[ROCm] Enable float8 utils and base tests on ROCm#3942
[ROCm] Enable float8 utils and base tests on ROCm#3942brucechanglongxu wants to merge 1 commit intopytorch:mainfrom
Conversation
- Remove @skip_if_rocm from test_round_scale_down_to_power_of_2 in test_float8_utils.py and test_linear_from_recipe in test_base.py - Fix expected value for "largest normal number" test case: _round_scale_down_to_power_of_2(2^127 * (2 - 2^-23)) returns 2^127 on ROCm vs inf on CUDA. This is because log2(FLT_MAX) rounds to exactly 127.0 on ROCm (giving 2^127) while CUDA rounds up slightly (giving 2^128 = inf). Both are valid behaviors for this edge case value (3.4028235e+38). Validated on MI300X (gfx942), PyTorch 2.9.0a0, ROCm 7.1.
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/3942
Note: Links to docs will display an error until the docs builds have been completed. This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
Thank you for your pull request and welcome to our community. Action RequiredIn order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks! |
meta-cla
Bot
added
the
CLA Signed
|
Gentle ping — approved and mergeable. Enables float8 utils and base tests on ROCm. |
|
@vkuzo Gentle ping — approved by BowenBao and mergeable. Enables float8 utils and base tests on ROCm. Would appreciate a merge when you get a chance. |
2 participants
Validated on MI300X (gfx942), PyTorch 2.9.0a0, ROCm 7.1.
cc: @BowenBao