News AMD CES 2022 Event to Highlight Ryzen 3D V-Cache, Radeon Ups

a Ryzen 9 5900X with 3D V-Cache can deliver up to a 15 percent uplift in gaming performance (on average) at 1080p resolution. That could give AMD the ammunition it needs to battle the Core i9-12900K and Core i7-12700K.

"could give"?
its a last gen cpu vs their flagship new gen.
it'll liekly match or beat it. That alone is a feat.

Zen 4 will likely wreck 12th gen (which is 12th gens real rival) given how each generation of zen improves.
Am I the only one that find it ridiculous we're up to 192mb of cache on AMD to compete with 20mb of cache on 12th gen Intel ?
better have the cache and not need it than have the need for it but not have it.

and it alone can give up to 15% gaming performance proves its useful.
 
better have the cache and not need it than have the need for it but not have it.

and it alone can give up to 15% gaming performance proves its useful.
It's going to have the same problem as 12th gen, there are no GPUs that will show the difference...Zen3d will catch up on games that they are behind but it's going to be really hard to find any game where it will be any faster, because of the GPUs.

And it's still going to be more expensive than intel, even if AMD keeps the prices at the same level as the versions without the cache it's still going to be too expensive.

Also it's 15% average of a grand total of 5 games, that's all AMD has shown, ...with the worst case only giving 4% more performance, what if the worst case is going to be the case most of the times?

Also it's 15% at locked 4Ghz which means that it could be that the final difference compared to normal 5xxx is going to be smaller, depending on how much power the cache takes away from the cores and how it will influence the heat dissipation.

Also you have to keep in mind that if this cache actually does work and gives good improvements all around then intel can do the same thing very easily and very fast, intel already did something similar with broadwell and now they do have foveros so the only thing stopping them is that they don't yet need to do it.
 
  • Like
Reactions: wifiburger
3D V-Cache will likely only widen the price delta between AMD and Intel’s enthusiast-class processors.

Intel is eating AMD's lunch by undercutting AMD on Performance : Price ratio with Alder Lake. It would take some serious guts for AMD to raise prices even more. Only reason I didn't upgrade is because AMD got greedy.
 
Also you have to keep in mind that if this cache actually does work and gives good improvements all around then intel can do the same thing very easily and very fast, intel already did something similar with broadwell and now they do have foveros so the only thing stopping them is that they don't yet need to do it.

Always a ray of sunshine for intel aren't you?

Intel is already stacking cores on top of cores. Adding v-cache might be a serious issue heat wise. So maybe not so easy for intel.

Also it's 15% average of a grand total of 5 games, that's all AMD has shown, ...with the worst case only giving 4% more performance, what if the worst case is going to be the case most of the times?

This statement is known as bias planting. Your assuming worse case and stating it. We don't know a thing until it's in the hands of reviewers. And Intel has been mighty guilty in the past of saying "Oh we got a 20% boost" and then it's only 2%-3%. So you have to wait for reviewers to have their say.

Also it's 15% at locked 4Ghz which means that it could be that the final difference compared to normal 5xxx is going to be smaller, depending on how much power the cache takes away from the cores and how it will influence the heat dissipation.
It's likely an engineering sample. So not apples to apples.

That said, I do believe AMD will come close to matching Intels performance. (Purely hypothetical guess) HOWEVER they won't be able to offer the same value.
 

salgado18

Distinguished
Feb 12, 2007
928
373
19,370
It's going to have the same problem as 12th gen, there are no GPUs that will show the difference...Zen3d will catch up on games that they are behind but it's going to be really hard to find any game where it will be any faster, because of the GPUs.
But the benchmarks already show the difference, so there are GPUs that can benefit from it. Also, some games can't reach 144 fps (to match a 144Hz monitor) because of the CPU.

And it's still going to be more expensive than intel, even if AMD keeps the prices at the same level as the versions without the cache it's still going to be too expensive.
That's tricky, because Alder Lake is probably already more expensive than Zen 3, but Intel can just shrink their profit margins to hold the overall market share. AMD can't do that easily, they need the profits to keep going.

Also, at the very least they will lower current gen prices to match Intel, like they did in the past (the FX-8120 launched a bit pricier than the i7-2600k, but quickly decreased to price parity and even a bit lower). Zen 3D will probably cost a bit more, but not much, otherwise they lose sales. Intel hit AMD very hard with these prices, and Ryzen prices will come down, but the question is how much can they do it and still keep making money.
 
Also you have to keep in mind that if this cache actually does work and gives good improvements all around then intel can do the same thing very easily and very fast, intel already did something similar with broadwell and now they do have foveros so the only thing stopping them is that they don't yet need to do it.
That's if the system and CPU architecture actually benefits from having more cache in the first place. If you have a decent enough memory subsystem, throwing on more cache brings diminishing returns. e.g., Athlon 64s got away with less cache than the Pentium 4 because of having an on-die memory controller.

Arguably having more cache on Zen 3 bringing performance improvements may be indicative that DDR4 wasn't going to be up to snuff to feed the higher end SKUs.
 
Always a ray of sunshine for intel aren't you?

Intel is already stacking cores on top of cores. Adding v-cache might be a serious issue heat wise. So maybe not so easy for intel.
They are stacking cores for desktop alder lake? Do we have a link for that?
Other than that, exactly my point, if the stacked cache is a heat issue it might be a shot in the foot for amd.
Notice the if ,I'm not saying it will for sure.
This statement is known as bias planting. Your assuming worse case and stating it. We don't know a thing until it's in the hands of reviewers. And Intel has been mighty guilty in the past of saying "Oh we got a 20% boost" and then it's only 2%-3%. So you have to wait for reviewers to have their say.
So you have previous experience with a company overstating the performance increase so that must mean that this company now is stating facts?!
What's your reasoning here?
Yes intel's 20% increase turned out to be 2-3% so why shouldn't the same be possible with AMD?
That's why we use things like "what if" as I did, I didn't state fact or even opinion, I just made a point about nothing being sure.
 

spongiemaster

Admirable
Dec 12, 2019
2,273
1,277
7,560
Where did I say that?

Well in any case, we also didn't have the technology to do such either.
Your post was one sentence long. Kind of narrows down where you said that.

3D stacking is not necessary to add more cache to a CPU. Server chips have had more cache than mainstream CPU's forever without stacking it above the die. The fact that AMD and Intel haven't added more cache without stacking which would be cheaper than stacking just drives home the lack of cost advantage point even further.
 
the profit from 5600-5950 is irrelevant.

The bulk of their profit CPU wise is on HEDT/server grade stuff i.e. threadripper and epyc
"Normal" hardware makes them more money although it includes gpus, while HEDT/server is bunched up with console cpu sales.
It's hard to tell, impossible even, if hedt/server or normal cpus make them more money since amd doesn't release separate numbers

https://ir.amd.com/news-events/pres...-reports-third-quarter-2021-financial-results
Computing and Graphics segment revenue was $2.4 billion, up 44 percent year-over-year and 7 percent quarter-over-quarter. The year-over-year and quarter-over-quarter increases were driven by higher Ryzen, Radeon and AMD Instinct processor sales.
Enterprise, Embedded and Semi-Custom segment revenue was $1.9 billion, up 69 percent year-over-year and 20 percent quarter-over-quarter. The increases were driven by higher EPYC processor and semi-custom product sales.
 
They are stacking cores for desktop alder lake? Do we have a link for that?
Other than that, exactly my point, if the stacked cache is a heat issue it might be a shot in the foot for amd.
Notice the if ,I'm not saying it will for sure.

So you have previous experience with a company overstating the performance increase so that must mean that this company now is stating facts?!
What's your reasoning here?
Yes intel's 20% increase turned out to be 2-3% so why shouldn't the same be possible with AMD?
That's why we use things like "what if" as I did, I didn't state fact or even opinion, I just made a point about nothing being sure.

Sorry I misspoke. I meant to say Raptor Lake which is rumored to be stacking the cpu core dies. So putting cache on top of that will be difficult.

And yes, AMD might be exaggerating. Raja for example always had over blown the products he released under AMD. Both companies are guilty of this. Remember Intel saying "real world applications don't matter" Just like I said before, wait for reviews before passing judgement. That's the only fair thing to do before declaring it's marketing fluff on either side.
 
Last edited:
3D stacking is not necessary to add more cache to a CPU. Server chips have had more cache than mainstream CPU's forever without stacking it above the die. The fact that AMD and Intel haven't added more cache without stacking which would be cheaper than stacking just drives home the lack of cost advantage point even further.
A cursory look at older Xeon processors (around 2005) tells me that if server chips did have more cache than mainstream CPUs, it wasn't much (maybe 2x at best). And if we're going outside of x86, like say to POWER, L3 cache was typically off-die at the time.

Cache also takes up a large chunk of real-estate on the die:
OiaaB.jpg

(note L2 cache is the top left block in each core)

Assuming this is a Skylake based CPU, say an i7-6700(K) which has 8MB of L3 cache, L3 is already taking up 15% of the die space. Ballooning this up to say 80MB, would make L3 cache be at least as big as the processor itself, which isn't practical using a planar, monolithic technique.

In any case, the context to my answer is addressing the concern of given a choice between adding more cache (now that there's technology to add more without the issue above), or pushing out a new architecture to compete with what Intel's offering, it's cheaper to add more cache because you don't have to design anything except the infrastructure to support more cache. And if all AMD has to do to compete with Adler Lake now is add more cache, then they can do that to buy more time to make sure Zen 4 is ready to hit the ground running.

EDIT: If you want more die shots showing just how much space cache takes up:
Zen 3: https://wccftech.com/amd-ryzen-5000...gh-res-die-shots-close-ups-pictured-detailed/ (L3 cache is taking up basically half a CCX chiplet's die area)
Penryn (Core 2 Duo 45nm): https://en.wikichip.org/wiki/intel/microarchitectures/penryn_(client) (also, the L2 cache basically takes up a third to half the die area)
Nehalem: https://www.anandtech.com/show/2658 (L3 takes up less space than the others, but still a significant amount)
 
Last edited: