Chris Wilson (of Intel) decided to check out the GLAMOR performance for the open-source Radeon driver under Cairo, his pet project. As shared a few days ago, AMD will be using GLAMOR as the means of 2D acceleration for the Radeon HD 7000 series. While HD 7000 series is the current target, this GLAMOR acceleration code should work on previous generations of Radeon hardware since GLAMOR is basically piping the 2D code over OpenGL.
Right now for the Radeon HD 7000 series, the default means of 2D acceleration is still using "ShadowFB" since the RadeonSI Gallium3D driver isn't even into shape yet for handling GLAMOR. ShadowFB is just using a shadow frame-buffer with the CPU.
Up to this point, GLAMOR was just being used with Intel's hardware/driver. On the Intel side, however, for Sandy Bridge and Ivy Bridge at least I've found the other experimental 2D acceleration option -- SNA -- to easily beat GLAMOR in terms of stability, reliability, and performance.
Anyhow, Chris Wilson's results for GLAMOR Radeon are quite interesting and can be found on his blog. Below is the main takeaway:
Perhaps the first thing that we notice is the raw power of the discrete graphics as exposed by using OpenGL directly from within cairo. Secondly, we notice the lack luster performance of the existing EXA driver for the Radeon chipset – remember everything below the lines implies that the GPU driver in Xorg is behaving worse than could be achieved just through client-side sowftware rendering, that using RENDER acceleration is nothing of the sort. And then our attention turns to the newcomer, glamor on radeon. It is still notably much slower than both the CPU and using OpenGL directly. However, it is of very similar performance to the existing EXA driver, sometimes slower, sometimes faster (if you look at the relative x11perf, then it reveals some areas where the EXA driver could do major improvements).Phoronix tests of GLAMOR Radeon acceleration will be conducted once I'm back to America from Nicaragua.