Looks like a small stride error during indexbuf building for triangles. PR has been submitted, but not sure if this fix would work universally for all cases. (As it may depend on if this code is…
So this is happening on Metal, but I imagine the same issue would occur in OpenGL if use_strip_prims
was also false in OpenGL.
`source/blender/draw/intern/draw_cache_impl_curves.cc - static…
Same issue as #121020 afaik, so will keep discussion there.
Currently a little tricky to see where this is going wrong. The backend seems to be rendering what it is given, although the geometry is doubling up on certain strands, when i'm unsure if it…
As an additional note, I've now been able to consistently reproduce this on my local machine using the Tree Creature asset in EEVEE Next, instead of via the render tests. The stracktrace is…
Thanks Brecht, then I do believe we should still land this patch regardless, as it should not be changing GPU dispatch behavior on this flag.
(The flag is still situationally used to wait and…
Thanks Jeroen, will resolve!
Good to confirm shadows working on Intel GPUs too. Will be performing a test pass on Intel later today to see where we are at overall.
Although, I do feel this patch makes sense to land regardless, as it should harden the category of errors and stalls when running in debug, as the flag now being disabled was not intended to be…
Hmm unfortunately timeout seems to still occur, will continue attempting to reproduce.
Please note that I was unable to reproduce the original crash, so this may not resolve the same issue, but is a best-guess at what may have been the root cause.
So looking, the change is correctly in, but unsure why this is still hanging, investigating now, but will see if I can reproduce easily.
Looking into this now, I do recall a separate fix for this, which should have resolved it. There was a small edge-case where if an event signal was in a single command buffer with no other…
PR issued with the fix. Strokes should now render the same as OpenGL. There were a few small issues present, thickness and color parameters were being incorrectly propagated, and the datatype…
Should this be closed? or is this still looked after?
Yeah we can close this one, unless it is useful as WIP to potentially inform a future solution? Although this particular implementation…
Able to reproduce in both EEVEE and EEVEE Next, which is odd as they both use separate shadow techniques. Will investigate what is going on, imagine its possibly an issue with precision and that…
Patch above should resolve. Render pass rendering was fine, but CPU readback was not taking into account the texture view layer during sampling. Resulting in the same top-most layer being read…