update VLLM_FLASH_ATTN_V1 to VLLM_USE_FLASH_ATTN_PA

vllm/attention/layer.py

@@ -75,7 +75,7 @@ class Attention(nn.Module):
             calculate_kv_scales = cache_config.calculate_kv_scales
         else:
             kv_cache_dtype = "auto"
-            block_size = 16 if not envs.VLLM_USE_FLASH_ATTN_PA or not envs.VLLM_FLASH_ATTN_V1 else 64
+            block_size = 16 if not envs.VLLM_USE_FLASH_ATTN_PA else 64
             is_attention_free = False
             calculate_kv_scales = False
         if num_kv_heads is None:
@@ -303,7 +303,7 @@ class MultiHeadAttention(nn.Module):
         attn_backend = get_attn_backend(head_size,
                                         dtype,
                                         kv_cache_dtype=None,
-                                        block_size=16 if not envs.VLLM_USE_FLASH_ATTN_PA or not envs.VLLM_FLASH_ATTN_V1 else 64,
+                                        block_size=16 if not envs.VLLM_USE_FLASH_ATTN_PA else 64,
                                         is_attention_free=False)
         backend = backend_name_to_enum(attn_backend.get_name())
         if current_platform.is_rocm():

vllm/config.py

@@ -1497,7 +1497,7 @@ PrefixCachingHashAlgo = Literal["builtin", "sha256"]
 class CacheConfig:
     """Configuration for the KV cache."""
 
-    block_size: BlockSize = 16 if not envs.VLLM_USE_FLASH_ATTN_PA or not envs.VLLM_FLASH_ATTN_V1 else 64  # type: ignore
+    block_size: BlockSize = 16 if not envs.VLLM_USE_FLASH_ATTN_PA else 64  # type: ignore
     """Size of a contiguous cache block in number of tokens. This is ignored on
     neuron devices and set to `--max-model-len`. On CUDA devices, only block
     sizes up to 32 are supported. On HPU devices, block size defaults to 128.

vllm/envs.py

@@ -154,7 +154,6 @@ if TYPE_CHECKING:
     VLLM_CUSTOM_ALLREDUCE_SUPPORTED_WORLDSIZE_MAX: int = 16
     VLLM_ENFORCE_EAGER_BS_THRESHOLD: Optional[int] = None
     VLLM_HAS_CONTEXT_DEFAULT: bool = False
-    VLLM_FLASH_ATTN_V1: bool = False
     VLLM_USE_NN: bool = False
     VLLM_ENABLE_TBO: bool = False
     VLLM_TBO_REQ_DELAY_MS: int = 0
@@ -1047,11 +1046,6 @@ environment_variables: dict[str, Callable[[], Any]] = {
     # to restore the default usage.
     "VLLM_HAS_CONTEXT_DEFAULT":
     lambda: bool(int(os.getenv("VLLM_HAS_CONTEXT_DEFAULT", "0"))),
-
-    # If set, vLLM will use FlashAttention Backend for v1 attention computation on rocm
-    "VLLM_FLASH_ATTN_V1":
-    lambda: (os.environ.get("VLLM_FLASH_ATTN_V1", "False").lower() in
-             ("true", "1")),
     
     # If set, vLLM will transpose weight to use nn layout
     "VLLM_USE_NN":

vllm/platforms/rocm.py

@@ -275,7 +275,7 @@ class RocmPlatform(Platform):
             #     logger.info_once("Using Triton backend on V1 engine.")
             #     return TRITON_ATTN_VLLM_V1
             
-            if envs.VLLM_FLASH_ATTN_V1 and block_size == 64:
+            if envs.is_set("VLLM_USE_FLASH_ATTN_PA") and envs.VLLM_USE_FLASH_ATTN_PA and block_size == 64:
                 logger.info_once("Using Flash Attention backend on V1 engine. (only supports block size 64)")
                 return FLASH_ATTN_V1