The Performance Impact Of Genoa-X's 3D V-Cache With The AMD EPYC 9684X

Written by Michael Larabel in Processors on 24 July 2023 at 01:30 PM EDT. Page 3 of 5. 30 Comments.
NAS Parallel Benchmarks benchmark with settings of Test / Class: BT.C. Default was the fastest.
NAS Parallel Benchmarks benchmark with settings of Test / Class: CG.C. Default was the fastest.
NAS Parallel Benchmarks benchmark with settings of Test / Class: IS.D. Default was the fastest.
NAS Parallel Benchmarks benchmark with settings of Test / Class: LU.C. Default was the fastest.
NAS Parallel Benchmarks benchmark with settings of Test / Class: SP.C. Default was the fastest.

The NAS Parallel Benchmarks also showed very nice improvements directly associated with the large L3 cache.

OpenVINO benchmark with settings of Model: Vehicle Detection FP16, Device: CPU. Default was the fastest.
OpenVINO benchmark with settings of Model: Vehicle Detection FP16, Device: CPU. Default was the fastest.
OpenVINO benchmark with settings of Model: Person Vehicle Bike Detection FP16, Device: CPU. Default was the fastest.
OpenVINO benchmark with settings of Model: Person Vehicle Bike Detection FP16, Device: CPU. Default was the fastest.
OpenVINO

And the very significant difference observed from 3D V-Cache with the OpenVINO AI toolkit.

WRF benchmark with settings of Input: conus 2.5km. Default was the fastest.
WRF benchmark with settings of Input: conus 2.5km. Default was the fastest.

The WRF weather forecasting workload enjoyed nice time-savings from 3D V-Cache that if frequently dealing with large models would make Genoa-X definitely worthwhile.


Related Articles