[mlir][SME] Update E2E test to show potential optimisation (NFC) #107585
Conversation
This file contains hidden or 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
nujaa
requested review from
banach-space,
dcaballe and
nicolasvasilache
as code owners
|
@llvm/pr-subscribers-mlir @llvm/pr-subscribers-mlir-linalg Author: Hugo Trachino (nujaa) ChangesIntroduces loop hoisting to ARM SME E2E tests to allow the hoisting of the tile load offering very important speedup. Discussed here : https://discourse.llvm.org/t/mlir-for-arm-sme-reducing-tile-data-transfers/80065/2 Full diff: https://github.com/llvm/llvm-project/pull/107585.diff 2 Files Affected:
diff --git a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul-transpose-a.mlir b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul-transpose-a.mlir
index a57348a543c3cf..886211b65efa2d 100644
--- a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul-transpose-a.mlir
+++ b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul-transpose-a.mlir
@@ -82,8 +82,16 @@ module attributes {transform.with_named_sequence} {
transform.apply_patterns to %func {
transform.apply_patterns.vector.lower_contraction lowering_strategy = "outerproduct"
transform.apply_patterns.vector.lower_masks
+ transform.apply_patterns.canonicalization
} : !transform.any_op
+ // Step 5: Hoist load of accumulator.
+ %func_h = transform.structured.hoist_redundant_vector_transfers %func
+ : (!transform.any_op) -> !transform.any_op
+ %all_loops = transform.structured.match interface{LoopLikeInterface} in %module
+ : (!transform.any_op) -> !transform.any_op
+ transform.apply_licm to %all_loops : !transform.any_op
+ transform.loop.hoist_loop_invariant_subsets %all_loops : !transform.any_op
transform.yield
}
}
diff --git a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul.mlir b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul.mlir
index 79c9fcac70604b..4b6b9a9c746499 100644
--- a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul.mlir
+++ b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul.mlir
@@ -88,8 +88,16 @@ module attributes {transform.with_named_sequence} {
transform.apply_patterns.vector.lower_contraction lowering_strategy = "outerproduct"
transform.apply_patterns.vector.lower_masks
transform.apply_patterns.vector.rank_reducing_subview_patterns
+ transform.apply_patterns.canonicalization
} : !transform.any_op
+ // Step 6: Hoist load of accumulator.
+ %func_h = transform.structured.hoist_redundant_vector_transfers %func
+ : (!transform.any_op) -> !transform.any_op
+ %all_loops = transform.structured.match interface{LoopLikeInterface} in %bufferize
+ : (!transform.any_op) -> !transform.any_op
+ transform.apply_licm to %all_loops : !transform.any_op
+ transform.loop.hoist_loop_invariant_subsets %all_loops : !transform.any_op
transform.yield
}
}
|
|
@llvm/pr-subscribers-mlir-sme Author: Hugo Trachino (nujaa) ChangesIntroduces loop hoisting to ARM SME E2E tests to allow the hoisting of the tile load offering very important speedup. Discussed here : https://discourse.llvm.org/t/mlir-for-arm-sme-reducing-tile-data-transfers/80065/2 Full diff: https://github.com/llvm/llvm-project/pull/107585.diff 2 Files Affected:
diff --git a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul-transpose-a.mlir b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul-transpose-a.mlir
index a57348a543c3cf..886211b65efa2d 100644
--- a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul-transpose-a.mlir
+++ b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul-transpose-a.mlir
@@ -82,8 +82,16 @@ module attributes {transform.with_named_sequence} {
transform.apply_patterns to %func {
transform.apply_patterns.vector.lower_contraction lowering_strategy = "outerproduct"
transform.apply_patterns.vector.lower_masks
+ transform.apply_patterns.canonicalization
} : !transform.any_op
+ // Step 5: Hoist load of accumulator.
+ %func_h = transform.structured.hoist_redundant_vector_transfers %func
+ : (!transform.any_op) -> !transform.any_op
+ %all_loops = transform.structured.match interface{LoopLikeInterface} in %module
+ : (!transform.any_op) -> !transform.any_op
+ transform.apply_licm to %all_loops : !transform.any_op
+ transform.loop.hoist_loop_invariant_subsets %all_loops : !transform.any_op
transform.yield
}
}
diff --git a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul.mlir b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul.mlir
index 79c9fcac70604b..4b6b9a9c746499 100644
--- a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul.mlir
+++ b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSME/matmul.mlir
@@ -88,8 +88,16 @@ module attributes {transform.with_named_sequence} {
transform.apply_patterns.vector.lower_contraction lowering_strategy = "outerproduct"
transform.apply_patterns.vector.lower_masks
transform.apply_patterns.vector.rank_reducing_subview_patterns
+ transform.apply_patterns.canonicalization
} : !transform.any_op
+ // Step 6: Hoist load of accumulator.
+ %func_h = transform.structured.hoist_redundant_vector_transfers %func
+ : (!transform.any_op) -> !transform.any_op
+ %all_loops = transform.structured.match interface{LoopLikeInterface} in %bufferize
+ : (!transform.any_op) -> !transform.any_op
+ transform.apply_licm to %all_loops : !transform.any_op
+ transform.loop.hoist_loop_invariant_subsets %all_loops : !transform.any_op
transform.yield
}
}
|
banach-space
approved these changes
Sep 6, 2024
MacDue
approved these changes
Sep 7, 2024
| } : !transform.any_op | ||
|
|
||
| // Step 5: Hoist load of accumulator. |
There was a problem hiding this comment.
Nit: It's both the load and store of the accumulator that's hoisted.
|
Typo optionnal -> optional (also maybe say optimization rather than optional) |
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.
Introduces loop hoisting to ARM SME E2E tests to allow the hoisting of the tile load offering very important speedup.
Discussed here : https://discourse.llvm.org/t/mlir-for-arm-sme-reducing-tile-data-transfers/80065/2