[AMD][BACKEND] Filter barriers in Membar between LocalLoads and AsyncCopy when pipelining
#6639
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Membardoes not deduce that two shared memory sub views do not alias if they are from the same allocation. When pipelining we therefore get barriers between theLocalLoadand the prefetch loads (AsyncCopy/BufferLoadToLocal) because they read/write into sub views of the same allocation.This PR adds a filter function to Membar analysis to avoid those barriers by ignoring the dependency of
LocalLoadif it consumes anAsyncTokenfrom anAsyncWait. For such cases the barrier of theAsyncWaitis enough to correctly synchronize the access.