mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-03-24 00:57:39 +00:00
4237aedf9a
=?UTF-8?q?Increase=20tolerance=20for=20FP16=20GEMM=20test?= =?UTF-8?q?s=20to=20handle=20non-deterministic=20ro=E2=80=A6=20(#4335)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit …unding Three tests were failing intermittently with small errors (0.01-1.5%) due to non-deterministic FP16 accumulation order from GPU thread scheduling: - test_ck_tile_batched_gemm - test_ck_tile_grouped_gemm_preshuffle - test_ck_tile_grouped_gemm_multi_d These tests use kbatch=1 (no split-K), so errors are from order-dependent rounding, not atomics. Increased tolerances from 1e-3 to 2e-3 (0.2%) to account for FP16 precision limits while still catching real bugs. - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com>