Running OpenCL On The CPU With POCL 1.0, Xeon & EPYC Testing
For those wondering how the dual Xeon Gold and EPYC POCL OpenCL performance compares to GPUs running OpenCL...
No real surprises all things considered...
In ViennaCL though, the EPYC 7601 and Xeon Gold Tyan servers did come out slightly ahead of the Radeon RX 580 with this test not scaling as well as other OpenCL kernels.
For those wondering if using Blender's OpenCL code-path with POCL on the CPU is faster than just using Blender's CPU renderer, unfortunately it's much slower:
The blender data in this result file. There are also i7-8700K Blender OpenCL/CPU results here.
Still working on some additional POCL OpenCL tests out of curiosity, stay tuned. Particularly around comparing CPU-based renderers versus OpenCL options and also OpenCL vs. OpenMP / MPI where such comparisons can be made.
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.