NVIDIA / cutlass

CUDA Templates for Linear Algebra Subroutines
Other
5.39k stars 907 forks source link

[BUG] std::numeric_limits<half_t>::digits #1701

Open akamiru opened 1 month ago

akamiru commented 1 month ago

Just a minor thing I noticed.

include/cutlass/half_t.h set std::numeric_limit::digits to 10 but it should be 11. The implicits 1 bit at the start is counted normally e.g. std::numeric_limits::digits == 24.

It appears to be wrong in line 565 and 628. Not sure if anything with cutlass depends on it being wrong so I'm not filing an pull request. Let me know if I should.

thakkarV commented 1 month ago

good eye! you are welcome to file a PR. You are right that no one depends on this right now AFAICT

github-actions[bot] commented 1 week ago

This issue has been labeled inactive-30d due to no recent activity in the past 30 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be labeled inactive-90d if there is no activity in the next 60 days.