Logan Adams

When just returning true like the code above in the `assert_no_cuda_mismatch` you have to specify `DS_BUILD_SPARSE_ATTN=0`. So that command becomes `DS_BUILD_OPS=1 DS_BUILD_SPARSE_ATTN=0 ./install.sh`. However, that still results in the following...

ROCm half type should be fixed in #3236. The fix on building the ops without finding cuda home is ongoing.

@adammoody - since I think we have all DeepSpeed related issues solved, I'm going to close this for now. If you hit other issues, please re-open.

Hi @JY-Ren - I believe the suggestion from @richarddwang will solve this for you. We always recommend a hostfile/using that if possible. Please let us know and re-open the issue...

This seems like the best answer so far, and the issue is fairly stale. Closing for now, if folks have other suggestions, please post here, if you have other questions,...

@mallorbc - Do things work fine with JIT compilation? You're just not able to build the ops? Especially since you can build the ops on 0.8.0 but not 0.8.1 or...

Makes sense, thanks. I'd probably recommend opening a new issue for the int8 dtype errors, so the right folks can look at that. But for the build issues, that's odd...

@mallorbc - correct, I was curious if there was one op that was the problem but seems to be a host of cuda type issues, which is why I was...

Interesting it also occurs with Cuda 11.6 and 11.8. Linking [this issue](https://github.com/microsoft/DeepSpeed/issues/2858) since it appears to be the same thing.

@mallorbc - So its likely [this is related, and this fix works for that user](https://github.com/microsoft/DeepSpeed/pull/3085) - we're compiling in a different environment than we are using, or at least can...