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

Page 609 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Gideon

Gideon

Platinum Member
Nov 27, 2007
2,030
5,035
136
TwistedAndy said:
There are some new Geekbench 6 results for AMD Ryzen 9 9900X (link):

View attachment 102582

It looks pretty good
Click to expand...
Just came to post it. This looks to finally be retail silicon.

The clocks are sky-high though. Essentially 5.65Ghz the entire run (usually there is some fluctuation).

If this is stock perf, it looks quite decent. Nothing out of this world though
 
  • Like
Reactions: lightmanek, Tlh97 and coercitiv
inf64

inf64

Diamond Member
Mar 11, 2011
3,884
4,692
136
ST is in line with what AMD presented (~17% faster) while MT is a bit lower, maybe because of a more limited PPT for 9900X?
 
  • Like
Reactions: Mopetar
Hail The Brain Slug

Hail The Brain Slug

Diamond Member
Oct 10, 2005
3,882
3,311
146
Nothingness said:
I don't see 7900X Linux scores 10% higher than Windows one. But I might be missing something; looking into GB database is such a pain :(
Click to expand...
Earlier in the thread I demonstrated a 5-6% ST uplift from W11 to Linux on my 7950X.

5% on Debian, 6% on Manjaro (Arch)

Page 586 - Discussion - Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)

Page 586 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
forums.anandtech.com forums.anandtech.com
 
Last edited:
  • Like
Reactions: Elfear, Tlh97, lightmanek and 8 others
poke01

poke01

Diamond Member
Mar 8, 2022
4,226
5,565
106
Nothingness said:
I don't see 7900X Linux scores 10% higher than Windows one. But I might be missing something; looking into GB database is such a pain :(
Click to expand...

My bad. I seen somewhere that Linux scores higher, should have double checked the %.
 
poke01

poke01

Diamond Member
Mar 8, 2022
4,226
5,565
106
Okay to spice things up. How much of the GB uplift is due to AVX-512?

Zen 5 should be a PS3 emulator king!
 
N

Nothingness

Diamond Member
Jul 3, 2013
3,301
2,373
136
Hail The Brain Slug

Hail The Brain Slug

Diamond Member
Oct 10, 2005
3,882
3,311
146
Nothingness said:
If you look at the result @yuri69 posted above you can guess ;)
The largest outlier is Object Remover.
Click to expand...
Yeah but how much does the weighting of the tests meaningfully affected change the actual score? We'd need a run with and without AVX512 enabled to really know.

I could do that on my 7950X but I'm not sure how valuable it would be over having a Zen 5 sample to run.
 
N

Nothingness

Diamond Member
Jul 3, 2013
3,301
2,373
136
Hail The Brain Slug said:
Yeah but how much does the weighting of the tests meaningfully affected change the actual score? We'd need a run with and without AVX512 enabled to really know.

I could do that on my 7950X but I'm not sure how valuable it would be over having a Zen 5 sample to run.
Click to expand...
Indeed. But that still helps estimating what the improvements are for tests that don't benefit from AVX-512. And here clang is a bit of a disappointment while HTML5 is a surprise. Let's wait a few more days to get a more accurate picture.
 
  • Like
Reactions: Tlh97, lightmanek and SarahKerrigan
N

Nothingness

Diamond Member
Jul 3, 2013
3,301
2,373
136
igor_kavinski said:
Isn't clang going to need compiler patches for Zen 5 support? That also means that GB6 may need a revision.
Click to expand...
Do you realize that'd mean all legacy code would need recompilation to be fast on Zen5?

That's not how things work. And that's why OoOE is required to have a fast CPU (until things change dramatically in the HW/SW space...): to let a beefed up core benefit from its extra ressources without having to recompile everything. And all modern CPU are good at that.

I can't remember when was the last time, recompiling with instruction scheduling for native CPU brought any significant speedup (>2%, that is not in measurement noise) for code I run. Of course native instruction selection is a different story but that doesn't apply to clang.
 
  • Like
Reactions: Tlh97, Makaveli, Mopetar and 1 other person
inf64

inf64

Diamond Member
Mar 11, 2011
3,884
4,692
136
A bit unrelated, but does mention Zen 5 :
www.techpowerup.com

AMD is Becoming a Software Company. Here's the Plan

Just a few weeks ago, AMD invited us to Barcelona as part of a roundtable, to share their vision for the future of the company, and to get our feedback. On site, were prominent AMD leadership, including Phil Guido, Executive Vice President & Chief Commercial Officer and Jack Huynh, Senior VP &...
www.techpowerup.com www.techpowerup.com

The slide AMD used shows the previously disclosed IPC figure, so it seems they are sticking to it
1720452648757.png
 
  • Like
Reactions: Elfear, Tlh97, lightmanek and 4 others
Hail The Brain Slug

Hail The Brain Slug

Diamond Member
Oct 10, 2005
3,882
3,311
146
Nothingness said:
Indeed. But that still helps estimating what the improvements are for tests that don't benefit from AVX-512. And here clang is a bit of a disappointment while HTML5 is a surprise. Let's wait a few more days to get a more accurate picture.
Click to expand...

W10 AVX 512 on/off. Hardly any difference. Ignore comparison to previous W11 scores, W10 seems to score slightly better for ST.
 
  • Like
Reactions: Tlh97, lightmanek, inf64 and 1 other person
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom