AMD EPYC 7773X "Milan-X" Performance Continues To Impress With Tremendous Opportunity For Large-Cache Server CPUs

Written by Michael Larabel in Processors on 12 July 2022. Page 1 of 9. 14 Comments

Back in March when AMD Milan-X rolled out I published a number of EPYC 7773X benchmarks as well as Milan-X benchmarks in the cloud. Since then there have been new Linux kernel improvements and other changes in the ever-advancing open-source world. Plus simply more time to conduct additional tests over the summer. Here is the latest round of my AMD EPYC 7773X 1P and 2P benchmarking compared to the Milan EPYC 7713/7763 SKUs as well as Intel's Xeon Platinum 8380 "Ice Lake" competition.

In my continued testing of the AMD EPYC 7773X processors over the past few months, the top-end Milan-X SKU continues showcasing the tremendous potential for large-cache server CPUs moving forward. As I am continually adding new/updated benchmarks to my harness as well as since the Milan-X debut being the formal launch of Ubuntu 22.04 LTS, new versions of the Linux kernel, and also a newer version of the AMD Daytona system firmware / BIOS, I figured it was due for a summer-time update on the AMD EPYC 7773X Linux performance.

The server processors tested for this Linux benchmarking comparison included:

- AMD EPYC 7713
- AMD EPYC 7713 2P
- AMD EPYC 7763
- AMD EPYC 7763 2P
- AMD EPYC 7773X
- AMD EPYC 7773X 2P
- Intel Xeon Platinum 8380
- Intel Xeon Platinum 8380 2P

Thus a nice top-of-the-line look at the current, released wares from AMD and Intel. The Intel and AMD reference server platforms were used for testing, each configuration was running 8 x 32B DDR4-3200 memory per CPU, and using a 3.8TB Micron 9300 NVMe solid-state drive for storage. Thanks again to Intel and AMD for having provided the processor review samples for making this testing possible.

On the software side all of these processors were tested on Ubuntu 22.04 LTS with the latest stable release updates. For a fresh look at things, the Linux 5.19 Git kernel was used for monitoring the bleeding-edge performance. All of the CPUs were run with the "performance" CPU frequency governor. During the benchmarking the CPU power consumption was monitored using the exposed RAPL/PowerCap sysfs interfaces.

On each CPU over 24 hours of benchmarks were run looking at a variety of server and technical computing areas. Now let's get looking at these fresh performance numbers with the latest Linux kernel, a number of new/updated benchmarks since the March Milan-X launch, and other system software updates.


Related Articles