Revert hip cooperative groups workaround

ROCm / DeepSpeed

DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective.

https://www.deepspeed.ai/

Apache License 2.0

4 stars 3 forks source link

Closed rraminen closed 3 years ago

rraminen commented 3 years ago

-This PR contains the following:

The files normalize_kernels.cu, softmax_kernels.cu and general_kernels.cu now use HIP cooperative groups APIs.

The condition " __CUDA_ARCH >= 700 || defined(HIP_PLATFORM_HCC__)" to enable the code on CUDA and ROCm systems

CC: @jithunnair-amd