Skip to content
This repository has been archived by the owner on Oct 31, 2024. It is now read-only.

[TASK] Understand the LLVM/AMD codenames for GPU architectures #1

Open
dipietrantonio opened this issue Sep 28, 2022 · 0 comments
Open

Comments

@dipietrantonio
Copy link
Collaborator

Many libraries allow optimizations based on GPU architecture. You can specify what architecture you want to compile the library for, by using the CMake flag AMDGPU_TARGETS=...
For the MI100, the code gfx908 seems valid (i.e. no errors thrown), but on AMD Readmes, gfx908:xnack- is found often.
We need to understand exactly what to ask for. The answer lies here: https://llvm.org/docs/AMDGPUUsage.html

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant