AMD EPYC 8324P / 8324PN Siena 32-Core Siena Linux Server Performance

Written by Michael Larabel in Processors on 10 October 2023 at 08:00 PM EDT. Page 8 of 9. 5 Comments.
TensorFlow benchmark with settings of Device: CPU, Batch Size: 16, Model: ResNet-50. Xeon Gold 6421N was the fastest.

The Intel Xeon Gold 6421N was faster in raw performance for TensorFlow, but wasn't that surprising when considering the higher turbo frequency and Advanced Matrix Extensions (AMX) with Sapphire Rapids that is able to be leveraged by TensorFlow.

TensorFlow benchmark with settings of Device: CPU, Batch Size: 32, Model: ResNet-50. Xeon Gold 6421N was the fastest.

For those considering EPYC Siena for AI at the edge, it was actually a bit surprising that the Xeon Gold 6421N didn't have a wider lead in TensorFlow due to Sapphire Rapids having AMX...

TensorFlow benchmark with settings of Device: CPU, Batch Size: 64, Model: ResNet-50. Xeon Gold 6421N was the fastest.
TensorFlow benchmark with settings of Device: CPU, Batch Size: 64, Model: ResNet-50. Xeon Gold 6421N was the fastest.
TensorFlow benchmark with settings of Device: CPU, Batch Size: 64, Model: ResNet-50. Xeon Gold 6421N was the fastest.
TensorFlow benchmark with settings of Device: CPU, Batch Size: 64, Model: ResNet-50. Xeon Gold 6421N was the fastest.

The TensorFlow performance was enough that on a value basis was comparable to the tested Siena processors, but still at much higher power consumption that will ultimately influence the TCO.


Related Articles