as well as a period of assimilation, since it will not feel natural until you’ve used it for awhile. Once you get the hang of it though, there are some compelling use cases. For example, you can have Maps open in the background for walking directions and the camera open in a pop-up window for quick access. You could also take notes while watching a video or surfing the Web, even copy text or an image from one window to another.

S Pen

One of the defining hardware features of the Galaxy Note line since the beginning has been the S Pen. Much more than a simple stylus, the S Pen integrates tightly with the Note 4 both at a hardware and software level. Simple actions like removing the S Pen from its silo will activate the Air command menu, which gives you quick access to some of the more common tasks you can perform with the stylus. The S Pen is so essential that the Note 4 gets separation anxiety if you walk away and forget it, alerting you to go back and retrieve the accessory. This attention to detail is apparent throughout a wide range of use cases for the S Pen and the Note 4 as a whole.

Many of the S Pen features revolve around OCR (Optical Character Recognition), the software that converts written text into digital type. The Note 4’s OCR software runs solely on the internal hardware—no network connection required—good news for privacy proponents and users who frequently function in Airplane Mode or other scenarios with limited connectivity.

The active nature of the S Pen allows you to interact with the display without physically touching the stylus to the screen. Samsung calls this Air view, and it is off by default. Hovering the stylus over the display will present a cursor that can be placed over certain hotspots, such as links, calendar items, or images in your gallery, to show a preview of the contents. The improved S Pen now recognizes 2048 pressure levels, which makes for a smooth writing experience and is particularly useful for drawing applications.

The S Pen overall is very usable from all perspectives. Though the cross section of the stylus is rectangular, the rounded edges and a subtle texture make it both comfortable and easy to use. The action button and pressure sensitivity are very responsive, and inking with the S Pen is very precise with no noticeable lag.

The Air command menu (activated when the stylus is removed or the stylus button is pressed) gives you quick access to four common tasks: action memo, smart select, image clip, and screen write. Action memo is quite versatile, allowing you to jot down an email address, phone number, or street address and place a phone call, start an email, or get directions to a location with just a few taps. Action memo items can be saved to S Note or pinned to your home screen for easy access later.

The remaining three options in the Air command menu are more focused in their use cases. Smart select can be used to make a rectangular selection of whatever is on the screen, either saving the image or converting detected text into easily usable characters. Image clip is also used to capture a region of the screen, either using a full rectangle, an ellipse, or a freeform selection. Screen write performs a screen capture of the full screen and imports it directly into an editor where you can mark up the image with notes and then share it through another app or service.

S Note

The utility of S Note is more than simply capturing thoughts and saving them to the cloud. S Note offers templates—everything from schedules to recipe cards—with which you can initiate a new note using the format that best fits your needs. If you find it difficult to write small enough to fit the given space in your chosen template, S Note can provide you with a zoomed in view, making it much easier to achieve the desired level of accuracy within the required space. Handwritten text can then be selected and converted to digital type as desired. Notes can be shared or exported using a number of different formats including S Note, PDF, text, or an image.

S Health

One of the more recent developments in the personal technology arena is the rise of wearables such as smart watches, fitness bands, and the like. While there is no shortage of devices that can track your activity and use that data to provide feedback on your health (Samsung already has multiple wearable products), this functionality is starting to make its way to smartphones in the form of pedometers, heart rate sensors, and similar technology.

Samsung S Health interfaces with the Note 4's pedometer and heart rate sensor as well as other fitness oriented wearables like the Galaxy Gear Fit (see middle screenshot above) in order to monitor sleep patterns and other health related measurements. S Health also supports the use of the heart rate sensor in order to calculate measurables such as stress level or SpO2 (oxygen saturation in your blood) levels. You can also manually enter other data pertaining to food consumption or your current weight.

CPU And System Performance

In this section, we evaluate system-level performance by running a series of synthetic and real-world workloads, along with some browser-based Web tests. There are several facets to overall device performance, including single- and multi-threaded CPU performance, memory and storage speed, and GPU rendering, all of which will be probed by our suite of benchmarks. If you're interested in learning more about how these benchmarks work, what versions we use, or our testing methodology, please read our article about how we test mobile device system performance.

The Galaxy Note 4 uses the same Snapdragon 805 SoC as the Nexus 6, so we should see similar levels of performance. In Basemark OS II, they do indeed perform about the same. The Exynos 7420's four Cortex-A57 CPUs in the Galaxy S6 outperform the Note 4's older Krait 450 cores by 34% in the single- and multi-threaded CPU System test despite Krait's clock speed advantage. Interestingly, the Adreno 420 GPU in the Note 4 offers similar performance to the PowerVR GX6450 in the iPhone 6 Plus and the Mali T760MP8 in the GS6 for this OpenGL ES 2.0 based graphics test.

The Note 4 performs well overall in AndEBench, pulling ahead of the Nexus 6 in the 3D Graphics and CPU-centric Platform tests. In CoreMark-HPC, another test for single- and multi-threaded CPU performance, the GS6 once again outperforms the Note 4 by 31%. The GS6 also outperforms both Snapdragon 805 devices in the Memory Bandwidth test even though they all have the same theoretical max bandwidth of 25.6 GB/s.

We previously discussed the Nexus 6's storage performance issue, which results from using full disk encryption. It's no surprise then to see the Note 4 pull well ahead of it in this test. It is a little surprising to see it match the performance of the UFS 2.0 based NAND in the GS6 though.

In Geekbench, both the Note 4 and Nexus 6 perform virtually the same. Compared to the Snapdragon 801 phones, the Note 4's Krait CPU is clocked 8% higher and performs 6% to 14% better in the two single-core CPU tests depending on which device we're comparing. There's a wider 20% gap in memory performance over the 801, but in everyday use the Snapdragon 805's advantage will not really be noticeable.

On the surface, the Cortex-A57 CPU seems to give the GS6 a noticeable advantage over the Note 4 in single-core performance, posting a 50% higher integer score and a 34% higher floating point score. Remember however, that the A57 uses the newer 64-bit AArch64 ISA, which includes additional SIMD cryptography instructions. This was discussed in our Snapdragon 810 Performance Preview, where we broke down the performance gains in each Geekbench Integer subtest. If we remove the encryption tests, notably for AES and SHA1, then the GS6's advantage in single-core integer performance drops to 20%. The floating point delta stands, since it does not include any encryption tests.

Turning away from the synthetic benchmarks to a more realistic workload, the GS6's performance advantage drops to 13% overall. Breaking this down further shows the newer GS6 with a substantial 64% advantage in the Web Browsing test and 26% in the Writing test. The Note 4 manages to best the GS6 in the Video Playback test however.

Performance between the Note 4 and Nexus 6 is pretty close, with the Note 4 holding a slim lead in most workloads. Compared to the previous generation Snapdragon 801 in the LG G3, the Note 4 performs 25% better overall. The 805's advantage diminishes when compared to other 801 devices however: 20% better than the Sony Z3, 15% better than the HTC One (M8), and only 2% better than the Galaxy S5. Differences in CPU governor behavior accounts for the variation. For the G3, LG shows a preference for battery life over performance (likely to offset the effects of the QHD display) maintaining a lower average CPU frequency than the higher performing Galaxy S5.

There's no real performance difference between the Note 4 and Nexus 6 for web browsing. The performance delta between the Note 4 and the GS6 varies between essentially nothing in JSBench to 26% in Browsermark.

The Note 4 is a bit slower than the Nexus 6 when it comes to scrolling a web page and UI responsiveness in general. Where browser scrolling is mostly a smooth affair on the Nexus 6, there's definitely some stutter with the Note 4. When scrolling within a web page, the Note 4 sets the CPU frequency to 1190MHz compared to 1497MHz for the Nexus 6 running Android 5.1 (it used 1728MHz with Android 5.0). The two devices also use different touch input boost frequencies: 1267MHz for the Note 4 and 1497MHz for the Nexus 6. The smoother experience on the Nexus 6 comes at the expense of battery life however.

Overall, the Galaxy Note 4's system performance is very good, matching or slightly exceeding the performance of the similarly equipped Nexus 6. Since the Krait CPUs in Snapdragon 805 receive only an 8% bump in frequency, we don't see a substantial difference in performance between the Note 4 and the older Snapdragon 801 phones. And while the newer Galaxy S6 definitely offers better performance, the Note 4 really is not that far behind.

GPU And Gaming Performance

Mobile GPU performance is becoming increasingly important as people begin to see their phones and tablets as portable gaming machines. This section explores GPU performance with several synthetic and real-world game engine tests. To learn more about how these benchmarks work, what versions we use, or our testing methodology, please read our article about how we test mobile device GPU performance.

The Adreno 420 GPU in both the Galaxy Note 4 and Nexus 6 perform well in this OpenGL ES 2.0 benchmark. Not surprisingly, it pulls ahead of the Snapdragon 801 based phones (Adreno 330 GPU and only 14.9GB/s memory bandwidth) in the Graphics category by 25% to 45% depending on the device. The bulk of this advantage is due to improvements in pixel shading and texturing.

It’s a bit more surprising to see the Note 4 top the Graphics score of the Mali T760MP8 in the Galaxy S6. Both devices perform about the same when focusing primarily on geometry setup. The Note 4 gains its advantage when doing pixel shading and texturing operations.

At the medium quality setting, Basemark X shows a similar 20% to 25% performance delta between the Note 4 and the Snapdragon 801 based phones (The LG G3 is an outlier, with lower than average graphics performance.) The Note 4 can't keep pace with the GS6, falling behind the newer phone by 28%.

At the higher quality setting, the two Snapdragon 805 devices pull ahead of the iPhone 6 Plus, otherwise the results remain the same.

GFXBench Manhattan is an OpenGL ES 3.0 based benchmark that puts an emphasis on pixel shaders. This was one of the things Qualcomm focused on when developing the Adreno 420 and it shows: The Note 4 scores ~60% higher than Adreno 330 equipped phones.

While the Note 4 is competitive with the GS6 in 3DMark: Ice Storm Unlimited, even outpacing the newer phone in pixel shading duties, it falls behind in this more demanding test by 31%. Perhaps the synthetic tests will provide an answer.

The Note 4 experiences some thermal throttling in T-Rex, which is why it falls behind the Nexus 6. Once again, the Snapdragon 805 shows a 25% to 30% advantage over the previous generation 801 (disregarding the G3).

The synthetic tests don't show any particular weak points for the Note 4, its extra memory bandwidth put to good use in the Alpha Blending and Fill tests.

It's curious to see the Galaxy S6 outperform the Note 4 in both Manhattan and T-Rex, but fall so far behind in the synthetic tests. Monitoring clock frequencies shows that the S6 hits its max GPU frequency of 772MHz in Manhattan as expected; however, it only ramps to 700MHz in the synthetic tests and 3DMark. A 10% deficit in clock speed can account for some of the deficit, but not all. Seeing the low scores in Alpha Blending and Fill, I thought the memory bus may not be reaching its max frequency. According to the GPU frequency table, however, the memory bus runs at max frequency as long as the GPU is clocked at 600MHz or above. There's no obvious explanation for this discrepancy.

The Adreno 420 in the Note 4 is not the top dog in the GPU race anymore, but still offers excellent performance running modern games. Rendering enough pixels to fill the QHD display with a complex 3D scene is still too much for the Adreno 420. Fortunately, most games render at reduced resolutions offscreen so this should not be an issue in most cases.

Where this limitation does come into play is when using the Samsung Gear VR. With your eyes so close to the screen, even a QHD resolution on an AMOLED panel is not enough to mask individual pixels. The need for higher resolution panels for VR/AR applications will continue to push GPU development. Until GPUs catch up, the complexity of VR content will be limited.

Battery Life And Thermal Throttling

Battery life may be the most important performance metric for a mobile device. After all, it doesn't matter how quickly a phone or tablet can load webpages or how many frames per second the GPU can crank through once the battery runs down and the device shuts off. To learn more about how we test this critical facet of mobile computing, please read our battery testing methodology article.

The Galaxy Note 4 manages to last an impressive 6 hours 49 minutes in PCMark, currently our best test for real-world battery life. It even outlasts the Nexus 6 despite using the same SoC. One reason for this has to do with CPU frequency. In the chart above, the Nexus 6 was still running Android 5.0, where it maintained a higher average CPU frequency, even holding two cores at max frequency (2649MHz) for the Writing test. The Note 4's CPU governor, in comparison, allows the frequency to bounce around, rarely going above 1497MHz.

Having full disk encryption enabled and handled solely in software also hurts performance and battery life on the Nexus 6. The PCMark workloads have several small read and write operations that add up over several hours, particularly in the Writing and Photo Editing tasks. The Android 5.1 update for the Nexus 6 disables the CPU thread migration boost feature, improving battery life to 354 minutes but slightly lowering performance.

At first glance, the Note 4 appears to do well in the GPU/gaming focused GFXBench battery test, lasting longer than the Nexus 6 and iPhone 6 Plus. It's not until we look at the performance chart that we understand why. The Note 4 shows a 40% reduction in performance relative to the Nexus 6 due to thermal throttling. With the GPU frequency scaled back, it uses less power and lasts longer.

Looking at the battery drain and performance graph above from the GFXBench battery test, we can see the Note 4 throttle back to less than 50% about 8.5 minutes into the test before recovering to about 70% of the original value over the last half. Heat dissipation is a design weak point, which is evident in the thermal image of the back cover taken during the same battery test. It's certainly obvious where the SoC is located (the yellow circle is the rear camera and the green rectangle in the upper-right is the battery). The chassis does a poor job of spreading and dissipating the heat generated by the GPU, making the Note 4 more susceptible to thermal throttling.

You should not have an issue getting through the day on a single charge with the Note 4; however, it does have a power saving mode that restricts background data transfers and performance, even an optional grayscale display mode, for stretching battery life a little further. If you're really desperate, the ultra power saving mode limits what apps can be used and switches to a simple black-and-white display mode.

The Note 4 comes with Samsung's Fast Charge feature, which is compatible with Qualcomm’s Quick Charge 2.0, that can charge the battery to 50% in about 30 minutes, according to Samsung.

Conclusion

Samsung’s Note series has gained a loyal following by combining premium hardware with phablet defining features. They aren’t flashy or hip or trendy. They are meant for power users and professionals who use their phones for getting real work done. Sure, you can still have fun with a Note too, but it’s not a phone designed for the masses.

The Galaxy Note 4 extends the Note tradition as a flagship phablet starting with its most recognizable feature, the 5.7-inch SAMOLED display. For this generation, Samsung bumps the resolution from a so-last-year 1080p (386 PPI) to 1440p (515 PPI). Some may question the need for so many pixels, but for AMOLED screens, whose PenTile matrices have a subpixel deficit compared to LCD panels, this is a good thing.

Looking at all of these pixels is a real joy thanks to a well-calibrated, proper sRGB display mode, making this one of the best looking screens currently on the market. We also appreciate Samsung continuing to provide multiple display modes for people who prefer the more vibrant and saturated colors that result from using an extended gamut. Choice is a good thing.

Surrounding the screen is an aluminum frame, a new material for the Note line. While the overall appearance remains similar to the Note 3, the metal frame, painted with a color-matched finish and highlighted by polished, chamfered edges, looks classy and gives the Note 4 a solid feel, free from any flexing or creaking.

The plastic back panel is still removable, providing access to the removable battery and SD card slot. A feature sure to make road warriors happy. It retains the faux leather finish of the Note 3, but the fake stitching is gone.

While no longer the fastest SoC on the block, the Snapdragon 805 packed inside is still very quick and capable. It’s paired with 3GB of RAM and 32GB of speedy internal storage. A category 4 LTE modem from Qualcomm keeps the data flowing.

The camera in the Note 4 is one of the best currently available. It's 16 MP Sony sensor with OIS delivers high-resolution images of good quality, and the phase detection autofocus is very fast. We occasionally saw small misses in white balance leading to images with a slight green cast, and the iPhone 6 Plus still holds a small edge in some lower-light scenarios. However, Samsung is the current leader when it comes to HDR, providing a live preview onscreen and producing great results with virtually no processing lag.

S Pen is another notable feature. Effectively doubling the number of pressure sensitivity levels from the Note 3, the S Pen delivers a smoother and more natural writing experience. The Pen is more than just a simple stylus, however. The integrated Wacom digitizer enables additional features like being able to hover the pen over the screen to show context sensitive menus. TouchWiz also makes good use of the pen via the Air Command menu that opens when it’s removed from the silo, providing quick access to common tasks, and functionality is extended through apps such as S Note and integrated OCR.

The Note 4 now runs Android Lollipop, but doesn’t use the more refined TouchWiz UI that comes with the new Galaxy S6. Although there are still a few rough edges when it comes to UI design and functionality, Samsung provides additional software features that make good use of the bigger screen. The ability to shrink the display size, place input controls within easy reach, and add shortcuts to the side key panel make it possible to still interact with the phone using a single hand when necessary. Multi Window is great for multi-tasking and receives some usability improvements for this generation.

The Note 4 is a great phone, but we do have a few minor quibbles. Despite the fast processor, browser scrolling and UI interactions still exhibit some stuttering due to Samsung’s conservative CPU governor settings. This does improve battery life though, so this might not be a negative depending on your priorities. The Note 4 is also more susceptible to thermal throttling than other phones when the GPU is pushed hard. Samsung needs to do a better job using the metal chassis to spread and dissipate heat. The rear mounted external speaker is less than ideal, reducing audio quality and producing muffled sound when sitting on a table.

Despite these few minor flaws, the Note 4 is a powerful phone, whose hardware and software features set it apart from all the other phablets flooding the market. For this reason, the Samsung Galaxy Note 4 earns our Editors’ Choice award.

Tim Ferrill is a Contributing Writer for Tom's Hardware. Follow him on Twitter.

Follow Tom's Hardware on Twitter, Facebook and Google+.

Mediatek announced its next-generation high-end SoC, the Helio X20. The chip is unique in its design, as it features not one, not two (like big.LITTLE), but three CPU clusters for a total of 10 CPU cores. Indeed, the Mediatek Helio X20 is the world's first deca-core mobile chip.

The Helio X20 is also Mediatek's first Cortex-A72-based chip. Cortex-A72 is ARM's next-generation high-end CPU core and successor to the Cortex-A57. It's more efficient and more powerful than its predecessor, and in the case of Helio X20, will be built on TSMC's 20nm process.

The Helio X20 chip comes with a dual-core 2.5 GHz Cortex-A72 cluster, a quad-core 2.0 GHz Cortex-A53 cluster for medium tasks, and another quad-core 1.4 GHz Cortex-A53 cluster for lighter and on-going tasks.

The Mediatek Coherent System Interconnect and the CorePilot 3.0 advanced scheduling algorithm are used to manage the three clusters and ensure that the right cluster is used at the right time and for the right task. Otherwise, the whole system could become inefficient, either in terms of power consumption or in terms of performance.

Mediatek seems to believe that different CPUs should be used for different tasks on a mobile device. For instance, a higher-performance mobile CPU should be used for gaming in order to provide strong graphics performance and an enjoyable user experience in general -- even at the cost of lower battery life.

Mediatek is running the high-end Cortex-A72 cores at 2.5 GHz each, which seems quite high. Even if Cortex-A72 is significantly more efficient than Cortex-A57, when playing games, those cores will have to run close to their peak performance and for a relatively long time (in mobile time). It will be interesting to see whether the chip can handle such high performance on a 20nm process.

Mediatek said it wanted to have a tri-cluster CPU because the dual-cluster big.LITTLE setup was just not enough -- neither in low-power nor peak performance.

This is an interesting claim because the Helio X20's low-end cluster also doesn't max out at a very low frequency. The lowest (maximum) clock speed of its clusters is 1.4 GHz, which historically is quite high for big.LITTLE's low-end cluster. That's without even counting that the Cortex-A53 is also slightly more powerful and power-hungry than its predecessor, the Cortex-A7. Even so, Mediatek showed that its tri-cluster CPU managed to improve on a dual-cluster design in terms of power consumption.

It's true that we haven't seen mobile chips go to this level of performance at the high end before, in part because Cortex-A72 is a brand-new next-generation core, and in part because even Samsung's 14nm chip only goes as high as 2.2 GHz. Therefore, Mediatek may indeed have the highest-performance big.LITTLE-like chip yet.

One of the reasons we're starting to see ARM chip makers push beyond even eight-core chips now is because ARM cores are so cheap. Adding a couple of extra cores doesn't add too much to the cost of the SoC, especially when we're talking about tiny Cortex-A53 cores.

This is not something other chip makers such as Intel can easily do, even with its mobile CPUs. Every extra core costs significantly more money, which is why Intel's mainstream PC chips are mainly dual-core and why its mobile Atom chips are still only quad-core.

Instead of creating a cluster that's highly optimized for a certain class of applications, Intel is likely to continue with its Turbo-Boost strategy (low base clock speeds, temporary high speed bursts). This has the advantage of keeping costs the same while also continuing to increase the burst speed of the CPU, which in turn helps it solve some tasks more quickly and perform well in most benchmarks.

However, it also means that the "low-end" tasks will be served by the same "wide" microarchitecture, which is less efficient than using a simpler CPU, even when at a low clock speed.

The Turbo Boost strategy also has a major disadvantage in gaming, because in games you're only really going to get the base clock speed, maxed out. The chip will not be able to keep Turbo Boost enabled for long. If you see a chip that promises 1.2 GHz base speed with 2.5 GHz Turbo-Boost, the lower speed is what games should be using over the period of time you're playing the game.

Mediatek is also pushing the chip's multimedia capabilities and its overall performance. The Helio X20 will support EIS, OIS and AIS (Automatic Image Stabilization) stabilization technologies for video and images, 120 Hz displays (as part of a new trend that's also going to be great for VR headsets), and advanced vision processing by CPU/GPU and ISP heterogeneous computing.

Mediatek's third major announcement for the Helio X20 was its own LTE modem, which supports worldwide coverage. It's a Cat. 6 LTE modem with download speeds up to 300 Mbps and 20+20 (MHz) carrier aggregation.

Mediatek's Helio X20 SoC is expected to ship in products by the end of the year. Most should only see it in Q1 2016, as the chip is only starting to sample in Q3 2015, and it takes a couple of quarters to reach the market.

The Helio X20 makes many promises, and it may even end up fulfilling most of them. However, it's uncertain how it will compete next year against 14nm and 16nm FinFET chips. If Mediatek would have been more aggressive about the process node it chose, it would have had a chip that competes not just against "2nd-tier" high-end chips, but the highest-end mobile chips on the market. At 20nm, that's unlikely to happen as it won't be able to beat an Exynos 8 or a Snapdragon 820/830 built on 14/16nm FinFET.

Either way, Mediatek has been making strong progress with its chips in a few short years, going from making SoCs only for low-end smartphones and tablets to competing with the best in the market. Mediatek may just rise to become the third-party chip making competitor Qualcomm needs (unless Samsung also decides to start selling its chips to others, but that seems unlikely for now).

Follow us @tomshardware, on Facebook and on Google+.

Samsung's $14.3 billion (15.6 billion won) chip plant, which is supposed to help the company increase revenues as smartphone sales slow, will begin production in the first half of 2017, according to Reuters. The new plant will be built in South Korea in the city of Pyeongtaek, south of Seoul.

The company expects to spend another $9.2 billion (10 billion won) investing in additional capacity.

"The Pyeongtaek semiconductor plant will play a central role in solidifying leadership in the mobile and server markets, which have shown rapid growth in demand recently, and securing share in the next-generation internet of things market," Samsung said in a statement.

Samsung will mostly produce DRM chips at the new plant, but some capacity will also be allocated to mobile processors depending on demand.

Samsung is the world's market leader for mobile DRAM with around 40 percent market share, and it's the second-ranked chip maker by revenue, after Intel. The company has seen increasing success lately with both its foundry and its application processor businesses.

With the success of the Galaxy S6, the Exynos chip has become more popular, as well. This has given Samsung a reason to keep investing significant resources into this AP division to continue improving the Exynos line of chips.

The Exynos chips will save Samsung money by not having to buy high-end SoCs from another company. It also gives Samsung the opportunity to start selling the chips to other OEMs. However, that's unlikely to happen for now, as Samsung will probably want to keep the Exynos chips for its own products as some sort of competitive advantage and a differentiating factor.

The Exynos 7420 was already built on the company's own 14nm FinFET process, while other chip makers such as Qualcomm had to compete on a 20nm planar process and wait another cycle until Samsung freed up its foundry for customers.

On the foundry side of the chip making business, Samsung became the leader in process technology for third-party SoC makers who are looking for foundries in which to manufacture their chips. Because of that, now Apple wants to make its chips in Samsung's foundries as well, giving Samsung even more of a reason to want to put more money in its chip plants.

We have reached out to Samsung for comment and confirmation, but thus far our queries have not been filled.

Update, 5/8/15, 6:20am PST: Samsung has confirmed to Tom's Hardware that the company's new $14 billion chip plant will begin production earlier than originally planned, in the first half of 2017. Samsung hasn't yet decided what specific products will be manufactured there, but the company expects the new chip plant to help it "solidify its industry leadership in the expanding mobile and server sectors, as well as respond to growing need for semiconductor components in the Internet of Things (IoT) marketplace."

Follow us @tomshardware, on Facebook and on Google+.

Intel announced that it's going to change its financial reporting structure by combining the operating results of its Mobile and Communications Group and its PC Client Group. From now on, Intel will be reporting its financial results only for the combined Client Computing Group, rather than report separately its mobile and its PC results. This change will be applied starting with the next earnings report on April 14.

According to the company, the new group was created to "address all aspects of the client computing market segment and utilize Intel's intellectual property to offer compelling customer solutions."

That makes some sense, as Intel, much like Microsoft and other tech companies, are now trying to unify their mobile and PC platforms. However, one consequence of this is also that shareholders as well as others may not be able to see as clearly how much money Intel is losing in mobile anymore.

Intel has been losing roughly $1 billion per quarter for the past two years (about $8 billion in total). That's an incredibly large sum of money for any of its competitors, whether we're talking about AMD, Nvidia, Qualcomm or other ARM chip makers. Those sort of losses would be unsustainable for them, but because Intel still makes so much money from its PC and server chip divisions, the company is able to eat that cost.

High Cost Structures

Why is Intel losing so much money in the first place? There are multiple reasons. One is because Intel used to be primarily a PC chip company, with higher cost structures. Higher cost structure basically means "cost bloat." That's not necessarily a bad thing in and of itself. PC chips have evolved a certain way and from a certain technology base (CISC) that has led to an increased base cost for designing such a chip, at least compared to ARM mobile chips.

Intel's x86 chips still contain some CISC baggage compared to RISC chips, which is made evident by the fact that ARM chips on one- or two-generations older process nodes can still be competitive against the latest Atom chips. If ARM chips were actually built on the same Intel process node, they would likely be far ahead of Atom chips.

To be competitive with ARM chips, Atom chips need "something more" such as other instructions, or a newer more expensive process node, and so on. All of that adds to Intel's cost structure for designing a mobile chip.

Intel has also enjoyed significantly larger profit margins than mobile chip makers. While Intel's profits generally measure in tens or even hundreds of dollars per chip unit, ARM chip makers usually measure their profits in single-digit dollar figures or low two-digit numbers at best for every chip unit. When Intel applies those high cost structures and profit margins to its mobile chips, it discovers that it can't be competitive in mobile with its "normal" price points.

This happened before with Intel's XScale division, and it's likely the reason why it was forced to eventually sell it and get out of the mobile market. Even though there are plenty of profitable (ARM-based) mobile chip makers in the industry, Intel couldn't be one. If it made any profit, it likely wasn't sufficient for a company such as Intel to justify the existence of its mobile business.

Heavy Subsidies

Another reason for why Intel is losing so much money is because to get to the price point where it's competitive against high-end ARM chips, the company needs to heavily subsidize its Atom chips.

An Atom chip would normally sell for over $50, but Intel can't compete at that price point (in the mobile market) because that's roughly twice as much as most ARM chip makers charge a similarly-performing high-end chip. Intel wants to be in the mobile market, so at least for now it's willing to eat that subsidy.

However, the ideal price point for Atom in the mobile market would be somewhere close to $80, if not more. Intel has already begun selling Atom-based chips under the brand name of Celeron and Pentium for $82 to $161. Intel doesn't seem to want to sell Atom chips as low as $30 (or less), which leads us to reason number three.

Pricey Partnerships

A third reason why Intel is losing money in mobile would be that the company is investing heavily in other areas of its mobile chip business to get into the market. That includes partnership programs (making it easier for OEMs to use Atom chips in their devices, paying for branding, etc.) or investments into other mobile chip companies in order to get them to use its Atom design in their SoCs.

This third reason is actually a consequence of the first (high cost structures). Instead of trying to sell $30 chips on its own at a loss, the company would rather license the Atom design to makers of inexpensive SoCs, such as Rockchip and Spreadtrum.

Intel seems willing to do this even though the royalty from those chips will at best gain Intel only millions of dollars a year (even ARM itself doesn’t make that much per year, and its IP dominates the mobile market). However, millions of dollars in profit is still better than billions in losses.

Another important thing to note here is that Intel is also outsourcing one of its core competencies, that of using its advanced process technologies to third-party foundries such as TSMC in order to build those inexpensive Atom-based Rockchip SoCs.

Intel wouldn't do this if it could profitably build its own Atom chips for the low-end on its own 14 nm, 22 nm or even 32 nm process nodes. Instead, it seems the better outcome is to have partners such as Rockchip and Spreadtrum build mobile chips on a competitor's 28 nm planar process technology (TSMC).

In its recent press release, Intel said it would post an $800 million turnaround in mobile this year. (That means posting a $200 million loss instead of $1 billion.) However, now that Intel is combining its reporting, it may be more difficult to parse out how mobile is doing compared to PCs.

If the company was losing $1 billion per quarter for the past two years, it must have had some big money-bleeding programs that it's now willing to cut. Those programs could include the subsidies, the partner sponsorships and so on. 

So the question is, with those programs gone, will that hurt Intel's growth in the mobile market, or is the company confident enough that its chips are on an unstoppable growth path regardless of their lack of subsidies?

Does Intel have another strategy up its sleeve that can offset $1 billion in dollar per quarter in mobile investments? Does that include counting the new $82 to $161 Atom-based Celerons and Pentiums, which mainly will be used in low-end PCs and Chromebooks, as mobile chips?

If that's so, that would explain a rapid shift from billion dollar losses to something closer to profitability, as the company could essentially shift the extremely high profit margins from $161 Atom chips to the subsidized $30 Atom chips for smartphones and tablets. It's not clear how exactly Intel is going to make up that $800 million this year, which is why these questions still remain unanswered.

Regardless of whether Intel will actually post better financials in mobile soon or not, it's obvious the company wants to play this game for the long haul. Uniting the mobile and PC divisions could create some synergy, making it easier to transfer technology between the two, but the effects of that probably won't be seen for another few years.

Follow us @tomshardware, on Facebook and on Google+.

In addition to the announcement surrounding Braswell SoCs, Intel had a boatload of news to share out of IDF Shenzen, primarily surrounding the company's global mobile strategy. Intel's SoFIA chips got the LTE-TDD treatment, meaning they'll be compatible with China's mobile networks, and the SoFIA roadmap has been expanded to include IoT applications.

There will also be a new RealSense camera coming, and this one will be available on smartphones.

SoFIA Goes To China

Intel made a strong play into the mobile market (particularly for emerging markets worldwide) with its Atom x3 (SoFIA) and Atom x5 and x7 (Cherry Trail) mobile chips at Mobile World Congress last month, and we're already seeing the platforms evolve.

First, Intel demoed an Atom x3 LTE (SoFIA) chip running an actual smartphone, equipped with LTE-TDD to work in China. The demo ran over China Mobile's network and is expected to be available 2H15. Intel did not specify which chip it was exactly, but we presume that it's the Intel Atom x3-C3340.

The quad-core 1.4 GHz, 64-bit x3-C3340 is built for entry-level and value smartphones and has a Mali-T720 MP2 GPU (which supports Full HD video, OpenGL ES3.0 and OpenCL 1.2). It offers LPDDR2/3 memory, eMMC 4.51 storage and has WiFi, Bluetooth, GLONASS and NFC for connectivity. Intel previously stated that the x3-C3340 should offer double the media editing performance of competitors' similar SoCs.

Already a launch partner, Rockchip CEO Min Li gladhanded with Intel CEO Brian Krzanich on the IDF Shenzen stage to promote the Intel Atom 3G-R (aka the x3-C3230RK, the Rockchip variant of SoFIA). He said that 10 or so ODMs are working on products running the 3G-R chip.

The x3-C2320RK is a 64-bit quad-core SoC clocked at 1.2 GHz with a Mali 450 MP4 GPU intended for lower-end tablets, phablets and smartphones and will be in products on the market worldwide in 2H15. Take "worldwide" with a grain of salt, though; the U.S. is most likely not going to see any of these products.  

Atom x3 And The Internet Of Things

Although in retrospect this makes a great deal of sense, it was perhaps a bit of a surprise to learn that Intel has expanded the Atom x3 roadmap to include the Internet of Things (IoT).

Intel is developing purpose-built Atom x3 SoCs to bring 3G and LTE capabilities, and their low power demands, to IoT for applications such as telematics, dash cams, fleet management, point-of-sale (POS), and specially-designed tablets for industry and healthcare. The idea is that the scope of potential applications could reach all the way to the connected home.

With the new Atom x3 chips for IoT, Intel promised longer life; the parts are guaranteed to last seven to 10 years. Further, they can handle operating conditions from -40 to 85 degrees C, and endure through all sorts of different weather conditions.

Smartphones And RealSense

There was but a quick mention of this last point, but Intel's Krzanich also showed off a new RealSense camera mounted into (onto?) a 6-inch smartphone. Intel said that the new RealSense camera will offer "longer range" than its predecessor.

It's difficult to say what that means, exactly; presumably, he's talking about the RealSense SnapShot Depth camera, which first rolled out on the Dell Venue Tab 7000. You can read our hands on with the camera here. Note that of the issues we examined, range was not one of them. However, it's possible that an increased "range" will lead to more accuracy and consistency.

In any case, the fact that Intel has managed to fit a RealSense camera into such a small device (6 inches versus the Venue Tab 7000's 8-inch form factor) is a notable achievement.

Seth Colaner is the News Director for Tom's Hardware. Follow Seth Colaner @SethColaner. Follow us @tomshardware, on Facebook and on Google+.

HTC recently launched the HTC One M9, but apparently, the company wants to follow Apple's strategy and release a bigger model as well. HTC will release the larger HTC One M9+ in China for now, but the company doesn't plan to launch it in Europe or North America.

What's different between the HTC One M9 and the HTC One M9+? For one, the M9+ comes with a slightly larger 5.2" screen, as opposed to the 5" screen from the M9. If you thought the 1080p resolution was already too low for a 5" flagship screen in 2015, then you won't be disappointed to hear that the HTC One M9's 5.2" screen has a 2560 x 1440 resolution (the same as the Galaxy S6 and the LG G3).

There have been many reports that the Snapdragon 810 in the HTC One M9 and other devices overheats, which is why it seems HTC went with Mediatek's latest flagship chip, the Helio X10 (Mediatek MT6795T). This chip is based on the 64-bit ARMv8 architecture and comes with eight Cortex-A53 cores with a frequency of 2.2 GHz.

The chip should be roughly as powerful as the Snapdragon 810, but it should also be much more efficient. That means the Mediatek MT6795T should throttle much less as well, allowing for higher sustained performance. Two of its unique features are support for 480 fps slow-motion video recording and 120 Hz screen refresh rates, which is twice as much as what other chips on the market support currently.

Of course, HTC also loaded this with its "BoomSound" speakers that are paired with Dolby Audio for quality surround sound simulation.

On the back, the HTC One M9+ has a 20 MP "Duo Camera" to give that extra depth in pictures (the second camera is 2.1MP), while on the front it comes with a 4MP "UltraPixel" camera that performs well even in low-light environments.

One disappointing aspect about the HTC One M9 was that it didn't have a fingerprint sensor, even though its main competitors, the Galaxy S6 and the iPhone 6, do. Although fingerprint authentication is not perfect, if done right and if the fingerprints (or hashes of the fingerprints) never leave your device, it should be a much more secure alternative to using no password at all or using weak "1234" PINs.

Fortunately, the HTC One M9+ does seem to come with a fingerprint reader, although of course it remains to be seen just how well it will work. Most fingerprint readers so far haven't had great accuracy, which has led to bad user experiences and therefore low adoption among users.

The device will come with Android Lollipop (likely version 5.0.2, just like the M9) on board along with HTC's Sense 7 customizations, which include HTC Themes and the Blinkfeed feature (news and social media posts). HTC hasn't announced a launch date or what it will cost when it arrives in China.

More on the HTC One M9:

HTC One M9 Debuts, Flashes Style, Promises Performance

HTC One M9, Hands-On At MWC 2015

HTC One M9 Available Tonight, Latest Flagship Goes Toe-To-Toe With Samsung

Follow us @tomshardware, on Facebook and on Google+.

Last year, it was rumored that Samsung would go back to making Apple's next A9 chip on its 14nm FinFET process. This was after TSMC had already won all the A8 and A8X orders and Apple stopped making its chips at Samsung. Later it was rumored that Samsung lost most of the A9 orders, and that TSMC would go back to making at least 70 percent of Apple's new chips.

A new report from Bloomberg said that Apple will use Samsung's foundry to build all of its A9 chips, even after TSMC already budgeted $12 billion to invest in its new 16nm FinFET process technology. The A9 chips will be built at Samsung's Giheung plant in South Korea.

Apple spent $25.8 billion buying its chips last year, which represents 7.6 percent of the industry's total chip purchases, according to Gartner, a market research company. Because Apple ended its chip manufacturing contract with Samsung last year, Samsung saw a loss of almost $1 billion. The new orders from Apple could give Samsung a $1 billion gain this year.

Any additional orders will be handled by its partner Global Foundries, which has licensed the same 14nm process technology from Samsung. The partnership will give Samsung access to Global Foundries' factories in Texas, New York state and South Korea.

“If Globalfoundries quickly adopts Samsung's most advanced technology and increases yield, it could also win orders from Qualcomm."

Qualcomm has also been rumored to use Samsung's 14nm FinFET process for its Snapdragon 820 flagship chip, as well as for other lower-end chips going forward. Samsung recently tried to distance itself from Qualcomm by refusing to use the Snapdragon 810 chips in the Galaxy S6. However, it's likely that Samsung will continue to build Qualcomm's chips in its factories, as that will increase the company's revenues.

It's less costly to build the same chip with one process technology, which is why companies such as Apple and Qualcomm would prefer to build a single generation of chips at the same foundry. Business politics and more aggressive negotiations can sometimes interfere with that, but the way for a company such as Apple to limit those issues is to build one generation of chips at one foundry, and every other generation at another.

This way, both foundries would continue to compete for the same customer, and at the same time Apple wouldn't have to abruptly end its business with either one of them. Apple tends to sell its previous-generation iPhones and iPads for a couple more years after a new generation comes out, which means the foundry building its chips would still continue to receive income from Apple (albeit significantly less for those chip generations).

Apple also tends to release its new iPhones and iPads in the fall, and that's when we should first see its new 14nm FinFET A9 chips. The A8 chip used the second generation Cyclone core, so the A9 may use a redesigned core based on the same ARMv8 instruction set. That means the A9 could benefit not just from the new, more energy-efficient FinFET process, but also from improvements in the CPU's own microarchitecture.

Follow us @tomshardware, on Facebook and on Google+.

We've had x86 Chromebooks and we've had ARM Chromebooks, but soon we may see Chrome OS work on yet another architecture: MIPS. The MIPS architecture, now owned by Imagination Technologies, appears to be supported in Chrome OS in some recent improvements made to Coreboot. Coreboot is Google's open source and lightweight alternative to the proprietary BIOS firmware we see in most computers.

The improvements done to Coreboot for the MIPS architecture mentioned the Pistachio SoC, which comes with a dual-core InterAptiv CPU. These improvements were written by Ionela Voinescu, who works at Imagination. Right now, we only have an indication of support for the InterAptiv core, which is one of Imagination's mid-range CPU lines. So far, Imagination has focused on selling its MIPS CPU designs to the lower end of the market. That's where the company has a better chance to make an entrance and displace a competing ARM CPU by undercutting it on price, on performance/price, or performance/die area.

MIPS Technologies, the company behind the MIPS architecture, had claimed for a long time that MIPS chips are higher performance for the same die area when built on the same process node. That may work in theory, but in practice, MIPS chips usually arrive on old process nodes, which removes their performance advantage. However, Imagination's partners, who build MIPS CPUs, can still use the lower-cost argument against the ARM chip competition.

Chrome OS, being a (mostly) architecture-agnostic operating system, can work on x86, ARM or MIPS chips. Although Google itself has mainly pushed Intel's x86 chips in Chromebooks, the company may finally be more willing to promote ARM or MIPS Chromebooks, as well. Intel already dominates the Chromebook market, and Google probably doesn't want a single company to hold a chip monopoly on Chromebooks.

Chrome OS can do one thing very well, which is run the browser -- very fast. That means it should be able to run well enough even on mid-range mobile hardware, such as Imagination's InterAptiv CPUs.

Imagination hasn't yet announced the next generation of its high-end MIPS64-based CPU cores for mobile, which could also be used for higher-performance Chromebooks. Those CPUs should offer better competition to ARM's Cortex A57 or Qualcomm's Kryos core. Just how much competition will also depend on the process node Imagination's partners will use.

Follow us @tomshardware, on Facebook and on Google+.

Over the past few years, Apple has tried to distance itself from using Samsung as a supplier, which makes sense considering that Samsung was also Apple's main competitor in the mobile market. Apple didn't want Samsung to gain any advantages it might not get otherwise if the company wasn't its supplier.

Apple managed to manufacture 100 percent of its chips on TSMC's processes as recently as last year. However, earlier this year, there were some rumors that Apple may go back to Samsung and its 14nm FinFET process, which is ahead of TSMC's 16nm FinFET process both in terms of performance and time to market speed. 

The Exynos 7420 chip inside the Galaxy S6 is already built on Samsung's 14nm process, while no TSMC 16nm chip is expected to come out until Q4 this year. That wouldn't be a major issue for Apple, though, as that's when the company plans to release its new iPhones and iPads anyway.

New reports say that Apple will continue to manufacture the bulk of its A9 and A9X chips at TSMC's foundries. As many as 70 percent of the chips are expected to be built on TSMC's 16nm process, with the A9X (the iPad chip) to be built exclusively on TSMC's process.

"We believe TSMC will earn most of the A9 orders thanks to its superior yield ramp and manufacturing excellence in mass-production. We expect TSMC to earn all of Apple's A9X orders (for the next generation iPad) and most of the A9 (for the next generation of iPhone), aggregating to an allocation of over 70%," said analysts from Daiwa Securities.

Even if TSMC's process has slightly lower performance than Samsung's 14nm process, that shouldn't be much of an issue inside the much larger iPads, compared to the smaller space inside the iPhones. However, considering the new report said that up to 70 percent of the chips will be built on 16nm, it's possible that Apple may build the A9 chip that goes inside the iPhone 6S Plus, as well. The iPad market is shrinking, and it was never that large compared to the iPhone market, so at least some of the iPhone chips will have to be built on 16nm in order to reach that high percentage.

Apple doesn't like to rely much on Samsung as a supplier anymore, which is probably why the company intends to use TSMC whenever possible, but this time there may have been a yield issue as well. This was the first time Samsung opted to use Exynos chips exclusively in its high-end flagship and its most popular device, which is already seeing great demand, so there may not have been too much room left on its foundries for Apple's chips.

Apple is also expected to sell many new next-generation iPhones this fall, and it's possible Samsung couldn't cover the manufacturing for both flagship devices in such a short period of time.

Follow us @tomshardware, on Facebook and on Google+.

Last summer, the Khronos Group announced its intention to replace the current OpenGL graphics API with a next-generation API that is much leaner and gives more direct control over gaming hardware such as the CPU and the GPU.

Some gaming developers have complained for years that while graphics APIs such as DirectX and OpenGL can standardize a variety of different CPUs and GPUs, that abstraction layer was too thick, and it significantly reduced the performance of that hardware. This is why it's game developers that have been pushing for an API like AMD's Mantle, which was then given to Khronos to modify and extend and turn it into the next-generation Vulkan API.

Although Khronos has just announced the new Vulkan API at GDC, Imagination has already created a proof of concept driver for its PowerVR GPUs. The company has also ported one of its OpenGL ES 3.0 demos to the new API, which it showed at the event.

As Vulkan is meant to cover all devices from embedded to desktop, it's no surprise that it should work on mobile GPUs as well. In fact, Khronos said that all hardware that supports the OpenGL ES 3.1 API right now will support Vulkan, too. In the demo, Imagination showed the following effects:

High-quality, physically-based shadingHDR (High dynamic range) rendering20 unique 2K PVRTC textures2 GiB of texture data compressed to 266 MiB using Imagination's PVRTC texture compression standard4 x MSAA (Multi-sample anti-aliasing)16 x Anisotropic texture filteringPhysically-correct material parametersLow CPU usage, very efficient GPU usageCorrect specular reflections on reflective materialsMore than 250,000 trianglesPost processing effects: saturation, exposure and tone mapping

Even though the OpenGL ES API was created with efficiency in mind so it's suitable for the mobile environment, the Vulkan API still manages to go beyond that with a much higher CPU efficiency due to the lower overhead and the more direct access to the hardware. Such a leaner API can lead to higher performance in games and drivers that are less complex and easier to write and update than ever before.

Imagination said that higher-level management of the GPU will now have to be performed in the application itself, rather than in the driver, which could be extra work for developers. However, this could be offset by the fact that developers don't have to work around the driver anymore, either. They can tell the GPU exactly what they want it to do. If that application uses an engine to do the rendering, the management will be done by that engine anyway, so most developers can get the Vulkan speed-up virtually for free, with no extra work.

Along with the free performance gain, Vulkan will also give games more consistent performance, which isn't possible right now because different GPU hardware can respond differently to certain GPU commands in OpenGL.

The Vulkan API, unlike OpenGL ES, is much more parallel as well, and it can do work in multiple threads more easily. Rendering commands can now be created on all of a CPU's cores, which can be useful to games that need to recreate their render commands often (games such as Minecraft).

Vulkan also has a much more intuitive and transparent design, which gives developers more control over their own applications.

“Vulkan gives you the advantage of knowing exactly the state that you are setting. Take for example the glActiveTexture() function in OpenGL ES: it is not obvious whether this function will change the state globally for all shaders or maybe change the state just for the current shader program.In Vulkan, this is explicitly defined: you know that when you bind your resources, it is changing the state for the bound command buffer because that is the first parameter to the function."

In OpenGL ES, the memory allocation process is a black box. In Vulkan, the application knows how much and what type of memory it is using. This type of control can help to further optimize resource-heavy applications.

Khronos promised to release the full Vulkan specs later this year and have hardware supporting the new API soon after that. Imagination also said that it will continue to work on porting the Vulkan API to its PowerVR GPUs and release example source code in the near future.

Follow us @tomshardware, on Facebook and on Google+.

According to a report from Taiwan, MediaTek is trying to become Samsung's new chip partner, now that Samsung has begun distancing itself from using Qualcomm's chips. MediaTek didn't say much about its deal with Samsung, but the company's Deputy General Manager, Zhu Shan-gzu, admitted their relationship with Samsung has been improving lately.

"We made progress little by little in some direction," Shan-gzu said.

MediaTek has already announced some new chips at MWC 2015 this week, one of which includes the brand-new Cortex-A72 CPU core, which is the successor to the Cortex-A57 core seen inside chips such as the Qualcomm Snapdragon 810 and Samsung Exynos 7420.

MediaTek has been making bigger advancements with new chips at the high-end of the market, as demand for these chips has increased. In China, MediaTek is already a strong competitor to Qualcomm, because many Chinese manufacturers prefer to use its chips. MediaTek's chips have also appeared in some low-end and mid-range phones from Sony, HTC, Asus and other known mobile companies, although MediaTek's chips are in much fewer smartphone models than those of Qualcomm.

As Samsung is looking to differentiate itself from other smartphone OEMs who tend to default to using Qualcomm's chips, the closeness to MediaTek may become inevitable. Samsung can build its own Exynos chips, but those usually target only the high-end of the market, and only its Galaxy S and Galaxy Note flagships have them.

It's unlikely that Samsung will begin producing chips for all performance and pricing levels anytime soon, so a partnership with another chip maker, like MediaTek, makes sense.

At the same time, this is a big opportunity for MediaTek to conquer more of the western markets, where we're seeing much of Samsung's success. A strong partnership with Samsung could help MediaTek sell many more chips considering Samsung's significant smartphone market share. It would also give MediaTek a stronger entry in the U.S. market where most smartphone customers have gotten used to either Qualcomm chips or Apple's own chips in iPhones.

Follow us @tomshardware, on Facebook and on Google+.

Today, Qualcomm announced that it's working on the next-generation Snapdragon 820 SoC that will include the company's own custom ARMv8 CPU core, named Kryo, and a brand-new cognitive platform called Zeroth. The company also said that the chip will be built on a "cutting edge" FinFET process.

Previously, Kryo was leaked under the codename "Taipan" and as being built on a 14nm process, which implied Qualcomm would use Samsung's process technology, because TSMC will only have a 16nm FinFET process node.

Kryo will be Qualcomm's first custom CPU core since it built Krait, which has been in use since 2012. It's generally considered that Qualcomm, as well as other mobile chip makers, were caught off-guard when Apple announced a new custom ARMv8-based CPU core in 2013, a full year before anyone expected ARMv8 chips to start appearing.

Chances are that Qualcomm also didn't believe it needed to build a 64-bit chip until the end of 2015 or early 2016. But after seeing Apple launch a 64-bit CPU so early, Qualcomm may have reconsidered its position on 64-bit chips and licensed the big.Little Cortex-A53/Cortex-A57 CPU IP from ARM in order to build the Snapdragon 810.

Qualcomm didn't give too many details about its Kryo CPU core today. However, it should be sampled to manufacturers in the second half of the year, which likely means we won't see the Snapdragon 820 on the market until early 2016.

Another announcement Qualcomm made about the Snapdragon 820 was in relation to the Zeroth cognitive computing platform. For the past several years, Qualcomm has been working on Neural Processing Units (NPUs), as the company calls them, which are processors that mimic human learning.

"The premium mobile experiences of the future will extend beyond traditional features and functionality and be defined by devices that have the ability to learn and adapt to the needs of the user, through fully harnessing the growing levels of compute, multimedia and connectivity in our mobile devices," said Cristiano Amon, executive vice president of Qualcomm Technologies and co-president of QCT. "At MWC 2015 we'll take the first steps towards realizing this vision with the Zeroth platform, and set the stage for a new level of intelligence and personalization for mobile devices. Zeroth intelligence will scale across a wide range of implementations from automobiles, wearables, smartphones and client computing and have a learned personalization that has the ability to transfer across devices and as a consumer upgrade to the next generation."

The idea is to get artificial intelligence to learn as a human would, through positive reinforcement. Back in 2013, Qualcomm envisioned that these NPUs will one day be part of a regular SoC, just like a CPU or a GPU. The Snapdragon 820 will be the first chip to make that vision a reality.

Qualcomm said that the Zeroth platform could be used for things such as more advanced behavioral-based authentication, recognizing objects in photos, more accurate handwriting, increased optimization for LTE connectivity, improved speech and audio recognition, and more.

Follow us @tomshardware, on Facebook and on Google+.

Earlier this year, it came out that the Galaxy S6 would support all three wireless charging standards: Qi, PMA (Power Matters Alliance) and Rezence. Today, Mediatek announced that it's going to offer the first multi-mode wireless charging ASIC (application-specific integrated circuit) to other OEMs interested in having the same kind of support for wireless charging as the Galaxy S6.

Over the past few years, there has been a fight for dominance of the two main wireless charging standards: Qi and PowerMat. Qi started appearing in popular devices owned by many people, while PMA was being pushed by other larger companies such as AT&T and Starbucks.

Then the self-proclaimed "next-generation" Rezence magnetic resonance standard appeared, which promised to completely discard the idea that you need to wirelessly charge your device on top of a fixed equipment.

In reality, Qi and PMA, which use magnetic induction charging, don't seem like revolutionary technologies that promise to set us free from the tyranny of cables, when you still have to put your device on top of something that has a cable. The difference in convenience between doing this and just introducing the cable into your phone's port isn't that big.

Rezence aimed to allow smartphone users to leave their devices in a general area close to the wireless charger. For instance, if a charger is installed into a table, a smartphone user could simply leave the device anywhere on the table, and it would charge. This could impact the charging rate and effectiveness, but the Rezence standard uses Bluetooth Smart connections to pinpoint where the devices are located and establish a more clear charging path to them.

The Power Matters Alliance (PMA) has recently merged with the Alliance For Wireless Power (A4WP), which is behind the Rezence technology, so that made the fragmentation issue a little smaller. The Qi standards still exists, though, and chip makers such as Mediatek and OEMs such as Samsung seem to have decided to just put the two together as well.

Mediatek's new MT3188 ASIC that will support all three standards is already in mass production, and we should see products incorporating it within a few months.

Follow us @tomshardware, on Facebook and on Google+.

Today, Qualcomm announced its new FIDO-compliant ultrasonic fingerprint scanning technology that can record a much more accurate 3D image of the fingerprint's outer skin layer, compared to fingerprint scanning technologies based on capacitive sensors (such as Apple's Touch ID).

"Mobile devices increasingly store our most valuable and sensitive information, while passwords alone do not provide the protection consumers deserve," said Raj Talluri, senior vice president, product management, QTI. "Snapdragon Sense ID 3D Fingerprint Technology's unique use of ultrasonic technology revolutionizes biometrics from 2D to 3D, allowing for greater accuracy, privacy and stronger authentication. We are very proud to bring the mobile industry's first ultrasonic-based biometric authentication technology to mobile device manufacturers and their customers, who will benefit from the improved and differentiated user experience."

Until Apple's Touch ID arrived, nobody cared enough to use fingerprint scanning for authentication in the consumer space, and the previous low-accuracy technologies proved unpopular.

Even though there have been fingerprint scanning technologies in other devices such as notebooks or even smartphones, Apple did it right not just by using a more advanced and accurate technology that isn't a hassle for consumers to use, but the company also put it exactly in the place where most iPhone users would put their fingers anyway -- the home button.

The high accuracy played a role, but what gave Touch ID a large adoption on the iPhones that supported it was the fact that the users didn't have to do anything differently than normal, other than setting it up initially.

Although there have been some scares about fingerprint data being easily copied even from online photos, right now that's not a huge concern, because the fingerprint data is tied to the phones themselves. Someone who would want to authenticate with your cloned fingerprint data would also need your phone.

If or when Apple will start letting iOS devices owners authenticate to online services with their fingerprint, that's when we can start getting worried about the fingerprint data being stolen, cloned and spoofed. Consumers would wish that the fingerprint scanning technology is much more advanced in order to make it almost impossible for thieves to replicate that fingerprint data. Fortunately, such a solution is already here, thanks to Qualcomm's just announced Sense ID technology.

Apple's Touch ID uses a capacitive sensor technology that measures the different capacitance values in the ridges and valleys of the user's fingerprint when a charge is applied to the Touch ID circuit. The sensor creates an image of all of those values, applies a cryptographic hashing algorithm to the data, and then stores that hash in the Secure Enclave, a secure hardware zone on the phone's chip.

Apple's Touch ID works quite well in general, although not always. When fingers are wet, sweaty or oily, the accuracy can drop significantly.

Qualcomm's Sense ID technology brings the following advantages over Apple's Touch ID:

It's compliant with the FIDO Alliance's UAF (Universal Authentication Framework). The FIDO standard brings much of the tech industry together behind a single authentication standard meant to replace passwords.

This may not be a major issue initially for Apple users, because they're only using Touch ID to authenticate to the phone and to Apple Pay, but when this sort of authentication is also used for online services, the online world will be forced to use two different standards instead of just one: FIDO and Apple's own standard. This will create fragmentation, and some developers may end up supporting only one of the two standards, to the detriment of the users of the unsupported standard.

Sense ID can work through glass, plastic or metal. This means OEMs can simply put it behind the phone's screen, instead of creating a special button for it. Because many Android OEMs have already removed their buttons from the front of the phone, they had to put the fingerprint scanner on the back of the device.

That's not an ideal place, and it can end up frustrating the phone's owner. If it can be put instead behind the screen's virtual "unlock button," this should work more smoothly. This sort of thing can be done only because Qualcomm uses ultrasonic technology that can penetrate most materials.

Perfectly clean fingers not necessary. For the same reason Sense ID works through glass and metal, the technology can also capture the fingerprint's image through sweat or hand lotion with no issues.

Cloning-proof fingerprint template. Because the ultrasonic technology can penetrate the finger's skin layers at a deeper level, cloning or spoofing the fingerprint should be much harder to achieve from a simple online picture.

There are a few similarities between Qualcomm's Sense ID technology and Apple's Touch ID. For one, they both seem based on a custom version of ARM's TrustZone technology, which represents a "secure world" separated from the "insecure" operating system, where things such as the fingerprint template, passcodes or DRM keys are kept securely. Qualcomm calls this zone the "SecureMSM," while Apple calls it the "Secure Enclave." Both technologies send the fingerprint data through a dedicated connection to the secure storage zone.

Qualcomm said that the fingerprint data also stays on the device (so it's not stored in a cloud somewhere), but it's not clear whether it's the actual fingerprint data that is stored, or a hash of it, like Apple's Touch ID.

Qualcomm's Sense ID technology will be available in devices in the second half of this year and will arrive with chips such as the Snapdragon 810 and Snapdragon 425. With the biggest mobile chip company making such technology, it shouldn't take long before many other smartphones come with fingerprint scanning technology that's at least as good as Apple's Touch ID, helping more people secure their devices in a easy way.

Follow us @tomshardware, on Facebook and on Google+.

The budget Sony Experia E4 was unveiled earlier this month as the successor to the Xperia E3, but there was one missing feature that may have disappointed some: 4G connectivity. Sony will rectify that with the new Xperia E4g, a smartphone that changes little from the Xperia E4 but adds 4G capabilities.

The Xperia E4g comes with a quad-core 1.5 GHz processor, which is a step up from the quad-core 1.3 GHz chip that the Xperia E4 had. It's not clear whether this is a Cortex-A53-based chip like the Snapdragon 410, which also comes with integrated LTE, or just another Cortex-A7-based chip that's higher clocked. If it's Cortex-A7 and it uses the same 28nm process node, it may be slightly less efficient in terms of battery life.

The device has a 5" qHD (960 x 540) screen, 1 GB of RAM, 8 GB of storage, 5 MP/1080p rear camera, a 2 MP/720p front camera, and the same 2,300 mAh battery as before; it will also come with Android KitKat out of the box. However, it will receive the Lollipop upgrade later on. Being a budget phone, it's also likely to be the last update the device will receive until Sony states otherwise, so that should also factor into the purchase decision.

Fortunately, although it has received a slightly faster chip and LTE connectivity, it will retail for the same price that the Xperia E4 did initially, 129 euro. The Xperia E4 has dropped to 120 euro in order to distinguish itself from the LTE-based Xperia E4g.

The device also comes with some interesting features such as the Xperia Transfer App, which allows users to transfer photos, bookmarks, apps, music and messages from Blackberrys, Windows phones or other Android smartphones.

The camera app comes with an Auto Scene Recognition mode, which can recognize up to 52 modes to help users capture better photos in whatever environment they're in. Thanks to NFC support, those photos can also be easily shared with other phones.

Sony advertises the device as having two days of battery life, which is believable if we're talking about moderate use, considering it doesn't have an especially high screen resolution and the processor, either Cortex A7 or Cortex A53, should be quite efficient.

However, the battery life can be further expanded with the STAMINA Mode, which Sony has had on its devices for some time, as well as with an Ultra STAMINA Mode that can keep the phone lasting for a full week with only the core phone functions enabled. That's a feature that many would probably like on other phones, as well.

Follow us @tomshardware, on Facebook and on Google+.

Today, Qualcomm announced four new mobile chips: the low-end Snapdragon 415 and Snapdragon 425, and the mid-range Snapdragon 618 and Snapdragon 620.

Snapdragon 415 & Snapdragon 425

The Snapdragon 415 will likely succeed the current Snapdragon 410. It's also based on Cortex A53 CPU cores, but it has eight of them instead of four. It seems Qualcomm is really starting to embrace this eight-core marketing, possibly influenced by high market demand in Asia for such chips, although it's not entirely clear that this many cores is needed today, even for gaming.  All cores will be clocked at 1.4 GHz, and the chip will be bundled with the X5 Cat. 4 LTE modem that supports download speeds of 150 Mbps and upload speeds of 50 Mbps.

The Snapdragon 425 is a little faster, with CPU cores clocked at 1.7 GHz, and it also has a more advanced Cat. 7 LTE modem that supports 300 Mbps download speeds and 100 Mbps upload speeds. The modem supports carrier aggregation, which should make it easier to automatically switch between carriers depending on signal strength. The signals can also be combined for improved performance. This modem is called the X8 and will be integrated in the Snapdragon 618 and the Snapdragon 620, as well.

Snapdragon 618 & Snapdragon 620

The Snapdragon 618 and the Snapdragon 620 seem to be categorized as Snapdragon 600-series chips, which have been mid-range chips so far. However, these chips are supposed to use ARM's next-generation Cortex-A72 CPU cores.

That's surprising for two reasons. For one thing, these two new chips are expected to come out this year, yet they are using a CPU core that wasn't supposed to come out until next year.

Second, this also implies that Qualcomm has something prepared for the high-end that's even faster than Cortex-A72, and it will come out this year as well. The Snapdragon 820 was already leaked earlier this year, and it's supposed to have a 14nm custom "Taipan" core made by Qualcomm.

The surprisingly fast release of Cortex-A72 chips, along with "midrange" branding for the 618 and 620, comes mainly as a consequence of building these chips on the old 28nm planar process (according to a Qualcomm spokesperson), rather than the recommended 16nm (or 14nm) FinFET process for the high-performance Cortex-A72 core. This could increase the power consumption of the Snapdragon 618 and Snapdragon 620 significantly, which is probably why Qualcomm decided to use more conservative clock speeds for them.

Both the Snapdragon 618 and the Snapdragon 620 will have their Cortex-A72 cores clocked at 1.8 GHz, and the Cortex-A53 cores will be clocked at 1.2 GHz. It's still not clear if these frequencies will be low enough for Cortex-A72 to be efficient, though. The main difference between the two SoCs is that the Snapdragon 618 will only have two Cortex-A72 cores, while the Snapdragon 620 will have four.

New Adreno GPU With Hardware Tessellation

The mid-range chips will also get a "next-generation GPU" that will support the latest graphics APIs (presumably referring to OpenGL ES 3.1 here), as well as hardware tessellation and geometry shading, which are features that Google introduced in its Android Expansion Pack to complement OpenGL ES APIs. Qualcomm didn't reveal the name of this new GPU, nor did it give any indication about how much faster it will be compared to current Adreno generations.

Other Features

Both the 618 and 620 will also support features such as Quick Charge 2.0, which we've seen on high-end Qualcomm chips before, as well as security features.

The Snapdragon 618 and the Snapdragon 620 also come with other features that have been available on high-end chips before, such as dual-ISP camera support, 4k video capturing at 30fps, and HEVC encoding. (There's no VP9 support, apparently.)

All four chips announced today come with 802.11ac Wi-Fi, which should make this new standard more mainstream in 2015. The chips only support LPDDR3 RAM, unlike the Snapdragon 810 which supports LPDDR4, but that's likely to be sufficient as they're only intended for the low-end and mid-range markets.

The Snapdragon 415 SoC will be in devices in the first half of the year, according to Qualcomm, while the other three chips will come to market in the second half of this year.

Update, 2/19/15, 2:45pm: We received slightly incorrect information on a feature and have now removed a reference to it.  

Follow us @tomshardware, on Facebook and on Google+.

According to a new report (Korean) from South Korea, LG has started working on a next-generation high-end SoC that's going to use the recently announced Cortex-A72 CPU and an unnamed Mali GPU (likely the high-end Mali-T880).

LG has spent almost $200 million developing the NUCLUN chip, which hit the market late last year with CPU cores such as Cortex A15 and Cortex A7 (big.LITTLE configuration), but it still ended up overheating and aggressively throttling. The sales of the one and only phone to use it, the LG G3 Screen, quickly plummeted because of that processor, and LG was forced to pull the plug on the NUCLUN.

The company then tried to build another chip that was going to use the Cortex A53 and Cortex A57 cores (also in big.LITTLE configuration). This new chip was meant to be used as an alternative to Qualcomm's Snapdragon 810. However, due to technical issues that lead to overheating, the company decided to scrap this project as well.

LG doesn't seem to want to give up on building chips, though, especially now that it sees Samsung wanting to focus even more on its own chips with its Exynos 7 series. Therefore, the company is once again trying to build a next-generation chip using Cortex-A72 cores, according to South Korean sources.

The Cortex-A72, built on 16nm FinFET (TSMC's process), should be three times faster than a 28nm Cortex A15 processor (such as LG's NUCLUN). However, according to the report, LG's chip will only be built on the 20nm process, which should be quite obsolete by the time Cortex-A72 chips start arriving in 2016. LG is also supposedly six months behind schedule already, which can't be very good for a project that has just begun.

Even if the chip arrives in a reasonable time frame (relative to the competition), it's still not clear whether LG now has the qualified engineering teams to help the company build a chip without any more technical issues. Even a long-time chip maker such as Qualcomm has had similar overheating problems this year, and it should be even more difficult for a new chip maker such as LG to figure out how to make good chips.

Follow us @tomshardware, on Facebook and on Google+.

Samsung announced that it has begun mass production of chips on its new 14nm FinFET process. This is the first time a company other than Intel has built FinFET-based three-dimensional chips. FinFETs are what will allow foundries to overcome performance and scaling limitations of the 20nm planar process.

The new process enables up to 20 percent faster speed, 35 percent less power consumption, and 30 percent productivity gain when compared to Samsung's own 20nm process, on which chips such as the Exynos 5430 (Galaxy Alpha) and Exynos 5433 (Galaxy Note 4) were built.

“Samsung's advanced 14nm FinFET process technology is undoubtedly the most advanced logic process technology in the industry," said Gabsoo Han, Executive Vice President of Sales & Marketing, System LSI Business, Samsung Electronics. “We expect the production of our 14nm mobile application processor to positively impact the growth of the mobile industry by enabling further performance improvements for cutting-edge smartphones."

The move to FinFETs also puts Samsung much closer to Intel than ever. This is due both to Samsung's aggressive research in the area for more than a decade, but also Intel's delays for its own 14nm process. Technically, though, Samsung's 14nm is not a pure 14nm process, but somewhere between a 20nm process and a 14nm process -- which still leaves Samsung in a much better position relative to Intel than in the past.

Formerly, Intel went to 22nm FinFET when others like Samsung and TSMC were only moving to 28nm planar. That's a difference of about a generation and a half in process node technology; now, that difference has seemingly shrunk to around half a generation. With any luck, that difference could disappear completely in the coming years, as Moore's Law will approach its end. It will become harder and less profitable to move to the next process mode for the pioneers of smallest process nodes, allowing those who are behind to catch up to the latest node technology.

The first chip to use the 14nm FinFET process from Samsung is probably going to be the Exynos 7420, the same chip that's supposed to appear in the Galaxy S6. The Galaxy S6 will likely be announced at Samsung's March 1 event in Barcelona.

Follow us @tomshardware, on Facebook and on Google+.

ARM hasn't been in the GPU IP market for long. It acquired Falanx, a GPU IP designer, less than 10 years ago (2006), after it was previously sub-licensing GPU IP from Imagination. Since then, ARM's Mali graphics IP has seen steady growth, and ARM is now the GPU IP leader in both the Android market (Apple only uses Imagination GPUs for iOS devices anyway) and in digital TVs market.

ARM's Mali GPU shipped in 550 million devices in 2014, which is 150 million more than in the previous year. It also gained 27 new licensees in 2014, totaling 110 Mali licensees.

ARM has managed to keep its customers happy by continuing to introduce competitive GPU IP. The latest introduction is the Mali-T880 GPU, which will be one level above the Mali-T860, which is in turn the Mali-T760 successor. The Mali-T760 was first seen in the Galaxy Note 4 (international Exynos version), so its successor, the Mali-T860, will likely arrive in the Galaxy Note 5, while the Mali-T880 will probably arrive in Galaxy S7.

The new Mali-T880 will be 80 percent faster compared to the Mali-T760, if power consumption remains at the same levels. If the performance is kept at the same level instead, then the GPU would consume 40 percent less power.

It will be up to chip makers such as Samsung to decide the best compromise between an increase in performance and a decrease in power consumption, but usually chip makers decide to keep the power consumption the same and go all-in with the performance increase. It's also one of the reasons why new smartphones don't seem to be dramatically better in battery life, despite having bigger batteries every year.

There is a new trend in making smartphones much thinner, such as the iPhone 6 or the Galaxy Alpha. These devices tend to have significantly smaller batteries, too. That leads OEMs to want to use more efficient processors, rather than chips that are much more powerful. However, the more efficient chips are only used to compensate for the smaller batteries, so the ultra-thin devices shouldn't get better battery life on average.

The Mali-T880 is highly configurable, and OEMs can use as many as 16 cores, which is up from eight for the Mali-T760 but the same maximum number of cores that the the Mali-T860 will have, as well. The Mali-T880 can also be complemented by ARM's other IP products such as the Mali-V550 video processor and the Mali-DP550 display processor, which can help deliver 4k video.

ARM's high-end product IP suite for 2016 will include the Cortex-A53 CPU core; the Cortex-A72 CPU core, which at 16nm FinFET can deliver three times the performance of the original 28nm Cortex A15; the Mali-T880 GPU, Mali-V550 and Mali-DP550 for media acceleration; and the next-generation CoreLink CCI-500 cache coherent interconnect that ties everything together.

Follow us @tomshardware, on Facebook and on Google+.

According to industry sources from South Korea, Samsung is getting ready to go to phase two of its plan to reduce its dependence on Qualcomm. Phase one seems to have been the elimination of Qualcomm's chips from some of its most popular devices, such as the Galaxy S series.

The Galaxy Note series is likely to follow soon and use only Exynos chips as well, now that we know that Samsung is already working on successors to its Exynos 7420 chip (the one apparently going into the Galaxy S6).

Reducing reliance on Qualcomm's chips is not easy, though, and not just because Qualcomm has had a history of making good chips that most other OEMs have used, but also because Qualcomm has been a pioneer in integrating its modems into its SoCs. In fact, this sort of bundling has made Qualcomm's chips the default option for smartphone OEMs, especially in LTE markets such as the U.S.

Integrating the modem into a chip means it costs less to make, and it can be delivered faster in products. Qualcomm's mobile chip leadership won't be truly threatened until competitors can build chips that rival Qualcomm's in terms of performance and efficiency and are also able to integrate their own modems into those chips.

Samsung has been building its own modems since last year, but so far it hasn't integrated them into a single-chip solution. The modems have been attached separately to its Exynos processors. The company has already been working on its own Cat. 10 LTE modem that it may introduce along with the Exynos 7420 in the Galaxy S6 smartphone, but the modem will likely not be integrated into the SoC.

Following the Galaxy S6, the company is expected to ship devices with single-chip solutions with integrated modems. One of the first beneficiaries of such a chip could be the Galaxy Note 5 later this year.

Follow us @tomshardware, on Facebook and on Google+.

According to Nvidia's earnings report for "fiscal 2015" (essentially the year 2014), which ended on January 25, Nvidia saw returns that exceeded the expectations of Wall Street analysts. The company made $4.68 billion in revenue, which was a 13.4 increase year over year. Profit also soared by 43.4 percent year over year to $631 million.

Nvidia still made most of its money from PC GPUs, which represented 85.8 percent of its total revenues in the last quarter. The company's GPU sales grew by 10.7 percent last year to $3.84 billion. Gaming was a strong driver for Nvidia last year, and the company saw an increase of 38 percent for PC and notebook GPUs.

Although its Tegra chips brought only a little over half a billion dollars ($579 million) for the whole year, the company saw its mobile chip sales increase by 45.5 percent. Nvidia credited this sharp rise in sales mainly thanks to its SHIELD tablet and the high interest auto makers seem to have in the company's chips. (Nvidia's automotive chip sales doubled year over year.)

Earlier this year at CES, Nvidia unveiled the Tegra X1, which for now specifically addresses the automotive market. Nvidia believes its visual computing expertise is critical for self-driving cars, because the cars will need to "see" the environment around them and make sense of it in order to drive safely through it.

Nvidia, being one of the few companies that can make chips that are low-power enough yet still deliver incredible graphics performance, seems to have become a good partner for car makers that are ready to embrace the future by taking advantage of advanced mobile technology.

Although the Tegra sales saw significant growth, Nvidia is still considered a relatively minor player in the mobile market, and the company has been almost non-existent in the smartphone market for the past two years. Nvidia has chosen to focus more on high-performance chips and "winning" against other mobile chip competitors through raw performance, but at a major cost in power consumption.

Other chip makers such as Qualcomm and Mediatek have become so successful in the smartphone market because they have made low power consumption a bigger priority than absolute performance. The companies still compete in performance, but only after they've ensured a certain threshold of power consumption which they cannot exceed.

Nvidia, on the other hand, is more concerned with winning in performance metrics at all costs. This has arguably been a strategic flaw for Nvidia that has kept it from gaining significant market share in the mobile market. With its newfound success in the automotive market, Nvidia's future in the smartphone and even the tablet market (which has been in decline lately) is uncertain, but we're looking forward to see what the company has prepared for the Mobile World Congress next month.

Follow us @tomshardware, on Facebook and on Google+.

ARM Holdings, the company behind the mobile chip architecture that's in the vast majority of smartphones and tablets, reported on Wednesday that it made 25 percent more profits than it expected in the last quarter.

The company's designs were used in 3.5 billion chips in the last quarter. Most of those chips appeared in embedded devices rather than smartphones, but the company saw strong adoption for its ARMv8 architecture and its Cortex-A57 and Cortex-A53 CPU cores, which go into mobile devices.

"Following the acceleration in royalty revenue growth in the second half of 2014, and with a wide range of OEMs (manufacturers) introducing products based on ARM's V8 architecture this year, the outlook for royalty revenues this year is very encouraging," said ARM's Chief Financial Officer, Tim Score.

Qualcomm, Nvidia, Samsung and Mediatek have all licensed ARM's 64-bit designs instead of going with their own this year. Although this may not last for too long, as Qualcomm (the number one mobile chip maker) is poised to release its own CPU core at the end of the year, it should give ARM a significant boost in profits and revenue over the next few quarters.

With the switch to the new architecture, ARM has also boosted its royalty from 1-1.5 percent of the chip's prices to 2 percent. Tim Score also said that the royalties will play a significant role in the company's revenues going forward. He also said that ARM signed 53 new licensing deals in the last quarter, including eight for the ARMv8 architecture. He expects that by the end of the year, half of the smartphones in the market will be powered by an ARMv8 chip.

In the last quarter, ARM made $182 million in profit and $348 million in revenue, so the company is highly profitable as more than half of its revenue is profit. The revenue is split evenly between licenses and royalties. ARM said that it's expecting to get a 10 percent increase in revenue year over year in the next quarter.

Follow us @tomshardware, on Facebook and on Google+.

With Samsung reducing its reliance on Qualcomm's chips and trying to differentiate from all the other mobile OEMs who seem to default to the same chips, the company has been focusing more on its in-house processors, such as the upcoming 64-bit Exynos 7420 that's supposed to arrive in the Galaxy S6.

According to a job description on Linkedin, Samsung is also working on a couple of other chips from the Exynos 7-series. The two chips will be called the Exynos 7890, which is presented as a 14nm chip, and the Exynos 7650, which is also a 64-bit ARMv8 chip, but it's not clear whether it will also be built on the 14nm processor.

The Exynos 7650 could be a mid-range or even a low-end 20nm chip that Samsung intends to use for its own lower-end devices. This could prove that Samsung is getting more and more interested in supplying its whole smartphone division with its own chips and not just the high-end devices.

Samsung has mainly used its in-house chips for high-end devices so far. The company has used some of its chips in mid-range devices, but usually those were chips Samsung used a year or two before in high-end devices. For instance, the Exynos 4 chip used in the Galaxy S3 was later used in some of Samsung's mid-range devices.

Samsung is still the world leader in smartphone sales, and the proliferation of its own chips in its devices could mean revenue hits for Qualcomm, as well as others that are currently supplying chips to Samsung. Although there isn't a high chance that Samsung will start selling its Exynos chips to other companies until it can properly supply its own mobile division, the company could decide to do so in the future to become an actual competitor to Qualcomm and Mediatek, or even Intel, if it ever decides to buy AMD.

The Exynos 7890 and Exynos 7650 chips don't have an arrival date, but it's likely we'll see both in devices this year. There's a high probability that the 14nm Exynos 7890 could land in the Galaxy Note 5 later this year.

Follow us @tomshardware, on Facebook and on Google+.

Today, ARM announced that it acquired the Dutch security software company Offspark, which is the main developer behind the PolarSSL crypto library for encrypted SSL connections on the web and on IoT devices.

According to ARM, the PolarSSL library is already "pervasive" in the IoT world, which is exactly why the company thought Offspark is a good match for it. ARM is getting ready to launch its mbed OS for IoT devices later this year. Security on IoT devices could become a huge issue in the future if not tackled from the beginning with good security designs and protocols.

ARM has already invested in the developing of the Thread mesh networking protocol for IoT, which should secure connections among devices at home. Now ARM wants to secure the Internet connections to the mbed OS-powered IoT devices, as well.

"We have always said that security must be the foundation of any IoT system and the acquisition of Offspark is evidence of us making that happen," said Krisztian Flautner, general manager, IoT business, ARM. "PolarSSL technology is already deployed by the leading IoT players. The fact that those same companies also utilize ARM Cortex processor and software technologies means we are now able to provide a complete bedrock solution for the industry to innovate from."

ARM will rename the PolarSSL crypto library to "mbed TLS" to better show the relationship between its OS and the library. The mbed TLS library will be available for both standalone use and as part of the mbed OS.

ARM will release the mbed OS, mbed TLS and Thread and other important technology the company has been developing for IoT under the open source Apache 2.0 license at the end of the year. The mbed TLS 1.3.10 library is already available under GPL on the PolarSSL website.

Follow us @tomshardware, on Facebook and on Google+.