You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Using the latest release of AMDVLK included in the 22.20 driver for Ubuntu 22.04 only gives single rate performance when testing with vkpeak. AMDGPU-PRO from the same package provides support for packed FP16 support, but not for int16. RADV recently had support for both packed fp16 and int16 merged. Merge Request
Are there plans to provide support for double rate 16 bit instructions in the Open Source driver?
llpc currently runs the scalarizer pass, which helps a lot in reducing register pressure and increasing occupancy but has the side-effect of preventing packed instructions. See this issue for more details and workarounds: GPUOpen-Drivers/llpc#1369
Using the latest release of AMDVLK included in the 22.20 driver for Ubuntu 22.04 only gives single rate performance when testing with vkpeak. AMDGPU-PRO from the same package provides support for packed FP16 support, but not for int16. RADV recently had support for both packed fp16 and int16 merged. Merge Request
Are there plans to provide support for double rate 16 bit instructions in the Open Source driver?
AMDVLK:
AMDGPU-Pro:
RADV (git-f533dff 2022-07-01 jammy-oibaf-ppa):
The text was updated successfully, but these errors were encountered: