[mlir][Vector] Add UB conversions to different tests and pipelines #125145
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
This PR adds the UB to LLVM/SPIR-V conversion pass to some pipelines and tests. This is in preparation to introducing the first generation of `ub.poison` in Vector dialect transformations. It should effectively be NFC at this point.
dcaballe
requested review from
aartbik,
PeimingLiu,
yinying-lisa-li,
matthias-springer,
antiagainst and
kuhar
as code owners
llvmbot
added
mlir:llvm
mlir:spirv
mlir:vectorops
mlir:sparse
|
@llvm/pr-subscribers-mlir-sparse @llvm/pr-subscribers-mlir-vector Author: Diego Caballero (dcaballe) ChangesThis PR adds the UB to LLVM/SPIR-V conversion pass to some pipelines and tests. This is in preparation to introducing the generation of Full diff: https://github.com/llvm/llvm-project/pull/125145.diff 9 Files Affected:
diff --git a/mlir/include/mlir/Dialect/SparseTensor/Transforms/Passes.h b/mlir/include/mlir/Dialect/SparseTensor/Transforms/Passes.h
index 2e9c297f20182a..212f7b6f13c260 100644
--- a/mlir/include/mlir/Dialect/SparseTensor/Transforms/Passes.h
+++ b/mlir/include/mlir/Dialect/SparseTensor/Transforms/Passes.h
@@ -13,10 +13,10 @@
#ifndef MLIR_DIALECT_SPARSETENSOR_TRANSFORMS_PASSES_H_
#define MLIR_DIALECT_SPARSETENSOR_TRANSFORMS_PASSES_H_
+#include "mlir/Dialect/Bufferization/Transforms/OneShotAnalysis.h"
#include "mlir/IR/PatternMatch.h"
#include "mlir/Pass/Pass.h"
#include "mlir/Transforms/DialectConversion.h"
-#include "mlir/Transforms/OneToNTypeConversion.h"
//===----------------------------------------------------------------------===//
// Include the generated pass header (which needs some early definitions).
diff --git a/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp b/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
index 887ab1d40cabae..018233dc767b42 100644
--- a/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
+++ b/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
@@ -6,23 +6,15 @@
//
//===----------------------------------------------------------------------===//
-#include "mlir/Dialect/SparseTensor/Pipelines/Passes.h"
-
-#include "mlir/Conversion/ArithToLLVM/ArithToLLVM.h"
-#include "mlir/Conversion/ControlFlowToLLVM/ControlFlowToLLVM.h"
-#include "mlir/Conversion/GPUToNVVM/GPUToNVVMPass.h"
#include "mlir/Conversion/Passes.h"
#include "mlir/Dialect/Arith/Transforms/Passes.h"
-#include "mlir/Dialect/Bufferization/Transforms/Bufferize.h"
-#include "mlir/Dialect/Bufferization/Transforms/OneShotAnalysis.h"
#include "mlir/Dialect/Bufferization/Transforms/Passes.h"
#include "mlir/Dialect/Func/IR/FuncOps.h"
#include "mlir/Dialect/GPU/IR/GPUDialect.h"
#include "mlir/Dialect/GPU/Transforms/Passes.h"
-#include "mlir/Dialect/LLVMIR/NVVMDialect.h"
#include "mlir/Dialect/Linalg/Passes.h"
#include "mlir/Dialect/MemRef/Transforms/Passes.h"
-#include "mlir/Dialect/SparseTensor/IR/SparseTensor.h"
+#include "mlir/Dialect/SparseTensor/Pipelines/Passes.h"
#include "mlir/Dialect/SparseTensor/Transforms/Passes.h"
#include "mlir/Pass/PassManager.h"
#include "mlir/Transforms/Passes.h"
@@ -107,6 +99,9 @@ void mlir::sparse_tensor::buildSparsifier(OpPassManager &pm,
pm.addPass(createGpuModuleToBinaryPass(gpuModuleToBinaryPassOptions));
}
+ // Convert poison values.
+ pm.addPass(createUBToLLVMConversionPass());
+
// Ensure all casts are realized.
pm.addPass(createReconcileUnrealizedCastsPass());
}
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/shuffle16x16.mlir b/mlir/test/Integration/Dialect/Vector/CPU/shuffle16x16.mlir
index 03bc2b5123c482..33a76b13c245ab 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/shuffle16x16.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/shuffle16x16.mlir
@@ -1,7 +1,9 @@
// RUN: mlir-opt %s -convert-vector-to-scf -convert-scf-to-cf \
// RUN: -transform-interpreter \
// RUN: -test-transform-dialect-erase-schedule \
-// RUN: -convert-vector-to-llvm -convert-func-to-llvm -convert-arith-to-llvm -convert-cf-to-llvm -reconcile-unrealized-casts | \
+// RUN: -convert-vector-to-llvm -convert-func-to-llvm \
+// RUN: -convert-arith-to-llvm -convert-cf-to-llvm \
+// RUN: -convert-ub-to-llvm -reconcile-unrealized-casts | \
// RUN: mlir-runner -e entry -entry-point-result=void \
// RUN: -shared-libs=%mlir_c_runner_utils | \
// RUN: FileCheck %s
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-1d.mlir b/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-1d.mlir
index c4222a744f0a96..2393bd1c79dea9 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-1d.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-1d.mlir
@@ -1,9 +1,9 @@
-// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf,expand-strided-metadata,lower-affine,convert-arith-to-llvm,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
+// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf,expand-strided-metadata,lower-affine,convert-arith-to-llvm,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-ub-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
// RUN: mlir-runner -e entry -entry-point-result=void \
// RUN: -shared-libs=%mlir_c_runner_utils | \
// RUN: FileCheck %s
-// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf{full-unroll=true},expand-strided-metadata,lower-affine,convert-arith-to-llvm,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
+// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf{full-unroll=true},expand-strided-metadata,lower-affine,convert-arith-to-llvm,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-ub-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
// RUN: mlir-runner -e entry -entry-point-result=void \
// RUN: -shared-libs=%mlir_c_runner_utils | \
// RUN: FileCheck %s
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-2d.mlir b/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-2d.mlir
index 61e25c2dee732a..009c1375beaab6 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-2d.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-2d.mlir
@@ -1,9 +1,9 @@
-// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf,lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
+// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf,lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,convert-ub-to-llvm,reconcile-unrealized-casts)" | \
// RUN: mlir-runner -e entry -entry-point-result=void \
// RUN: -shared-libs=%mlir_c_runner_utils | \
// RUN: FileCheck %s
-// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf{full-unroll=true},lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
+// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf{full-unroll=true},lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,convert-ub-to-llvm,reconcile-unrealized-casts)" | \
// RUN: mlir-runner -e entry -entry-point-result=void \
// RUN: -shared-libs=%mlir_c_runner_utils | \
// RUN: FileCheck %s
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-3d.mlir b/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-3d.mlir
index 6c120d1b28c380..d41d9c93bc2b33 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-3d.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/transfer-read-3d.mlir
@@ -1,9 +1,9 @@
-// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf,lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
+// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf,lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,convert-ub-to-llvm,reconcile-unrealized-casts)" | \
// RUN: mlir-runner -e entry -entry-point-result=void \
// RUN: -shared-libs=%mlir_c_runner_utils | \
// RUN: FileCheck %s
-// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf{full-unroll=true},lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
+// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf{full-unroll=true},lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,convert-ub-to-llvm,reconcile-unrealized-casts)" | \
// RUN: mlir-runner -e entry -entry-point-result=void \
// RUN: -shared-libs=%mlir_c_runner_utils | \
// RUN: FileCheck %s
diff --git a/mlir/test/Integration/Dialect/Vector/CPU/transfer-to-loops.mlir b/mlir/test/Integration/Dialect/Vector/CPU/transfer-to-loops.mlir
index 1e74bc72e5b7f4..a7fec69b694921 100644
--- a/mlir/test/Integration/Dialect/Vector/CPU/transfer-to-loops.mlir
+++ b/mlir/test/Integration/Dialect/Vector/CPU/transfer-to-loops.mlir
@@ -1,9 +1,9 @@
-// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf,lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
+// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf,lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,convert-ub-to-llvm,reconcile-unrealized-casts)" | \
// RUN: mlir-runner -e main -entry-point-result=void \
// RUN: -shared-libs=%mlir_runner_utils,%mlir_c_runner_utils | \
// RUN: FileCheck %s
-// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf{full-unroll=true},lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,reconcile-unrealized-casts)" | \
+// RUN: mlir-opt %s -pass-pipeline="builtin.module(func.func(convert-vector-to-scf{full-unroll=true},lower-affine,convert-scf-to-cf),convert-vector-to-llvm,finalize-memref-to-llvm,convert-func-to-llvm,convert-arith-to-llvm,convert-cf-to-llvm,convert-ub-to-llvm,reconcile-unrealized-casts)" | \
// RUN: mlir-runner -e main -entry-point-result=void \
// RUN: -shared-libs=%mlir_runner_utils,%mlir_c_runner_utils | \
// RUN: FileCheck %s
diff --git a/mlir/test/lib/Conversion/ConvertToSPIRV/TestSPIRVVectorUnrolling.cpp b/mlir/test/lib/Conversion/ConvertToSPIRV/TestSPIRVVectorUnrolling.cpp
index 0bad43d5214b18..85245ea50782d6 100644
--- a/mlir/test/lib/Conversion/ConvertToSPIRV/TestSPIRVVectorUnrolling.cpp
+++ b/mlir/test/lib/Conversion/ConvertToSPIRV/TestSPIRVVectorUnrolling.cpp
@@ -6,15 +6,12 @@
//
//===-------------------------------------------------------------------===//
-#include "mlir/Dialect/Arith/IR/Arith.h"
-#include "mlir/Dialect/Func/IR/FuncOps.h"
#include "mlir/Dialect/SPIRV/IR/SPIRVDialect.h"
#include "mlir/Dialect/SPIRV/Transforms/SPIRVConversion.h"
+#include "mlir/Dialect/UB/IR/UBOps.h"
#include "mlir/Dialect/Vector/IR/VectorOps.h"
-#include "mlir/Dialect/Vector/Transforms/LoweringPatterns.h"
#include "mlir/Pass/Pass.h"
#include "mlir/Pass/PassManager.h"
-#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
namespace mlir {
namespace {
@@ -30,7 +27,8 @@ struct TestSPIRVVectorUnrolling final
}
void getDependentDialects(DialectRegistry ®istry) const override {
- registry.insert<spirv::SPIRVDialect, vector::VectorDialect>();
+ registry
+ .insert<spirv::SPIRVDialect, vector::VectorDialect, ub::UBDialect>();
}
void runOnOperation() override {
diff --git a/mlir/test/lib/Dialect/LLVM/TestLowerToLLVM.cpp b/mlir/test/lib/Dialect/LLVM/TestLowerToLLVM.cpp
index 253748eb79ae11..fca58b6a170628 100644
--- a/mlir/test/lib/Dialect/LLVM/TestLowerToLLVM.cpp
+++ b/mlir/test/lib/Dialect/LLVM/TestLowerToLLVM.cpp
@@ -20,14 +20,12 @@
#include "mlir/Conversion/MemRefToLLVM/MemRefToLLVM.h"
#include "mlir/Conversion/ReconcileUnrealizedCasts/ReconcileUnrealizedCasts.h"
#include "mlir/Conversion/SCFToControlFlow/SCFToControlFlow.h"
+#include "mlir/Conversion/UBToLLVM/UBToLLVM.h"
#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h"
#include "mlir/Conversion/VectorToSCF/VectorToSCF.h"
#include "mlir/Dialect/Func/IR/FuncOps.h"
-#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
#include "mlir/Dialect/Linalg/Passes.h"
#include "mlir/Dialect/MemRef/Transforms/Passes.h"
-#include "mlir/IR/DialectRegistry.h"
-#include "mlir/Pass/Pass.h"
#include "mlir/Pass/PassManager.h"
#include "mlir/Pass/PassOptions.h"
#include "mlir/Transforms/Passes.h"
@@ -81,6 +79,8 @@ void buildTestLowerToLLVM(OpPassManager &pm,
pm.addPass(createConvertControlFlowToLLVMPass());
// Convert Index to LLVM (always needed).
pm.addPass(createConvertIndexToLLVMPass());
+ // Convert UB to LLVM (always needed).
+ pm.addPass(createUBToLLVMConversionPass());
// Convert remaining unrealized_casts (always needed).
pm.addPass(createReconcileUnrealizedCastsPass());
}
|
kuhar
approved these changes
Jan 31, 2025
|
Thanks! Two (non-blocking) Qs:
-Andrzej |
I separated these changes from the upcoming PRs I’m working on and is triggered by test failures. I detached it from the rest because it involves other components, like sparse, and I didn’t want to overwhelm reviewers with a large PR spanning across multiple components. In general, we’ll need to introduce UB conversion more broadly over time. These changes are just the minimal ones needed for now. |
aartbik
approved these changes
Jan 31, 2025
Makes sense - separating changes like this does simplify reviewing, thanks! I'd be tempted to expand the summary (so that visiting GitHub is not required to get the full context). |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/89/builds/15838 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/130/builds/9482 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/80/builds/9565 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/138/builds/9832 Here is the relevant piece of the build log for the reference |
|
@dcaballe Seems like there are some link issues |
|
Fixed @ d810c74 |
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 the UB to LLVM/SPIR-V conversion pass to some pipelines and tests. This is in preparation to introducing the generation of
ub.poisonin Vector dialect transformations. It should effectively be NFC at this point.