AMD Zen 4 AVX-512 Performance Analysis On The Ryzen 9 7950X
While I've only had a few days to look at the AVX-512 performance with Zen 4 on the desktop, the results have been looking well so far from this comparison with software able to make use of AVX-512.
For a variety of different workloads like JSON parsing, CPU mining, 3D rendering with OSPRay Studio, AI workloads like OpenVINO, and various other tests, the AVX-512 execution was yielding sizable speed-ups over falling back to AVX2 instead thanks to the workaround with clearcpuid=304 and compiling the open-source software with no AVX-512 extensions enabled (but leaving AVX / AVX2).
Here is a look at the CPU peak frequency across the entire span of benchmarks tested... No real change compared to without AVX-512, unlike early Intel AVX-512 CPUs that downclocked notoriously when executing AVX-512 instructions.
Likewise, the CPU power consumption was similar when running the AVX-512 enabled software. For nearly all the tests the AVX2 vs. AVX-512 results were almost identical. Among the exceptions were with the Intel oneDNN and OpenVINO software where the AVX-512 usage led to a 10~20 Watt increase for the Ryzen 9 7950X, but the rest of the time tended to deliver nearly the same results.
During these demanding workloads and even with the Cooler Master ML280 does get quite warm at least as reported by the Linux k10temp driver. AMD has detailed that running at a 95 Celsius TJMax for the higher core count parts isn't surprising and "intended and by design" as it's a max safe operating temperature and is being effectively handled. In any event, making use of AVX-512 hadn't led to any noticeable increase in the CPU temperature. AMD cites typical load temperatures for the Ryzen 7000 series as 70 to 90 degrees Celsius.
On average for the tested AVX-512 workloads, making use of the AVX-512 instructions led to around 59% higher performance compared to when artificially limiting the Ryzen 9 7950X to AVX2 / no-AVX512.
From these results I am rather impressed by the AVX-512 performance out of the AMD Ryzen 9 7950X. While initially being disappointed when hearing of their "double pumping" approach rather than going for a 512-bit data path, these benchmark results speak for themselves. For software that can effectively make use of AVX-512 (and compiled so), there is significant performance uplift to enjoy while no negative impact in terms of reduced CPU clock speeds / higher power consumption (with oneDNN being one of the only exceptions seen so far in terms of higher power draw).
AVX-512 is looking good on the Ryzen 7000 series and I'll continue running more benchmarks over the weeks ahead. These AVX-512 results make me all the more excited for AMD EPYC "Genoa" where AVX-512 can be a lot more widely-used among HPC/server workloads. Thanks to AMD for providing the Ryzen 7000 series review kit for enabling this launch-day Linux benchmarking.
Now continue on to read today's complete AMD Ryzen 9 7900X + 7950X Linux review for all of the CPU benchmarks I prepared for the launch day. If you enjoy all of my Linux hardware testing over the past 18+ years, consider joining Phoronix Premium during the autumn special.
If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal or Stripe tips are also graciously accepted. Thanks for your support.