SMT Remains Very Advantageous For 5th Gen AMD EPYC Performance
In total I ran more than 120 benchmarks on this AMD EPYC 9575F 1P Supermicro server for this fresh SMT comparison.
On a geo mean basis for all of the benchmarks in total, having SMT enabled was a 13% improvement than running the EPYC 9575F processor with SMT disabled. SMT typically was of measurable benefit to the 5th Gen AMD EPYC processor with the exception of some HPC workloads that perform better with SMT disabled or otherwise limited by memory bandwidth. SMT also hurt the OpenVINO inference latency but by and large Simultaneous Multi-Threading remains an important and valuable feature for AMD processors. For HPC clusters and others with workloads where SMT isn't of benefit, it can be simply disabled.
There are many real-world applications from ray-tracing and rendering workloads to databases, compression, server workloads like Memcached / Nginx / PostgreSQL, and many others where SMT provides very significant performance benefits.
SMT enabled on the AMD EPYC 9575F on average led to just a 2 Watt increase to the CPU power consumption than when it was disabled. Regardless of SMT, the EPYC 9575F was running up to its 400 Watt TDP.
With the CPU power consumption not significantly different, SMT didn't have any impact on the server CPU temperature.
With SMT enabled there was a small CPU peak frequency difference but again in the vast majority of the workloads having SMT enabled means better performance.
Those wanting to go through all the data in full can see this result page for all 120+ benchmarks for this fresh round of 5th Gen AMD EPYC SMT benchmarking.
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.