GLAMOR Acceleration Continues To Be Cleaned Up
Now that X.Org Server 1.17 RC1 has been released with a focus on improving GLAMOR and integrating the xf86-video-intel DDX, Keith Packard has written a blog post about the work that has gone on so far since GLAMOR's inception for optimizing and cleaning up this 2D-over-OpenGL acceleration method.
Keith (and Eric Anholt who has also been heavily involved with GLAMOR) have had to deal with removing Intel fallback paths, removing redundant pointers, reducing pixmap private complexity, and other changes.
For most readers, the interesting part will be GLAMOR's future work. Keith is lanning to work on issues surrounding X vs. OpenGL color formats, finishing new CompositeGlyphs code, creating pure shader-based gradients, rewriting Composite to use the GPU for more computations, and trying again at GPU-accelerated trapezoids.
More information on GLAMOR's ongoing clean-ups can be found via Keith Packard's blog at KeithP.com.
Keith (and Eric Anholt who has also been heavily involved with GLAMOR) have had to deal with removing Intel fallback paths, removing redundant pointers, reducing pixmap private complexity, and other changes.
For most readers, the interesting part will be GLAMOR's future work. Keith is lanning to work on issues surrounding X vs. OpenGL color formats, finishing new CompositeGlyphs code, creating pure shader-based gradients, rewriting Composite to use the GPU for more computations, and trying again at GPU-accelerated trapezoids.
More information on GLAMOR's ongoing clean-ups can be found via Keith Packard's blog at KeithP.com.
13 Comments