[ROCm] Change default settings for ROCm

tests/v1/attention/test_rocm_attention_backends_selection.py

@@ -47,7 +47,7 @@ def mock_on_gfx9():
         (
             {},
             None,
-            AttentionBackendEnum.TRITON_ATTN.get_path(),
+            AttentionBackendEnum.ROCM_ATTN.get_path(),
         ),
         # Test Case 2: Explicit TRITON_ATTN backend
         (
@@ -74,11 +74,10 @@ def mock_on_gfx9():
             AttentionBackendEnum.ROCM_AITER_UNIFIED_ATTN.get_path(),
         ),
         # Test Case 6: VLLM_ROCM_USE_AITER=1
-        # (defaults to AITER FA when MHA not explicitly disabled)
         (
             {"VLLM_ROCM_USE_AITER": "1"},
             None,
-            AttentionBackendEnum.ROCM_AITER_FA.get_path(),
+            AttentionBackendEnum.ROCM_ATTN.get_path(),
         ),
         # Test Case 7: VLLM_ROCM_USE_AITER=1 + VLLM_ROCM_USE_AITER_MHA=1
         (
@@ -102,11 +101,10 @@ def mock_on_gfx9():
             AttentionBackendEnum.TRITON_ATTN.get_path(),
         ),
         # Test Case 10: VLLM_ROCM_USE_AITER=1 + VLLM_ROCM_USE_AITER_MHA=0
-        # (explicitly disabled)
         (
             {"VLLM_ROCM_USE_AITER": "1", "VLLM_ROCM_USE_AITER_MHA": "0"},
             None,
-            AttentionBackendEnum.TRITON_ATTN.get_path(),
+            AttentionBackendEnum.ROCM_ATTN.get_path(),
         ),
         # Test Case 11: VLLM_ROCM_USE_AITER=1 + explicit ROCM_ATTN
         (

vllm/envs.py

@@ -112,7 +112,6 @@
     VLLM_DISABLED_KERNELS: list[str] = []
     VLLM_DISABLE_PYNCCL: bool = False
     VLLM_ROCM_USE_AITER: bool = False
-    VLLM_ROCM_USE_AITER_PAGED_ATTN: bool = False
     VLLM_ROCM_USE_AITER_LINEAR: bool = True
     VLLM_ROCM_USE_AITER_MOE: bool = True
     VLLM_ROCM_USE_AITER_RMSNORM: bool = True
@@ -281,6 +280,14 @@ def maybe_convert_bool(value: str | None) -> bool | None:
     return bool(int(value))
 
 
+def use_aiter() -> bool:
+    from vllm._aiter_ops import is_aiter_found_and_supported
+
+    return is_aiter_found_and_supported() and os.getenv(
+        "VLLM_ROCM_USE_AITER", "True"
+    ).lower() in ("true", "1")
+
+
 def disable_compile_cache() -> bool:
     return bool(int(os.getenv("VLLM_DISABLE_COMPILE_CACHE", "0")))
 
@@ -951,14 +958,7 @@ def _get_or_set_default() -> str:
     ),
     # Disable aiter ops unless specifically enabled.
     # Acts as a parent switch to enable the rest of the other operations.
-    "VLLM_ROCM_USE_AITER": lambda: (
-        os.getenv("VLLM_ROCM_USE_AITER", "False").lower() in ("true", "1")
-    ),
-    # Whether to use aiter paged attention.
-    # By default is disabled.
-    "VLLM_ROCM_USE_AITER_PAGED_ATTN": lambda: (
-        os.getenv("VLLM_ROCM_USE_AITER_PAGED_ATTN", "False").lower() in ("true", "1")
-    ),
+    "VLLM_ROCM_USE_AITER": use_aiter,
     # use aiter linear op if aiter ops are enabled
     # The following list of related ops
     # - scaled_mm (per-tensor / rowwise)
@@ -980,9 +980,9 @@ def _get_or_set_default() -> str:
         os.getenv("VLLM_ROCM_USE_AITER_MLA", "True").lower() in ("true", "1")
     ),
     # Whether to use aiter mha ops.
-    # By default is enabled.
+    # By default is disabled.
     "VLLM_ROCM_USE_AITER_MHA": lambda: (
-        os.getenv("VLLM_ROCM_USE_AITER_MHA", "True").lower() in ("true", "1")
+        os.getenv("VLLM_ROCM_USE_AITER_MHA", "False").lower() in ("true", "1")
     ),
     # Whether to use aiter fp4 gemm asm.
     # By default is disabled.

vllm/platforms/rocm.py

@@ -321,6 +321,11 @@ def get_attn_backend_cls(
                 vllm_config is not None
                 and vllm_config.attention_config.use_prefill_decode_attention
             ):
+                logger.warning_once(
+                    "use_prefill_decode_attention is deprecated and will be removed in "
+                    "future releases. "
+                    "Use --attention_config.backend to select the desired backend"
+                )
                 logger.info("Using Rocm Attention backend.")
                 return AttentionBackendEnum.ROCM_ATTN.get_path()
 
@@ -334,9 +339,9 @@ def get_attn_backend_cls(
                 logger.info("Using Aiter Flash Attention backend.")
                 return AttentionBackendEnum.ROCM_AITER_FA.get_path()
 
-            # Default: Triton Unified Attention
-            logger.info("Using Triton Attention backend.")
-            return AttentionBackendEnum.TRITON_ATTN.get_path()
+            # Default: ROCm split Attention
+            logger.info("Using ROCm Attention backend.")
+            return AttentionBackendEnum.ROCM_ATTN.get_path()
 
         raise RuntimeError(
             f"Attention backend {selected_backend.name} is not supported on "