NVIDIA / TransformerEngine

A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs, to provide better performance with lower memory utilization in both training and inference.

https://docs.nvidia.com/deeplearning/transformer-engine/user-guide/index.html

Apache License 2.0

1.85k stars 309 forks source link

Update minimum CMake version #1037

Closed timmoon10 closed 2 months ago

timmoon10 commented 2 months ago

Description

We have had users run into build errors using CMake 3.20 (see https://github.com/NVIDIA/TransformerEngine/issues/803#issuecomment-2128275904), so this PR bumps the minimum version to 3.21 and uses it in the core-only build test. I may bump the version again if the build test fails.

I've also removed logic that links NVTX. https://github.com/NVIDIA/TransformerEngine/pull/1025 upgraded to NVTX 3, which is header-only.

Closes https://github.com/NVIDIA/TransformerEngine/issues/803. Also see https://github.com/NVIDIA/TransformerEngine/pull/943.

Type of change

[ ] Documentation change (change only to the documentation, either a fix or a new content)
[ ] Bug fix (non-breaking change which fixes an issue)
[ ] New feature (non-breaking change which adds functionality)
[ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
[x] Infra/Build change
[ ] Code refractor

Changes

Increase minimum CMake version
No longer link to NVTX

Checklist:

[x] I have read and followed the contributing guidelines
[x] The functionality is complete
[x] I have commented my code, particularly in hard-to-understand areas
[x] I have made corresponding changes to the documentation
[ ] My changes generate no new warnings
[x] I have added tests that prove my fix is effective or that my feature works
[ ] New and existing unit tests pass locally with my changes

ksivaman commented 2 months ago

/te-ci