News Core i9-12900HK Beats Threadripper 1950X In Cinebench R20 Benchmark

Neilbob

Distinguished
Mar 31, 2014
202
235
19,620
It is a stupid comparison, but as the article points out, this is the first mobile CPU to beat the 1950x. This is a supposed 115W (max turbo) mobile beating a 180W HEDT CPU.
180W on an arguably inferior GloFo 14nm process and really rather poor clock speeds.
Not intending to be confrontational or anything, but I just felt that was worth pointing out. (y)

I do agree that, regardless of the pointlessness of the comparison, it is still fairly impressive nonetheless (although with that said, it seems crazy to me that a 10nm /'Intel 7' mobile chip is using 115W under any circumstances).
 
  • Like
Reactions: Makaveli

jgraham11

Distinguished
Jan 15, 2010
54
21
18,535
180W on an arguably inferior GloFo 14nm process and really rather poor clock speeds.
Not intending to be confrontational or anything, but I just felt that was worth pointing out. (y)

I do agree that, regardless of the pointlessness of the comparison, it is still fairly impressive nonetheless (although with that said, it seems crazy to me that a 10nm /'Intel 7' mobile chip is using 115W under any circumstances).

Agreed, its 4 years later and with the 12900HK consuming 115W of power for the majority of the Cinebench run, its only beating 180W CPU from 4 years ago and is 36% more efficient. Good job Intel, you did it!
 
  • Like
Reactions: VforV and Makaveli

spongiemaster

Admirable
Dec 12, 2019
2,276
1,280
7,560
Agreed, its 4 years later and with the 12900HK consuming 115W of power for the majority of the Cinebench run, its only beating 180W CPU from 4 years ago and is 36% more efficient. Good job Intel, you did it!
Seems the irony is lost on the people trashing this achievement that AMD themselves hasn't come close to achieving it yet. Let us know when an AMD mobile CPU at any TDP beats a 1950X.
 

VforV

Respectable
BANNED
Oct 9, 2019
578
287
2,270
Intel is desperate for positive AL press (like it does not have enough already), since it's still not selling like a hot cake at all!

For over a month now since December I've been watching a local store in EU and it has (I kid you not) the 12600K at $273 and 12600KF at $266 and they were always "in stock" non-stop (while Ryzen 5600x is at $304, mid you). And this is a pretty small shop, so if it were the case of selling good they should have run out of stock fast, yet it did not happen once...

I find it very, very funny. No wonder the press is pushing the click bait like crazy to generate more (fake) AL interest. Pffff.
 
Apr 1, 2020
1,431
1,083
7,060
I was about to post how it's 25% faster than the 5900HX in Blender and that's an achievement, but it takes the 12900HK much more power to accomplish that feat, but I don't think it's a valid comparison here.

In a mobile workstation where power and battery life isn't as relevant as it's being connected to AC power, the 12900HK provides a massive performance benefit over the 5900HK, assuming the performance scales as well in that software as it does in Cinebench and Blender. In a gaming notebook it's a different story as we saw with the ASUS G14.

I'm no fan of Intel, but the fact they're putting this kind of performance in a mobile processor is GOOD for the market and puts AMD at a disadvantage. This also goes back to my post on another topic that I am disappointed that AMD doesn't seem to be raising core counts with Zen 4. I remember seeing articles saying that AMD has patents on the same kind of mixed core strategy Intel is using, but they haven't implemented them in a retail product yet, and AMD may be forced to do that, even if it's just in the mobile market.
 
  • Like
Reactions: jacob249358

larkspur

Distinguished
For over a month now since December I've been watching a local store in EU and it has (I kid you not) the 12600K at $273 and 12600KF at $266 and they were always "in stock" non-stop (while Ryzen 5600x is at $304, mid you). And this is a pretty small shop, so if it were the case of selling good they should have run out of stock fast, yet it did not happen once...
Not sure how you come to these conclusions. Those are great prices for AL vs Zen 3. More real-world evidence that Intel is absolutely dominating price/perf. They are clearly mass producing the heck out of these chips - you do realize they have a LOT more manufacturing capacity than TSMC right? They have a superior platform, a zillion different SKUs hitting the market addressing every single consumer performance bracket and budget and they have availability(something we all used to take for granted)! Stating the obvious here but - we wouldn't see such availability if Intel relied on TSMC's fabs to produce these chips.

Ultimately I like Rembrandt in the mobile space (that iGPU looks fantastic) but until we see reviews, AL mobile appears to be top notch in CPU performance. Near 12600k performance in a mobile system? Caveats aside - YES PLEASE!

With nothing new announced by AMD for AM4 other than the expensive and gaming-centric 5800X3D and with it becoming increasingly apparent that AM5 will be DDR5-only in a time of massive DDR5 price inflation, it is more obvious than ever that AL is the go to for new systems now and for at least another half year. HUB mentions this in the 12700 vs 5800X video. I wish I could argue otherwise - I was hoping to see AMD drop some prices or come out with something new to compete. Oh well, I guess. When you are supply-constrained by TSMC, making such price cuts is difficult.
 

Makaveli

Splendid
I was about to post how it's 25% faster than the 5900HX in Blender and that's an achievement, but it takes the 12900HK much more power to accomplish that feat, but I don't think it's a valid comparison here.

In a mobile workstation where power and battery life isn't as relevant as it's being connected to AC power, the 12900HK provides a massive performance benefit over the 5900HK, assuming the performance scales as well in that software as it does in Cinebench and Blender. In a gaming notebook it's a different story as we saw with the ASUS G14.

I'm no fan of Intel, but the fact they're putting this kind of performance in a mobile processor is GOOD for the market and puts AMD at a disadvantage. This also goes back to my post on another topic that I am disappointed that AMD doesn't seem to be raising core counts with Zen 4. I remember seeing articles saying that AMD has patents on the same kind of mixed core strategy Intel is using, but they haven't implemented them in a retail product yet, and AMD may be forced to do that, even if it's just in the mobile market.

Where has it been posted AMD is not raising core counts on zen 4?
 
Apr 1, 2020
1,431
1,083
7,060
Where has it been posted AMD is not raising core counts on zen 4?

AMD’s Ryzen 7000 Zen 4 16-Core, 8-Core CPUs Break Cover | Tom's Hardware (tomshardware.com)

If that's true, then AMD isn't going to be raising core counts on Ryzen series processors. Desktop wise it's not a problem, but in the mobile market if Intel is going to trump AMD's core count, granted it might be with a mix of P and E cores, AMD's going to be at a disadvantage.

Also something that got me was the closing statement in WCCFTech's article on this issue:

AMD had the guts to compare their Zen 3D CPU to Intel's Alder Lake CPUs but they didn't even compare their Ryzen 6000 APUs to Intel's mobility lineup and their H-series was almost missing from the show (in terms of performance benchmarks). Even the Ryzen 6000U benchmarks were deceptive in comparison as they used a 15W (Ryzen 5000U) APU against a 28W (Ryzen 6000U) APU. AMD's Zen 3+ might make for some nice mainstream and entry-level designs but Intel has a really strong notebook line with Alder Lake that will target them aggressively at all segments.

And that disturbs me a bit. Why wouldn't AMD show up Intel?
 

teodoreh

Distinguished
Sep 23, 2007
315
13
18,785
Firstly, Threadripper 1950X is a 2017 processor build in 14nm. In case you haven't noticed, we are living in the year 2022. I am surprised, Intel hasn't tried to compare Core i9-12900HK with a Phenom II :D

Secondly, Threadripper is a desktop workhorse, Core i9-12900HK is a laptop processor with a crazy 110W TDP envelope. Try pushing this processor to its 110W limit and you will end up with throttling at best or a dead laptop at worse.

Third and most important, I don't trust Intel's leaks. It's the same company that forgot to mention the overclock and cooling of their 28-core CPU on the... same Cinebench . Picking up a single piece of software (which can be better optimized or more memory efficient) and leaking that your 110W TDP CPU is faster than a 4 year 180W TDP processor, is just, pathetic. How desperate Intel feels right now?
 

spongiemaster

Admirable
Dec 12, 2019
2,276
1,280
7,560
Firstly, Threadripper 1950X is a 2017 processor build in 14nm. In case you haven't noticed, we are living in the year 2022. I am surprised, Intel hasn't tried to compare Core i9-12900HK with a Phenom II :D

Secondly, Threadripper is a desktop workhorse, Core i9-12900HK is a laptop processor with a crazy 110W TDP envelope. Try pushing this processor to its 110W limit and you will end up with throttling at best or a dead laptop at worse.

Third and most important, I don't trust Intel's leaks. It's the same company that forgot to mention the overclock and cooling of their 28-core CPU on the... same Cinebench . Picking up a single piece of software (which can be better optimized or more memory efficient) and leaking that your 110W TDP CPU is faster than a 4 year 180W TDP processor, is just, pathetic. How desperate Intel feels right now?
Based on the specs AMD announced for the 6980HX, and comparing them to the 5900HX, the 6980HX isn't going to come close to matching the 12900HK in multi-threaded performance. What's going to be the excuse then for AMD's latest and greatest not being able to beat a lowly 4 year old 1950X?

This wasn't an Intel leak. The benchmarks were done by a 3rd party using a test platform they got their hands on.
 
  • Like
Reactions: sivaseemakurthi

Makaveli

Splendid
AMD’s Ryzen 7000 Zen 4 16-Core, 8-Core CPUs Break Cover | Tom's Hardware (tomshardware.com)

If that's true, then AMD isn't going to be raising core counts on Ryzen series processors. Desktop wise it's not a problem, but in the mobile market if Intel is going to trump AMD's core count, granted it might be with a mix of P and E cores, AMD's going to be at a disadvantage.

That link isn't really proof that they will not increase core counts past 16.
 
Meanwhile in other news, Intel is going to release the 13th-gen Raptor this fall, which according to reports in i9 form will have twice the E-cores of the i9 12900K. To think I just built an 11th-gen gaming rig last summer!

I tip my hat to AMD for waking up Intel and thankfully we still have competition. Time for AMD to step it back up now.
 
Apr 1, 2020
1,431
1,083
7,060
That link isn't really proof that they will not increase core counts past 16.

Hence why I said in my post "AMD doesn't seem to be raising core counts with Zen 4" and in my second post "If that's true". At no time did I say AMD has stated they will not be increasing core counts. We KNOW AMD will be increasing core counts on the EPYC line to 96 this year and 128 next year, which means they must be either increasing core count per CCX or using additional CCXs.

This picture emerged a couple of days ago on HotHardware and is supposed to be a sample of AMD's Genoa 96 core CPU. If that's genuine and is representative of the final product, it appears that core count per CCX will not be increasing and therefore core counts on Ryzen, and possibly even Threadripper, products will not be increasing either. It's not hard to imagine why this might be the case, it's the same reason CPUs moved to an MCM design from monolithic. With Zen 4 being manufactured on a bleeding edge node, getting 12 fully functional cores per CCX would no doubt prove a very daunting task which would generate a large number of failed dies.

But as it stands now it appears AMD will be focusing on IPC and higher clocks vs increasing core count in Ryzen models.

Resize
 

VforV

Respectable
BANNED
Oct 9, 2019
578
287
2,270
Not sure how you come to these conclusions. Those are great prices for AL vs Zen 3. More real-world evidence that Intel is absolutely dominating price/perf. They are clearly mass producing the heck out of these chips - you do realize they have a LOT more manufacturing capacity than TSMC right? They have a superior platform, a zillion different SKUs hitting the market addressing every single consumer performance bracket and budget and they have availability(something we all used to take for granted)! Stating the obvious here but - we wouldn't see such availability if Intel relied on TSMC's fabs to produce these chips.

Ultimately I like Rembrandt in the mobile space (that iGPU looks fantastic) but until we see reviews, AL mobile appears to be top notch in CPU performance. Near 12600k performance in a mobile system? Caveats aside - YES PLEASE!

With nothing new announced by AMD for AM4 other than the expensive and gaming-centric 5800X3D and with it becoming increasingly apparent that AM5 will be DDR5-only in a time of massive DDR5 price inflation, it is more obvious than ever that AL is the go to for new systems now and for at least another half year. HUB mentions this in the 12700 vs 5800X video. I wish I could argue otherwise - I was hoping to see AMD drop some prices or come out with something new to compete. Oh well, I guess. When you are supply-constrained by TSMC, making such price cuts is difficult.
Alder Lake is not selling good despite better prices. How did I came to these conclusions? Based on facts and watching the shops sales.

My observations were in the past 2 months on local EU stores, like I said.

My facts are based on reports from media outlets and 1+1 = 2.

Here you go:
Alder Lake boosts Intel's CPU market share, but AMD remains dominant

Intel Alder Lake CPUs Chip Away At AMD's Dominant Sales Lead At Major Retailer

Intel Core i7-12700K Leads Alder Lake Sales as AMD’s Ryzen 5000 Keeps 12th Gen Out of Top 5

You can google for more info, but reports 1 and 2 are from December and 3 is from January. They all say the same thing: Alder Lake sells better than the failure Rocket Lake, but worse than Zen3 which still tops the charts, despite being more expensive than AL.

The issues are many with AL and the reasons why the slow adoption and not so great sales. They have been stated and discussed a lot of times by now: much more expensive motherboards (only Z690, until the lower tiers come), DDR5 issues, Win11 issues, CPU coolers compatibility with LGA 1700 issues, some games not working with AL issues and the reluctance to buy a whole new system for AM4 users when a simple CPU upgrade is possible for them, etc.

That's why they don't sell well. That's why AMD does not need to lower prices yet (it does not mean I support higher prices).

I do admit (yes, I can admit intel positives, despite preferring AMD, unlike some intel fanbois that can't do the same for AMD) that AL sales will probably improve this year (by how much? who knows...) when at least the non-K SKUs come (especially the 12400/f) and the much cheaper lower tier motherboards. Then AL will be the best choice for someone wanting a new PC (unless AMD drops prices then), but until then even now at the start of January, months later after it's launch AL is not selling better than Zen3, contrary to the media narrative and push for constant AL promo and praise.

These are facts, not sentiments.
 
Last edited:
These are facts, not sentiments.
Intel doesn't rely on selling one CPU at a time to end users...
OEM sales are in the ballpark of 70-80 million PCs in a quarter with the vast majority of those being intel.
If AL sold well until now or not can only be cleared up by intel's next quarterly report, until we see these official numbers everything anybody says is going to be 100% sentiment and 0% facts.
https://www.canalys.com/newsroom/Canalys-global-PC-market-Q2-2021

Not selling well to the DIY crowd and not selling well in general are completely different things.
 
  • Like
Reactions: Why_Me

teodoreh

Distinguished
Sep 23, 2007
315
13
18,785
Based on the specs AMD announced for the 6980HX, and comparing them to the 5900HX, the 6980HX isn't going to come close to matching the 12900HK in multi-threaded performance. What's going to be the excuse then for AMD's latest and greatest not being able to beat a lowly 4 year old 1950X?

This wasn't an Intel leak. The benchmarks were done by a 3rd party using a test platform they got their hands on.

As I wrote before, whoever benchmarked this thing (which I cannot accept or reject that has Intel connections), took one (1) specific piece of software (Cinebench) and wow suddenly 12900HK beats desktop processors.
Don't twist my sayings and mostly, don't rely on assumptions over assumptions. If 12900HK can be for 5' a better processor than a 4 year desktop monster, let it be. If it is the best mobile processor, let it be. But there are many things that need to be specified:

  1. Benchmark the CPU on a whole range of software not just on Cinebench.
  2. Ensure that there will be no overclocking or if it is, ALL processors involved will also get benchmarked on OC mode.
  3. Record possible throttling issues. Laptop with 110W processor working over 15'? It will either have huge fan and noise, reduction in speed or very high temps that will reduce its lifespan.
  4. It's all about the cost. If this processor can deliver what it promises and has the same or slightly bigger price than the competition, kudoz intel. If it costs twice the price of 6980HX, who cares?

It's a 10nm processor and soon it will be competing 6nm processors with 45W TDP. Who wants a 110W TDP laptop, and why?
 
  • Like
Reactions: Makaveli and VforV

larkspur

Distinguished
Alder Lake is not selling good despite better prices. How did I came to these conclusions? Based on facts and watching the shops sales.
These are facts, not sentiments.
Bwahaha, really? You're talking about two articles that use November data from etailers and one that looks at "most popular" at etailers. OEMs, hello? You do realize that pre-builts are the most popular systems out there right? We DIYers are a tiny minority of CPU sales. Not to mention the vast majority of people don't even buy a k series chip or a z-series board. I'm actually surprised to see how popular k-series and (by default) z-series has sold so well on Amazon and Newegg although such sources are generally not indicative of the market as a whole.

But there are many things that need to be specified:

  1. Benchmark the CPU on a whole range of software not just on Cinebench.
  2. Ensure that there will be no overclocking or if it is, ALL processors involved will also get benchmarked on OC mode.
  3. Record possible throttling issues. Laptop with 110W processor working over 15'? It will either have huge fan and noise, reduction in speed or very high temps that will reduce its lifespan.
  4. It's all about the cost. If this processor can deliver what it promises and has the same or slightly bigger price than the competition, kudoz intel. If it costs twice the price of 6980HX, who cares?
It's a 10nm processor and soon it will be competing 6nm processors with 45W TDP. Who wants a 110W TDP laptop, and why?
You do realize that you can click on the link and look at the benches they performed yourself right? Hint: They did more benches than Cinebench. I can't read the language but I can see the numbers on the charts. This is an independent site - it isn't Intel's internal benchmarks. It is, however an Intel reference laptop provided by Intel for the purpose of 3rd party reviews. And if you took the time to look at the review (they actually call it a "preview"), you would see they compare this laptop CPU with other laptop CPUs. The TR is just mentioned as a "Wow!".
 

VforV

Respectable
BANNED
Oct 9, 2019
578
287
2,270
Bwahaha, really? Cop-out and denial text
Careful not to hurt yourself there.

I'm sure that you and your kind saying that AL is selling well will turn into magic sale numbers for intel and the actual worse than Zen3 sale numbers reported already are just fictitious.

Welcome to my IGNORE list. You're now part of a select club of [***] and [**].

I don't care to waste my time with people that can't understand what "facts" mean or ignore them when presented to them. Equally so, I don't care to waste my time with those who always twist everything up to make up excuses and distort reality to their own liking. Bye.