NVIDIA 24-Way GPU Comparison With Many OpenCL, CUDA Workloads
As part of re-testing all hardware prior to major GPU/driver launches, here is a look at the latest NVIDIA OpenCL/CUDA performance on Linux -- complementing the recent Blender 2.90 benchmarks and the latest NVIDIA vs. AMD Linux gaming performance. In still waiting to find out when we will get any NVIDIA Ampere hardware for Linux testing, I have been having some benchmarking fun and extended this to a 24-way graphics card comparison back to Maxwell in looking at not only the raw GPU compute performance but also the performance-per-Watt / power consumption and GPU thermal values.
From the NVIDIA GeForce GTX 950 through GeForce RTX 2080 Ti and TITAN RTX, twenty-four Maxwell / Pascal / Turing graphics cards were tested for this fresh comparison in getting the current numbers when testing using Ubuntu 20.04 LTS with the latest driver (NVIDIA 450.66). But in still waiting to find out if/when there will be any Ampere hardware for Linux testing, I went back further than usual in some of the graphics cards for testing given the extra time in providing this reference article today. Sadly another NVIDIA launch where Linux doesn't appear to have much (or any?) emphasis.
The graphics cards (all at their default speeds) for testing based on availability this round of OpenCL and CUDA testing on Ubuntu 20.04 + NVIDIA 450.66 included:
- GTX 950
- GTX 960
- GTX 970
- GTX 980
- GTX 980 Ti
- GTX TITAN X
- GTX 1050 Ti
- GTX 1060
- GTX 1070
- GTX 1070 Ti
- GTX 1080
- GTX 1650
- GTX 1650 SUPER
- GTX 1660
- GTX 1660 SUPER
- GTX 1660 Ti
- RTX 2060
- RTX 2060 SUPER
- RTX 2070
- RTX 2070 SUPER
- RTX 2080
- RTX 2080 SUPER
- RTX 2080 Ti
- TITAN RTX
Via the Phoronix Test Suite a wide range of OpenCL and CUDA benchmarks were carried out, complementing the recent Linux gaming and CUDA/OptiX Blender 2.90 testing.
The Phoronix Test Suite was monitoring the GPU power consumption on a per-test basis for generating performance-per-Watt metrics as well as collecting the GPU thermal data.
As this is basically updated data and nothing particularly new besides being fresh metrics on the latest Ubuntu releases and newest public driver, those interested in going through the large mound of data can head on over to the OpenBenchmarking.org result file. From the OpenBenchmarking.org result file you can also easily generate performance-per-dollar results as well based upon your local used/new GPU pricing. In this article are just a few highlights while on OpenBenchmarking.org is all of the data in full. Or see how your own system(s) compare by running phoronix-test-suite benchmark 2009116-FI-GPUCOMPUT20 against this data.
Updated NVIDIA vs. AMD OpenCL figures are currently running.