The Xeon vs. EPYC Performance With Intel's oneAPI Embree & OSPray Render Projects
Written by Michael Larabel in Software on 30 September 2019. Page 2 of 4. 14 Comments

The OSPray testing started off like we're used to seeing out of the EPYC 7742 2P... Generally measurably better performance compared to the 8280 processors. The EPYC 7742 2P with a combined 256 threads generally means superb render performance.

There was good EPYC Rome performance with OSPray out of both the Path Tracer and SciVis renderers included in the project.

In some of the scenes, the EPYC 7742 2P server was 64% faster over the dual Platinum 8280 setup. It was also staggering to see the generational improvement compared to the EPYC 7601 2P.

With the "magnetic reconnection" demo with OSPray was the only case we tested with this open-source ray-tracing engine where the Xeon Platinum 8280 performance was ahead of the EPYC 7742. In this context it didn't seem to scale as nicely or a combination with the AVX-512 optimizations kicking in.

But aside from the magnetic reconnection demo, OSPray was running very well on the AMD EPYC Rome server.

Related Articles

Trending Linux News