NVIDIA GH200 Grace CPU vs. AMD EPYC 9005 Turin CPU Performance

Written by Michael Larabel in Processors on 7 November 2024 at 12:37 PM EST. Page 4 of 5. 20 Comments.

👁 Algebraic Multi-Grid Benchmark benchmark with settings of Performance Per Thread. NVIDIA GH200 Grace CPU was the fastest.

👁 miniBUDE benchmark with settings of Performance Per Thread, Implementation: OpenMP, Input Deck: BM2. EPYC 9575F was the fastest.

Some HPC benchmarks aren't as well optimized for AArch64 as they are for x86_64.

👁 LAMMPS Molecular Dynamics Simulator benchmark with settings of Performance Per Thread, Model: Rhodopsin Protein. NVIDIA GH200 Grace CPU was the fastest.

👁 LAMMPS Molecular Dynamics Simulator benchmark with settings of Performance Per Thread, Model: 20k Atoms. NVIDIA GH200 Grace CPU was the fastest.

With the LAMMPS molecular dynamics software the NVIDIA Grace CPU was very competitive to the EPYC Turin CPUs in raw performance and in turn leading by a significant margin performance-per-Watt.

👁 GROMACS benchmark with settings of Performance Per Thread, Implementation: MPI CPU, Input: water_GMX50_bare. EPYC 9575F was the fastest.

Across other CPU-based HPC workloads the EPYC 9005 series was delivering the best performance.

👁 QuantLib benchmark with settings of Performance Per Thread, Configuration: Multi-Threaded. NVIDIA GH200 Grace CPU was the fastest.

With the QuantLib quantitative finance software, the Grace CPU narrowly fell behind the EPYC 9575F but did have a narrow lead in providing the best performance-per-Watt of the tested parts.

20 Comments - Next Page

Page: 1 2 3 4 5 Next Page

URL: https://www.phoronix.com/review/nvidia-grace-epyc-turin/4

⇱ NVIDIA GH200 Grace CPU vs. AMD EPYC 9005 Turin CPU Performance Review - Phoronix

NVIDIA GH200 Grace CPU vs. AMD EPYC 9005 Turin CPU Performance