Block a user
Enable Distribute memory across device Denoise Optix Error
Fix #119959: Enabling "Distribute memory between devices" for Cycles results in error
OPTIX_ERROR_INTERNAL_ERROR when rendering using Cycles with OptiX
OptiX failed to load a required library to function. Possibly because you have a leftover nvoptix.dll or nvrtum64.dll in C:\Windows\system32 that should not be there for DCH drivers (they have the…
OPTIX_ERROR_INTERNAL_ERROR when rendering using Cycles with OptiX
Please run blender_debug_gpu.cmd
, reproduce and attach the generated blender_debug_output.txt
.
Fix #119959: Enabling "Distribute memory between devices" for Cycles results in error
WIP: Rework CUDA kernel build to use separable compilation
8945b29762
EEVEE: Overscan/Border mixed resolution rendering
ebf12765a6
Merge branch 'blender-v4.1-release'
7a395e2e7f
Revert changes from main commits that were merged into blender-v4.1-release
516912f741
Merge branch 'blender-v4.1-release'
4842bf1101
Cleanup: make format
Patrick Mours
created branch cuda_separable_compilation in pmoursnv/blender
2024-03-21 10:54:02 +01:00
0ca1c55c70
Rework CUDA kernel build to use separable compilation
8945b29762
EEVEE: Overscan/Border mixed resolution rendering
ebf12765a6
Merge branch 'blender-v4.1-release'
7a395e2e7f
Revert changes from main commits that were merged into blender-v4.1-release
516912f741
Merge branch 'blender-v4.1-release'
Cycles: Use default CUDA context instead of creating a new one
But shouldn't that be handled by the second condition in if (result != CUDA_SUCCESS && result != CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE)
?
Cycles: Use default CUDA context instead of creating a new one
cuDevicePrimaryCtxRetain
and cuDevicePrimaryCtxGetState
are thread-safe, per device (the driver has a mutex per device which is held during the call). So should still be safe even if two…
The render crashes with the OptiX denoiser option
OptiX Denoising does not work on multiple frames