AMD EPYC Genoa/Genoa-X & Bergamo vs. Intel Xeon Sapphire Rapids On Ubuntu 23.10

Written by Michael Larabel in Processors on 22 November 2023 at 02:34 PM EST. Page 6 of 7. 4 Comments.

oneDNN benchmark with settings of Harness: Deconvolution Batch shapes_1d, Data Type: bf16bf16bf16, Engine: CPU. Xeon Platinum 8490H was the fastest.

oneDNN benchmark with settings of Harness: Recurrent Neural Network Inference, Data Type: bf16bf16bf16, Engine: CPU. Xeon Platinum 8490H was the fastest.

TensorFlow benchmark with settings of Device: CPU, Batch Size: 16, Model: ResNet-50. Xeon Platinum 8490H was the fastest.

Where the Intel Sapphire Rapids Xeon Platinum 8490H proved most competitive remains with the AI workloads where the software is ready to capitalize upon the Advanced Matrix Extensions (AMX) support found with 4th Gen Xeon Scalable.

Though for TensorFlow with ResNet-50, the EPYC 9554 delivered better performance-per-Watt than the Xeon Platinum 8490H and the other Bergamo and Genoa(X) parts were only slightly behind the SPR CPU with power efficiency here.

TensorFlow benchmark with settings of Device: CPU, Batch Size: 256, Model: ResNet-50. EPYC 9554 2P was the fastest.

With larger batch sizes, the Intel SPR advantage eroded.

OpenVINO benchmark with settings of Model: Face Detection FP16, Device: CPU. Xeon Platinum 8490H 2P was the fastest.

OpenVINO benchmark with settings of Model: Person Detection FP16, Device: CPU. EPYC 9684X 2P - Power was the fastest.

OpenVINO benchmark with settings of Model: Face Detection FP16-INT8, Device: CPU. Xeon Platinum 8490H 2P was the fastest.

OpenVINO benchmark with settings of Model: Face Detection FP16-INT8, Device: CPU. EPYC 9554 2P was the fastest.

OpenVINO benchmark with settings of Model: Face Detection Retail FP16, Device: CPU. EPYC 9754 2P - Power was the fastest.

OpenVINO benchmark with settings of Model: Face Detection Retail FP16, Device: CPU. EPYC 9554 2P was the fastest.

OpenVINO benchmark with settings of Model: Weld Porosity Detection FP16, Device: CPU. Xeon Platinum 8490H 2P was the fastest.

OpenVINO benchmark with settings of Model: Road Segmentation ADAS FP16-INT8, Device: CPU. EPYC 9754 2P - Power was the fastest.

OpenVINO benchmark with settings of Model: Road Segmentation ADAS FP16-INT8, Device: CPU. EPYC 9554 2P was the fastest.

The Intel Sapphire Rapids processors tend to perform excellent with the OpenVINO AI toolkit, again thanks to AMX. But in some of the OpenVINO tests, the AMD EPYC 9684X Genoa-X processor can still outperform the Xeon Platinum 8490H. It will be very interesting to see how much generational uplift comes with Emerald Rapids.

ClickHouse benchmark with settings of 100M Rows Hits Dataset, First Run / Cold Cache. EPYC 9554 2P was the fastest.

ASKAP benchmark with settings of Test: tConvolve MPI, Degridding. EPYC 9684X 2P - Power was the fastest.

ASKAP benchmark with settings of Test: tConvolve MPI, Gridding. EPYC 9684X 2P - Power was the fastest.

Graph500 benchmark with settings of Scale: 26. EPYC 9754 2P was the fastest.

GROMACS benchmark with settings of Implementation: MPI CPU, Input: water_GMX50_bare. EPYC 9754 2P - Power was the fastest.

Blender benchmark with settings of Blend File: Barbershop, Compute: CPU-Only. EPYC 9754 2P - Power was the fastest.

4 Comments - Next Page

Page: 1 2 3 4 5 6 7 Next Page