search

ROCm / aotriton

Ahead of Time (AOT) Triton Math Library
MIT License
42 stars 15 forks source link

[Queued PR] Port fixes from 0.7.2b #56

Closed xinyazhang closed 1 week ago

xinyazhang commented 2 weeks ago

Major Changes

  1. varlen support fixes
  2. Fix the numerical error due to rounding differences b/w FMA and MUL+SUB
    • The test is added test_large_bf16_nan_values in {test,tritonsrc}/test_backward.py

Note: "Fix NaN created by 0.0 (from sm_scale) * -inf (from masking)." is firstly developed on main and then ported to 0.7.2b

xinyazhang commented 1 week ago

Close due to 0.7.2b fix adds new bug. Will include these in the new PR