Zen5 only 16 core.

[ty_ger]

Wow! 7 FPS?!

 

[InVasMani]

I hope X3D gets buried

From what I heard the top Alchemy GPU sku might be PCIE 5.0 with a massive L cache.

 

[P4-630]

Im almost triggered to test it on the 8700K but we havent got same gpu.. curious now

I'm sure you'll see a boost in fps with various of your games with a new gen CPU/platform upgrade.

 

[Vayra86]

I'm sure you'll see a boost in fps with various of your games with a new gen CPU/platform upgrade.

Oh yeah no doubt. Im just saving that a-ha erlebnis until this 8700K is worn out. 7800X3D or its successor is the idea... or maybe Intel will release a sane CPU by then, who knows.

 

[dirtyferret]

Cause the 13600k has 14 cores, not 6. Compare the 7700x to the 7600, or the 10700k to the 10600k, there are differnces of around 16-18% in lows.

yet the ryzen 7600 performs better than the 5900x despite half the cores/threads, clearly a performance based game when it comes to the CPU that doesn't seem too reliant on increased AMD cache in X3D chips

 

[P4-630]

or maybe Intel will release a sane CPU by then, who knows.

I'm waiting for Arrow Lake, and see if it's worth it. A somewhat weaker NPU, I don't think I care about that if it doesn't help boost fps with gaming.
https://www.techpowerup.com/forums/threads/possible-near-future-2024-arrow-lake-build.321793/

I won't be upgrading my GPU before GTA 6 PC is out anyway, the games I play run fine @1440p with my 2070 Super.

 

[JustBenching]

yet the ryzen 7600 performs better than the 5900x despite half the cores/threads, clearly a performance based game when it comes to the CPU that doesn't seem too reliant on increased AMD cache in X3D chips

And how does that dispel the fact that the game likes extra cores? You can't compare different architectures - obviously - it's called ceteris paribus.

 

[Onasi]

Cause the 13600k has 14 cores, not 6. Compare the 7700x to the 7600, or the 10700k to the 10600k, there are differnces of around 16-18% in lows.

Okay. Comparing. The 7600 runs a significant frequency and power budget deficit compared to the 7700X. The story is similar with 8700K/9900K.
And the differences are all, while noticeable in graphs, hardly deal breaking and a far cry from your “Just looking at 0.2% lows, the full chip gets 3 times the lows.” statement. At no point in this test we see different CPUs of the same generation displaying anything even close to such a delta.

 

[dirtyferret]

You can't compare different architectures - obviously - it's called ceteris paribus.

this is why people fail to understand the basics of "performance", they simply don't get it and never will.

You can't compare different architectures on a CPU test that literally lists and compares different CPU architecture

And how does that dispel the fact that the game likes extra cores?

the article I linked dispels the fact that judging CPUs solely based on cores is one of ignorance. Your stance simply cements that "fact"

 

[wheresmycar]

In all frankness, being a regular on PCPP for years i haven't once come across anyone bonkers and bent on buying higher core count processor purely for gaming. The ones who do stretch beyond 8c/16t are usually gaming and streaming simultaneously or gamers who are also involved with content creation (or any other type of multi-threading performance workload). Although i admit there are rare OVERKILL instances with "BIG WALLET little brains" who just simply want the best of the best hardware regardless of cost or whether the selected hardware carries merit for the intended workload. I've seen a few of those and usually if you challenge their intended purchase they're pretty upfront in confirming their negligence which is fine in my book. There are some rare instances with some CPU bound titles which do benefit with more than 8 cores for that added layer of desired smoothness - for example a handful of 1000+ unit CPU-hog RTS titles or possibly some other odd-ball titles with similar CPU resource heavy and core scalability demands. For the vast majority, based on current and previous architectures we don't need more than 6 cores for gaming with 8 already being a luxury.

All of that was leading to a question... is there anyone here who is actually suggesting we need more than 8? Or is this a simple case of suggesting some games can scale into additional cores/threads regardless of real-time benefits?

 

[JustBenching]

this is why people fail to understand the basics of "performance", they simply don't get it and never will.

Agreed, they just can't get it. If number of cores doesn't matter then why is the 7700x faster than the 7600?

Okay. Comparing. The 7600 runs a significant frequency and power budget deficit compared to the 7700X. The story is similar with 8700K/9900K.
And the differences are all, while noticeable in graphs, hardly deal breaking and a far cry from your “Just looking at 0.2% lows, the full chip gets 3 times the lows.” statement. At no point in this test we see different CPUs of the same generation displaying anything even close to such a delta.

That's because all of his CPUs are chocked with slower memory. Still a 30% delta is huge, but good thing is he is sharing the location he run the bench. So you can test it out for yourself - it's right at the start of the game

 

[Onasi]

Agreed, they just can't get it. If number of cores doesn't matter then why is the 7700x faster than the 7600?

Indeed, why is a CPU that easily boosts above 5.5 Ghz and has a power budget of almost 150 watts faster than the one on the same architecture that goes only up to 5.1 and has almost half as much power allowance? Itz a miztary. We will never know.

That's because all of his CPUs are chocked with slower memory. Still a 30% delta is huge, but good thing is he is sharing the location he run the bench. So you can test it out for yourself - it's right at the start of the game

First, that’s his testing methodology, nothing wrong with that. Secondly, we’re throwing memory into the mix now? Should we also consider manually tuned in timings? And 30% is still not “3 times”, especially when we talk about 0.1% lows.

 

[JustBenching]

Indeed, why is a CPU that easily boosts above 5.5 Ghz and has a power budget of almost 150 watts faster than the one on the same architecture that goes only up to 5.1 and has almost half as much power allowance? Itz a miztary. We will never know.


First, that’s his testing methodology, nothing wrong with that. Secondly, we’re throwing memory into the mix now? Should we also consider manually tuned in timings? And 30% is still not “3 times”, especially when we talk about 0.1% lows.

You know you can just...test it right? You have a 12 core cpu, disasble half and run the game

 

[dirtyferret]

then why is the 7700x faster than the 7600

because padawan the 7700x is a better gaming CPU than the 7600 or 5950x or 10900k based on performance

If number of cores doesn't matter

never said it does not matter, said judging CPUs solely on core count as opposed to performance is ignorant or as the article rightfully stated

What We Learned​

So there you have it, core count tells you very little about a CPU's gaming performance, and while we're sure this data won't come as a surprise to many of you, this does seem like a topic that we regularly need to revisit...To say "6 cores is all you need for gaming" or inversely "6 cores aren't enough for gaming" grossly oversimplifies the situation, and really, both are wrong, because, you guessed it, you don't want to be talking about core count when it comes to gaming performance.

 

[Onasi]

You know you can just...test it right? You have a 12 core cpu, disasble half and run the game

That would require me to buy the game, which is a pointless expenditure since I have no interest in it whatsoever and definitely not going to pay Sony my hard earned shekels to prove something to a sentient wall.
Secondly, yeah, if I turn off a whole CCD on my 5900X and lose 32 megs of cache I WOULD turn it into a parody of a 5600X and it probably will perform in that ballpark.
Thirdly, I have a goddamn 1070 in my personal rig, I will be GPU bound in that abomination of a console port even if I replace the CPU with a literal potato. So I am not sure that will lead to any conclusive evidence.

 

[Assimilator]

When people say more cores don't make a difference they don't usually mean " because you are gpu bound"

Of course they do. The only person in this thread who has interpreted it any other way is you.

 

[JustBenching]

never said it does not matter, said judging CPUs solely on core count as opposed to performance is ignorant or as the article rightfully stated

And i don't disagree with that. In order to see if cores matter you compare across cpus with the same architecture, that's captain obvious stuff, no? Otherwise you might as well say core count doesn't matter for Cinebench cause the 7700x is faster than the 3900x. Duh

 

[Onasi]

And i don't disagree with that. In order to see if cores matter you compare across cpus with the same architecture, that's captain obvious stuff, no? Otherwise you might as well say core count doesn't matter for Cinebench cause the 7700x is faster than the 3900x. Duh

Ah yes, Cinebench. My favorite video game. Truly, a real benchmark of gaming performance for all other titles on the market.

I am still struggling to see what your actual point is since you seem to change it every other post.

Let us get down to brass tacks. Can additional cores be helpful in games? Yes, in some instances. Is there a point of diminishing returns? Definitely. Are core counts a definite indicator of CPUs gaming performance? Obviously not. Finally, to answer the driving question of this thread - is there a point, for games specifically, to go above the entrenched configuration of 8/16 provided that these 8 cores are of contemporary architecture? No, there is no evidence to suggest so. Any performance advantages displayed by higher core count parts of the same architecture are explainable by other characteristics of said parts. Even more so when widely accepted fastest gaming CPU on the market today is a 8/16 SKU.

 

[dirtyferret]

Otherwise you might as well say core count doesn't matter for Cinebench cause the 7700x is faster than the 3900x. Duh

I'm not going to reply to the former of your last post because it would simply be a regurgitation of what @Onasi has already stated. As for your latter statement quoted above, simply replace "Cinebench" with gaming and you finally get it because your argument has been a contradiction to it and Ive yet to see one person agree with you.

 

[rv8000]

1 CCD limit for gaming.

Other CCD useful for pinning every other process onto though.

What would be more impactful is more than eight cores on a single CCD, faster IF/newer IO die (2000 MHz IF is no faster on Zen 4 AM5 than Zen 3 AM4, hence 1.5/1 not 1/1 ratio, or less latency penalty from cross CCD workloads. Doesn't matter if you have 64 cores from eight CCDs, still won't have better gaming performance than six or eight on a single CCD.

If it hasn’t be mentioned already, the current rumor is the IF clock on Zen 5 will see a bump from 2000 > 2400.

 

[freeagent]

If they made a tuneable CPU with phat cache I would be ok with that. I just dislike how locked down they are currently. I am not denying their performance, it’s there for all to see..

I know they will grant my wish at some point.

 

[JustBenching]

Ah yes, Cinebench. My favorite video game. Truly, a real benchmark of gaming performance for all other titles on the market.

I am still struggling to see what your actual point is since you seem to change it every other post.

Let us get down to brass tacks. Can additional cores be helpful in games? Yes, in some instances. Is there a point of diminishing returns? Definitely. Are core counts a definite indicator of CPUs gaming performance? Obviously not. Finally, to answer the driving question of this thread - is there a point, for games specifically, to go above the entrenched configuration of 8/16 provided that these 8 cores are of contemporary architecture? No, there is no evidence to suggest so. Any performance advantages displayed by higher core count parts of the same architecture are explainable by other characteristics of said parts. Even more so when widely accepted fastest gaming CPU on the market today is a 8/16 SKU.

My actual point is pretty simple, get a 6c/12t CPU, add 2 or 4 extra cores (not on a different CCD) and it will be faster in lots of games cause a lot of games scale beyond 6cores. I can't make it simpler than that.

I've tested the exact same area as PCGH did in your link, 6/12 vs 8/16 + 8. The difference in averages is around 25% (143 vs 114 fps) but for example, at the last heavy scene with the explosions the 6core dropped to 92 fps with 0.2% of 49 while the full chip dropped to 138 with 0.2% of 77. That's 50% higher average and 57% higher lows. Of course the game is very playable on both but the point is that more cores more better even for games.

Spoiler

 

[oxrufiioxo]

My actual point is pretty simple, get a 6c/12t CPU, add 2 or 4 extra cores (not on a different CCD) and it will be faster in lots of games cause a lot of games scale beyond 6cores. I can't make it simpler than that.

I've tested the exact same area as PCGH did in your link, 6/12 vs 8/16 + 8. The difference in averages is around 25% (143 vs 114 fps) but for example, at the last heavy scene with the explosions the 6core dropped to 92 fps with 0.2% of 49 while the full chip dropped to 138 with 0.2% of 77. That's 50% higher average and 57% higher lows. Of course the game is very playable on both but the point is that more cores more better even for games.

Spoiler

While I agree with you, it doesn't make the 7900X3D a bad cpu and in a lot of scenarios it performs just fine and in others better than expected.

I honestly feel about the same with E cores though and even though they give less issues typically they are not completely trouble free either. Regardless it looks like hybrid architectures are here to stay so my hope is that both intel/amd improve them at a rapid pace.

I too hope that the IF is massively improved if not on this socket at the very least on AM6.

 

[JustBenching]

While I agree with you, it doesn't make the 7900X3D a bad cpu and in a lot of scenarios it performs just fine and in others better than expected.

I honestly feel about the same with E cores though and even though they give less issues typically they are not completely trouble free either. Regardless it looks like hybrid architectures are here to stay so my hope is that both intel/amd improve them at a rapid pace.

I too hope that the IF is massively improved if not on this socket at the very least on AM6.

Wasn't talking about the 7900x 3d, but any 6core in general. 8cores is where diminishing returns start kicking in hard right now, but 6cores ain't *enough for heavy areas of games (tom's dinner, the above area on TLOU etc.).

*By enough, I mean maintaining very high average and low fps. The game is obviously still very playable.

 

[dirtyferret]

My actual point is pretty simple, get a 6c/12t CPU, add 2 or 4 extra cores (not on a different CCD) and it will be faster in lots of games cause a lot of games scale beyond 6cores. I can't make it simpler than that.

techspot actually ran this test

What We Learned​

Cache matters. Often more so than cores, when it comes to PC gaming performance. So, there you have it. As we found with Intel's 10th-gen series two years ago, and in fact, probably more than we realized two years ago. The arrival of AMD's 3D V-Cache processors has proven this beyond a shadow of a doubt, causing significant challenges for Intel's gaming performance, something the company deeply cares about.

This data also supports recommendations we made years ago. For example, we appreciated the value delivered by the Ryzen 5 5600 series and recommended it despite it having only 6 cores, which many at the time believed would be insufficient for future gaming.


Today, the 6 and 8-core Zen 3 parts are still delivering comparable performance, and of course, parts like the Ryzen 5 5600X are still very usable.

So, until games fully saturate the 5600X, you won't see an improvement with the 5800X, and by that time, we expect both CPUs will be struggling. Of course, there are instances where 8-core models of the same architecture are faster than their 6-core counterparts, but in those examples, the 6-core processors still deliver highly playable performance, making the core count argument moot, especially considering the cost difference.


As a more modern example, the Ryzen 5 7600 costs $210 and delivers comparable gaming performance to the 5800X3D. The Ryzen 7 7700 costs $310 – almost 50% more – and it would be challenging to find a game where the 7700 is even 20% faster than the 7600; in fact, such a scenario might not even exist. Looking over our most recent data from the 5700X3D review, the 7700X is just 4% faster than the 7600X on average, so in terms of value, the 6-core model is significantly better for gaming.