[mlir][vector] Rename ConvertVectorToLLVM to NaiveConvertVectorToLLVM #160055
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
banach-space
requested review from
PeimingLiu,
aartbik,
dcaballe,
fabianmcg,
ftynse,
makslevental,
matthias-springer,
nicolasvasilache,
rolfmorel,
stellaraccident and
yinying-lisa-li
as code owners
llvmbot
added
mlir:llvm
mlir:linalg
mlir:gpu
mlir:vectorops
mlir:sparse
Member
|
@llvm/pr-subscribers-flang-fir-hlfir @llvm/pr-subscribers-mlir-sparse Author: Andrzej Warzyński (banach-space) ChangesFollowing the discussion in the Tensor Compiler Working Group (see The new name better reflects the intent of the pass: it serves as a basic Patch is 115.20 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/160055.diff 113 Files Affected:
diff --git a/mlir/include/mlir/Conversion/Passes.h b/mlir/include/mlir/Conversion/Passes.h
index da061b269daf7..cc247ac424ee0 100644
--- a/mlir/include/mlir/Conversion/Passes.h
+++ b/mlir/include/mlir/Conversion/Passes.h
@@ -78,7 +78,7 @@
#include "mlir/Conversion/VectorToAMX/VectorToAMX.h"
#include "mlir/Conversion/VectorToArmSME/VectorToArmSME.h"
#include "mlir/Conversion/VectorToGPU/VectorToGPU.h"
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h"
#include "mlir/Conversion/VectorToSCF/VectorToSCF.h"
#include "mlir/Conversion/VectorToSPIRV/VectorToSPIRVPass.h"
#include "mlir/Conversion/VectorToXeGPU/VectorToXeGPU.h"
diff --git a/mlir/include/mlir/Conversion/Passes.td b/mlir/include/mlir/Conversion/Passes.td
index 1a37d057776e2..fd1656956dfab 100644
--- a/mlir/include/mlir/Conversion/Passes.td
+++ b/mlir/include/mlir/Conversion/Passes.td
@@ -1427,7 +1427,7 @@ def ConvertArmSMEToLLVM : InterfacePass<"convert-arm-sme-to-llvm", "FunctionOpIn
// VectorToLLVM
//===----------------------------------------------------------------------===//
-def ConvertVectorToLLVMPass : Pass<"convert-vector-to-llvm"> {
+def NaiveConvertVectorToLLVMPass : Pass<"naive-convert-vector-to-llvm"> {
let summary = "Lower the operations from the vector dialect into the LLVM "
"dialect";
let description = [{
diff --git a/mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h b/mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h
similarity index 81%
rename from mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h
rename to mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h
index cfb6cc313bc63..1e700e880dbd0 100644
--- a/mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h
+++ b/mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h
@@ -5,8 +5,8 @@
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
-#ifndef MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVM_H_
-#define MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVM_H_
+#ifndef MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVM_H_
+#define MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVM_H_
#include "mlir/Transforms/DialectConversion.h"
@@ -24,4 +24,4 @@ void registerConvertVectorToLLVMInterface(DialectRegistry ®istry);
}
} // namespace mlir
-#endif // MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVM_H_
+#endif // MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVM_H_
diff --git a/mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h b/mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h
similarity index 61%
rename from mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h
rename to mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h
index 410b881db7959..395104498293d 100644
--- a/mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h
+++ b/mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h
@@ -5,16 +5,16 @@
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
-#ifndef MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVMPASS_H_
-#define MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVMPASS_H_
+#ifndef MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVMPASS_H_
+#define MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVMPASS_H_
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Dialect/Vector/Transforms/VectorTransforms.h"
namespace mlir {
class Pass;
-#define GEN_PASS_DECL_CONVERTVECTORTOLLVMPASS
+#define GEN_PASS_DECL_NAIVECONVERTVECTORTOLLVMPASS
#include "mlir/Conversion/Passes.h.inc"
} // namespace mlir
-#endif // MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVMPASS_H_
+#endif // MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVMPASS_H_
diff --git a/mlir/include/mlir/Dialect/SparseTensor/Pipelines/Passes.h b/mlir/include/mlir/Dialect/SparseTensor/Pipelines/Passes.h
index efbe5c56a219b..ac5c3fa281383 100644
--- a/mlir/include/mlir/Dialect/SparseTensor/Pipelines/Passes.h
+++ b/mlir/include/mlir/Dialect/SparseTensor/Pipelines/Passes.h
@@ -13,7 +13,7 @@
#ifndef MLIR_DIALECT_SPARSETENSOR_PIPELINES_PASSES_H_
#define MLIR_DIALECT_SPARSETENSOR_PIPELINES_PASSES_H_
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h"
#include "mlir/Dialect/SparseTensor/Transforms/Passes.h"
#include "mlir/Pass/PassOptions.h"
@@ -162,8 +162,8 @@ struct SparsifierOptions : public PassPipelineOptions<SparsifierOptions> {
}
/// Projects out the options for `createConvertVectorToLLVMPass`.
- ConvertVectorToLLVMPassOptions convertVectorToLLVMOptions() const {
- ConvertVectorToLLVMPassOptions opts{};
+ NaiveConvertVectorToLLVMPassOptions convertVectorToLLVMOptions() const {
+ NaiveConvertVectorToLLVMPassOptions opts{};
opts.reassociateFPReductions = reassociateFPReductions;
opts.force32BitVectorIndices = force32BitVectorIndices;
opts.armNeon = armNeon;
diff --git a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
index e516118f75207..b946844f0070d 100644
--- a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
+++ b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
@@ -24,7 +24,7 @@
#include "mlir/Conversion/GPUCommon/GPUToLLVM.h"
#include "mlir/Conversion/LLVMCommon/Pattern.h"
#include "mlir/Conversion/MemRefToLLVM/MemRefToLLVM.h"
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Dialect/Async/IR/Async.h"
#include "mlir/Dialect/GPU/IR/GPUDialect.h"
#include "mlir/Dialect/GPU/Transforms/Passes.h"
diff --git a/mlir/lib/Conversion/VectorToLLVM/CMakeLists.txt b/mlir/lib/Conversion/VectorToLLVM/CMakeLists.txt
index 8e2620029c354..1106799cf6904 100644
--- a/mlir/lib/Conversion/VectorToLLVM/CMakeLists.txt
+++ b/mlir/lib/Conversion/VectorToLLVM/CMakeLists.txt
@@ -1,6 +1,6 @@
add_mlir_conversion_library(MLIRVectorToLLVM
PARTIAL_SOURCES_INTENDED
- ConvertVectorToLLVM.cpp
+ NaiveConvertVectorToLLVM.cpp
ADDITIONAL_HEADER_DIRS
${MLIR_MAIN_INCLUDE_DIR}/mlir/Conversion/VectorToLLVM
@@ -27,7 +27,7 @@ add_mlir_conversion_library(MLIRVectorToLLVM
add_mlir_conversion_library(MLIRVectorToLLVMPass
PARTIAL_SOURCES_INTENDED
- ConvertVectorToLLVMPass.cpp
+ NaiveConvertVectorToLLVMPass.cpp
ADDITIONAL_HEADER_DIRS
${MLIR_MAIN_INCLUDE_DIR}/mlir/Conversion/VectorToLLVM
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp b/mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.cpp
similarity index 99%
rename from mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
rename to mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.cpp
index e7266740894b1..339227bd49393 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.cpp
@@ -1,12 +1,14 @@
-//===- VectorToLLVM.cpp - Conversion from Vector to the LLVM dialect ------===//
+//===- VectorToLLVM.cpp ---------------------------------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
+// (Naive) Conversion from Vector to the LLVM dialect
+//===----------------------------------------------------------------------===//
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Conversion/ArithCommon/AttrToLLVMConverter.h"
#include "mlir/Conversion/ConvertToLLVM/ToLLVMInterface.h"
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp b/mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.cpp
similarity index 91%
rename from mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
rename to mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.cpp
index 0b44ca7ceee42..ba2875acf6d5c 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.cpp
@@ -1,12 +1,15 @@
-//===- VectorToLLVM.cpp - Conversion from Vector to the LLVM dialect ------===//
+//===- NaiveConverVectorToLLVM.cpp
+//-----------------------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
+// (Naive) Conversion from Vector to the LLVM dialect
+//===----------------------------------------------------------------------===//
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h"
#include "mlir/Conversion/LLVMCommon/ConversionTarget.h"
#include "mlir/Conversion/LLVMCommon/TypeConverter.h"
@@ -28,7 +31,7 @@
#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
namespace mlir {
-#define GEN_PASS_DEF_CONVERTVECTORTOLLVMPASS
+#define GEN_PASS_DEF_NAIVECONVERTVECTORTOLLVMPASS
#include "mlir/Conversion/Passes.h.inc"
} // namespace mlir
@@ -36,8 +39,9 @@ using namespace mlir;
using namespace mlir::vector;
namespace {
-struct ConvertVectorToLLVMPass
- : public impl::ConvertVectorToLLVMPassBase<ConvertVectorToLLVMPass> {
+struct NaiveConvertVectorToLLVMPass
+ : public impl::NaiveConvertVectorToLLVMPassBase<
+ NaiveConvertVectorToLLVMPass> {
using Base::Base;
@@ -60,7 +64,7 @@ struct ConvertVectorToLLVMPass
};
} // namespace
-void ConvertVectorToLLVMPass::runOnOperation() {
+void NaiveConvertVectorToLLVMPass::runOnOperation() {
// Perform progressive lowering of operations on slices and all contraction
// operations. Also materializes masks, lowers vector.step, rank-reduces FMA,
// applies folding and DCE.
diff --git a/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp b/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
index dabbea1bdec63..37f050390c492 100644
--- a/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
+++ b/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
@@ -70,14 +70,14 @@ void mlir::sparse_tensor::buildSparsifier(OpPassManager &pm,
pm.addPass(memref::createExpandStridedMetadataPass());
pm.addPass(createLowerAffinePass());
pm.addPass(
- createConvertVectorToLLVMPass(options.convertVectorToLLVMOptions()));
+ createNaiveConvertVectorToLLVMPass(options.convertVectorToLLVMOptions()));
pm.addNestedPass<func::FuncOp>(createConvertComplexToStandardPass());
pm.addNestedPass<func::FuncOp>(arith::createArithExpandOpsPass());
pm.addNestedPass<func::FuncOp>(createConvertMathToLLVMPass());
pm.addPass(createConvertMathToLibmPass());
pm.addPass(createConvertComplexToLibm());
pm.addPass(
- createConvertVectorToLLVMPass(options.convertVectorToLLVMOptions()));
+ createNaiveConvertVectorToLLVMPass(options.convertVectorToLLVMOptions()));
// Finalize GPU code generation.
if (gpuCodegen) {
diff --git a/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp b/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
index 18f105ef62e38..b471b38cc9c3a 100644
--- a/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
+++ b/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
@@ -9,7 +9,7 @@
#include "mlir/Dialect/Vector/TransformOps/VectorTransformOps.h"
#include "mlir/Conversion/LLVMCommon/TypeConverter.h"
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Conversion/VectorToSCF/VectorToSCF.h"
#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
#include "mlir/Dialect/Transform/IR/TransformDialect.h"
diff --git a/mlir/lib/RegisterAllExtensions.cpp b/mlir/lib/RegisterAllExtensions.cpp
index 69a85dbe141ce..4ca27f0305460 100644
--- a/mlir/lib/RegisterAllExtensions.cpp
+++ b/mlir/lib/RegisterAllExtensions.cpp
@@ -31,7 +31,7 @@
#include "mlir/Conversion/PtrToLLVM/PtrToLLVM.h"
#include "mlir/Conversion/SCFToEmitC/SCFToEmitC.h"
#include "mlir/Conversion/UBToLLVM/UBToLLVM.h"
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Conversion/XeVMToLLVM/XeVMToLLVM.h"
#include "mlir/Dialect/AMX/Transforms.h"
#include "mlir/Dialect/Affine/TransformOps/AffineTransformOps.h"
diff --git a/mlir/test/Conversion/VectorToLLVM/pass-option-serialization.mlir b/mlir/test/Conversion/VectorToLLVM/pass-option-serialization.mlir
index 323d86ac40988..7f43604a7e08c 100644
--- a/mlir/test/Conversion/VectorToLLVM/pass-option-serialization.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/pass-option-serialization.mlir
@@ -2,8 +2,8 @@
// This test also allows us to exercise these options (to some extent) even if we
// don't use them in other Vector to LLVM conversion tests. This is quite relevant
-// for the `Vector` Dialect (and `--convert-vector-to-llvm` pass) as in many cases
-// we use the Transform Dialect (TD) rather than `--convert-vector-to-llvm` for
+// for the `Vector` Dialect (and `--naive-convert-vector-to-llvm` pass) as in many cases
+// we use the Transform Dialect (TD) rather than `--naive-convert-vector-to-llvm` for
// testing. So here we don't check the correctness of the passes, as they're
// covered by other tests that use TD, but we still provide some test coverage of
// these pass options.
@@ -11,13 +11,13 @@
// We don't need to actually parse any IR to print the pass options. We just need
// to provide --dump-pass-pipeline
-// RUN: mlir-opt --convert-vector-to-llvm --dump-pass-pipeline %s 2>&1 | FileCheck %s --check-prefix=DEFAULT
+// RUN: mlir-opt --naive-convert-vector-to-llvm --dump-pass-pipeline %s 2>&1 | FileCheck %s --check-prefix=DEFAULT
-// RUN: mlir-opt --convert-vector-to-llvm='vector-contract-lowering=matmul vector-transpose-lowering=flat' \
+// RUN: mlir-opt --naive-convert-vector-to-llvm='vector-contract-lowering=matmul vector-transpose-lowering=flat' \
// RUN: --dump-pass-pipeline %s 2>&1 | FileCheck %s --check-prefix=NON-DEFAULT
// CHECK: builtin.module(
-// CHECK-SAME: convert-vector-to-llvm{
+// CHECK-SAME: naive-convert-vector-to-llvm{
// CHECK-SAME: enable-amx={{[aA-zZ0-9]+}}
// CHECK-SAME: enable-arm-neon={{[aA-zZ0-9]+}}
// CHECK-SAME: enable-arm-sve={{[aA-zZ0-9]+}}
diff --git a/mlir/test/Conversion/VectorToLLVM/use-vector-alignment.mlir b/mlir/test/Conversion/VectorToLLVM/use-vector-alignment.mlir
index 12fe3552ce1b7..80100541f164e 100644
--- a/mlir/test/Conversion/VectorToLLVM/use-vector-alignment.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/use-vector-alignment.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt %s --convert-vector-to-llvm='use-vector-alignment=0' --split-input-file | FileCheck %s --check-prefix=MEMREF-ALIGN
-// RUN: mlir-opt %s --convert-vector-to-llvm='use-vector-alignment=1' --split-input-file | FileCheck %s --check-prefix=VEC-ALIGN
+// RUN: mlir-opt %s --naive-convert-vector-to-llvm='use-vector-alignment=0' --split-input-file | FileCheck %s --check-prefix=MEMREF-ALIGN
+// RUN: mlir-opt %s --naive-convert-vector-to-llvm='use-vector-alignment=1' --split-input-file | FileCheck %s --check-prefix=VEC-ALIGN
//===----------------------------------------------------------------------===//
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir b/mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir
index 91e5358622b69..8eea6ff6a312e 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt %s --convert-vector-to-llvm='force-32bit-vector-indices=1' | FileCheck %s --check-prefix=CMP32
-// RUN: mlir-opt %s --convert-vector-to-llvm='force-32bit-vector-indices=0' | FileCheck %s --check-prefix=CMP64
+// RUN: mlir-opt %s --naive-convert-vector-to-llvm='force-32bit-vector-indices=1' | FileCheck %s --check-prefix=CMP32
+// RUN: mlir-opt %s --naive-convert-vector-to-llvm='force-32bit-vector-indices=0' | FileCheck %s --check-prefix=CMP64
// CMP32-LABEL: @genbool_var_1d(
// CMP32-SAME: %[[ARG:.*]]: index)
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-reduction-to-llvm.mlir b/mlir/test/Conversion/VectorToLLVM/vector-reduction-to-llvm.mlir
index c7d9e22fb2423..0dc7d101648a2 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-reduction-to-llvm.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-reduction-to-llvm.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt %s -convert-vector-to-llvm -split-input-file | FileCheck %s
-// RUN: mlir-opt %s -convert-vector-to-llvm='reassociate-fp-reductions' -split-input-file | FileCheck %s --check-prefix=REASSOC
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm='reassociate-fp-reductions' -split-input-file | FileCheck %s --check-prefix=REASSOC
// CHECK-LABEL: @reduce_add_f32(
// CHECK-SAME: %[[A:.*]]: vector<16xf32>)
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-scalable-memcpy.mlir b/mlir/test/Conversion/VectorToLLVM/vector-scalable-memcpy.mlir
index 80e6caa05db5e..b24b1c72f8a1f 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-scalable-memcpy.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-scalable-memcpy.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -convert-vector-to-llvm | mlir-opt | FileCheck %s
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm | mlir-opt | FileCheck %s
// CHECK: vector_scalable_memcopy([[SRC:%arg[0-9]+]]: memref<?xf32>, [[DST:%arg[0-9]+]]
func.func @vector_scalable_memcopy(%src : memref<?xf32>, %dst : memref<?xf32>, %size : index) {
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-to-llvm-interface.mlir b/mlir/test/Conversion/VectorToLLVM/vector-to-llvm-interface.mlir
index 5973c2ba2cbd0..2e73aa8d066e0 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-to-llvm-interface.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-to-llvm-interface.mlir
@@ -1,6 +1,6 @@
// RUN: mlir-opt --convert-to-llvm="filter-dialects=vector" --split-input-file %s | FileCheck %s
// RUN: mlir-opt --convert-to-llvm="filter-dialects=vector allow-pattern-rollback=0" --split-input-file %s | FileCheck %s
-// RUN: mlir-opt %s -convert-vector-to-llvm -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm -split-input-file | FileCheck %s
//===========================================================================//
// Basic tests for Vector-to-LLVM conversion
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir b/mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir
index 2d33888854ea7..816e4c4ddb8d5 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -convert-vector-to-llvm -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm -split-input-file | FileCheck %s
//===========================================================================//
// Complex tests for Vector-to-LLVM conversion
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-xfer-to-llvm.mlir b/mlir/test/Conversion/VectorToLLVM/vector-xfer-to-llvm.mlir
index d3f6d7eca90b4..ca3224525bbe5 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-xfer-to-llvm.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-xfer-to-llvm.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt %s -convert-vector-to-llvm -split-input-file | FileCheck '-D$IDX_TYPE=i32' %s
-// RUN: mlir-opt %s --convert-vector-to-llvm='force-32bit-vector-indices=0' | FileCheck '-D$IDX_TYPE=i64' %s
+// RUN: mlir-opt %...
[truncated]
|
Member
|
@llvm/pr-subscribers-mlir-linalg Author: Andrzej Warzyński (banach-space) ChangesFollowing the discussion in the Tensor Compiler Working Group (see The new name better reflects the intent of the pass: it serves as a basic Patch is 115.20 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/160055.diff 113 Files Affected:
diff --git a/mlir/include/mlir/Conversion/Passes.h b/mlir/include/mlir/Conversion/Passes.h
index da061b269daf7..cc247ac424ee0 100644
--- a/mlir/include/mlir/Conversion/Passes.h
+++ b/mlir/include/mlir/Conversion/Passes.h
@@ -78,7 +78,7 @@
#include "mlir/Conversion/VectorToAMX/VectorToAMX.h"
#include "mlir/Conversion/VectorToArmSME/VectorToArmSME.h"
#include "mlir/Conversion/VectorToGPU/VectorToGPU.h"
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h"
#include "mlir/Conversion/VectorToSCF/VectorToSCF.h"
#include "mlir/Conversion/VectorToSPIRV/VectorToSPIRVPass.h"
#include "mlir/Conversion/VectorToXeGPU/VectorToXeGPU.h"
diff --git a/mlir/include/mlir/Conversion/Passes.td b/mlir/include/mlir/Conversion/Passes.td
index 1a37d057776e2..fd1656956dfab 100644
--- a/mlir/include/mlir/Conversion/Passes.td
+++ b/mlir/include/mlir/Conversion/Passes.td
@@ -1427,7 +1427,7 @@ def ConvertArmSMEToLLVM : InterfacePass<"convert-arm-sme-to-llvm", "FunctionOpIn
// VectorToLLVM
//===----------------------------------------------------------------------===//
-def ConvertVectorToLLVMPass : Pass<"convert-vector-to-llvm"> {
+def NaiveConvertVectorToLLVMPass : Pass<"naive-convert-vector-to-llvm"> {
let summary = "Lower the operations from the vector dialect into the LLVM "
"dialect";
let description = [{
diff --git a/mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h b/mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h
similarity index 81%
rename from mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h
rename to mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h
index cfb6cc313bc63..1e700e880dbd0 100644
--- a/mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h
+++ b/mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h
@@ -5,8 +5,8 @@
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
-#ifndef MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVM_H_
-#define MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVM_H_
+#ifndef MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVM_H_
+#define MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVM_H_
#include "mlir/Transforms/DialectConversion.h"
@@ -24,4 +24,4 @@ void registerConvertVectorToLLVMInterface(DialectRegistry ®istry);
}
} // namespace mlir
-#endif // MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVM_H_
+#endif // MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVM_H_
diff --git a/mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h b/mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h
similarity index 61%
rename from mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h
rename to mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h
index 410b881db7959..395104498293d 100644
--- a/mlir/include/mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h
+++ b/mlir/include/mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h
@@ -5,16 +5,16 @@
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
-#ifndef MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVMPASS_H_
-#define MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVMPASS_H_
+#ifndef MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVMPASS_H_
+#define MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVMPASS_H_
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Dialect/Vector/Transforms/VectorTransforms.h"
namespace mlir {
class Pass;
-#define GEN_PASS_DECL_CONVERTVECTORTOLLVMPASS
+#define GEN_PASS_DECL_NAIVECONVERTVECTORTOLLVMPASS
#include "mlir/Conversion/Passes.h.inc"
} // namespace mlir
-#endif // MLIR_CONVERSION_VECTORTOLLVM_CONVERTVECTORTOLLVMPASS_H_
+#endif // MLIR_CONVERSION_VECTORTOLLVM_NAIVECONVERTVECTORTOLLVMPASS_H_
diff --git a/mlir/include/mlir/Dialect/SparseTensor/Pipelines/Passes.h b/mlir/include/mlir/Dialect/SparseTensor/Pipelines/Passes.h
index efbe5c56a219b..ac5c3fa281383 100644
--- a/mlir/include/mlir/Dialect/SparseTensor/Pipelines/Passes.h
+++ b/mlir/include/mlir/Dialect/SparseTensor/Pipelines/Passes.h
@@ -13,7 +13,7 @@
#ifndef MLIR_DIALECT_SPARSETENSOR_PIPELINES_PASSES_H_
#define MLIR_DIALECT_SPARSETENSOR_PIPELINES_PASSES_H_
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h"
#include "mlir/Dialect/SparseTensor/Transforms/Passes.h"
#include "mlir/Pass/PassOptions.h"
@@ -162,8 +162,8 @@ struct SparsifierOptions : public PassPipelineOptions<SparsifierOptions> {
}
/// Projects out the options for `createConvertVectorToLLVMPass`.
- ConvertVectorToLLVMPassOptions convertVectorToLLVMOptions() const {
- ConvertVectorToLLVMPassOptions opts{};
+ NaiveConvertVectorToLLVMPassOptions convertVectorToLLVMOptions() const {
+ NaiveConvertVectorToLLVMPassOptions opts{};
opts.reassociateFPReductions = reassociateFPReductions;
opts.force32BitVectorIndices = force32BitVectorIndices;
opts.armNeon = armNeon;
diff --git a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
index e516118f75207..b946844f0070d 100644
--- a/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
+++ b/mlir/lib/Conversion/GPUCommon/GPUToLLVMConversion.cpp
@@ -24,7 +24,7 @@
#include "mlir/Conversion/GPUCommon/GPUToLLVM.h"
#include "mlir/Conversion/LLVMCommon/Pattern.h"
#include "mlir/Conversion/MemRefToLLVM/MemRefToLLVM.h"
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Dialect/Async/IR/Async.h"
#include "mlir/Dialect/GPU/IR/GPUDialect.h"
#include "mlir/Dialect/GPU/Transforms/Passes.h"
diff --git a/mlir/lib/Conversion/VectorToLLVM/CMakeLists.txt b/mlir/lib/Conversion/VectorToLLVM/CMakeLists.txt
index 8e2620029c354..1106799cf6904 100644
--- a/mlir/lib/Conversion/VectorToLLVM/CMakeLists.txt
+++ b/mlir/lib/Conversion/VectorToLLVM/CMakeLists.txt
@@ -1,6 +1,6 @@
add_mlir_conversion_library(MLIRVectorToLLVM
PARTIAL_SOURCES_INTENDED
- ConvertVectorToLLVM.cpp
+ NaiveConvertVectorToLLVM.cpp
ADDITIONAL_HEADER_DIRS
${MLIR_MAIN_INCLUDE_DIR}/mlir/Conversion/VectorToLLVM
@@ -27,7 +27,7 @@ add_mlir_conversion_library(MLIRVectorToLLVM
add_mlir_conversion_library(MLIRVectorToLLVMPass
PARTIAL_SOURCES_INTENDED
- ConvertVectorToLLVMPass.cpp
+ NaiveConvertVectorToLLVMPass.cpp
ADDITIONAL_HEADER_DIRS
${MLIR_MAIN_INCLUDE_DIR}/mlir/Conversion/VectorToLLVM
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp b/mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.cpp
similarity index 99%
rename from mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
rename to mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.cpp
index e7266740894b1..339227bd49393 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVM.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.cpp
@@ -1,12 +1,14 @@
-//===- VectorToLLVM.cpp - Conversion from Vector to the LLVM dialect ------===//
+//===- VectorToLLVM.cpp ---------------------------------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
+// (Naive) Conversion from Vector to the LLVM dialect
+//===----------------------------------------------------------------------===//
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Conversion/ArithCommon/AttrToLLVMConverter.h"
#include "mlir/Conversion/ConvertToLLVM/ToLLVMInterface.h"
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp b/mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.cpp
similarity index 91%
rename from mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
rename to mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.cpp
index 0b44ca7ceee42..ba2875acf6d5c 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.cpp
@@ -1,12 +1,15 @@
-//===- VectorToLLVM.cpp - Conversion from Vector to the LLVM dialect ------===//
+//===- NaiveConverVectorToLLVM.cpp
+//-----------------------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
+// (Naive) Conversion from Vector to the LLVM dialect
+//===----------------------------------------------------------------------===//
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVMPass.h"
#include "mlir/Conversion/LLVMCommon/ConversionTarget.h"
#include "mlir/Conversion/LLVMCommon/TypeConverter.h"
@@ -28,7 +31,7 @@
#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
namespace mlir {
-#define GEN_PASS_DEF_CONVERTVECTORTOLLVMPASS
+#define GEN_PASS_DEF_NAIVECONVERTVECTORTOLLVMPASS
#include "mlir/Conversion/Passes.h.inc"
} // namespace mlir
@@ -36,8 +39,9 @@ using namespace mlir;
using namespace mlir::vector;
namespace {
-struct ConvertVectorToLLVMPass
- : public impl::ConvertVectorToLLVMPassBase<ConvertVectorToLLVMPass> {
+struct NaiveConvertVectorToLLVMPass
+ : public impl::NaiveConvertVectorToLLVMPassBase<
+ NaiveConvertVectorToLLVMPass> {
using Base::Base;
@@ -60,7 +64,7 @@ struct ConvertVectorToLLVMPass
};
} // namespace
-void ConvertVectorToLLVMPass::runOnOperation() {
+void NaiveConvertVectorToLLVMPass::runOnOperation() {
// Perform progressive lowering of operations on slices and all contraction
// operations. Also materializes masks, lowers vector.step, rank-reduces FMA,
// applies folding and DCE.
diff --git a/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp b/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
index dabbea1bdec63..37f050390c492 100644
--- a/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
+++ b/mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp
@@ -70,14 +70,14 @@ void mlir::sparse_tensor::buildSparsifier(OpPassManager &pm,
pm.addPass(memref::createExpandStridedMetadataPass());
pm.addPass(createLowerAffinePass());
pm.addPass(
- createConvertVectorToLLVMPass(options.convertVectorToLLVMOptions()));
+ createNaiveConvertVectorToLLVMPass(options.convertVectorToLLVMOptions()));
pm.addNestedPass<func::FuncOp>(createConvertComplexToStandardPass());
pm.addNestedPass<func::FuncOp>(arith::createArithExpandOpsPass());
pm.addNestedPass<func::FuncOp>(createConvertMathToLLVMPass());
pm.addPass(createConvertMathToLibmPass());
pm.addPass(createConvertComplexToLibm());
pm.addPass(
- createConvertVectorToLLVMPass(options.convertVectorToLLVMOptions()));
+ createNaiveConvertVectorToLLVMPass(options.convertVectorToLLVMOptions()));
// Finalize GPU code generation.
if (gpuCodegen) {
diff --git a/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp b/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
index 18f105ef62e38..b471b38cc9c3a 100644
--- a/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
+++ b/mlir/lib/Dialect/Vector/TransformOps/VectorTransformOps.cpp
@@ -9,7 +9,7 @@
#include "mlir/Dialect/Vector/TransformOps/VectorTransformOps.h"
#include "mlir/Conversion/LLVMCommon/TypeConverter.h"
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Conversion/VectorToSCF/VectorToSCF.h"
#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
#include "mlir/Dialect/Transform/IR/TransformDialect.h"
diff --git a/mlir/lib/RegisterAllExtensions.cpp b/mlir/lib/RegisterAllExtensions.cpp
index 69a85dbe141ce..4ca27f0305460 100644
--- a/mlir/lib/RegisterAllExtensions.cpp
+++ b/mlir/lib/RegisterAllExtensions.cpp
@@ -31,7 +31,7 @@
#include "mlir/Conversion/PtrToLLVM/PtrToLLVM.h"
#include "mlir/Conversion/SCFToEmitC/SCFToEmitC.h"
#include "mlir/Conversion/UBToLLVM/UBToLLVM.h"
-#include "mlir/Conversion/VectorToLLVM/ConvertVectorToLLVM.h"
+#include "mlir/Conversion/VectorToLLVM/NaiveConvertVectorToLLVM.h"
#include "mlir/Conversion/XeVMToLLVM/XeVMToLLVM.h"
#include "mlir/Dialect/AMX/Transforms.h"
#include "mlir/Dialect/Affine/TransformOps/AffineTransformOps.h"
diff --git a/mlir/test/Conversion/VectorToLLVM/pass-option-serialization.mlir b/mlir/test/Conversion/VectorToLLVM/pass-option-serialization.mlir
index 323d86ac40988..7f43604a7e08c 100644
--- a/mlir/test/Conversion/VectorToLLVM/pass-option-serialization.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/pass-option-serialization.mlir
@@ -2,8 +2,8 @@
// This test also allows us to exercise these options (to some extent) even if we
// don't use them in other Vector to LLVM conversion tests. This is quite relevant
-// for the `Vector` Dialect (and `--convert-vector-to-llvm` pass) as in many cases
-// we use the Transform Dialect (TD) rather than `--convert-vector-to-llvm` for
+// for the `Vector` Dialect (and `--naive-convert-vector-to-llvm` pass) as in many cases
+// we use the Transform Dialect (TD) rather than `--naive-convert-vector-to-llvm` for
// testing. So here we don't check the correctness of the passes, as they're
// covered by other tests that use TD, but we still provide some test coverage of
// these pass options.
@@ -11,13 +11,13 @@
// We don't need to actually parse any IR to print the pass options. We just need
// to provide --dump-pass-pipeline
-// RUN: mlir-opt --convert-vector-to-llvm --dump-pass-pipeline %s 2>&1 | FileCheck %s --check-prefix=DEFAULT
+// RUN: mlir-opt --naive-convert-vector-to-llvm --dump-pass-pipeline %s 2>&1 | FileCheck %s --check-prefix=DEFAULT
-// RUN: mlir-opt --convert-vector-to-llvm='vector-contract-lowering=matmul vector-transpose-lowering=flat' \
+// RUN: mlir-opt --naive-convert-vector-to-llvm='vector-contract-lowering=matmul vector-transpose-lowering=flat' \
// RUN: --dump-pass-pipeline %s 2>&1 | FileCheck %s --check-prefix=NON-DEFAULT
// CHECK: builtin.module(
-// CHECK-SAME: convert-vector-to-llvm{
+// CHECK-SAME: naive-convert-vector-to-llvm{
// CHECK-SAME: enable-amx={{[aA-zZ0-9]+}}
// CHECK-SAME: enable-arm-neon={{[aA-zZ0-9]+}}
// CHECK-SAME: enable-arm-sve={{[aA-zZ0-9]+}}
diff --git a/mlir/test/Conversion/VectorToLLVM/use-vector-alignment.mlir b/mlir/test/Conversion/VectorToLLVM/use-vector-alignment.mlir
index 12fe3552ce1b7..80100541f164e 100644
--- a/mlir/test/Conversion/VectorToLLVM/use-vector-alignment.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/use-vector-alignment.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt %s --convert-vector-to-llvm='use-vector-alignment=0' --split-input-file | FileCheck %s --check-prefix=MEMREF-ALIGN
-// RUN: mlir-opt %s --convert-vector-to-llvm='use-vector-alignment=1' --split-input-file | FileCheck %s --check-prefix=VEC-ALIGN
+// RUN: mlir-opt %s --naive-convert-vector-to-llvm='use-vector-alignment=0' --split-input-file | FileCheck %s --check-prefix=MEMREF-ALIGN
+// RUN: mlir-opt %s --naive-convert-vector-to-llvm='use-vector-alignment=1' --split-input-file | FileCheck %s --check-prefix=VEC-ALIGN
//===----------------------------------------------------------------------===//
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir b/mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir
index 91e5358622b69..8eea6ff6a312e 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-mask-to-llvm.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt %s --convert-vector-to-llvm='force-32bit-vector-indices=1' | FileCheck %s --check-prefix=CMP32
-// RUN: mlir-opt %s --convert-vector-to-llvm='force-32bit-vector-indices=0' | FileCheck %s --check-prefix=CMP64
+// RUN: mlir-opt %s --naive-convert-vector-to-llvm='force-32bit-vector-indices=1' | FileCheck %s --check-prefix=CMP32
+// RUN: mlir-opt %s --naive-convert-vector-to-llvm='force-32bit-vector-indices=0' | FileCheck %s --check-prefix=CMP64
// CMP32-LABEL: @genbool_var_1d(
// CMP32-SAME: %[[ARG:.*]]: index)
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-reduction-to-llvm.mlir b/mlir/test/Conversion/VectorToLLVM/vector-reduction-to-llvm.mlir
index c7d9e22fb2423..0dc7d101648a2 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-reduction-to-llvm.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-reduction-to-llvm.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt %s -convert-vector-to-llvm -split-input-file | FileCheck %s
-// RUN: mlir-opt %s -convert-vector-to-llvm='reassociate-fp-reductions' -split-input-file | FileCheck %s --check-prefix=REASSOC
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm='reassociate-fp-reductions' -split-input-file | FileCheck %s --check-prefix=REASSOC
// CHECK-LABEL: @reduce_add_f32(
// CHECK-SAME: %[[A:.*]]: vector<16xf32>)
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-scalable-memcpy.mlir b/mlir/test/Conversion/VectorToLLVM/vector-scalable-memcpy.mlir
index 80e6caa05db5e..b24b1c72f8a1f 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-scalable-memcpy.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-scalable-memcpy.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -convert-vector-to-llvm | mlir-opt | FileCheck %s
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm | mlir-opt | FileCheck %s
// CHECK: vector_scalable_memcopy([[SRC:%arg[0-9]+]]: memref<?xf32>, [[DST:%arg[0-9]+]]
func.func @vector_scalable_memcopy(%src : memref<?xf32>, %dst : memref<?xf32>, %size : index) {
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-to-llvm-interface.mlir b/mlir/test/Conversion/VectorToLLVM/vector-to-llvm-interface.mlir
index 5973c2ba2cbd0..2e73aa8d066e0 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-to-llvm-interface.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-to-llvm-interface.mlir
@@ -1,6 +1,6 @@
// RUN: mlir-opt --convert-to-llvm="filter-dialects=vector" --split-input-file %s | FileCheck %s
// RUN: mlir-opt --convert-to-llvm="filter-dialects=vector allow-pattern-rollback=0" --split-input-file %s | FileCheck %s
-// RUN: mlir-opt %s -convert-vector-to-llvm -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm -split-input-file | FileCheck %s
//===========================================================================//
// Basic tests for Vector-to-LLVM conversion
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir b/mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir
index 2d33888854ea7..816e4c4ddb8d5 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -convert-vector-to-llvm -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -naive-convert-vector-to-llvm -split-input-file | FileCheck %s
//===========================================================================//
// Complex tests for Vector-to-LLVM conversion
diff --git a/mlir/test/Conversion/VectorToLLVM/vector-xfer-to-llvm.mlir b/mlir/test/Conversion/VectorToLLVM/vector-xfer-to-llvm.mlir
index d3f6d7eca90b4..ca3224525bbe5 100644
--- a/mlir/test/Conversion/VectorToLLVM/vector-xfer-to-llvm.mlir
+++ b/mlir/test/Conversion/VectorToLLVM/vector-xfer-to-llvm.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt %s -convert-vector-to-llvm -split-input-file | FileCheck '-D$IDX_TYPE=i32' %s
-// RUN: mlir-opt %s --convert-vector-to-llvm='force-32bit-vector-indices=0' | FileCheck '-D$IDX_TYPE=i64' %s
+// RUN: mlir-opt %...
[truncated]
|
nicolasvasilache
approved these changes
Sep 22, 2025
kuhar
approved these changes
Sep 22, 2025
Contributor
|
More than the pass renaming, I think it's important to set the expectations in the pass documentation which, at this point, are not 100% clear to me. I'm not sure the goal is to make the conversion "naive" but more target and "algorithmically-opinionated" :) independent, moving hardware specific decisions (how to lower a contraction/transpose, etc...) and algorithmic decisions (vector unrolling, vector linearization, etc.) to independent passes. With that in mind, I feel we should work on creating those new passes and moving the corresponding transformations out of this gigantic pass before thinking about the final name. WDYT? |
Contributor
My 2 cents on this is while I would love to see it happen, the reality is driven by priorities. |
Member
|
Agree with @dcaballe on breaking it down, on future prs. My view is that naive is always opinionated anyway, so the name isn't too bad. Targets usually name these simplistic lowering as C semantics (e.g. C API), which is functional but no performance promises. Usually exist for portability, which is close enough to this intent, but we have no C semantics here. Hence, I think naive is a reasonable name. Basic, simple, direct, literal are all synonyms with the same intent, so to me, any of those names work just as well. |
Contributor
|
My point was more about doing a large renaming now and having to revisit it again after materializing the actual changes that we agreed upon than about the actual naming itself. For context, we agreed that we'd move target specific or opinionated lowerings like vector contractions, transposes, vector unrolling, vector linearization... to separate passes and keep the vector conversion to LLVM simpler. However, this PR is renaming the conversion pass but the pass is doing exactly the same conversions, including the ones that we plan to move out (definitely nothing naive yet). My suggestion was to materialize the actual conversion changes first and think about the name later as it's unclear the resulting pass is going to be naive or we want to call it something else. I'm not blocking anything here but that's my view. |
Member
+1 If something is named "Naive" we should have documentation on how you can build a not "Naive" version. When we discussed this, my understand was to rename it to "Naive" when we have all the pieces broken down and can actually say how you could build your own pipeline for your own usecase. |
Following the discussion in the Tensor Compiler Working Group (see https://discourse.llvm.org/t/tcdg-tensor-compiler-design-group-meeting-notes-2025-09-17), this patch renames ConvertVectorToLLVM to NaiveConvertVectorToLLVM. The new name better reflects the intent of the pass: it serves as a basic reference implementation for lowering from the Vector dialect to the LLVM dialect, without any consideration for performance. In practice, performance tuning requires knowledge of the target architecture, which this pass deliberately avoids.
banach-space
force-pushed
the
andrzej/vector/rename_vector_to_llvm
branch
from
f78e616 to
1c6e71b
Compare
Contributor
Author
|
Thank you all for your feedback, some really good points have been raised!
Naming is hard 😅 … However, I would like to start signalling to folks ASAP that this pass is intended as trivial/unoptimised/POC. We may need to adjust again, but this step improves clarity today.
Agreed. I’ve added a high-level description in the Vector dialect docs. Happy to iterate on this if more detail would be useful.
Indeed!
My intent is to signal early that this pass is limited and will become even more so as functionality is split out. I want to make this clear to our users as well as maintainers ;-) -Andrzej |
llvmbot
added
flang
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
8 participants
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.
Following the discussion in the Tensor Compiler Working Group (see
https://discourse.llvm.org/t/tcdg-tensor-compiler-design-group-meeting-notes-2025-09-17),
this patch renames ConvertVectorToLLVM to NaiveConvertVectorToLLVM.
The new name better reflects the intent of the pass: it serves as a basic
reference implementation for lowering from the Vector dialect to the LLVM
dialect, without any consideration for performance. In practice, performance
tuning requires knowledge of the target architecture, which this pass
deliberately avoids.
Note, the changes in this PR are mostly editorial and can be split into two
groups:
--convert-vector-to-llvm->--naive-convert-vector-to-llvm,ConvertVectorToLLVM* ->NaiveConvertVectorToLLVM*`