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

[PyTorch] Replace `int8_t` in Pybind11 extensions with `int64_t` #882

Closed timmoon10 closed 4 months ago

timmoon10 commented 4 months ago

Description

https://github.com/NVIDIA/TransformerEngine/pull/772 added int8_t arguments to some of the PyTorch extensions. PyTorch support for int8_t in Pybind11 extensions was added in PyTorch 2.3.0 (see https://github.com/pytorch/pytorch/pull/119639), resulting in TE build errors with older PyTorch versions (https://github.com/NVIDIA/TransformerEngine/issues/881).

This PR replaces these int8_t with int64_t. These are cast to int anyways, and it doesn't seem inconceivable to set the SM margin above 127.

Type of change

[ ] Documentation change (change only to the documentation, either a fix or a new content)
[x] 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)

Changes

Replace int8_t in PyTorch extensions with int64_t

Checklist:

[x] I have read and followed the contributing guidelines
[x] The functionality is complete
[ ] I have commented my code, particularly in hard-to-understand areas
[ ] 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 4 months ago

/te-ci pytorch