Sign in with
Sign up | Sign in

Results: GPU Core Benchmarks

Samsung Galaxy Note 10.1 (2014 Edition) Review: Wi-Fi Vs. LTE
By

3DMark Ice Storm Unlimited

Futuremark has become a name synonymous with benchmarking, and the company's latest iteration of 3DMark offers three main graphical benchmarks: Ice Storm, Cloud Gate, and Fire Strike. Currently, the DirectX 9-level Ice Storm tests are cross-platform for Windows, Windows RT, Android, and iOS.

Ice Storm simulates the demands of OpenGL ES 2.0 games using shaders, particles, and physics via the company's in-house engine. Although it was just released in May of last year, the on-screen portions of Ice Storm have already been outpaced by modern mobile chipsets, with Nvidia's Tegra 4 and Qualcomm's Snapdragon 800 both easily maxing out the Extreme version (1080p with high-quality textures). However, Ice Storm Unlimited, which renders the scene off-screen at 720p, is still a good gauge of GPU-to-GPU performance.

The 3DMark results are quite surprising. For one, Overall score doesn't vary as much as we expected among the Snapdragon 800, Exynos 5 Octa, Tegra 4, and Apple A7. The Tegra Note 7 finishes on top with 16,567 points, followed closely by the iPad Air. The Snapdragon 800-powered Note 10.1 (2014 Edition) LTE scores roughly 15,200 points, followed by the Exynos 5 Octa-powered Note 10.1 (2014 Edition) Wi-Fi and 2013 Nexus 7.

While the Overall scores are close, the Graphics and Physics scores are all over the place. In Graphics, the iPad Air and its PowerVR G6430 engine obliterate the competition. The Tegra 4 is surprisingly far behind in second place, while the Adreno 330 follows in third.

The Physics scores yield the biggest surprise, though. Nvidia's Tegra 4 does well, helping it secure the best Overall score. The Snapdragon 800-powered Note 10.1 (2014 Edition) LTE finishes with the second-best Physics result, while Apple's iPad registers the worst Physics score among all five tablets (which is why the iPad finishes in second when the Overall number is tabulated). Since the Physics test is CPU-bound, these results imply that the iPad's dual-core A7 likely limits performance in this particular metric.

Anomaly 2 Benchmark

Anomaly 2 is a game published by 11 bit studios and available in the Google Play store. The title's creators also publish a benchmarking tool, Anomaly 2 Benchmark, which lets you see your device will fare in the full game.

Despite its real-world pedigree, the Anomaly 2 Benchmark is quite graphics-dependent, and the Tegra Note 7 absolutely smashes its competition in this metric. That's a bit surprising, given that 11 bit studios has stated that while the Tegra 4-powered Nvidia Shield is a good performer, the Galaxy S4, Galaxy Note 3, and even Xperia Z1 perform better.

Nevertheless, the Note 7 finishes with almost three times the score (~700,000) of its closest competitors. In almost a dead heat, the Adreno 320-powered Nexus 7 (2013) finishes in second place,closely followed by the Adreno 330-powered Note 10.1 (2014 Edition) LTE. The Note 10.1 Wi-Fi's Mali-T628 ends up in last place.

Basemark X 1.1

Basemark X 1.1 is a benchmarking tool by Rightware that utilizes the popular Unity game engine, allowing for device comparisons across a wide range of platforms, including Android, iOS, and Windows Phone 8. This also has the advantage of making the benchmarks correlate well with the real-world performance of numerous games available on the market.

In Basemark X 1.1, the iPad Air dominates its competition; the Adreno 330-equipped Note 10.1 (2014 Edition) LTE follows behind by a significant margin, and the Tegra Note 7 trails it. The Mali-T626-powered Note 10.1 (2014 Edition) Wi-Fi does fall slightly behind in the high-quality test compared to the medium measurement—enough for the Nexus 7 (2013) to surpass it.

In the Basemark X 1.1 Medium Quality On Screen test, Apple's iPad Air beats the two Galaxy Note 10.1s. In the High Quality version of this test, the Tegra Note 7 surpasses not only the LTE Snapdragon-based Note 10.1, but even the iPad Air. Meanwhile, the Nexus 7 beats the Exynos-based Note 10.1 to reinforce how well this benchmark shows the punishing nature of driving a high-resolution output.

In the Off Screen test, you can see that once resolution is taken out of the equation, both versions of the Galaxy Note 10.1 (2014 Edition) fall within striking distance of the competition.

Once again, the Wi-Fi model places significantly and consistently behind the LTE model.

GFXBench 3.0

Kishonti GFXBench 3.0 is a cross-platform GPU benchmark supporting both the OpenGL ES 2.0 and OpenGL ES 3.0 standards. It comprises game-like scenarios, as well as lower-level tests designed to measure specific subsystems. See GFXBench 3.0: A Fresh Look at Mobile Benchmarking for a complete test-by-test breakdown of this suite.


In the GFXBench 3.0 On Screen tests, we clearly see the impact of running at 2560x1600, as both Note 10.1 (2014 Edition) tablets perform pretty poorly compared to the iPad Air and Tegra Note 7.

Both Note 10.1s make a significant rebound in the Off Screen test, once their resolution handicap is taken out of the equation. The iPad Air still manages to finish ahead of the Mali-powered Note 10.1 (2014 Edition) Wi-Fi in the Manhattan and T-Rex tests, but is then dominated by the Snapdragon 800-powered LTE model.

In the Driver Overhead tests, we clearly see that the Wi-Fi version of the Note 10.1, powered by Samsung's own Exynos SoC, is optimized much better than the Snapdragon-based LTE model. Meanwhile, the Fill tests show the Snapdragon 800's Adreno 330 has the most raw horsepower out of any of the assembled GPU cores.

In GFXBench 3.0's Render Quality test, the Exynos 5 Octa's Mali GPU finally finds a benchmark where it outshines not only the Snapdragon SoC, but the other competitors as well. With 10 to 30 percent better scores, the Note 10.1 (2014 Edition) Wi-Fi beats the competition handily. Surprisingly, the Tegra Note 7 finishes in second place in Standard, but last in High Precision. Since it's being billed as a productivity tablet, as opposed to a purely consumption-oriented device, it makes sense that the more widely available Note 10.1 (2014 Edition) Wi-Fi model uses the Exynos 5, as render quality would be more important than frame rate in most business applications. It could also speak to why the Exynos is so favored among the no-frills entry-level ChromeOS devices, where hardcore 3D performance is practically a non-point.

Display all 21 comments.
  • -3 Hide
    blackmagnum , July 8, 2014 1:05 AM
    The only possible way Samsung mobile devices can be a worthwhile alternative to Apple is when they ditch the every-man Android and create their own tightly controlled/ managed OS like Apple. Do that and consumers might not feel like they're just buying the Samsung for the hardware.
  • 0 Hide
    Farrwalker , July 8, 2014 8:33 AM
    On page 7. Results: CPU Core Benchmarks:
    Your bar graph "MobileXPRT 2013" seems to be in error.
    For example, the text says, "Note 10.1 (2014 Edition) Wi-Fi leads with 300 points . . ."
    but the bar is the shortest and indicates less than 150 points.
  • 0 Hide
    blueer03 , July 8, 2014 9:30 AM
    You need to proofread this big time. From page 9:
    Samsung's Exynos-based Note 10.1 (2014 Edition) Wi-Fi holds its own against the Tegra Note 7, while the LTE version of the Note 10.1 (2014 Edition) again falls significantly behind its Snapdragon 800-powered Wi-Fi counterpart.

    And this happens all throughout. The LTE is a Snapdragon, the Wifi is an Exynos. Keep repeating that to yourself as you re-write the descriptions and it will make this easier to read.
  • 1 Hide
    Kevin Harrelson , July 8, 2014 9:54 AM
    We got one of these for my son (age 13) to help with his school work. We got it from Best Buy and got the extended warranty. Both sound dumb, but it was actually a good move! The backlight has gone out on this thing TWICE. I happen to think that it is a lovely little tablet, but the backlight is a major reliability issue.
  • 1 Hide
    adamovera , July 8, 2014 10:48 AM
    Quote:
    On page 7. Results: CPU Core Benchmarks:
    Your bar graph "MobileXPRT 2013" seems to be in error. For example, the text says, "Note 10.1 (2014 Edition) Wi-Fi leads with 300 points . . ." but the bar is the shortest and indicates less than 150 points.

    This benchmark's sub-tests produce scores in seconds (lower is better), and the overall score is given as a typical higher-is-better score, so the lowest bar indicates the fastest completion. Sorry about the confusion, I'll look into other ways to represent this test.
  • 2 Hide
    adamovera , July 8, 2014 10:51 AM
    Quote:
    You need to proofread this big time. From page 9:
    Samsung's Exynos-based Note 10.1 (2014 Edition) Wi-Fi holds its own against the Tegra Note 7, while the LTE version of the Note 10.1 (2014 Edition) again falls significantly behind its Snapdragon 800-powered Wi-Fi counterpart.
    And this happens all throughout. The LTE is a Snapdragon, the Wifi is an Exynos. Keep repeating that to yourself as you re-write the descriptions and it will make this easier to read.

    Good catch, thanks! Fixed.
  • 0 Hide
    Tomtompiper , July 8, 2014 1:30 PM
    The 2014 is almost 5 months old and has been superseded by the Amoled screened Galaxy Tab S 10.1 which wipes the floor with the opposition. http://www.gsmarena.com/samsung_galaxy_tab_s_105-review-1097.php Do try to keep up!
  • 3 Hide
    adamovera , July 8, 2014 2:06 PM
    Quote:
    The 2014 is almost 5 months old and has been superseded by the Amoled screened Galaxy Tab S 10.1 which wipes the floor with the opposition. http://www.gsmarena.com/samsung_galaxy_tab_s_105-review-1097.php Do try to keep up!

    Actually, you'd think this has been phased out, but it's the current 10-inch Galaxy "Note" product, meaning it has the S Pen. The Galaxy "Tab" S does not - still unclear to me what makes the "S" stand out. I believe Samsung is literally attempting to offer an alternative product to every single other device in existence - complete mobile domination. I lost count of their current "Galaxy" line at 11 products, and that was awhile back.

    The duo of Note 10.1 (2014)'s came in very handy both as comparison data in other articles and as testbeds for compiling our benchmark suite. Unfortunately, the article had to be pushed back several times, but the huge hardware difference between products carrying the same name was always something we wanted to illustrate, initially for the chipset-vs-chipset angle, but later for the optimization aspect as well. We're currently working through a small backlog of mobility articles, but each will be more timely than the last. My apologies.
  • 1 Hide
    Blazer1985 , July 8, 2014 6:22 PM
    I really wish Samsung could help the end user distinguish what they are buying. The 2 Note are classified as "LTE" and "WIFI ONLY" which sounds like 2 equal devices except for the sim card slot while they actually have a totally different soc.
  • 0 Hide
    vaughn2k , July 8, 2014 9:03 PM
    ... and Samsung prices are now on par with Apple's. others are okay. Will wait for Asus and other competition. Should be cheaper though...
  • -2 Hide
    Ninjawithagun , July 9, 2014 6:53 AM
    These benchmarks are fundamentally flawed. No where is there mention that both the Nexus and Apple iPad Air use lower resolution displays, which dramatically reduces the overall load on the CPU and GPU vs. the Samsung Note devices. The benchmarks are completely useless since an "apples-to-oranges" comparison is not a valid methodology for comparing these uniquely different tablets. And why was the new Samsung Tab S used instead of the year old 2014 Edition versions? Another "fail" on behalf of the author.
  • 0 Hide
    godnodog , July 9, 2014 11:23 AM
    As a user of the Note 10.1 2012, I have to strongly disagree with the new multitasking sytem, as it is clearly a step backwards from what it used to be, now you can ONLY have 2 windows opened, as I, by not installing the update, can have 4 / 5 / 6 opened, obviously I don´t have them opened, but I frequently have 3 windows opened simultaniously, wich now I can´t have. Also it apperars that the multitask is no longer floating, agains mine that still does. I have experienced no problem whatsoever like you described "were slow, unresponsive, or just plain bad".
  • 1 Hide
    wiltjk , July 10, 2014 7:18 PM
    A key point of your review is that this is about a "pen" tablet. Not having used a physical keyboard since 2003, I am pleased to see some alternatives to the Intel/Windows dominated offerings.

    It is a good review for apples to apples on the Samsung hardware options.

    Would you suggest that this is "as-good-as-it-gets" for the few who prefer pen-based tablets? Any foresight in other pen based tablets?
  • 0 Hide
    Zarathvstra , July 11, 2014 2:23 AM
    Quote:
    These benchmarks are fundamentally flawed. No where is there mention that both the Nexus and Apple iPad Air use lower resolution displays, which dramatically reduces the overall load on the CPU vs. the Samsung Note devices. The benchmarks are completely useless since an "apples-to-oranges" comparison is not a valid methodology for comparing these uniquely different tablets. And why was the new Samsung Tab S used instead of the year old 2014 Edition versions? Another "fail" on behalf of the author.


    Ninjawithagun,
    Tablets are not modular like a pc... whilst it may have been good for the author the resolution issue with these benchmarks, an apples to apples comparison of the individual components would NOT help the user decide which unit as a whole is the fastest as indeed, they are NOT modular.
  • 0 Hide
    Ninjawithagun , July 12, 2014 8:46 AM
    Quote:
    Quote:
    These benchmarks are fundamentally flawed. No where is there mention that both the Nexus and Apple iPad Air use lower resolution displays, which dramatically reduces the overall load on the CPU vs. the Samsung Note devices. The benchmarks are completely useless since an "apples-to-oranges" comparison is not a valid methodology for comparing these uniquely different tablets. And why was the new Samsung Tab S used instead of the year old 2014 Edition versions? Another "fail" on behalf of the author.


    Ninjawithagun,
    Tablets are not modular like a pc... whilst it may have been good for the author the resolution issue with these benchmarks, an apples to apples comparison of the individual components would NOT help the user decide which unit as a whole is the fastest as indeed, they are NOT modular.


    Not true. Most (not all) tablets are in fact modular in that the CPU and GPU operate independently of one another. Case in point, the Apple A7 uses a dual-core 64-bit SOC processor and PowerVR G6430 graphics chip. The Samsung uses the Exynos 5 Octa processor with a Mali-T628 MP6 graphics chip. Your definition of modular is fundamentally flawed in that you think it pertains to physical ability for individual items to removed/replaced. Modular infers to the actual architecture of the system in that several different parts from different manufactures are integrated together to function as a whole unit.
  • 0 Hide
    Zarathvstra , July 12, 2014 6:04 PM
    Quote:
    Quote:
    Quote:
    These benchmarks are fundamentally flawed. No where is there mention that both the Nexus and Apple iPad Air use lower resolution displays, which dramatically reduces the overall load on the CPU vs. the Samsung Note devices. The benchmarks are completely useless since an "apples-to-oranges" comparison is not a valid methodology for comparing these uniquely different tablets. And why was the new Samsung Tab S used instead of the year old 2014 Edition versions? Another "fail" on behalf of the author.


    Ninjawithagun,
    Tablets are not modular like a pc... whilst it may have been good for the author the resolution issue with these benchmarks, an apples to apples comparison of the individual components would NOT help the user decide which unit as a whole is the fastest as indeed, they are NOT modular.


    Not true. Most (not all) tablets are in fact modular in that the CPU and GPU operate independently of one another. Case in point, the Apple A7 uses a dual-core 64-bit SOC processor and PowerVR G6430 graphics chip. The Samsung uses the Exynos 5 Octa processor with a Mali-T628 MP6 graphics chip. Your definition of modular is fundamentally flawed in that you think it pertains to physical ability for individual items to removed/replaced. Modular infers to the actual architecture of the system in that several different parts from different manufactures are integrated together to function as a whole unit.


    This does not argue to the point. Here's a fact and the crux of the matter... An apples to apples comparison of gpu's and processors does NOT tell the user how apps will perform on tablet x when constrained by their other components be it amount of ram, screen res, gpu, cpu etc.

    i.e. benchmarks are not "fundamentally flawed" (nor is my understanding and use of the term modular)
  • 1 Hide
    apertotes , July 14, 2014 4:36 AM
    I think it is deceiving to not point out on the table on first page that memory can easily be expanded with a cheap microsd card. It is not 16/32 gb, it's 16/32+128 gb.
  • 0 Hide
    szalkerous , July 16, 2014 9:21 AM
    At first I thought there was something new, and I realized this is a review for a tablet I bought last year.

    FYI, the KitKat update makes the SD card practically useless. Rooting these tablets is frustrating at best, and the KNOX system is the most terrible idea Samsung ever came up with.

    Looking back, I should have gotten a Nexus.
  • 0 Hide
    apertotes , July 16, 2014 10:06 AM
    Quote:

    FYI, the KitKat update makes the SD card practically useless.


    How can it be useless having 128 gb for music and videos/pictures? Maybe you can afford spotify and paying 15 dollars every month for an upgraded dropbox account, but many people don't, or even if they do, they do not have a reliable LTE signal 24/7.

  • 0 Hide
    SoPhat VaThana , July 17, 2014 1:29 AM
    good
Display more comments
React To This Article