The OpenGL Speed & Perf-Per-Watt From The Radeon HD 2000/3000 Series Through The R9 Fury

Written by Michael Larabel in Graphics Cards on 14 January 2016 at 11:00 AM EST. Page 1 of 10. 19 Comments.

What's the best way to beat the winter blues? Benchmarking, of course! For starting off our 2016 of graphics card benchmarking under Linux, I've been working on a large round-up of re-testing AMD Radeon graphics cards from the HD 2900XT (R600) graphics card through the latest R9 Fury (Fiji) graphics card while running Ubuntu and using the very latest open-source graphics driver stack. Here's an interesting look at how the OpenGL graphics performance has evolved on the AMD side over the past decade while also looking at the performance-per-Watt.

Using all of the graphics cards I had available and weren't busy in the dozens of other benchmarking systems at Phoronix, I set to carry out an OpenGL comparison -- looking at the raw frame-rates as well as the performance-per-Watt -- of every graphics card from the latest R9 200/300 series and R9 Fury all the way back to the R600 series from ~2006. I didn't test any hardware prior to the HD 2000 (R600) series for being too slow and it was with the R600 GPUs that support OpenGL 3.3 at least. However, not all of the graphics cards I tried to run were in a workable state on the current Radeon Linux driver due to various bugs as I'll go over shortly.

For this large comparison going from the HD 2000/3000 series through the R9 Fury, I used the very latest open-source AMD Linux graphics driver. Using Catalyst wasn't an option due to AMD having dropped older generations of hardware support from their proprietary driver. With the Catalyst legacy drivers not being maintained for modern Linux kernel / xorg-server releases, to use the same driver and test from the Fiji to R600 meant using the open-source driver stack. Anyways, besides the legacy limitation, using the open-source driver makes the most sense since that's the future direction of Radeon GPUs on Linux.

Due to going back and using older Radeon graphics processors, it meant having to largely focus on OpenGL 2 and OpenGL 3 era games, while on the newer graphics cards where supported I also ran some OpenGL 4 titles. In total the games/benchmarks used for this large performance comparison were BioShock Infinite, Counter-Strike: Global Offensive, Furmark, Metro Last Light Redux, OpenArena, Tesseract, Team Fortress 2, Unigine Valley, and Xonotic.

Before getting too far into this article, I'd just like to remind you that if you find these sort of large comparisons interesting and wish to see more of them in the future, please disable ad-block on this site or get a low-cost Phoronix Premium subscription to enjoy this site on our ad-free special service. It's only via ads and subscriptions that this work can continue, especially when many of the graphics cards used for this comparison were previously purchased by Phoronix Media as opposed to being free review samples with AMD having not sent over any in a few years. Thanks for your support and any PayPal tips or Bitcoin contributions are also much appreciated in ensuring many more interesting articles to come. Phoronix Premium also lets you view entire articles -- like this long 10 page article -- all on a single-page.


Related Articles