Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)

Page 815 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

gdansk

Diamond Member
Feb 8, 2011
3,276
5,186
136
I hope people compare power draw before/after on benchmarks that show improvement. The increased latency might have been due to power savings (whether intentional or not) so improving the latency may cost power. Maybe not enough to matter, but it would still be good for someone to check that instead of just running benchmarks and saying wowww higherrr numberzzzz!
No one is testing that because no one with a 2 CCD part cares. What they want are higher numbers...
It seems more relevant for Strix Point, but I doubt the new firmware is available there (if it even makes a similar change).
 

Hail The Brain Slug

Diamond Member
Oct 10, 2005
3,557
2,546
136
All the examples I saw in this thread, when you looked closer, showed the same regression with the 7950x versus the 7700x.
Yes, I saw you already deemed any peculiar results invalid as they were posted, so I won't press any further.

If any of these outlets that primarily do real application benchmarking re-benchmark, I look forward to seeing how the results change for the 9900X and 9950X.
 

Hitman928

Diamond Member
Apr 15, 2012
6,391
11,392
136
Yes, I saw you already deemed any peculiar results invalid as they were posted, so I won't press any further.

If any of these outlets that primarily do real application benchmarking re-benchmark, I look forward to seeing how the results change for the 9900X and 9950X.

Peculiar results are fine, but it needs to actually show the latency is the problem. The results I dismissed had no consistency in there being a cross-CCD penalty. For example, results where the 9900x performed just fine while the 9950x struggled. That, to me, means it's not the cross-CCD latency causing the weird performance issue, otherwise how could the 9900x be significantly outperforming the 9950x? You would need results where the 9950x under performs (and the 9900x, if tested, doesn't) compared to the 7950x. It's not an easy thing to show without trying to test for it, but so far, there hasn't really been any results that seem to be showing an issue with the additional latency, which may not have even existed in real work loads.
 

Hail The Brain Slug

Diamond Member
Oct 10, 2005
3,557
2,546
136
Peculiar results are fine, but it needs to actually show the latency is the problem. The results I dismissed had no consistency in there being a cross-CCD penalty. For example, results where the 9900x performed just fine while the 9950x struggled. That, to me, means it's not the cross-CCD latency causing the weird performance issue, otherwise how could the 9900x be significantly outperforming the 9950x? You would need results where the 9950x under performs (and the 9900x, if tested, doesn't) compared to the 7950x. It's not an easy thing to show without trying to test for it, but so far, there hasn't really been any results that seem to be showing an issue with the additional latency, which may not have even existed in real work loads.
Even if AGESA 1.2.0.2 shows the scaling to 9950X corrected for these workloads, there is no way to credit it to the reduced latency, either. Could be any number of things tweaked, changed, or fixed.

What matters in the end is if AMD fixes these peculiar results and strange scaling scenarios or not.
 

Josh128

Senior member
Oct 14, 2022
511
865
106
Agesa 1.2.0.2
This is with my bad 9950X
View attachment 107693

So the theory was that the high latencies were due to the second CCX being put in an aggressive sleep mode to save idle power vs Zen 4. Someone even posted examples from TPU 9950X review. Can you check idle power draws before and after?

*EDIT - reviewing TPUs review, I dont see the much lower idle results that I believe were posted somewhere earlier in this thread.
 

positivedoppler

Golden Member
Apr 30, 2012
1,140
236
116
Yeah, Im familiar with that, but that is not AMD saying they will fix it. Thats a rumor from some unknown Asian person on X.


since 9000 release we now have 1. windows update 2. increase power limit 3. latenacy improvement. Sloppy release, i lost track of all the changes so i feel bad for the reviewers who have to constantly rebenchmark
 

fastandfurious6

Senior member
Jun 1, 2024
214
311
96
WOW!

After hearing rumors about newest beta agesa 1.2.0.2 improving both performance and latency i decided to put it to the test!
I downloaded and flashed to newest ASUS bios 2401 and ran capframeX core-to-core latency

And wow indeed, the wispers were true, with agesa 1.2.0.2 AMD have finally fixed the core-to-core latency plaguing the 9000 from launch

This is with agesa 1.2.0.1A bios 2303 = cross CCD latency is ~180ns
View attachment 107682

Now with agesa 1.2.0.2 bios 2401 = cross CCD latency is ~75ns
View attachment 107683

Otherwise 100% same settings used for both runs, will be interesting to see how this affect the gaming performance / latency bound benchmarks

+40% IPC is achievable after all!!!!

PBO x20
disable SMT
disable windows security mitigations
windows kernel patches
fix parking behaviour
microcode updates
dual decoder fix
🙏 pray to Lisa Su morning, noon, afternoon, evening and in your sleep 🙏


the prophecy is becoming real !!!!!!!
 

Det0x

Golden Member
Sep 11, 2014
1,346
4,545
136
Well done to AMD for addressing the inter-CCD latency penalty. It would be nice if everyone could go back and rebenchmark Zen5 after all these OS updates and AGESA tweaks.
Like i wrote 1 page back, i hope/think Zen5 will get re-benched when X870E launches at the end of this month.
Now that the reviewers have learned how to setup their windows install for Ryzen + more mature agesa, Zen5 and especially the dual CCD cpus should look less underwhelming 👍
 

Det0x

Golden Member
Sep 11, 2014
1,346
4,545
136
Hardwarelux also found the same as me
As soon as the Windows Update 24H2 has been rolled out and the motherboard manufacturers have officially incorporated the latest AGESA updates into their BIOS versions, we will test all Ryzen processors again. This should mean we are well prepared for Arrow Lake.
 

tsamolotoff

Senior member
May 19, 2019
218
446
136
Games should have stayed on one CCD, if everything else was working right.
There are games (like CA engine games (Total War, Warhammer etc)) that utilize all cores, and there was a significant regression there according to the last AT tests (rip). Maybe someone can do the comparison and see if this update fixes abysmal fps in that benchmark
 

MS_AT

Senior member
Jul 15, 2024
365
798
96
Hardwarelux also found the same as me
interestingly they get higher absolute values. I guess you are using a tuned sample with higher IF clock? Or maybe different test frameworks.

And more and more comments on the forums seems to take the easy assumption that improvements to this synthetic benchmark is what is driving all the performance improvements.
There are games (like CA engine games (Total War, Warhammer etc)) that utilize all cores, and there was a significant regression there according to the last AT tests (rip). Maybe someone can do the comparison and see if this update fixes abysmal fps in that benchmark
There are other outlets where 7950x vs 9950x score the same in Total War Warhammer III in launch day reviews.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
26,389
15,513
136
@Det0x I can't find the screenshots where you displayed the effects of interccd latency. What tool did you use ? or what post number was it that showed these ? I want to test mine. I went back several pages and could not find it.
 

blackangus

Member
Aug 5, 2022
160
217
86
@Det0x I can't find the screenshots where you displayed the effects of interccd latency. What tool did you use ? or what post number was it that showed these ? I want to test mine. I went back several pages and could not find it.

I think this is what you were looking for good sir!
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
26,389
15,513
136
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |