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

1.6k stars 255 forks source link

Update required CMake version to 3.25 #943

Open timmoon10 opened 2 weeks ago

timmoon10 commented 2 weeks ago

Description

CMake 3.25 changes NVTX support by deprecating the CUDA::nvToolsExt target and adding a CUDA::nvtx3 target. I suspect this is related to the build error in https://github.com/NVIDIA/TransformerEngine/issues/879.

We also had another user run into strange build errors with CMake 3.20.4, which they fixed by upgrading to 3.24.3 (see https://github.com/NVIDIA/TransformerEngine/issues/803#issuecomment-2128275904). This makes me think our current requirement of 3.18 is inaccurate. CMake 3.25 was released in November 2022, so I don't think it's an unreasonable minimum.

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

Please list the changes introduced in this PR:

Update required CMake version to 3.25

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
[ ] 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

timmoon10 commented 2 weeks ago

/te-ci

timmoon10 commented 2 weeks ago

/te-ci