Add env var to disable FA4 warmup (#12430)

2025-10-31 12:25:00 -07:00
parent f7f9e41b36
commit a80bcb5a68
2 changed files with 11 additions and 0 deletions
@@ -28,6 +28,7 @@ SGLang supports various environment variables that can be used to configure its
 | `SGLANG_SKIP_P2P_CHECK` | Skip P2P (peer-to-peer) access check | `false` |
 | `SGL_CHUNKED_PREFIX_CACHE_THRESHOLD` | Sets the threshold for enabling chunked prefix caching | `8192` |
 | `SGLANG_FUSED_MLA_ENABLE_ROPE_FUSION` | Enable RoPE fusion in Fused Multi-Layer Attention | `1` |
+| `SGLANG_DISABLE_FA4_WARMUP` | Disable Flash Attention 4 warmup passes (set to `1`, `true`, `yes`, or `on` to disable) | `false` |

 ## DeepGEMM Configuration (Advanced Optimization)