Intel has finally launched its latest generation of new desktop chips, known as 10th-gen Comet Lake-S. They were announced months ago, and in that time, leaks with nearly every spec have been posted online. Intel still has a couple of tricks up its sleeve, though.
These processors, ranging from Core i3 up to Core i9, aim to up the ante against AMD’s Ryzen 3000 processors. Despite being another iteration of the same microarchitecture, Intel has brought some meaningful changes to the lineup that add some extra value to its chips, as well as some disappointments.
But is it all enough to hold back the rising tide of Ryzen dominance?
The Core i9-10900K claims 5.3GHz
Intel calls its 10th-gen desktop chips the “fastest gaming processors.” It bases that claim primarily on these chips’ most impressive spec, clock speed. At the top of the stack is the new Core i9-10900K, an unlocked 10-core processor with a maximum of 5.3GHz. You won’t find another chip that even claims to be that fast. In fact, AMD’s chips still don’t make it over the 5.0GHz hurdle.
Of course, there are a couple of caveats with that frequency claim. One is that 5.3GHz is only possible with Intel’s Thermal Velocity Boost (TVB). It’s an A.I.-powered frequency boost that opportunistically finds chances to push to higher frequencies. It isn’t guaranteed, and it can only be achieved in ideal circumstances. TVB is also only a single-core boost, whereas the Core i9-10900K only hits 4.8GHz on all cores.
The other big change is core count. In response to AMD’s aggressive addition of cores and threads, Intel has moved its top Core i9 from eight cores to 10. This makes for a clear distinction between Core i9 and Core i7, and also makes the Core i9 a bit more competitive. It sits right in between the 8-core Ryzen 9 3800X and 12-core Ryzen 9 3900X, despite being more expensive at $488. It’s still trounced by the 16-core Ryzen 9 3950X, but that’s a $700 processor.
The price disparity and lack of a high-end competitor is still notable, though. Before the rise of Ryzen, Intel used to call its new generation of chips the fastest processors overall. But that hasn’t been true for a couple of years now, and it’s still not true with 10th-gen.
The 10th-gen lineup
Beyond the the Core i9 chips, you’ll a wide variety of Core i7, Core i5, and Core i3. Here’s what the lineup looks like:
|Cores/Threads||Base clock||Intel Turbo Boost 2.0/3.0||Boost clock (all-core)||TDP||Price|
|Core i5-10600K||6/8||4.1GHz||4.8GHz/ —||4.5GHz||125w||$262|
|Core i5-10600||6/8||3.3GHz||4.8GHz/ —||4.4GHz||65w||$213|
|Core i5-10500||6/8||3.1GHz||4.5GHz/ —||4.2GHz||65w||$192|
|Core i5-10400||6/8||2.9GHz||4.3GHz/ —||4.0GHz||65w||$182|
|Core i3-10320||4/8||3.8GHz||4.6GHz/ —||4.4GHz||65w||$154|
|Core i3-10300||4/8||3.7GHz||4.4GHz/ —||4.2GHz||65w||$143|
|Core i3-10100||4/8||3.6GHz||4.3GHz/ —||4.1GHz||65w||$122|
Outside of the Core i3 chips, each model includes an unlocked “K” variant for overclocking. The 9th-gen predecessor to these were rated at 95 watts, but this year’s chips go all the way up to 125 watts. That’s a big jump in heat, even compared to AMD’s notoriously hot chips, which don’t go over 105 watts. The primary reason is due to the increase in cores — at least that’s what an Intel representative told me. According to Intel, the shared thermal design power (TDP) of these unlocked “K” chips is aimed at simplifying manufacturing process and cooling solution recommendations.
The other major change is hyperthreading. It’s been turned on across the board, bringing more threads to the entire stack. That brings it competitive with what AMD offers on these midrange and budget-tier parts.
As shown in the chart above, Intel is debuting the third generation of Intel Turbo Boost, though only on Core i7 and Core i9 processors. Turbo Boost 3.0 allows the processor to identify its two highest-performing cores and allows for short bursts of improved turbo performance. The result is the possibility for an extra 0.1GHz on these top processors. If it’s not obvious yet, Intel intends to eke out every possible clock speed gain it can from these aging processors. In the case of the Core i7-10700, hyperthreading and Turbo Boost 3.0 are the only changes over 2019’s offerings.
Intel has also introduced new overclocking features to squeeze out more performance from these chips. You can now disable hyperthreading per core, which Intel says can slightly extend time in turbo. Intel also introduced enhanced voltage and frequency curve controls, as well as graphical enhancements to the Intel Extreme Tuning Utility.
Intel’s 10th-gen line also includes a number of “F” variants that don’t include discrete graphics for a discounted price. Intel has also updated its 35-watt T-series of chips, which are commonly used in all-in-ones and other restricted form factors.
Specs are specs, but what about real-world performance? Well, Intel didn’t make the direct comparison to AMD, instead relying on comparisons versus older Intel chips. Unfortunately, there aren’t many specifics provided about how Intel arrived at these percentages. I can comfortably say that the quoted “up to 18-percent faster performance in 4K video editing” is likely due to the increase in cores in the Core i9 and hyperthreading throughout.
Intel also makes the comparison to a three-year old PC to show how big of an increase you could expect if you’ve been waiting to upgrade your CPU for a few years. Again, the details on what systems were used in this comparison weren’t given.
And then, there’s game performance. Many players will consider it cheating, but Intel relies on deep integration and optimization to achieve higher performance in titles. That means in some specific applications or games, Intel’s chips will win over AMD. That’s not because of Intel’s raw processing strength, but because of Intel’s huge reach. It’s built up that engineering partnership over the years, and though it doesn’t apply across the gamut, there are real performance differences.
Intel used Total War: Three Kingdoms and Remnant: From The Ashes as examples. In Total War, Intel optimized features like A.I. animations and physics specifically for its own platform, which the company claims results in six times more soldiers on screen at a time.
For real performance comparisons, we will need to test these chips ourselves. However, I expect results similar to last year’s comparison.
No PCIe 4 or 10nm
Intel’s long struggle with thw 10nm process was supposed to be over in 2019 with the introduction of 10th-gen processors. That’s true for mobile, where Intel has launched its 10th-gen Ice Lake chips. The selection is limited to only low-powered U and Y-series chips, topping out at 15 watts. Anything higher still uses the 14nm process, including these new desktop chips.
It’s not something Intel talks much about, but it’s proving to be a barrier to staying competitive with AMD. Shrinking the size of the die is all about transistor count. Intel used to double the transistor count every year or two, but has been stuck on 14nm for almost six years now. AMD, meanwhile, has already moved to a 7nm node, in 2019, which is equivalent to Intel’s 10nm.
As Intel’s major desktop release in 2020, 10th-gen Comet Lake-S confirms that we won’t get 10nm desktop chips until 2021 at the earliest. The successor to Comet Lake-S will be Rocket Lake, which is rumored to use a hybrid chiplet design, but still not a full embrace of 10nm. Meanwhile, AMD has its next-generation Ryzen 4000 desktop chips set for release later this year.
Intel’s 10th-gen still also doesn’t support PCIe 4.0. The latest generation of PCIe provides double the bandwidth of the previous generation, up to 32GB per second. That won’t affect performance of graphics cards much for now, but you can expect M.2 SSDs to get a bump in speed thanks to PCIe 4.0. AMD moved to this new standard with Ryzen 3000 as well, and we’re already seeing the first wave of PCIe 4.0 SSDs.