[mlir][Vector] Add support for sub-byte transpose emulation #80110
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
This PR adds patterns to convert a sub-byte vector transpose into a sequence of instructions that perform the transpose on i8 vector elements. Whereas this rewrite may not lead to the absolute peak performance, it should ensure correctness when dealing with sub-byte transposes.
|
@llvm/pr-subscribers-mlir-vector @llvm/pr-subscribers-mlir Author: Diego Caballero (dcaballe) ChangesThis PR adds patterns to convert a sub-byte vector transpose into a sequence of instructions that perform the transpose on i8 vector elements. Whereas this rewrite may not lead to the absolute peak performance, it should ensure correctness when dealing with sub-byte transposes. Full diff: https://github.com/llvm/llvm-project/pull/80110.diff 5 Files Affected:
diff --git a/mlir/include/mlir/Dialect/Vector/TransformOps/VectorTransformOps.td b/mlir/include/mlir/Dialect/Vector/TransformOps/VectorTransformOps.td
index 3ac6f28dcb938..ce88360aa52e9 100644
--- a/mlir/include/mlir/Dialect/Vector/TransformOps/VectorTransformOps.td
+++ b/mlir/include/mlir/Dialect/Vector/TransformOps/VectorTransformOps.td
@@ -151,7 +151,7 @@ def ApplyLowerMaskedTransfersPatternsOp : Op<Transform_Dialect,
"apply_patterns.vector.lower_masked_transfers",
[DeclareOpInterfaceMethods<PatternDescriptorOpInterface>]> {
let description = [{
- Apply opt-in patterns that lower vector.mask operations surrounding
+ Apply opt-in patterns that lower vector.mask operations surrounding
side-effecting ops:
- MaskedTransferReadOpPattern
- MaskedTransferWriteOpPattern
@@ -376,7 +376,7 @@ def ApplyVectorReductionToContractPatternsOp : Op<Transform_Dialect,
- ReorderCastOpsOnBroadcast
- ReorderElementwiseOpsOnTranspose
- These patterns have the effect of rewriting a vector.multi_reduce into a
+ These patterns have the effect of rewriting a vector.multi_reduce into a
vector.contract.
}];
diff --git a/mlir/include/mlir/Dialect/Vector/Transforms/VectorRewritePatterns.h b/mlir/include/mlir/Dialect/Vector/Transforms/VectorRewritePatterns.h
index 49b74c0c466d2..f5941d32e683f 100644
--- a/mlir/include/mlir/Dialect/Vector/Transforms/VectorRewritePatterns.h
+++ b/mlir/include/mlir/Dialect/Vector/Transforms/VectorRewritePatterns.h
@@ -371,6 +371,10 @@ FailureOr<Value> rewriteExtOfBitCast(RewriterBase &rewriter, Operation *extOp,
void populateVectorNarrowTypeRewritePatterns(RewritePatternSet &patterns,
PatternBenefit benefit = 1);
+/// Appends patterns for emulating a sub-byte vector transpose.
+void populateVectorTransposeNarrowTypeRewritePatterns(
+ RewritePatternSet &patterns, PatternBenefit benefit = 1);
+
} // namespace vector
} // namespace mlir
diff --git a/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp b/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
index 37127ea70f1e5..19922c4295fe0 100644
--- a/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
+++ b/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
@@ -162,6 +162,7 @@ void transform::ApplyLowerTransposePatternsOp::populatePatterns(
void transform::ApplyRewriteNarrowTypePatternsOp::populatePatterns(
RewritePatternSet &patterns) {
populateVectorNarrowTypeRewritePatterns(patterns);
+ populateVectorTransposeNarrowTypeRewritePatterns(patterns);
}
void transform::ApplySplitTransferFullPartialPatternsOp::populatePatterns(
diff --git a/mlir/lib/Dialect/Vector/Transforms/VectorEmulateNarrowType.cpp b/mlir/lib/Dialect/Vector/Transforms/VectorEmulateNarrowType.cpp
index 0110a8df89aee..193c9a6182b49 100644
--- a/mlir/lib/Dialect/Vector/Transforms/VectorEmulateNarrowType.cpp
+++ b/mlir/lib/Dialect/Vector/Transforms/VectorEmulateNarrowType.cpp
@@ -1052,6 +1052,52 @@ struct RewriteAlignedSubByteIntSignedExt : OpRewritePattern<ConversionOpType> {
}
};
+/// Rewrite a sub-byte vector transpose into a sequence of instructions that
+/// perform the transpose on wider (byte) element types.
+/// For example:
+/// %0 = vector.transpose %a, [1, 0] : vector<8x16xi4> to vector<16x8xi4>
+///
+/// is rewritten as:
+///
+/// %0 = arith.extsi %arg0 : vector<8x16xi4> to vector<8x16xi8>
+/// %1 = vector.transpose %0, [1, 0] : vector<8x16xi8> to vector<16x8xi8>
+/// %2 = arith.trunci %1 : vector<16x8xi8> to vector<16x8xi4>
+///
+struct RewriteVectorTranspose : OpRewritePattern<vector::TransposeOp> {
+ using OpRewritePattern<vector::TransposeOp>::OpRewritePattern;
+
+ RewriteVectorTranspose(MLIRContext *context, PatternBenefit benefit)
+ : OpRewritePattern<vector::TransposeOp>(context, benefit) {}
+
+ LogicalResult matchAndRewrite(vector::TransposeOp transposeOp,
+ PatternRewriter &rewriter) const override {
+ // Precondition: sub-byte integer transpose.
+ constexpr unsigned minNativeBitwidth = 8;
+ VectorType srcSubByteVecType = transposeOp.getSourceVectorType();
+ if (srcSubByteVecType.getElementTypeBitWidth() >= minNativeBitwidth)
+ return rewriter.notifyMatchFailure(transposeOp,
+ "not a sub-byte transpose");
+
+ // Perform the rewrite.
+ Location loc = transposeOp.getLoc();
+ // Signed/unsigned interpretation shouldn't matter here as we are just
+ // transposing the elements and truncating them back to the original size.
+ // TODO: Use unsigned extension (more efficient) when emulation or backend
+ // support is available.
+ auto srcNativeVecType =
+ srcSubByteVecType.cloneWith(std::nullopt, rewriter.getI8Type());
+ Value extOp = rewriter.create<arith::ExtSIOp>(loc, srcNativeVecType,
+ transposeOp.getVector());
+ Value newTranspose = rewriter.create<vector::TransposeOp>(
+ loc, extOp, transposeOp.getPermutation());
+ VectorType dstSubByteVecType = transposeOp.getResultVectorType();
+ rewriter.replaceOpWithNewOp<arith::TruncIOp>(transposeOp, dstSubByteVecType,
+ newTranspose);
+ return success();
+ }
+};
+
+
} // namespace
//===----------------------------------------------------------------------===//
@@ -1080,3 +1126,8 @@ void vector::populateVectorNarrowTypeRewritePatterns(
RewriteAlignedSubByteIntSignedExt<arith::SIToFPOp>>(
patterns.getContext(), benefit.getBenefit() + 1);
}
+
+void vector::populateVectorTransposeNarrowTypeRewritePatterns(
+ RewritePatternSet &patterns, PatternBenefit benefit) {
+ patterns.add<RewriteVectorTranspose>(patterns.getContext(), benefit);
+}
diff --git a/mlir/test/Dialect/Vector/vector-rewrite-narrow-types.mlir b/mlir/test/Dialect/Vector/vector-rewrite-narrow-types.mlir
index c4fbb4c219b91..02063a81664b8 100644
--- a/mlir/test/Dialect/Vector/vector-rewrite-narrow-types.mlir
+++ b/mlir/test/Dialect/Vector/vector-rewrite-narrow-types.mlir
@@ -226,6 +226,26 @@ func.func @aligned_sitofp(%a: vector<8xi4>) -> vector<8xf32> {
return %0 : vector<8xf32>
}
+// CHECK-LABEL: func.func @i4_transpose(
+// CHECK-SAME: %[[A:[0-9a-z]*]]
+func.func @i4_transpose(%a: vector<8x16xi4>) -> vector<16x8xi4> {
+ // CHECK: %[[EXT:.*]] = arith.extsi %[[A]] : vector<8x16xi4> to vector<8x16xi8>
+ // CHECK: %[[TRANS:.*]] = vector.transpose %[[EXT]], [1, 0] : vector<8x16xi8> to vector<16x8xi8>
+ // CHECK: %[[TRUNC:.*]] = arith.trunci %[[TRANS]] : vector<16x8xi8> to vector<16x8xi4>
+ %0 = vector.transpose %a, [1, 0] : vector<8x16xi4> to vector<16x8xi4>
+ return %0 : vector<16x8xi4>
+}
+
+// CHECK-LABEL: func.func @i7_transpose(
+// CHECK-SAME: %[[A:[0-9a-z]*]]
+func.func @i7_transpose(%a: vector<8x16xi7>) -> vector<16x8xi7> {
+ // CHECK: %[[EXT:.*]] = arith.extsi %[[A]] : vector<8x16xi7> to vector<8x16xi8>
+ // CHECK: %[[TRANS:.*]] = vector.transpose %[[EXT]], [1, 0] : vector<8x16xi8> to vector<16x8xi8>
+ // CHECK: %[[TRUNC:.*]] = arith.trunci %[[TRANS]] : vector<16x8xi8> to vector<16x8xi7>
+ %0 = vector.transpose %a, [1, 0] : vector<8x16xi7> to vector<16x8xi7>
+ return %0 : vector<16x8xi7>
+}
+
module attributes {transform.with_named_sequence} {
transform.named_sequence @__transform_main(%module_op: !transform.any_op {transform.readonly}) {
%f = transform.structured.match ops{["func.func"]} in %module_op
@@ -237,3 +257,4 @@ module attributes {transform.with_named_sequence} {
transform.yield
}
}
+
|
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
hanhanW
reviewed
Jan 31, 2024
| // TODO: Use unsigned extension (more efficient) when emulation or backend | ||
| // support is available. | ||
| auto srcNativeVecType = | ||
| srcSubByteVecType.cloneWith(std::nullopt, rewriter.getI8Type()); |
There was a problem hiding this comment.
I think rewriter.getIntegerType(minNativeBitwidth) is better for consistency.
hanhanW
approved these changes
Jan 31, 2024
carlosgalvezp
pushed a commit
to carlosgalvezp/llvm-project
that referenced
this pull request
Feb 1, 2024
This PR adds patterns to convert a sub-byte vector transpose into a sequence of instructions that perform the transpose on i8 vector elements. Whereas this rewrite may not lead to the absolute peak performance, it should ensure correctness when dealing with sub-byte transposes.
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 PR adds patterns to convert a sub-byte vector transpose into a sequence of instructions that perform the transpose on i8 vector elements. Whereas this rewrite may not lead to the absolute peak performance, it should ensure correctness when dealing with sub-byte transposes.