[ROCm] Synchronize sparse MLA metadata before graph replay (#47404)
Signed-off-by: zihaomu <zmu@amd.com>
This commit is contained in:
@@ -566,6 +566,9 @@ class ROCMAiterMLASparseMetadataBuilder(
|
||||
uni_seqlen_qo=1,
|
||||
fast_mode=True,
|
||||
)
|
||||
# The persistent metadata buffers are read by graph replay. Order
|
||||
# the async metadata write before the graph-captured decode kernel.
|
||||
torch.cuda.current_stream(self.device).synchronize()
|
||||
self._prev_metadata_key = metadata_key
|
||||
|
||||
metadata = ROCMAiterMLASparseMetadata(
|
||||
|
||||
Reference in New Issue
Block a user