Hi all, Just thought I'd bump this with a little more info.
- I've tested this with the new i965 vec4 nir backend without regressions. - I've got subroutine and varying AoA working I just need to do some tessellation shader tests and then desktop support will be ready**. I'll send a follow-up patchset for this no changes are needed to the current series. For anyone interested these as the remaining unreviewed patches: 7, 8, 9, 10, 13, 14 (tested not reviewed), 15, 17, 18, 19 **interface support will still be missing but niether Nvidia or AMD ship it either, I have some work in progress patches for this. On Wed, 2015-07-29 at 23:56 +1000, Timothy Arceri wrote: > If useful the series is in the 'gles8' branch of the repo here: > https://github.com/tarceri/Mesa_arrays_of_arrays.git > > Changes in v3: > - Rebased on the tessellation and subroutine changes > - Atomic counter AoA now working including indirect indexing (previously I > had broken piglit tests that were always passing as long as the shader > compiled) > - Included fix for atomic counters with binding other than 0 > - Split the program interface query chages for easier review > > Changes in v2: > - Reworked how uniform AoA are handled which simplified things greatly. > - Transform feedback and program interface query now working with AoA. > - Speed up glsl optimisation pass via improvement to dead code elimination. > - Some smaller reviewed patches pushed. > > _______________________________________________ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/mesa-dev _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
