Timur Kristóf

OOOPS, looks like I forgot to revert the other patch before applying this one. Will do another test with just the si_pipe patch.

Rebuilt mesa with the correct patch and updated my last comment.

While the `si_pipe` patch does help a lot, it's not perfect. Even though Gallium says that the GPU utilization is 99%, `radeontop` disagrees, and says the utilization is ~50% (up...

I built the latest master (as of today) from the ixit repo and ran SC2 with the mesa built from there. The results are definitely an improvement, but not as...

Investigated a bit further. I'm not 100% sure that the bottleneck is VS constants. I made the following changes: * `si_pipe.c` in `si_create_context` added the `SI_RESOURCE_FLAG_32BIT` to `stream_uploader` which allows...

Thanks for the reply. I understand that, but I think an exact match should be higher priority than any partial match.

Hi, > So the first thing to ask is whether there is a commitment to implement this on the part of someone actually writing Mesa drivers. Yes, there is interest...

> >WG has some review comments pending. Also will wait to ship until nvidium is at least semi-working with this (pending) to ensure things are usable as expected. > >...

> Are there any guarantees about how/where the mesh shader workgroups are launched? @MCRcortex Not sure if this thread is the right one to discuss GPU specific implementation details, but...

With regards to rasterization order: it seems that both AMD and NVidia do guarantee the "strict" rasterization order in their currently released GPUs. I haven't got any info about other...