- Mar 3, 2017
- 1,684
- 6,227
- 136
I agree but this doesnt explain the discrepancy in the tests posted by Geddagod, the only explanation of this discrepancy is that RPL gain nothing with AVX2 in Povray while the 7950X sure gain 18%, i find this contradictory, if Zen 4 gain such a percentage there s no reason that RPL wouldnt.
Now we dont know if his source has recompiled the tests such that they use AVX2 since that s not an official source.
The big redesign of the Zen 5 architecture. A big front side and backside in the core design.MT increase is typically < ST increase when TDP is unchanged
Where are you getting this
Please don't tell on me to my employer, for I'm ALWAYS using the $8000 server for checking websites that are stupidly blocked by the corporate firewallThat s not the point, no workstation doing critical work should be used to say browse the net
It's ughhh the opposite.From compiling all of the online speculation over the months. The conclusion for Zen 5 seems to be a large uplift in multicore processing power over Zen4 and as little as 10% uplift in single core performance. Not sure if that also means about 10% uplift in IPC.
We apparently don't have the same World Wide Web.From compiling all of the online speculation over the months. The conclusion for Zen 5 seems to be a large uplift in multicore processing power over Zen4 and as little as 10% uplift in single core performance.
See for yourself: https://www.anandtech.com/show/18747/the-amd-ryzen-9-7950x3d-review-amd-s-fastest-gaming-processorIs there a remote possibility that these benchmarks included workloads which are power-limited?
+62% 1T IPC is looking like reality with each passing day!It's ughhh the opposite.
You've been pushing "it's not a real real 1T bump" for eons. Maybe you should stop.
How does one even arrive at that conclusion.The big redesign of the Zen 5 architecture. A big front side and backside in the core design.
Are you suggesting that the Pov-ray binary that computerbase uses doesn't have any SSE/AVX code for AMD CPUs?
+62% 1T IPC is looking like reality with each passing day!
I'm more amazed by his legendary "there is no ST growth, only MT"How does one even arrive at that conclusion.
I do not think you are a real person. I think you are a bot. Compute that..It's ughhh the opposite.
You've been pushing "it's not a real real 1T bump" for eons. Maybe you should stop.
Zen 5 isn't your typical CPU now, is it?MT increase is typically < ST increase when TDP is unchanged
No AVX2 on AMD CPUs in Povray contrary to Intel.
That was proved by a member in this very forum who stated that it isnt implemented for AMD, only for Intel, as already said he recompiled Povray to enable AVX2 for AMD and the 7950X is 18% faster than with the regular version wich is the one used by Computerbase for ST and MT tests.
He also made tests with older versions of Zen and they were faster as well, even Zen 1 despite no AVX2 support.
Having your username and writing something like that is grounds for everyone to evacuate the website before the FBI assault teams break in...This doesn’t address my question. . .
here:This doesn’t address my question. . .
This doesn’t address my question. . .
That was what i m talking about, anything not related to this point has nothing do do with my statements, wich is that AVX2 is not used in Povray for AMD CPUs, i didnt talk of any other binaries, let alone of Computerbase ones.
The only relevance with Computerbase is that the Povray version they re using doesnt use AVX2 with AMD chips since they use the 3.7 version wich is the last official one.
So your comments about povray in SPEC are meaningless since comparing AVX to AVX2 compiled code versus SPEC compiled code is pointless.
So your comments about povray in SPEC are meaningless since comparing AVX to AVX2 compiled code versus SPEC compiled code is pointless.
Every 60 seconds in AnandTech, a 1% improvement to Zen 5 passes.+ 87% IPC. That's a bold claim cotton, lets see if it pans out.
GRRRR+ 87% IPC. That's a bold claim cotton, lets see if it pans out.