AMD's 4000-Series "Renoir" Ryzen Mobile chips arrived last week to critical acclaim, notching an impressive step forward for AMD's notebook ambitions as it takes the process node leadership position against Intel in laptops for the first time in history, but Intel isn't sitting still – its retort comes in the form of the H-Series Comet Lake chips for high-performance gaming and creativity-focused laptops.
Intel's new lineup pushes clock rates up to a maximum of 5.3 GHz, a new peak for laptops, and the company also brings its Core i7 lineup that comprises the bulk of its sales volume up to a peak of eight cores and 5.1 GHz for the first time. That's a notable clock speed advantage over AMD's Ryzen lineup, but unlike AMD's chips that come with a denser 7nm process and a new architecture, Intel's new lineup isn't really all that new: The chips are an iterative update of the previous-gen Coffee Lake processors and come with yet another flavor of the highly-refined 14nm process, so they still span from four to eight cores and fit within a 45W TDP.
Intel positions its chips as the leaders for desktop gaming, which often relies upon high clock speeds, but the company's peak 5.0 GHz (and beyond) clock speeds come in the form of its Thermal Velocity Boost (TVB) technology that requires the processor to operate in optimal temperature ranges, so performance will obviously vary based on the laptop's cooling capabilities. That means you'll likely only see those peaks in thicker and heavier laptops.
In either case, Intel touts that it has 30 thin-and-light gaming designs (sub-20mm thickness) coming to market, and over 100 other new laptop designs on the way. The company also recently touted that it has 70+ Project Athena laptops in the works, so it will obviously rely upon tight integration to wring out extra performance from other platform-level advantages, like Wi-Fi 6, AX201 Gig+, and Thunderbolt 3 connections.
|Processor Number||Base / Boost||Cores / Threads||TVB||L3 Cache||Memory|
|Intel Core i9-10980HK||2.4 / 5.3||8 / 16||x||16 MB||DDR4-2933|
|Core i9-9980HK||2.4 / 5.0||8 / 16||x||16 MB||DDR4-2666|
|Core i9-9980H||2.3 / 4.8||8 / 16||x||12 MB||DDR4-2666|
|Intel Core i7-10875H||2.3 / 5.1||8 / 16||x||16 MB||DDR4-2933|
|Intel Core i7-10850H||2.7 / 5.1||6 / 12||x||12 MB||DDR4-2933|
|Core i7-9850H||2.6 / 4.6||6 / 12||Row 5 - Cell 3||12 MB||DDR4-2666|
|Intel Core i7-10750H||2.6 / 5.0||6 / 12||x||12 MB||DDR4-2933|
|Core i7-9750H||2.6 / 4.5||6 / 12||Row 7 - Cell 3||12 MB||DDR4-2666|
|Intel Core i5-10400H||2.6 / 4.6||4 / 8||Row 8 - Cell 3||8 MB||DDR4-2933|
|Core i5-9400H||2.5 / 4.3||4 / 8||Row 9 - Cell 3||8 MB||DDR4-2666|
|Intel Core i5-10300H||2.5 / 4.5||4 / 8||Row 10 - Cell 3||8 MB||DDR4-2933|
|Core i5-9300H||2.4 / 4.1||4 / 8||Row 11 - Cell 3||8 MB||DDR4-2666|
Intel bumped up the peak dual-core clock speeds to 5.3 GHz for the flagship Core i9-10980HK, a nice 300 MHz increase, but as mentioned, this is only under optimal thermal and electrical conditions and likely won't last for any extended period of time. Intel says the standard Turbo Boost 3.0 / 2.0 frequencies are 200 MHz less than listed, and you'll see those more often during normal use. Intel also divulged that the -10980HK offers an all-core boost of 4.4 GHz and several of the chips have a configurable TDP that peaks at 65W.
Intel says the Thermal Velocity Boost feature contributes 200 MHz to the top boost frequency if the processor is at 65C or below, or 100 MHz if the CPU is from 65C to 85C. This approach leads to a two-tiered maximum power consumption approach, with non-TVB boost clocks pulling 107W (PL2), and peak power weighing in at 135W (an increase of 10W over the previous gen) with a 56 second default tau (boost duration).
Intel also brought its Core i7 lineup to a maximum of eight cores and sixteen threads with the -10875H, and while the company hasn't shared pricing, we expect this to come with a similar price tag to the previous-gen six-core model. Intel also brought clocks speeds up to a peak 5.1 GHz for its Core i7 lineup, and given that the i7 range comprises the bulk of sales, the company expects more than 60% of its new lineup to ship with 5.0 GHz (or beyond) capabilities.
Aside from those changes, the bulk of the improvements lies in the incremental clock speed improvements across the stack. Intel also improved memory from dual-channel DDR4-2666 to DDR4-2933 (up to 128GB) and maintained Optane memory support.
As before, the HK model is overclockable, while the Core i7-10850H is partially overclockable (a limit of 400 extra MHz). Intel also touts its new Adaptix Tuning Technology that is an automated overclocking software similar to its one-click Intel Performance Maximizer software for desktop chips. This software automatically dials in the best possible overclock based on the power and thermal capabilities of the laptop.
The first image above shows the Comet Lake die, while the second shows the previous-gen Coffee Lake die. As you can see, the die appear identical to the naked eye. Intel has made improvements to the fifth iteration of the 14nm process and perhaps reworked some critical pathways to improve performance, but there appears to be little to no modification on the overall architecture. Intel hasn't disclosed if there are new hardware-based mitigations for the new chips.
Intel does have its 10nm Tiger Lake processors coming to market this year, but said that it continues to use 14nm for the 45W chips due to its higher frequency capabilities.
|Processor Number||Base||Boost||Cores / Threads||TDP||L3 Cache||Memory|
|Intel Core i9-10980HK||2.4||5.3||8 / 16||45||16 MB||Dual DDR4-2933|
|AMD Ryzen 7 4900H||3.3||4.4||8 / 16||45||8 MB||Dual DDR4-3200|
|Intel Core i7-10875H||2.3||5.1||8 / 16||45||16 MB||Dual DDR4-2933|
|AMD Ryzen 7 4800H||2.9||4.2||8 / 16||45||8 MB||Dual DDR4-3200|
|Intel Core i7-10850H||2.7||5.1||6 / 12||45||12 MB||Dual DDR4-2933|
|Intel Core i7-10750H||2.6||5.0||6 / 12||45||12 MB||Dual DDR4-2933|
|AMD Ryzen 5 4600H||3.0||4.0||6 / 12||45||8 MB||Dual DDR4-3200|
|Intel Core i5-10400H||2.6||4.6||4 / 8||45||8 MB||Dual DDR4-2933|
|Intel Core i5-10300H||2.5||4.5||4 / 8||45||8 MB||Dual DDR4-2933|
The true test comes when Intel lines up against AMD's ultra-competitive Ryzen 4000-series. As we can see, Intel matches AMD core-for-core across the stack, but slathers in higher clock speeds to compete. However, AMD's chips will likely maintain their integrated graphics superiority, and the denser 7nm node brings power improvements that are hard to match.
Intel provided benchmarks comparing its new chips against three-year-old laptops, which is standard fare when the company compares performance based on upgrade cycles. They aren't very useful, though, and as with all vendor-provided benchmarks, you should take these with a grain of salt. Intel acknowledged that this is an unusually slim selection of benchmarks, but representatives said that production delays from China due to the coronavirus have left the company without many shipping models for testing. These projections also don't have any of the new Ryzen 4000 series models listed, but that's likely due to their arrival a mere two days ago. These factors combined make these benchmarks largely useless for any real comparisons.
Intel's new lineup leverages the company's tried-and-true tactic of using superior clock speeds, which often equate to improved gaming performance, as its major weapon against AMD's latest chips, but we won't know the real story until laptops hit the market. That's largely because Intel's boost mechanisms are limited by the thermal performance of the laptop, so performance will vary widely between different systems, especially as Intel turns up the power dial to fight off AMD.
Intel has vPro-equipped and Xeon models of these same chips coming later this year to complement a broad ecosystem of consumer laptops. Intel also continues to develop its Project Athena initiative that grants the company some advantages in the premium design space, but the areas where the company has a clear win are growing narrower. Meanwhile, AMD's laptop roster, not to mention the quality of the Ryzen 4000 laptops' secondary components, continue to improve at a rapid pace. That's critical as AMD continues to push into the laptop market that comprises 60% of the overall consumer chip market. We can't wait to see how the two lineups compare once they hit our labs.
Intel numbers, as AMD, Nvidia, and the likes have to be (always) taken with a grain of salt.
Its the i7-10875H that is 8c/16t, other than the 10980HK.
Second there is no way we see laptops with these boosts. See this from Ian at Anandtech. There is virtually no way a laptop maker will implement a PSU capable of producing the amperage needed.
"It does mean that in order to hit 5.3 GHz, the Core i9 is by default allowed to take 135 W across two cores, or 67.5 W per core. Even at 60W per core, you're looking at 50A of current per core... in a laptop. "