Intel Linux Kernel Optimizations Show Huge Benefit For High Core Count Servers

Written by Michael Larabel in Software on 29 March 2023 at 01:00 PM EDT. Page 3 of 5. 19 Comments.

OpenVKL benchmark with settings of Benchmark: vklBenchmark ISPC. Clear Linux: 240 Threads was the fastest.

With Intel's open-source OpenVKL project that is part of oneAPI, the performance of the two Linux distributions was similar until hitting 60+ threads at which point the Clear Linux kernel really shows off its advantages. When the Xeon Platinum 8490H dual socket server was running without limitations, it meant 25% faster performance compared to the stock Ubuntu 23.04.

C-Blosc benchmark with settings of Test: blosclz shuffle. Clear Linux: 60 Threads was the fastest.

C-Blosc benchmark with settings of Test: blosclz bitshuffle. Clear Linux: 60 Threads was the fastest.

The C-Blosc data store shows some performance issues when above 60 threads across the two sockets but show Clear Linux's optimizations overall becoming very significant at the higher core counts. Ubuntu 23.04 at 120/240 threads shows C-blosc performing worse than at 4 threads, which was not the case when using Clear Linux with Intel's latest kernel optimizations.

Blender benchmark with settings of Blend File: BMW27, Compute: CPU-Only. Clear Linux: 240 Threads was the fastest.

GROMACS benchmark with settings of Implementation: MPI CPU, Input: water_GMX50_bare. Clear Linux: 240 Threads was the fastest.

It's important to keep in mind that Intel's optimizations don't magically apply universally to all workloads out there.

19 Comments - Next Page

Page: 1 2 3 4 5 Next Page