[DSv4] Adding TRTLLM gen attention kernel (#43827)

This commit is contained in:
Yongye Zhu
2026-06-04 07:35:09 -07:00
committed by GitHub
parent 3e77036768
commit b5235fca2e
20 changed files with 2659 additions and 86 deletions
+5 -3
View File
@@ -547,10 +547,12 @@ class SlidingWindowMLASpec(SlidingWindowSpec):
@property
def real_page_size_bytes(self) -> int:
if self.model_version == "deepseek_v4":
# DeepseekV4: 448B NoPE + 128B RoPE + 8B fp8 scale = 584B per token.
if self.model_version == "deepseek_v4" and self.cache_dtype_str == "fp8_ds_mla":
# DeepseekV4 FlashMLA: 448B NoPE + 128B RoPE + 8B fp8 scale = 584B
# per token. FlashInfer's contiguous bf16/fp8 cache falls through to
# the element-size formula below.
return self.storage_block_size * 584
assert self.model_version is None, (
assert self.model_version in (None, "deepseek_v4"), (
f"Unsupported model version: {self.model_version}"
)
return (