Add comment to fp8 gemm

pytorch-labs · Nov 22, 2024 · b05a1f3 · b05a1f3
1 parent 5f2a2d1
commit b05a1f3
Showing 1 changed file with 2 additions and 1 deletion.
diff --git a/test/test_gpu/skip_tests_h100_triton_main.yaml b/test/test_gpu/skip_tests_h100_triton_main.yaml
@@ -16,8 +16,9 @@ fp8_attention:
 # fb-only kernels
 fp8_fused_quant_gemm_rowwise:
 fp8_gemm:
-  # FIXME: triton_*_persistent kernels are broken
+  # FIXME: out of shared memory
   - triton_persistent_fp8_gemm
+  # FIXME: out of shared memory
   - triton_tma_persistent_fp8_gemm
 gemm:
   # FIXME: out of shared memory