Intel Xeon Gold 5220R + Xeon Gold 6226R Linux Performance

Written by Michael Larabel in Processors on 10 April 2020. Page 4 of 8. 21 Comments

To little surprise, the Cascade Lake Refresh CPUs are performing well in Intel's own deep learning library DNNL (recently renamed to oneDNN as part of their oneAPI umbrella). DNNL/oneDNN is well optimized for AVX-512 among other Intel microarchitecture capabilities but even so AMD EPYC tends to hold up well in these Intel open-source software packages.

It will be interesting to see the Ice Lake Xeon performance in due course as for some workloads like the dav1d AV1 video decoder, AMD EPYC Rome is dominating Cascade Lake.

With Intel's OSPray ray-tracing engine the Xeon Gold 5220R/6226R CPUs are coming in neck-and-neck to the similarly priced AMD EPYC 7352 processors.

Related Articles