set default block_size to 16

vllm/config.py

@@ -1630,7 +1630,7 @@ PrefixCachingHashAlgo = Literal["builtin", "sha256", "sha256_cbor_64bit"]
 class CacheConfig:
     """Configuration for the KV cache."""
 
-    block_size: BlockSize = None if not envs.VLLM_USE_FLASH_ATTN_PA 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.