[AMDAIEFoldDmaWaits] Fold DMA wait operations across multi columns #986
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Yu-Zhewen
requested review from
MaheshRavishankar,
nirvedhmeshram,
yzhang93,
Abhishek-Varma and
jtuyls
as code owners
jtuyls
requested changes
Dec 13, 2024
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEControlCodeToTransaction.cpp
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEControlCodeToTransaction.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/test/controlcode_to_transaction.mlir
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/test/fold_dma_waits.mlir
Outdated
Show resolved
Hide resolved
jtuyls
reviewed
Dec 17, 2024
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/test/fold_dma_waits.mlir
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
jtuyls
reviewed
Dec 17, 2024
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
jtuyls
reviewed
Dec 18, 2024
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEFoldDmaWaits.cpp
Outdated
Show resolved
Hide resolved
Yu-Zhewen
force-pushed
the
zhewen_batch_sync
branch
from
1cefca4 to
672e29a
Compare
Yu-Zhewen
force-pushed
the
zhewen_batch_sync
branch
from
672e29a to
e47e0f3
Compare
jtuyls
reviewed
Dec 18, 2024
| using DmaBdIdKey = std::pair<AMDAIE::TileOp, AMDAIE::ConnectionOp>; | ||
| using DmaBdIdPair = std::pair<DmaBdIdKey, uint32_t>; | ||
|
|
||
| FailureOr<DmaBdIdPair> retriveDmaBdIdPair( |
There was a problem hiding this comment.
Suggested change
| FailureOr<DmaBdIdPair> retriveDmaBdIdPair( | |
| FailureOr<DmaBdIdPair> retrieveDmaBdIdPair( |
| const Operation *batchParentOp, | ||
| const DenseSet<AMDAIE::ConnectionOp> &connectionOps, | ||
| const DenseMap<DmaBdIdKey, DenseSet<uint32_t>> &dmaBdIdsMap, | ||
| AMDAIE::NpuHalfDmaCpyNdOp currHalfDmaCpyNdOp, DmaBdIdPair &currBdIdPair) { |
There was a problem hiding this comment.
Suggested change
| AMDAIE::NpuHalfDmaCpyNdOp currHalfDmaCpyNdOp, DmaBdIdPair &currBdIdPair) { | |
| AMDAIE::NpuHalfDmaCpyNdOp currHalfDmaCpyNdOp, DmaBdIdPair currBdIdPair) { |
Non-expensive pairs are preferably passed by value as it's more simple, avoids reference issues (e.g. dangling, currBdIdPair shouldn't be changed ) and is as fast/faster if the contents are small.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This is an enhancement for #962.
In the previous PR, DMA waits on the same
connection(and the same tile) could be folded, exploiting the fact that each DMA channel has a queue size of 4.In this PR, DMA waits across multiple
columnscan also be folded, provided their correspondingrow,channel, anddirectionare the same. This optimization leverages the ability to specifycolNuminTCTSync, where the range[col, col + colNum)can be addressed.The numbers in the following table show the instruction size in words.