Intel Demoes i9-11900K Against Ryzen 9 5900X

[Random_Person1234]

Summary

Intel has demoed the 8c/16t i9-11900K against the 12c/24t Ryzen 9 5900X during CES. Paired with the RTX 3080, the 11900K appears to beat the 5900X by 2-8% in 7 games at 1080p high/ultra. 

Spoiler

 

Quotes

Quote

Intel has demoed its new Rocket Lake-S Core i9-11900K against AMD Ryzen 9 5900X 12-core CPU. According to the screenshots published by Geeknetic, Intel CPU appears to be faster in both gaming scenarios. The system has been paired with NVIDIA GeForce RTX 3080 graphics card, however, it is not stated if the system had Resizable BAR enabled.

 

My thoughts

I wonder how thermals look like given it's still on 14nm. If priced right, this could be a good CPU compared to the 5800X or 5900X in terms of gaming.

 

Sources

https://videocardz.com/newz/intel-demoes-core-i9-11900k-against-ryzem-9-5900x

https://www.geeknetic.es/Noticia/20888/Intel-anuncia-los-procesadores-de-sobremesa-Rocket-Lake-S-de-11a-generacion-con-un-14-mas-de-IPC-y-PCIe-40.html

 

[Jurrunio]

Sort of worrying for Intel. All companies cheery pick benchmarks to varying degree and even then, Intel only shows single digit gains for games

 

Also hope supply wont be as bad as AMD's. In PCPP the Zen 3 CPUs dont even have a price tag on half the time because stock is gone so quickly.

[WhitetailAni]

2 minutes ago, Random_Person1234 said:

Summary

Intel has demoed the 8c/16t i9-11900K against the 12c/24t Ryzen 9 5900X during CES. Paired with the RTX 3080, the 11900K appears to beat the 5900X by 2-8% in 7 games at 1080p high/ultra. 

  Reveal hidden contents

 

I wonder how thermals look like given it's still on 14nm. If priced right, this could be a good CPU compared to the 5800X or 5900X in terms of gaming.

Interesting.

I wonder if it will hold true.

[porina]

1 minute ago, Jurrunio said:

Sort of worrying for Intel. All companies cheery pick benchmarks to varying degree and even then, Intel only shows single digit gains for games

It'll be interesting what the gain is compared to Intel's previous gen product. This is a comparison with AMD, who claimed the gaming performance throne when they released Zen 3. If this is seen in wider testing by the usual tech sites when the CPU is released then Intel would recapture the gaming throne. I also note they're using the 5900X, I think it would be more interesting to go against 5800X here, at least, for most games.

[Jurrunio]

1 minute ago, porina said:

I think it would be more interesting to go against 5800X here, at least, for most games.

yup, not only for matching core count but 5800X also doesnt have multiple CCDs to worry about.

[Brooksie359]

4 minutes ago, porina said:

It'll be interesting what the gain is compared to Intel's previous gen product. This is a comparison with AMD, who claimed the gaming performance throne when they released Zen 3. If this is seen in wider testing by the usual tech sites when the CPU is released then Intel would recapture the gaming throne. I also note they're using the 5900X, I think it would be more interesting to go against 5800X here, at least, for most games.

Part of me thinks the only reason they would compare it to the 5900x is if they are planning on selling it for a similar price at which point I would take the 4 extra cores over the slightly improved gaming performance at 1080p. 

[Helpful Tech Witch]

Ok, but there’s a good chance these are very cherry picked. (As always)
Be wary of first party benchmarks.

[cm992]

Crap now I need a Z590 mobo and an i9-11900K to maintain my epic gamer clout

[porina]

9 minutes ago, Jurrunio said:

yup, not only for matching core count but 5800X also doesnt have multiple CCDs to worry about.

I'm digging a hole here...

 

I just looked at TechPowerUp's testing on 5800X. 

https://www.techpowerup.com/review/amd-ryzen-7-5800x/22.html

They claim the 10900k was about 3% faster than the 5800X at 1080p gaming.

 

Repeat at TechSpot's average of 11 games at 1080p Ultra:

https://www.techspot.com/review/2134-amd-ryzen-5800x/

About 1% difference here, in favor of the 10900k.

 

Did anyone fact check AMD's "fastest gaming CPU" claims? I have to admit, I didn't, since I saw the pricing and checked out. Or did AMD "win" using higher core count CPUs? 5900X and 5950X ever so slightly edges out ahead, but they're all within spitting distance of each other. Probably still more GPU limited than CPU.

 

5 minutes ago, Brooksie359 said:

Part of me thinks the only reason they would compare it to the 5900x is if they are planning on selling it for a similar price at which point I would take the 4 extra cores over the slightly improved gaming performance at 1080p. 

A more negative reason that springs to mind is that 5900X might actually perform worse than 5800X in these titles. I've not done the digging to see if that might be the case, but the 5900X would have two obvious disadvantages: two CCDs, and possible lower running clock if loaded. This could go either way. A highly threaded load should be more efficient when split to more cores (at same total power budget). And we've seen good scaling when fewer threads are active. Also I think AMD bins the better dies for the higher core CPUs anyway.

[Tristerin]

14 minutes ago, Jurrunio said:

Sort of worrying for Intel. All companies cheery pick benchmarks to varying degree and even then, Intel only shows single digit gains for games

 

Also hope supply wont be as bad as AMD's. In PCPP the Zen 3 CPUs dont even have a price tag on half the time because stock is gone so quickly.

I finally broke and just got a R5 3600.  Should be here tomorrow.

[CommanderAlex]

21 minutes ago, Random_Person1234 said:

Intel has demoed the 8c/16t i9-11900K against the 12c/24t Ryzen 9 5900X during CES. Paired with the RTX 3080, the 11900K appears to beat the 5900X by 2-8% in 7 games at 1080p high/ultra. 

Shouldn't this demonstrate, of course depending on how well each game's engine utilizes multi-core CPUs, how much "stronger" (in terms of IPC) Intel's cores are compared to AMD Ryzen's?? If they did CPU benchmarks that shows Intel's i9-11900K scored more than a 5950X, then that's impressive. 

 

Note: I'm just reading about the news coming out of CES so please excuse me if I left out something.

[SeriousDad69]

If it's 1-8% at 1080p it's probably nothing at 1440p and 4k. Knowing Intel, they'll charge $550 for it and claim it's the best thing since sliced bread. I guess it'll at least make a good 800FPS CSGO CPU.

[Jurrunio]

48 minutes ago, porina said:

I'm digging a hole here...

 

I just looked at TechPowerUp's testing on 5800X. 

https://www.techpowerup.com/review/amd-ryzen-7-5800x/22.html

They claim the 10900k was about 3% faster than the 5800X at 1080p gaming.

 

Repeat at TechSpot's average of 11 games at 1080p Ultra:

https://www.techspot.com/review/2134-amd-ryzen-5800x/

About 1% difference here, in favor of the 10900k.

 

Did anyone fact check AMD's "fastest gaming CPU" claims? I have to admit, I didn't, since I saw the pricing and checked out. Or did AMD "win" using higher core count CPUs? 5900X and 5950X ever so slightly edges out ahead, but they're all within spitting distance of each other. Probably still more GPU limited than CPU.

I reckon the claim comes from a few traditionally Intel-biased games like CSGO switching to in favour of AMD. Also Ryzen CPUs performed better in frame rate stability in some games (though that's not what AMD claimed, but what reviewers found out).

[Taf the Ghost]

7 games, and we know Intel & benchmarks, so some quick thoughts.

 

Total War (older version), Metro, Gears 5 and Far Cry still favor Intel SoCs even with Zen3 out. Mostly because they favor higher clocks, especially Far Cry & Total War.

 

Cyberpunk completely matters about where you test it, so they can easily find a spot that would favor them over AMD.

 

Watch Dog Legion & Assassin's Creed Valhalla both favor AMD GPUs, so a 3080 should be inducing an early bottleneck allowing for cherrypicking benchmark locations.

[Vishera]

Wait for Alder Lake

[RejZoR]

Oh no! Anyway...

 

Basically all Ryzen 5000 users XD

[Brooksie359]

1 hour ago, porina said:

I'm digging a hole here...

 

I just looked at TechPowerUp's testing on 5800X. 

https://www.techpowerup.com/review/amd-ryzen-7-5800x/22.html

They claim the 10900k was about 3% faster than the 5800X at 1080p gaming.

 

Repeat at TechSpot's average of 11 games at 1080p Ultra:

https://www.techspot.com/review/2134-amd-ryzen-5800x/

About 1% difference here, in favor of the 10900k.

 

Did anyone fact check AMD's "fastest gaming CPU" claims? I have to admit, I didn't, since I saw the pricing and checked out. Or did AMD "win" using higher core count CPUs? 5900X and 5950X ever so slightly edges out ahead, but they're all within spitting distance of each other. Probably still more GPU limited than CPU.

 

A more negative reason that springs to mind is that 5900X might actually perform worse than 5800X in these titles. I've not done the digging to see if that might be the case, but the 5900X would have two obvious disadvantages: two CCDs, and possible lower running clock if loaded. This could go either way. A highly threaded load should be more efficient when split to more cores (at same total power budget). And we've seen good scaling when fewer threads are active. Also I think AMD bins the better dies for the higher core CPUs anyway.

When I looked at reviews when the cpus first came out I believe the 5900x basically did better or the same as the 5800x for the most part in games. I mean the only games that there were significant differences between the 5800x and the 5900x were ones that favored the 5900x. Anyways o still think if these numbers Intel have released are any indication of what to expect of the 11 gen cpus then I would prefer the 5900x over the 11900k. 

[Taf the Ghost]

 

 

125w vs 105w, lol.

[SteveGrabowski0]

So dropping two cores and presumably ramping clockspeed like hell to hit the 125W power envelope? No thanks I'd take the 5900X in a second in that price segment.

[Kisai]

6 hours ago, HelpfulTechWizard said:

Ok, but there’s a good chance these are very cherry picked. (As always)
Be wary of first party benchmarks.

 

If you need 20% more power, with 4 less cores, it's not even a fair comparison. It would be fairer to compare the i7 part.

[Zmimgo]

Actually suprised they didn't have RDR2 in there, since I remember the i9-10900K beating the R9 5900X in it

[Taf the Ghost]

33 minutes ago, Zongohihello said:

Actually suprised they didn't have RDR2 in there, since I remember the i9-10900K beating the R9 5900X in it

The backported design might not be quite as good in RDR2. Some games have some really finely tuned optimizations that don't necessarily carry over. Also, RDR2 performance is a little wonky in general, so they could have easily had issues getting it to perform better.

[porina]

7 hours ago, Taf the Ghost said:

125w vs 105w, lol.

At least for Zen 2, and I'm not aware of it changing for Zen 3, they run at a power limit PPT higher than TDP. 105W TDP CPUs limit at 142W PPT from memory. On Intel side, it is unclear if they were allowing boost, which is a max of 250W with tau of 56 seconds from memory.

 

5 hours ago, SteveGrabowski0 said:

So dropping two cores and presumably ramping clockspeed like hell to hit the 125W power envelope? No thanks I'd take the 5900X in a second in that price segment.

Dropping cores is likely due to the increased transistor count per core. This is a 10nm design being built on 14nm, so it is going to be big. Clocks are essentially same as 10th gen so unchanged.

[williamcll]

I'll wait for the reviews.

[TetraSky]

Temperature difference?

Power Consumption ?

 

Simply showing a graph with numbers on it means very little when you consider everything else.

2 to 8% better... sure, but if you need a cryo chiller to cool it down and consume 400W, it may as well be dead on arrival (just an example).

Then, according to these, they are comparing their 8 cores/16 thread CPU to the AMD 12 cores/24 threads.

You lose 4 cores and you gain as little as 2% at 1080p ?  Not sure it's worth it when you consider EVERYTHING ELSE we can do with a PC. Then again, single core clock has always been Intel's "forte" in the last couple of years, which is typically what games still demand instead of properly using multicores.

 

 

Wouldn't be the first time Intel lied on their graphs, too.

Better wait for independent reviews...