Blender 2.80 Performance With Intel Xeon Platinum 8280 vs. AMD EPYC 7742

The Blender 2.80 release arrived at the end of July that unfortunately was too late for using that big new release in our launch-day testing of AMD's EPYC 7002 "Rome" processors but as a follow-up here are AMD EPYC 7742 performance benchmarks up against the Intel Xeon Platinum 8280 Cascade Lake as well as the AMD EPYC 7601 2P. Blender 2.80 performance is the focus of this article along with some other renderer benchmarks.

Blender 2.80 is now available from the Phoronix Test Suite and OpenBenchmarking.org for facilitating tests from this major update to Blender. For those not familiar with the changes brought by Blender 2.80, see the write-up at Blender.org.

The three servers used for testing (AMD Naples, AMD Rome, Intel Cascade Lake) were using Intel Optane 900p NVMe SSDs for storage and RAM satisfying all available channels and at each processor's optimal frequencies. Ubuntu 19.04 was running on each server with the Linux 5.2 kernel. The Phoronix Test Suite test profile makes use of Blender 2.80's official binaries and some of the common benchmark scenes, similar to the long-standing Blender 2.79 test.

In addition to the Blender 2.80 results are also some numbers for Tungsten, Appleseed, and POV-Ray renderers. Unfortunately no results for the proprietary V-RAY and IndigoBench renderers since they were crashing with 256 threads due to shortcomings in those applications.