Radeon ROCm 2.0 OpenCL Benchmarks With Linux 5.0 On Ubuntu 18.10 vs. NVIDIA's Linux Driver

Written by Michael Larabel in Display Drivers on 30 January 2019 at 11:26 AM EST. Page 1 of 4. 13 Comments.

With yesterday having posted fresh OpenGL/Vulkan Linux gaming benchmarks for the current NVIDIA GeForce and AMD Radeon graphics cards, in this article is the freshest OpenCL GPU compute data for that set of 14 graphics cards on the very latest Linux graphics driver stack. In the case of AMD Radeon open-source compute, it was tested using the new ROCm 2.0 atop the mainline Linux 5.0 kernel and Ubuntu 18.10.

There were some questions recently by Phoronix readers how well ROCm 2.0 is working (if at all) with the current mainline Linux kernel. Long story short, ROCm 2.0 was running well and straight-forward with the mainline kernel state, at least in going with Linux 5.0 for the bleeding-edge support. Additionally, the ROCm 2.0 stack was using the AMD-built Debian packages intended for Ubuntu 18.04 LTS but did work just fine on Ubuntu 18.10 paired with the mainline kernel. The Ubuntu repository information and more information pertaining to ROCm 2.0 can be found via their GitHub.

With ROCm 2.0 running on the Linux 5.0 kernel, the Radeon RX 580, RX 590, RX Vega 56, and RX Vega 64 graphics cards were benchmarked. On the NVIDIA side was the 415.27 binary driver running on the Linux 4.20.3 kernel. The same Intel Core i9 9900K + Ubuntu 18.10 system was used obviously throughout all testing. The current Pascal and Turing line-ups were tested on the NVIDIA side for comparison to the current AMD Polaris and Vega parts. Those green graphics cards were the GeForce GTX 1060, GTX 1070, GTX 1070 Ti, GTX 1080, GTX 1080 Ti, RTX 2060, RTX 2070, RTX 2080, RTX 2080 Ti, and TITAN RTX.

NVIDIA vs. Radeon OpenCL Linux GPU Compute

Coming up next week will be additional benchmarks as well as the performance-per-dollar, performance-per-Watt / system power consumption, etc. These fresh OpenCL benchmarks were carried out in a fully-automated and reproducible manner using the open-source Phoronix Test Suite.


Related Articles