64K Kernel Page Size Performance Benefits For HPC Shown With NVIDIA's GH200 Grace CPU

Written by Michael Larabel in Software on 27 February 2024 at 12:00 PM EST. Page 3 of 5. 9 Comments.
GraphicsMagick benchmark with settings of Operation: Sharpen. GPTshop.ai GH200 + Linux 6.8 64k was the fastest.
GraphicsMagick benchmark with settings of Operation: Enhanced. GPTshop.ai GH200 + Linux 6.8 64k was the fastest.

In some workloads the 64K kernel use was enough to bump the GPTshop.ai GH200 positioning to the front of the race against the tested x86_64 server processors.

ACES DGEMM benchmark with settings of Sustained Floating-Point Rate. EPYC 9754 2P was the fastest.

ACES DGEMM was another nice HPC workload showing off the big uplift possible if running a 64K page size kernel on AArch64 hardware.

7-Zip Compression benchmark with settings of Test: Compression Rating. EPYC 9754 2P was the fastest.

Going from Linux 6.5 to 6.8 alone wasn't much of a difference but the 64K kernel page size continues to prove very beneficial for large ARM servers/HPC.

Timed Godot Game Engine Compilation benchmark with settings of Time To Compile. EPYC 9684X 2P was the fastest.
Timed LLVM Compilation benchmark with settings of Build System: Ninja. EPYC 9554 2P was the fastest.
Timed Node.js Compilation benchmark with settings of Time To Compile. EPYC 9684X 2P was the fastest.

Code compilation workloads also improved as well with the 64K kernel page size.


Related Articles
Intel Core Ultra "Meteor Lake" Performance Improves With Linux 6.9
GCC 14 vs. LLVM Clang 18 Compiler Performance On Fedora 40
Intel Xeon Max Sees Some Performance Gains For OpenVINO & ONNX With Linux 6.9
Linux 6.9 Drives AMD 4th Gen EPYC Performance Even Higher For Some Workloads
Intel Core Ultra "Meteor Lake" Yields Faster Performance With Linux 6.9
Google Cloud's C3D Instances Provide Strong Performance Value For PingCAP's TiDB