Tuesday, July 20th 2021
Intel Core i9-12900K Allegedly Beats AMD Ryzen 9 5950X at Cinebench R20
With qualification samples of the upcoming Intel Core i9-12900K "Alder Lake-S" processors and companion Socket LGA1700 motherboards hitting the black-market, expect a deluge of benchmarks on social media. One such that stands out makes a fascinating claim that the i9-12900K beats AMD's current flagship Ryzen 9 5950X processor at Cinebench R20, which has been AMD's favorite multi-threaded benchmark. At stock speeds, with liquid cooling, the i9-12900K allegedly scores 810 points in the single-threaded test, and 11600 points in multi-threaded.
To put these numbers into perspective, a retail Ryzen 9 5950X scores 641 points in the single-threaded test, and 10234 points in multi-threaded, in our own testing. The i9-12900K is technically a 16-core processor, just like the 5950X, but half its cores are low-power "Gracemont." The "Alder Lake-S" chip appears to be making up ground on the single-threaded performance of the "Golden Cove" P-core, that's a whopping 25% higher than the "Zen 3" core on the 5950X. This is aided not just by higher IPC, but also the max boost frequency of 5.30 GHz for 1~2 cores, and 5.00 GHz "all-core" boost (for the P-cores).Given the multi-threaded scores, it's safe to assume that either Intel or Microsoft has figured out a way to leverage the P-cores and E-cores simultaneously in peak multi-threaded workloads. This is possible when both the "Golden Cove" and "Gracemont" cores have the ISA capability needed by the workload, which in case of Cinebench R20, is AVX. "Gracemont" is Intel's first low-power core to support AVX, AVX2, and AVX-VNNI instruction sets. "Golden Cove" features a more lavish ISA that includes AVX-512 (select client-relevant instructions).
Sources:
OneRaichu (Twitter), VideoCardz
To put these numbers into perspective, a retail Ryzen 9 5950X scores 641 points in the single-threaded test, and 10234 points in multi-threaded, in our own testing. The i9-12900K is technically a 16-core processor, just like the 5950X, but half its cores are low-power "Gracemont." The "Alder Lake-S" chip appears to be making up ground on the single-threaded performance of the "Golden Cove" P-core, that's a whopping 25% higher than the "Zen 3" core on the 5950X. This is aided not just by higher IPC, but also the max boost frequency of 5.30 GHz for 1~2 cores, and 5.00 GHz "all-core" boost (for the P-cores).Given the multi-threaded scores, it's safe to assume that either Intel or Microsoft has figured out a way to leverage the P-cores and E-cores simultaneously in peak multi-threaded workloads. This is possible when both the "Golden Cove" and "Gracemont" cores have the ISA capability needed by the workload, which in case of Cinebench R20, is AVX. "Gracemont" is Intel's first low-power core to support AVX, AVX2, and AVX-VNNI instruction sets. "Golden Cove" features a more lavish ISA that includes AVX-512 (select client-relevant instructions).
155 Comments on Intel Core i9-12900K Allegedly Beats AMD Ryzen 9 5950X at Cinebench R20
You only really need little cores for laptops where battery life is an issue !
Or they still can't keep their brand new architecture thermals under a reasonable limit? Me wonders whether the intel trickery is the avx512 vnni instructions ?
Probably best to take it with a mountain of salt until retail silicon is properly tested.
If the results are even remotely real, then those new Atom cores must have gained some AVX or AVX2 capability though.
R20 doesn't throw enough avx to trigger an -avx offset so chances of any avx 512 trickery is pretty slim
R23 is suposed to have avx but still isn't enough to trigger use of an -avx offset "if set to in bios", never has for me anyway.
Not sure why you don't get the offset, probably you got good cooling.
Or will it be a paper launch???? ;)
Whether that will be enough not to be a paper launch is a different question though, Ice Lake Xeons are also now out and these are massive and probably sell well.
All that "backwards support" on AM4 has been a trainwreck so far. Some people buy AMD hardware due to this feature (that AMD loves to gimp more and more as time goes on), but to me it made Ryzen unsellable (along with many other reasons). As a buyer, I have no idea with what BIOS board comes and what chips it supports out of box, there's no way to flash board to newer BIOS version if you need it with any Ryzen chip, therefore you depend on your local retailer to be kind enough and deal with it. So far, only one retailer actually messed with that as far as I know and if you bought from someone else, well you might have been completely screwed and had to buy e-waste A series chip, which makes already overpriced and overhyped Ryzen platform even more expensive. So I went with Intel, mostly because it was cheaper and because I was certain, that I won't have such stupid problems that need expensive solutions. And another problem was that AMD used to release new generation chips almost a year before natively compatible boards were released and then you had to wait, before those new boards that did exactly the same shit actually were priced like the older ones. Maybe in USA and other real first world countries something like that actually works, but not so much here and in actual 3rd world. And more than that, why to this day boards can't be flashed without CPU in it? Really how expensive could it be to put some cheapo ARM chip to do that or even better some e-waste tier MIPS chip from router? All in all, this backwards compatibility feature became a major PITA. I would have been fine with new board for each CPU generation, rather than something like that. But even if it was mostly a lie (because it didn't work out at all), I guess it helped AMD to sell first gen Ryzen and put it on media, so people would know that FX era ended and Ryzen is now cool (although it was unstable, buggy and in so many other ways unmarketable). I'm afraid, that until AMD fixes such a major issue, they won't sell me any of that Zen.
However, on topic, I'll wait on reviews ,Intel hype trains aren't new after all.
Here are benches on HWUB:
i5 performed really well, just like every other chip in that test and at least locally i5 is two tiems cheaper than 3600, not only that but has been for nearly two years without any price changes. If you want even cheaper CPU, then there's i3 10100(F), which is quite fast too:
It's not much slower than i5:
And here you go, your all expenses could have been minimized to just 100 USD and if you bought 3 AM4 chips, you literally spent over 1k USD for minimal gains. The entire i3 platform would have been 220-280 USD, literally less than 5800X alone.
The best case scenario is that you only used those AMD chips for a while and sold them for close to what you paid, then you could have minimized your losses, but even then, you ended up with 5800X and overpriced X570 board (there's no such things as reasonably priced X570 board, if you budget, you buy B550 stuff).
I know a thing or two about keeping a cpu for years I was on a i7-970 for 10 years before going AM4. And i'm not even sure why i'm wasting my time responding since you wrote a novel based on no history at all from my side. I paid $440 for the 3800X in dec 2019. Sold it for $400 in Jan 2021 so for a year of use $40. I picked up the 5800X for $650 CAD however because of the prior sale from old cpu cost to me was only $250. So that was what I paid to go from Zen 2 to Zen 3.
The rest of your video's irrelevant I was already on a AM4 platform I was not going intel.
Moral of the story you don't get to decide what is a solid investment for someone else when you aren't putting any funds towards their purchase. This whole post was just you telling me I should have gone intel......
The other thing with your post is 100% gaming. I'm an adult boss I do more than play games on my PC, gaming performance is not the primary factor in my builds.
Intel platform, over 6 years, would have cost you just 42 USD per year. And your whole Zen adventure is easily double that, closer to 90 USD per year.