[Bugfix] Disable tma_aligned_scales in test_fusions_e2e (#32916)

Signed-off-by: Xin Yang <xyangx@amazon.com>
2026-01-23 06:34:30 -08:00
parent d95d650762
commit 90c2007932
3 changed files with 9 additions and 1 deletions
--- a/vllm/model_executor/layers/quantization/utils/fp8_utils.py
+++ b/vllm/model_executor/layers/quantization/utils/fp8_utils.py
@@ -379,7 +379,7 @@ class W8A8BlockFp8LinearOp:
                False,
                self.act_quant_group_shape,
                column_major_scales=True,
-                tma_aligned_scales=True,
+                tma_aligned_scales=envs.VLLM_USE_DEEP_GEMM_TMA_ALIGNED_SCALES,
                use_ue8m0=self.use_deep_gemm_e8m0,
            )
            if self.is_deep_gemm_supported