AMD 4th Gen EPYC 9654 "Genoa" AVX-512 Performance Analysis
With the great AMD 4th Gen EPYC Linux performance showing significant generational uplift and dominating against the current Xeon Scalable "Ice Lake" competition, it's a combination of the twelve channels of DDR5 system memory support, up to 96 cores per socket, introduction of AVX-512, and other Zen 4 micro-architectural improvements. As follow-up testing articles to all of the Genoa data delivered thus far, over the weeks ahead I have additional benchmark results to share looking more closely at these different areas of improvement for AMD 4th Gen EPYC. In today's article is a look at the EPYC 9654 2P performance with AVX-512 on/off while also looking at the CPU power consumption impact and the affect on CPU clock frequencies and thermals.
For both the AMD Ryzen 7000 series and EPYC 9004 series, Zen 4's AVX-512 implementation has proven to be rather robust and nifty. Initially when learning about Zen 4's AVX-512 implementation for the Ryzen 7000 series launch at Lost Pines, I was a bit weary as they were talking up their "double pumping" approach rather than going with a 512-bit FPU data path. However, once hands-on testing I was left rather impressed with Zen 4's AVX-512 performance and much better off than the early generations of AVX-512 on the Intel side. I looked at the AVX-512 performance on the Ryzen 9 7950X and other comparisons since then on the desktop side while more recently being able to look at the Genoa AVX-512 performance.
To little surprise given the AVX-512 results with the Ryzen 7000 series, the AVX-512 performance with AMD 4th Gen EPYC is in great standing as I've come to enjoy over the past number of weeks. Genoa's AVX-512 performance has proven to be very efficient and without the negative impact on the clock frequencies and power consumption like commonly criticized during Intel's early AVX-512 days. Plus on the server side there is a much more diverse spectrum of workloads able to benefit from AVX-512.
For today's benchmarking article is a look at the AMD EPYC 9654 2P performance both with and without AVX-512 usage for seeing the impact on raw performance, performance-per-Watt, overall CPU power consumption as monitored via the Linux RAPL interfaces, the CPU0 core temperature, and the peak frequency being achieved each second of testing out of any of the combined 192 CPU cores.
This AMD EPYC 9654 AVX-512 testing was carried out on Ubuntu 22.10 with GCC 12.2 for the latest upstream Linux at the performance while riding the newly-released Linux 6.1 kernel.