The Interesting Tale Of AMD's FirePro Drivers
Written by Michael Larabel in Display Drivers on 30 December 2010. Page 7 of 8. 48 Comments

With OpenBenchmarking.org we are also able to further analyze the Phoronix Test Suite results such as looking at the aggregate sum, harmonic mean, and geometric mean across all of the tests as a whole. When looking at the aggregate sum of all the Linux OpenGL tests used, we see a drop between Catalyst 9.2 and Catalyst 9.3 that was caused by X-Plane 9, SPECViewPerf 9 EnSight, Maya, and other workstation tests. In 2010 there was then AMD's much-publicized workstation driver optimizations, but as expressed through this entire article, they were only temporary improvements. Only in a few of the SPECViewPerf 9/10 benchmarks have these improvements been sustained since the FirePro driver optimizations from this summer and in those cases they are largely restoring the performance to where it was at in early 2009 when this hardware was first released.

Looking at the geometric mean overview again clearly illustrates the Catalyst 10.12 driver as not being much faster than it was in Catalyst 9.2 and well below where it was earlier in the year when AMD announced it made some great strides in optimizing its FirePro driver.


Related Articles
Trending Linux News