I am transforming a ~7month old wip branch into something upstream-able. Lot more GLES3 tests are passing with these commits. But I need to spend some more hours to test 2-3 different GPU models, before I will create a MR.

#mesa3d #etnaviv #mntreform #librem5

dEQP-GLES31.functional.image_load_store.* 699/747 pass

With the help of some freshly reverse-engineered ISA opcodes :)

#etnaviv #mesa3d #rust #mntreform #librem5

dEQP-GLES3, old etnaviv compiler vs new one:

+6140 passes, −3004 crashes. 🚀

#etnaviv #mesa3d #rust #mntreform #librem5

You've probably seen kmscube before - but every screenshot of it is a milestone someone reached. Last night I hit one of those moments, after many years (again).

For the past months I shifted focus away from GLES3 to some foundation work that had to be done.

On the photo: kmscube running on the long-in-development new compiler backend, used for both the vertex and fragment shader.

#etnaviv #mesa3d #rust #mntreform #librem5

When your newly written NIR pass causes an infinite optimization pass loop, you know it’s time to go to bed.

#mesa3d #etnaviv

subnormal IEEE 754 pain

#etnaviv #mesa3d

Yeah… cleaning up a long-lived git branch with more than 200 commits will take some time. There are reverts, hacks, and all sorts of fun issues.

Pro tip: don’t keep your feature branches alive for too long - in my case, 2 years. Catching up with upstream is painful.

#mesa3d #etnaviv