Discussion Speculation: Zen 4 (EPYC 4 "Genoa", Ryzen 7000, etc.)

Page 442 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Vattila

Senior member
Oct 22, 2004
809
1,412
136
Except for the details about the improvements in the microarchitecture, we now know pretty well what to expect with Zen 3.

The leaked presentation by AMD Senior Manager Martin Hilgeman shows that EPYC 3 "Milan" will, as promised and expected, reuse the current platform (SP3), and the system architecture and packaging looks to be the same, with the same 9-die chiplet design and the same maximum core and thread-count (no SMT-4, contrary to rumour). The biggest change revealed so far is the enlargement of the compute complex from 4 cores to 8 cores, all sharing a larger L3 cache ("32+ MB", likely to double to 64 MB, I think).

Hilgeman's slides did also show that EPYC 4 "Genoa" is in the definition phase (or was at the time of the presentation in September, at least), and will come with a new platform (SP5), with new memory support (likely DDR5).

Untitled2.png


What else do you think we will see with Zen 4? PCI-Express 5 support? Increased core-count? 4-way SMT? New packaging (interposer, 2.5D, 3D)? Integrated memory on package (HBM)?

Vote in the poll and share your thoughts! :)
 
Last edited:
  • Like
Reactions: richardllewis_01

Hitman928

Diamond Member
Apr 15, 2012
6,187
10,694
136
Although you're absolutely correct that we have limited data to truly pinpoint what's happening, logic and process of elimination points to the GPU, because none of those things you mentioned can account for that much power draw. Both memory modules use the same voltage (1.10v) and probably use less than 10w as DDR5 has very low power draw. Everything else in the system was the same except for the CPU, motherboard and memory. Also, since I have that particular game, I know for a fact that it is heavily GPU bound and an RTX 4090 can consume up to 450w by itself at 4K maxed settings. I can post screenshots on demand if you want.

Even though HWUB's system power draw test is done at 2K, the Zen 4 system's power consumption is absurdly low and the Raptor Lake system is closer to what I would expect being an owner of the game and an RTX 4090.

In their 13600K review, HWUB measures the system power draw at 315w for the 13600K and 226w for the 7600x with blender. 90w difference, which is still less than the much lower power draw gaming workload LOL!

Powerr.png




I have the same platform and I've never noticed anything of the sort myself personally. Since I built my system, it has performed beyond my expectations even when it's underclocked and behaves as expected.

That said, I'm not saying we should discount HWUB's results completely. I'm only saying there are anomalous results in their round up which may point to problems with their methodology.

You are comparing 2 very different workloads and trying to come to a conclusion with only 2 data points and loads of variables. This is not possible. For instance, GN only found a 44.4W difference in CPU power being used during blender between a 13600k and a 7600x. If we were to try and use that in a cross comparison, it would show that the Intel platform has an unexplained 44.6 W of extra power consumption in the system that isn't being used by the CPU or GPU (Edit: JT's actually shows 135W difference in power between the 2 at the system level during Cinebench which would show an even greater amount of unaccounted for power [90.6 W] for the Intel platform). That's without considering PCIe links being active or how often the GPU would need to go to memory in the gaming scenario or how often it would have to load data through a high speed PCIe bus. So with almost no data and no controls over any variables, we can't reach any kind of conclusion here. You can say it looks strange, sure, but with zero evidence to the contrary or any actual flaw you can identify in their testing setup, I see no reason to say we should toss out this result just because it doesn't look like you would expect. If they look into it and say, we made a mistake here and update the results, great, but so far they seem pretty confident in their results so I feel comfortable in treating it as valid as any other data point from any other reviewer.

I'm glad you are very happy with your system, but unless you have actually replicated HWUB's system, software, and test methodology, it doesn't really give any insight into the validity of their results.
 
Last edited:

Yosar

Member
Mar 28, 2019
28
136
106
What makes those other reviewers crap?
Btw, the metareviews from 3Dcenter.org are usually supported by large-data set benchmarks from other popular (maybe you consider them "good") benchmarkers such as hardware unboxed.
For example the 5800x3d is ~15% faster than the 5800x in both hardware unboxed 41 game benchmark, as WELL as the meta review from 3Dcenter.

And 7600XT is 5% faster than 13600 K according to HU 54 games benchmark.
And according to 'metareview' based on 6-8 usually the same tests it's 13600 K which is 5% faster. That's why this 'metareview' is crap. You take already distorted results (too little sample of games) and multiply them 10 times using almost exactly the same data.
Not every benchmark is worth the same, and especially if they use the same tests. It's not how stochastic works. Quality of samples matter.
One HU 54 games benchmark is better and more trustworthy than 10 6-8 games benchmarks.
 

Geddagod

Golden Member
Dec 28, 2021
1,296
1,368
106
And 7600XT is 5% faster than 13600 K according to HU 54 games benchmark.
And according to 'metareview' based on 6-8 usually the same tests it's 13600 K which is 5% faster. That's why this 'metareview' is crap. You take already distorted results (too little sample of games) and multiply them 10 times using almost exactly the same data.
Not every benchmark is worth the same, and especially if they use the same tests. It's not how stochastic works. Quality of samples matter.
One HU 54 games benchmark is better and more trustworthy than 10 6-8 games benchmarks.
A couple things:
HU unboxed battle field V benchmarked is bugged because of Intel E-cores glitch, which he himself said in the video, shrinking the margin to 4 percent. And if you actually WATCHED the video, he says margins that are within 5 percent are effectively a tie.
The meta review does include a repeat of many of the same games, sure, but that's because the games tested are supposed to be representative of CPU-limited scenarios and popular games.
But more importantly, I also know that 3dcenter also includes data like 99% lows and such, other than just averages, though I'm not sure if they did that on the raptor lake meta review, ik they did it in previous meta reviews. They also include data from lower resolutions to accentuate the performance differences.
Additionally HU also tests games in a methodology that many don't exactly agree with, and result in different percentages in games than basically every other reviewer. For example, Ik this was already discussed above in this thread, but the Plague Requiem anomaly.
So to wrap it up:
HUB themselves admitted the difference between the two CPUs was within the margin of error in their own video.
HUB has some testing methodology flaws that results in some games having varying results such as Plague and Battlefield 5
3Dcenter.org includes data from lower resolutions and also data such as 99 percent lows and such.
Overall I would STILL say 3dcenter is the "higher quality" source. Because even IF they have a lower aggregate game pool, the fact that the impact of flawed testing methodology could be minimized by having so many reviewers, means that it ends up being more accurate in the end.
 

Yosar

Member
Mar 28, 2019
28
136
106
If HWUB was known for reasonably representative methodology, that might make sense. But all you're doing here is amplifying the impact of obviously flawed testing.

And you know the methology of those tests from 'metareview'? And they are clearly better? You have that certainty? Like Computerbase shown above? Definitely much better methodology. 3 basically the same benchmarks, and one crippling internally one of the tested processors.
If anything the 'representative' methodology for game test with 6-8 is much more suspicious, due to being done in time constraints or simply lack of knowledge or whatever reasons.
13 tests for 21 in 'metareview' are 6-8 games tested benchmarks, the rest are equal or below 12, it's a pure joke not 'representative' game testing.
 
  • Like
Reactions: Markfw

Exist50

Platinum Member
Aug 18, 2016
2,452
3,102
136
HWUB is a credible source of information on PC Hardware and technology. You are just some random dude on the internet.
So, they're credible not only despite clearly inexplicable results, but you insist that they're more credible than almost every other outlet combined? Is that seriously the argument you want to make?

Or let's put it another way. An outlet is only credible if they produce credible results. You're reversing cause and effect.
And you know the methology of those tests from 'metareview'? And they are clearly better? You have that certainty?
You realize the irony of this, right? You're defending an outlet with very clearly flawed methodology, and insisting that instead it must be everyone else that's wrong. So yes, I don't believe that they're better than almost every other outlet combined.

It's incredible the flaws some people are willing to overlook when then produce results that conform to their preexisting biases.
 
  • Like
Reactions: Carfax83

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
errr no. A site testing ST perfomance with 4 benchmarks, three of them from the same tool (Cinebench), and the other (Pov-Ray) hurting Ryzen perfomance deactivating AVX2 when a Ryzen CPU is detected (*). Yeah great approach to testing.

Wait a sec, are you implying that Computerbase.de intentionally deactivated AVX2, or that theCinebench and Pov-Ray applications themselves are deactivating AVX2 when a Ryzen CPU is detected.

I've never heard anything like what you are saying. I'm not that familiar with POV-Ray, but is AVX2 even activated by default? You would think it is, as AVX2 does come in handy for rendering workloads.

And wouldn't Intel CPUs be affected by this as well? Intel CPUs also support AVX2 as well you know, and actually, Raptor Lake has better AVX2 performance core for core compared to Zen 4.

So when you guys see pov-ray results (like the ComputerBase one), you have to know Ryzen perf is being damaged avoiding the use of AVX2 instruction set (15-16% dif). It wouldn't be slower than Raptor Lake, but actually faster:

Until we know whether AVX2 is on by default and has to be deactivated, or if it's off by default and has to be activated, I'm reserving judgement. I find it odd that AVX2 wouldn't be on by default as it is with Cinebench.

And these benches you uploaded are obviously edited:

povrayst-png.72463
 
Last edited:

Mopetar

Diamond Member
Jan 31, 2011
8,114
6,770
136
Oh no, not this again ☠☠☠

Apparently I missed something. Normally I'd just drop it to avoid rekindling a flame that's been smothered, but right now it'd only be like tossing a smoke bomb into a forest fire.

It looks like another post pointed out that it was a bug due to e-cores. Steve just needs to disable the e-cores to get an accurate result. Obviously it's important to report results like that though because someone running that game who doesn't know there's a bug is losing a lot of performance.

AMD really needs to hurry up and release Zen 4D if only to beat the pants off of everything else so this thread has a chance of getting back on track.
 
  • Like
Reactions: Kaluan

MarkPost

Senior member
Mar 1, 2017
322
616
136
Wait a sec, are you implying that Computerbase.de intentionally deactivated AVX2, or that theCinebench and Pov-Ray applications themselves are deactivating AVX2 when a Ryzen CPU is detected.

I've never heard anything like what you are saying. I'm not that familiar with POV-Ray, but is AVX2 even activated by default? You would think it is, as AVX2 does come in handy for rendering workloads.

And wouldn't Intel CPUs be affected by this as well? Intel CPUs also support AVX2 as well you know, and actually, Raptor Lake has better AVX2 performance core for core compared to Zen 4.



Until we know whether AVX2 is on by default and has to be deactivated, or if it's off by default and has to be activated, I'm reserving judgement. I find it odd that AVX2 wouldn't be on by default as it is with Cinebench.

I thought I was pretty clear: when binary detects a Ryzen CPU, it only uses AVX, avoiding AVX2 despite Ryzen as we all know includes AVX2. When binary detects an Intel CPU with AVX2 support, uses AVX2.

So for Ryzen to use AVX2, its necessary to activate it in source code (as I said, its open source) and compiling a new binary (i did that, and tested it, as can be seen above).

And these benches you uploaded are obviously edited:
povrayst-png.72463

Obviously. Its including Ryzen perfomance with a proper binary supporting AVX2 not only for Intel CPUs, but for Ryzen too.
 

maddogmcgee

Senior member
Apr 20, 2015
391
343
136
Bought myself a 5800x3d upgrade to my current mobo based on the super cheap cost compared to the alternatives. Tested a couple of games and saw a huge increase from what I had before (literally turned one game from sub 30 chop fest to 60fps beast). Less than a week later, I am looking at this thread and wondering about the 7000 series x3d release date......:weary: It's hard to be content as a nerd.
 

MarkPost

Senior member
Mar 1, 2017
322
616
136
If anyone with a Ryzen CPU is interested to test it, Ive uploaded here https://drive.google.com/file/d/13mfiQup_disHpgmb7TekN2Ge0FWq3I3Q/view?usp=sharing the povray binary compiled by me with AVX2 ON for all CPUs with AVX2 support (Ryzen included). And here can be downloaded the program itself to compare perfomance: Release POV-Ray Beta Release v3.8.0-beta.2 · POV-Ray/povray · GitHub

Test it with the "official" version, and then just replace the binary to test with the recompiled one

EDIT: forgot to mention. Pass to extract the binary: AnandTechForums
 
Last edited:

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
I thought I was pretty clear: when binary detects a Ryzen CPU, it only uses AVX, avoiding AVX2 despite Ryzen as we all know includes AVX2. When binary detects an Intel CPU with AVX2 support, uses AVX2.

So for Ryzen to use AVX2, its necessary to activate it in source code (as I said, its open source) and compiling a new binary (i did that, and tested it, as can be seen above).

Interesting. Of course, I have no way of verifying whether what you say is accurate. Pov-Ray has been around for decades. I remember Pov-Ray from the Pentium IV era, or even possibly before then.

AVX2 has been supported by AMD for a long time now as well, so I have no idea why it would be deactivated on Ryzen CPUs assuming what you say is true.

Doing a quick Google search seems to indicate that AVX2 acceleration IS SUPPORTED by Pov-Ray for AMD CPUs. There are even posts on PovRay.org that speak of AMD CPUs using AVX2.

At any rate, I'm not a programmer so I can't investigate this problem properly. But I do know that some compilers are faster than others due to better vectorization. So maybe the compiler you're using has better vectorized code output perhaps.

Just a guess.
 

jamescox

Senior member
Nov 11, 2009
644
1,105
136
Interesting. Of course, I have no way of verifying whether what you say is accurate. Pov-Ray has been around for decades. I remember Pov-Ray from the Pentium IV era, or even possibly before then.

AVX2 has been supported by AMD for a long time now as well, so I have no idea why it would be deactivated on Ryzen CPUs assuming what you say is true.

Doing a quick Google search seems to indicate that AVX2 acceleration IS SUPPORTED by Pov-Ray for AMD CPUs. There are even posts on PovRay.org that speak of AMD CPUs using AVX2.

At any rate, I'm not a programmer so I can't investigate this problem properly. But I do know that some compilers are faster than others due to better vectorization. So maybe the compiler you're using has better vectorized code output perhaps.

Just a guess.

I remember POV-ray being around a very long time ago. Wikipedia says initial release in 1991; I probably didn't see anything about it until the late 1990s. The first pentium was 1993, so we are talkign 386 or 486 days. The wikipedia article list the last stable release as 2018, so it wouldn't be surprising that that release doesn't have proper support for Zen processors. They should check the feature flag such that things like this does not happen. There probably are non-official versions with it fixed for Ryzen processors. I don't know if POV-ray is actually used much these days if it hasn't had a release in 4 years.
 
  • Like
Reactions: Carfax83

Exist50

Platinum Member
Aug 18, 2016
2,452
3,102
136
I remember POV-ray being around a very long time ago. Wikipedia says initial release in 1991; I probably didn't see anything about it until the late 1990s. The first pentium was 1993, so we are talkign 386 or 486 days. The wikipedia article list the last stable release as 2018, so it wouldn't be surprising that that release doesn't have proper support for Zen processors. They should check the feature flag such that things like this does not happen. There probably are non-official versions with it fixed for Ryzen processors. I don't know if POV-ray is actually used much these days if it hasn't had a release in 4 years.
If you want to do ray tracing today for any practical use case, you'd use a GPU renderer. Pretty much the same argument as can be used against Cinebench.
 

uzzi38

Platinum Member
Oct 16, 2019
2,705
6,427
146
Official supported memory standard = strange choice of RAM got it!

Anything beyond the official memory standard is technically overclocked and not stock. While I myself never ever run purely stock settings, I respect the reviewer's decision to stick to standardized settings.



You like to snipe from the shadows, but not one of you has attempted to explain any of these anomalies. Your take is, "just accept it." Well, maybe you're that naive and want to assume that HWUB is infallible but I'm sure as hell not.
Enabling XMP is also technically overclocking. Would you like reviewers to use JEDEC spec only from here on out?
 
  • Haha
Reactions: Kaluan

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
Enabling XMP is also technically overclocking. Would you like reviewers to use JEDEC spec only from here on out?

I already said that while I personally would never use stock memory speeds (I prefer faster memory as I have it in my own rig), I respect the reviewer's decision to do so. The official supported memory was the status quo for reviews for a long time before AMD "persuaded" reviewers to use overclocked settings, and not just on the memory, but on the fabric interconnect as well which was a big departure from previous review standards.

But I have to ask, how far are you willing to go? I said on a previous page, that while Zen 4 may benefit from DDR5 6200, Intel can go much higher. I bet many AMD supporters here and elsewhere would cry foul if a reviewer tested Zen 4 with DDR5 6200 and Raptor Lake with DDR5 8000 or some such.

For reasons of fairness, I suppose most reviewers still try to have similar memory speeds between Zen 4 and Raptor Lake. Only two reviewers that I can recall used significantly faster RAM with Raptor Lake compared to Zen 4, Linus Tech Tips and Tech Yes City. Both used DDR5 6800 CL34. One thing I found interesting, is that with the faster DDR5 RAM, the 13900K beat the 7950x in compression performance quite handily, and the 12900K edged out the 5950x.

Anyway, there are much faster DDR5 speeds available now than when those reviews were made.


 

MarkPost

Senior member
Mar 1, 2017
322
616
136
I remember POV-ray being around a very long time ago. Wikipedia says initial release in 1991; I probably didn't see anything about it until the late 1990s. The first pentium was 1993, so we are talkign 386 or 486 days. The wikipedia article list the last stable release as 2018, so it wouldn't be surprising that that release doesn't have proper support for Zen processors. They should check the feature flag such that things like this does not happen. There probably are non-official versions with it fixed for Ryzen processors. I don't know if POV-ray is actually used much these days if it hasn't had a release in 4 years.

AVX2 isn't implemented in stable version (3.7.0). It's implemented in beta versions (3.7.1 and 3.8.0, latest Aug 9, 2021)
 

LightningZ71

Golden Member
Mar 10, 2017
1,798
2,156
136
According to clipka at news.povray.org, this is a deliberate decision that POVRAY made AT THE REQUEST OF AMD in 2017. AMD did code-path testing on a bunch of machines and determined that the AVX codepath was faster on their hardware than AVX2. This is not a surprise as Zen1 handles 256 bit AVX2 operations as a pair of 128 bit ones and has roughly half the throughput. The link for the records:

CLIPKA discusses AVX code path code in POVRAY

Since there have been no major updates to POVRAY since then, it makes sense that it hasn't been addressed.
 

MarkPost

Senior member
Mar 1, 2017
322
616
136
According to clipka at news.povray.org, this is a deliberate decision that POVRAY made AT THE REQUEST OF AMD in 2017. AMD did code-path testing on a bunch of machines and determined that the AVX codepath was faster on their hardware than AVX2. This is not a surprise as Zen1 handles 256 bit AVX2 operations as a pair of 128 bit ones and has roughly half the throughput. The link for the records:

CLIPKA discusses AVX code path code in POVRAY

Since there have been no major updates to POVRAY since then, it makes sense that it hasn't been addressed.


yeah, that explanation would make sense if it would be true. But it isn't. Anyone with a Zen1 CPU can check that Zen1 is actually faster with binaries including AVX2 support for all CPUs. Proof:

Latest Beta version (3.8.0 beta 2, dated Aug 9, 2021) AVX

1700povray38b2avx.PNG

Latest Beta version (3.8.0 beta 2, dated Aug 9, 2021) AVX2

1700povray38b2avx2.PNG


And more importantly, FIRST beta build WITH AVX2 implemented (3.7.1-beta.3, dated Feb 19, 2017). Ironically AVX2 in this build isnt limited to Intel CPUs only, but AVX2 is activated for any AVX2 capable CPU, no need to recompile it

1700povray371b3avx2.PNG

... compared with the immediately previous beta build WITHOUT AVX2 implemented, just AVX here (3.7.1-beta.2, dated Jan 11, 2017):

1700povray371b2avx.PNG


In fact AVX2 was removed for non Intel CPUs with v3.7.1 Beta 6 (dated May 7, 2017). This is the result with this build (capped to generic AVX for non Intel CPUs):

1700povray371b6avx.PNG


... compared with the "official" v3.7.1 beta 5, the latest including AVX2 for all capable CPUs (dated Mar 25, 2017):

1700povray371b5avx2.PNG


As can be seen, Zen1 is ever faster with AVX2 (remember, more is better). Oh and since 2017 there are a bunch of beta releases, so no, doesn't make sense that it hasn't been addressed.
 
Last edited:

Exist50

Platinum Member
Aug 18, 2016
2,452
3,102
136
That's only a worse case scenario for CPU power differences though, not necessarily system power. They can be two different things.
Sure, but you can't explain 100W difference as run to run variation or any of the typical sources of statistical noise. 100W, either the CPU or GPU or both are responsible.
That's way more power than can be explained by CPU power alone. I already agreed 100W seems high and it's something HWUB should look into, but again, assuming that all of that power is being consumed by the GPU is a bad assumption and just throwing out their data because it looks high is an overreaction.
So why didn't HWUB actually investigate? There're plenty of applications that can tell you approximate numbers for the CPU, GPU, etc., and that should be plenty sufficient to diagnose the cause here. Instead they take the lazy way out and publish clearly flawed data without any effort to explain themselves.
They can put a caveat on it explaining it's probably a bug, which is exactly what they did, but witholding the data is not the prudent thing to do. If it is an e-core issue, they're not the only ones to still face random e-core issues from time to time.
They're the only ones that seem to have this particular issue. The scheduler isn't random, it's deterministic. If it's actually a bug with the scheduler it should be reproducible by others.

But the exact cause of the issue is besides the point. You have a hypothesis right here - that it's a scheduling issue. It would be absolutely trivial for HWUB to disable E-cores, rerun the test, and make a conclusion on that hypothesis. Yet they couldn't be bothered to put in that minimal amount of effort?

I don't understand why people insist they're such a reputable outlet while defending their unwillingness to put even a minimal amount of effort into validating their results.
 

Abwx

Lifer
Apr 2, 2011
11,557
4,349
136
Sure, but you can't explain 100W difference as run to run variation or any of the typical sources of statistical noise. 100W, either the CPU or GPU or both are responsible.

So why didn't HWUB actually investigate?

Because there s no need to investigate.

The 13900K take something like 35-40W more than the 7950X in games, assuming the 13900K has 5% better perf this imply that the GPU , wich is say at 390W with the 7950X, will madate 15% more power for thoses 5% better GPU perf, that s about 100W more GPU + CPU power and something like 130W more at the main.
 

Exist50

Platinum Member
Aug 18, 2016
2,452
3,102
136
Because there s no need to investigate.

The 13900K take something like 35-40W more than the 7950X in games, assuming the 13900K has 5% better perf this imply that the GPU , wich is say at 390W with the 7950X, will madate 15% more power for thoses 5% better GPU perf, that s about 100W more GPU + CPU power and something like 130W more at the main.
The comparison system was a 13600K vs a 7600X, where not only was the 13600K system consuming 100W extra, but it also was performing substantially worse than the 7600X. So if anything, the trend you describe would go in the opposite direction.
 
  • Like
Reactions: Geddagod