There were huge SNA performance gains on Ironlake over UXA in the most recent testing that happened last night. Curious to see how the SNA 2D acceleration architecture is working for Sandy Bridge graphics hardware, for which it was originally intended, here are some new benchmarks.
These benchmarks aren't the extensive SNA vs. UXA benchmarks that span several driver versions as talked about in the Ironlake/Arrandale article, but rather some straightforward 2D benchmarks like what happened on the Core i3 330M system. This time around though the testing commenced on an Intel Core i5 2520M "Sandy Bridge" mobile notebook running Ubuntu 13.04 with the Linux 3.8 kernel.
The Git code of the xf86-video-intel DDX as of this morning was utilized for this quick benchmarking roundabout. Details in full -- along with all of the Linux 2D test results and the software/hardware details -- can be found via OpenBenchmarking.org within 1302271-FO-INTELSAND34.
SNA is generally a hell of a lot faster than the default UXA acceleration, only in select scenarios (mostly GTK micro-benchmarks) is it "disappointing" when the performance is only comparable to that of UXA.
UXA is still the default means of 2D acceleration for the xf86-video-intel driver, but at least with Ubuntu 13.04 the driver packager is defaulting to SNA and other distribution vendors are moving there too.