News Intel Core i9-12900K and Core i5-12600K Review: Retaking the Gaming Crown

Mar 19, 2021
10
10
15
How is this a scientific test? No PBO on any of the AMD processors? A AMD 5600x still wrecks these new intel chips. Literally with the click of the button in the bios you enanble PBO and safely run AMD overlocking without any issues and you go straight to the leaderboards. Would love to see some more scientific and fair testing.
 
Dec 22, 2020
4
1
15
Here, the review article is offline (Error 404) from the page, the following message appears: "Sorry! Page not found.
The page you are looking for has been moved or removed from the site.

Please try searching our website or start again from our home page. "
 
Here, the review article is offline (Error 404) from the page, the following message appears: "Sorry! Page not found.
The page you are looking for has been moved or removed from the site.

Please try searching our website or start again from our home page. "
Sorry, we had a little glitch. It's back up.
 
We put Intel's new Core i9-12900K and Core i5-2600K through the wringer in Windows 10 and 11 testing with DDR4 and DDR5 memory.

Intel Core i9-12900K and Core i5-12600K Review: Retaking the Gaming Crown : Read more

We're sticking with our standard policy of allowing the motherboard to exceed Intel's recommended power limits, provided the chip remains within warrantied operating conditions. Our tests use the default lifted PL1 and PL2 restrictions. Almost all enthusiast-class motherboards come with similar settings, so this reflects the out-of-box experience with a high-end motherboard. Naturally, these lifted power limits equate to more power consumption, and thus more heat, as we'll cover in detail later in the review.
Why? Everybody is on the power conscious trip right now, why not show both so that people can actually choose which one is better for them?
The 12900k shows so much improvement that running it with base power would still be fine.
People need both PBP and MTP to make an informed decision.
 

InvalidError

Titan
Moderator
The test setup table incorrectly lists Alder Lake as LGA1200 and the memory specs doesn't make sense for either board:
"2 x16GB G.Skill Ripjaws S5, DDR5-5200 @ DDR4-4400 36-36-36-72 "
Can't use DDR5 DIMMs on the DDR4 board, can't spec DDR4 on a DDR5 board, timings don't make sense for DDR4-4400 which typically has 18-22 CAS latency. Wrong description copy-pasted to both places.
 

Howardohyea

Commendable
May 13, 2021
259
64
1,790
personally I really like this one long page split into multiple smaller ones, makes scrolling way easier.

Also is this review a bit early? I thought the embargo only lifts at 10 AM, while this review is 10 minutes before that
 

m3city

Reputable
Sep 17, 2020
25
18
4,535
Synthetic tests show Intel is faster, real world tests: application - lets say its on par, or Intel slower. Gaming - TLDR shows 9% more FPS. But:
  • with much higher power draw -> performance /watt lower by ~30 to 40%? power draw means bigger/costly cooling solution, more powerfull PSU etc.
  • you get to pay extra for MB. And then one is penealted by Intel that will require a new MB for next generation (correct me if I'm wrong here)
Personally, what matters for me is low power, silence, efficiency and performance at real world tasks. As it goes for gaming, there is really not much to show from intel, especially these differences will show up only with the monster-type GPU. And I couldnt care more for difference between 140 and 150 fps in Far Cry.

And on top of that, Intel matched CPUs released 11.2020 to 01.2021.
 
How is this a scientific test? No PBO on any of the AMD processors? A AMD 5600x still wrecks these new intel chips. Literally with the click of the button in the bios you enanble PBO and safely run AMD overlocking without any issues and you go straight to the leaderboards. Would love to see some more scientific and fair testing.
Scientific and fair would be stock performance first, follow up with overclocking. Time is limited, games like to lock you out when you run them on too many different CPUs, and Windows 10 + Windows 11, DDR4 + DDR5, Stock + OC ends up being a LOT of permutations. PBO is good for about 2-5% more performance at best. It's not even remotely going to close the gap between 12600K and 5600X. It's okay for Intel to retake the gaming performance crown. The fact that AMD had the CPU lead for several years, at least in some respects, was a good run. Zen 4 might make a good counterattack. But Alder Lake is a very impressive set of architectural and software improvements and Intel deserves credit for that.
 
The test setup table incorrectly lists Alder Lake as LGA1200 and the memory specs doesn't make sense for either board:
"2 x16GB G.Skill Ripjaws S5, DDR5-5200 @ DDR4-4400 36-36-36-72 "
Can't use DDR5 DIMMs on the DDR4 board, can't spec DDR4 on a DDR5 board, timings don't make sense for DDR4-4400 which typically has 18-22 CAS latency. Wrong description copy-pasted to both places.
Fixed. Obviously that should have said "@ DDR5-4400..."
 
  • Like
Reactions: Howardohyea

King_V

Illustrious
Ambassador
So, I'm seeing that Intel is performing better, yes. And I am definitely intrigued by the hybrid design.

But, I'm also still seeing that, even with their 10nm process (which they call "7"), they still rely on higher clocks, and greater power consumption. I definitely do appreciate that they list explicitly "base power" vs "turbo power" rather than trying to hide using base power as their TDP listing.

If they're going to claim their 10nm process is the equivalent of TSMC's 7nm, then shouldn't their performance/watt be on par with it? A big improvement over Rocket Lake, certainly, but, still not up to AMD's current offerings.
 

InvalidError

Titan
Moderator
Fixed. Obviously that should have said "@ DDR5-4400..."
Doesn't look fixed to me: the table still says the MSI Z690A WiFi DDR4 is using DDR5.

If they're going to claim their 10nm process is the equivalent of TSMC's 7nm, then shouldn't their performance/watt be on par with it?
It is, when you aren't in a scenario where it needs to be permanently on max turbo on all cores. In most games, power draw is usually within +/- 10% of Ryzen for the same or greater performance.
 
Jan 21, 2021
10
6
15
Any word on what Linux support for thread scheduling is like? I'm curious how this will affect server and workstation folks.

Alder Lake does not support AVX-512 under any condition (fused off in P cores, not supported in E cores) to ensure an even ISA application.

Looking at the power and performance numbers, is BIG.little spelling the end of AVX-512 for the foreseeable future? Seeing the maximum power numbers backing down from 300W is good to see.
 
  • Like
Reactions: ezst036 and kinney
Finally a review!

Seems intel was able to take some of the test first place, using way more power, but its there finally.

For gaming the 12900K does not seems a good choice, I agree with the review better wait for 12700K results before speding the amount of money and cooling a 12900K need.

Speaking of that, is it me or there are no temperature test anywhere to be found?


The 12600K seems a decent choice for gaming, but the fact you have to get new mobo and windows 11 for it to "shine" its really a turn down for me. At least for the time been and specially for gaming at 1440p.

After all unless you have such a powefull GPU (RTX 3090) the margins between Ryzen 5xxx and Alder Laker will get smaller.


As for productivity for the next few months I think Ryzen 5xxx and Core 10th Gen + Windows 10 should still be the better choice for stability. But still impresive results on this few tests.

Cheers!
 
  • Like
Reactions: Roland Of Gilead

TJ Hooker

Titan
Ambassador
The power draw numbers for the 11900K in the Blender tests seem off. Why are they 100+W higher than the 11900K power in the other tests, and why are they 100+W higher than the power draw measured in the same Blender tests from the 11900K review? Even the OC results from that review have much lower Blender power draw than what is listed here.

Edit: The 11700K blender power draw is also significantly higher than for the same Blender test in the 11700K review. 11600K power is a little higher than in previous reviews.
/Edit

On a slightly different note, I agree that it'd be interesting to see what power draw numbers for Alder Lake look like with a boost time limit enabled (either 56 s, or whatever the default will be for non-K ADL chips). Unrestricted boost is usually pretty bad for efficiency, past generations of Intel usually show only small performance differences with significantly higher power. Similar to how AMD PBO usually drops efficiency by a noticeable amount.
 
Last edited:
  • Like
Reactions: kinney

kinney

Distinguished
Sep 24, 2001
2,262
17
19,785
How is this a scientific test? No PBO on any of the AMD processors? A AMD 5600x still wrecks these new intel chips. Literally with the click of the button in the bios you enanble PBO and safely run AMD overlocking without any issues and you go straight to the leaderboards. Would love to see some more scientific and fair testing.

While I agree with Jerrad's response, it wouldn't cover that sort of gap, PBO isn't covered under warranty so it shouldn't be in a launch day review. All CPUs should be tested at settings that are covered under warranty, at least that's what I'd expect to see.
 

kinney

Distinguished
Sep 24, 2001
2,262
17
19,785
It seems that the 11900K has Adaptive Boost disabled unless specified in the charts? If so that's a huge disservice and misleading. What about MCE? By default on most if not all Z590 boards MCE is enabled and neither are considered overclocking by Intel thus is covered under warranty. I had both a 11900K and 5900X rig and fell in love with the 11900K with MCE left on and ABT enabled to where I sold the 5900X. I don't care about power usage, I'm idling 90% of the hours the machine is on and just want the best performance when I need it. It hits 5.3GHz on 7 of 8 cores simultaneously (Imgur), and holds 5.1Ghz on all 8 cores while gaming, never dipping below. For a daily driver rig, it's quick, the fastest I've ever used.

Regardless, I'm looking forward to upgrading to the 12900K. Great chip, it'll be interesting to see how this evolves.
 
Last edited:
  • Like
Reactions: bolweval
I am honestly a bit disappointed. I was hoping for a larger gap. Something like 15-20%, at least then it would more justify the cost of the z690 motherboards and ddr5. Either way, now I will bide my time until the 192mb L3 cache 5000 refresh comes out. Hopefully they will punch back at this ~9% loss to intel in gaming.
 

enewmen

Distinguished
Mar 6, 2005
2,247
3
19,815
Thanks for the review. Maybe a good suggestion is another test with performance per watt as the focus. Alder Lake vs Zen3 and big vs little cores - when available.
 

kinney

Distinguished
Sep 24, 2001
2,262
17
19,785
Any word on what Linux support for thread scheduling is like? I'm curious how this will affect server and workstation folks.



Looking at the power and performance numbers, is BIG.little spelling the end of AVX-512 for the foreseeable future? Seeing the maximum power numbers backing down from 300W is good to see.

The loss of AVX512 is unfortunate because Intel engineering did an AMA not long ago where they stated that they're going to continue pushing AVX512. I'm expecting this to return on future iterations of this architecture once they sort out the additional complexity in the ISA. Intel is not going to confirm or deny that today. 🙂I'm going to build one of these anyway since I keep two rigs, but for someone looking to keep a machine further out, AVX512 has proven to be extremely potent when it's used.
 
  • Like
Reactions: Flemkopf