Skip to content

Ifu 2023 01 03 #33

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 0 commits into from
Feb 8, 2023
Merged

Ifu 2023 01 03 #33

merged 0 commits into from
Feb 8, 2023

Conversation

liligwu
Copy link
Collaborator

@liligwu liligwu commented Jan 3, 2023

Disabled bf16 test for ROCm since the counterparts of __nv_bfloat16 and __nv_bfloat162 are not supported.

@liligwu liligwu requested a review from amathews-amd January 3, 2023 23:33
@amathews-amd
Copy link

amathews-amd commented Jan 4, 2023

Should we disable BF16 support (pytorch#1498, pytorch#1503) ?

hip_bfloat16 should be replacement of nv_bfloat16.
nv_bfloat162 seems it can implemented as array of 2 hip_bfloat16 for most cases (may need to be 4-byte aligned), and there is an override for conversion for older CUDA versions (https://github.com/pytorch/FBGEMM/pull/1498/files#diff-9e1398cad4b84bd53e53898d8d31f96e5821ade7e658e5914970639d9e497a4aR1476).

cc: @dllehr-amd

@liligwu liligwu merged commit 4216052 into main Feb 8, 2023
abojarov pushed a commit to abojarov/FBGEMM that referenced this pull request Oct 15, 2024
…_test

EmbeddingBag asm test.
Regression and new tests for asm forward kernel passed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants