mlir/MathExtras: consolidate with llvm/MathExtras #95087

artagnon · 2024-06-11T09:11:31Z

This patch is part of a project to move the Presburger library into LLVM.

llvmbot · 2024-06-11T09:12:04Z

@llvm/pr-subscribers-mlir-tensor
@llvm/pr-subscribers-mlir-func
@llvm/pr-subscribers-mlir-scf
@llvm/pr-subscribers-mlir-presburger
@llvm/pr-subscribers-mlir-memref
@llvm/pr-subscribers-mlir-core
@llvm/pr-subscribers-mlir-linalg

@llvm/pr-subscribers-mlir

Author: Ramkumar Ramachandra (artagnon)

Changes

This patch is part of a project to move the Presburger library into LLVM.

Patch is 27.71 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/95087.diff

30 Files Affected:

(modified) llvm/include/llvm/Support/MathExtras.h (+38-2)
(modified) llvm/unittests/Support/MathExtrasTest.cpp (+19-3)
(modified) mlir/include/mlir/Analysis/Presburger/Fraction.h (-1)
(modified) mlir/include/mlir/Analysis/Presburger/MPInt.h (+6-4)
(modified) mlir/include/mlir/Analysis/Presburger/SlowMPInt.h (-1)
(modified) mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h (+2-2)
(removed) mlir/include/mlir/Support/MathExtras.h (-51)
(modified) mlir/lib/Analysis/FlatLinearValueConstraints.cpp (-1)
(modified) mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp (-1)
(modified) mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp (+5-5)
(modified) mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp (+3-3)
(modified) mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp (-1)
(modified) mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp (+3-2)
(modified) mlir/lib/Dialect/Affine/IR/AffineOps.cpp (+5-1)
(modified) mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp (-1)
(modified) mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp (-1)
(modified) mlir/lib/Dialect/Func/IR/FuncOps.cpp (-1)
(modified) mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp (+2-2)
(modified) mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp (+2-2)
(modified) mlir/lib/Dialect/SCF/IR/SCF.cpp (-1)
(modified) mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp (+2-2)
(modified) mlir/lib/Dialect/SCF/Utils/Utils.cpp (+2-2)
(modified) mlir/lib/Dialect/Tensor/IR/TensorOps.cpp (+5-1)
(modified) mlir/lib/Dialect/Utils/StaticValueUtils.cpp (+2-2)
(modified) mlir/lib/Dialect/Vector/Transforms/VectorUnroll.cpp (-1)
(modified) mlir/lib/Dialect/Vector/Utils/VectorUtils.cpp (-1)
(modified) mlir/lib/IR/AffineExpr.cpp (+7-3)
(modified) mlir/lib/IR/AffineMap.cpp (+5-1)
(modified) mlir/unittests/Support/CMakeLists.txt (-1)
(removed) mlir/unittests/Support/MathExtrasTest.cpp (-31)

diff --git a/llvm/include/llvm/Support/MathExtras.h b/llvm/include/llvm/Support/MathExtras.h
index f0e4ee534ece3..c5cac51130b1d 100644
--- a/llvm/include/llvm/Support/MathExtras.h
+++ b/llvm/include/llvm/Support/MathExtras.h
@@ -424,11 +424,47 @@ template <uint64_t Align> constexpr inline uint64_t alignTo(uint64_t Value) {
   return (Value + Align - 1) / Align * Align;
 }
 
-/// Returns the integer ceil(Numerator / Denominator).
-inline uint64_t divideCeil(uint64_t Numerator, uint64_t Denominator) {
+/// Returns the integer ceil(Numerator / Denominator). Unsigned integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE uint64_t divideCeil(uint64_t Numerator,
+                                                 uint64_t Denominator) {
   return alignTo(Numerator, Denominator) / Denominator;
 }
 
+/// Returns the integer ceil(Numerator / Denominator). Signed integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t ceilDiv(int64_t Numerator,
+                                             int64_t Denominator) {
+  assert(Denominator);
+  if (!Numerator)
+    return 0;
+  // C's integer division rounds towards 0.
+  int64_t X = (Denominator > 0) ? -1 : 1;
+  bool SameSign = (Numerator > 0) == (Denominator > 0);
+  return SameSign ? ((Numerator + X) / Denominator) + 1
+                  : -(-Numerator / Denominator);
+}
+
+/// Returns the integer floor(Numerator / Denominator). Signed integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t floorDiv(int64_t Numerator,
+                                              int64_t Denominator) {
+  assert(Denominator);
+  if (!Numerator)
+    return 0;
+  // C's integer division rounds towards 0.
+  int64_t X = (Denominator > 0) ? -1 : 1;
+  bool SameSign = (Numerator > 0) == (Denominator > 0);
+  return SameSign ? Numerator / Denominator
+                  : -((-Numerator + X) / Denominator) - 1;
+}
+
+/// Returns the remainder of the Euclidean division of LHS by RHS. Result is
+/// always non-negative.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t mod(int64_t Numerator,
+                                         int64_t Denominator) {
+  assert(Denominator >= 1);
+  return Numerator % Denominator < 0 ? Numerator % Denominator + Denominator
+                                     : Numerator % Denominator;
+}
+
 /// Returns the integer nearest(Numerator / Denominator).
 inline uint64_t divideNearest(uint64_t Numerator, uint64_t Denominator) {
   return (Numerator + (Denominator / 2)) / Denominator;
diff --git a/llvm/unittests/Support/MathExtrasTest.cpp b/llvm/unittests/Support/MathExtrasTest.cpp
index 67239a24c47ad..b32094e0acdfe 100644
--- a/llvm/unittests/Support/MathExtrasTest.cpp
+++ b/llvm/unittests/Support/MathExtrasTest.cpp
@@ -434,8 +434,25 @@ TEST(MathExtras, IsShiftedInt) {
   EXPECT_FALSE((isShiftedInt<6, 10>(int64_t(1) << 15)));
 }
 
-template <typename T>
-class OverflowTest : public ::testing::Test { };
+TEST(MathExtras, CeilDivTest) {
+  EXPECT_EQ(ceilDiv(14, 3), 5);
+  EXPECT_EQ(ceilDiv(14, -3), -4);
+  EXPECT_EQ(ceilDiv(-14, -3), 5);
+  EXPECT_EQ(ceilDiv(-14, 3), -4);
+  EXPECT_EQ(ceilDiv(0, 3), 0);
+  EXPECT_EQ(ceilDiv(0, -3), 0);
+}
+
+TEST(MathExtras, FloorDivTest) {
+  EXPECT_EQ(floorDiv(14, 3), 4);
+  EXPECT_EQ(floorDiv(14, -3), -5);
+  EXPECT_EQ(floorDiv(-14, -3), 4);
+  EXPECT_EQ(floorDiv(-14, 3), -5);
+  EXPECT_EQ(floorDiv(0, 3), 0);
+  EXPECT_EQ(floorDiv(0, -3), 0);
+}
+
+template <typename T> class OverflowTest : public ::testing::Test {};
 
 using OverflowTestTypes = ::testing::Types<signed char, short, int, long,
                                            long long>;
@@ -560,5 +577,4 @@ TYPED_TEST(OverflowTest, MulResultZero) {
   EXPECT_FALSE(MulOverflow<TypeParam>(0, -5, Result));
   EXPECT_EQ(Result, TypeParam(0));
 }
-
 } // namespace
diff --git a/mlir/include/mlir/Analysis/Presburger/Fraction.h b/mlir/include/mlir/Analysis/Presburger/Fraction.h
index c07bb767f50bf..f76ba3f006d07 100644
--- a/mlir/include/mlir/Analysis/Presburger/Fraction.h
+++ b/mlir/include/mlir/Analysis/Presburger/Fraction.h
@@ -15,7 +15,6 @@
 #define MLIR_ANALYSIS_PRESBURGER_FRACTION_H
 
 #include "mlir/Analysis/Presburger/MPInt.h"
-#include "mlir/Support/MathExtras.h"
 
 namespace mlir {
 namespace presburger {
diff --git a/mlir/include/mlir/Analysis/Presburger/MPInt.h b/mlir/include/mlir/Analysis/Presburger/MPInt.h
index 12ab0598d10d9..343f50251c7dc 100644
--- a/mlir/include/mlir/Analysis/Presburger/MPInt.h
+++ b/mlir/include/mlir/Analysis/Presburger/MPInt.h
@@ -17,7 +17,8 @@
 #define MLIR_ANALYSIS_PRESBURGER_MPINT_H
 
 #include "mlir/Analysis/Presburger/SlowMPInt.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/ADT/ArrayRef.h"
+#include "llvm/Support/MathExtras.h"
 #include "llvm/Support/raw_ostream.h"
 #include <numeric>
 
@@ -31,9 +32,10 @@ namespace presburger {
 /// from the mlir::presburger namespace. So to access the 64-bit overloads, an
 /// explict call to mlir::ceilDiv would be required. These using declarations
 /// allow overload resolution to transparently call the right function.
-using ::mlir::ceilDiv;
-using ::mlir::floorDiv;
-using ::mlir::mod;
+using ::llvm::ArrayRef;
+using ::llvm::ceilDiv;
+using ::llvm::floorDiv;
+using ::llvm::mod;
 
 namespace detail {
 /// If builtin intrinsics for overflow-checked arithmetic are available,
diff --git a/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h b/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
index c70306761c549..482581c573cea 100644
--- a/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
+++ b/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
@@ -18,7 +18,6 @@
 #ifndef MLIR_ANALYSIS_PRESBURGER_SLOWMPINT_H
 #define MLIR_ANALYSIS_PRESBURGER_SLOWMPINT_H
 
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/APInt.h"
 #include "llvm/ADT/Hashing.h"
 #include "llvm/Support/raw_ostream.h"
diff --git a/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h b/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
index 7a24c201a39a7..0e39928b1d6a9 100644
--- a/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
+++ b/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
@@ -17,7 +17,7 @@
 #include "mlir/IR/SymbolTable.h"
 #include "mlir/Interfaces/InferTypeOpInterface.h"
 #include "mlir/Interfaces/SideEffectInterfaces.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 namespace mlir {
 namespace mesh {
@@ -114,7 +114,7 @@ inline int64_t shardDimension(int64_t dimSize, int64_t shardCount) {
     return ShapedType::kDynamic;
 
   assert(dimSize % shardCount == 0);
-  return ceilDiv(dimSize, shardCount);
+  return llvm::ceilDiv(dimSize, shardCount);
 }
 
 // Get the size of an unsharded dimension.
diff --git a/mlir/include/mlir/Support/MathExtras.h b/mlir/include/mlir/Support/MathExtras.h
deleted file mode 100644
index 17a747393d26e..0000000000000
--- a/mlir/include/mlir/Support/MathExtras.h
+++ /dev/null
@@ -1,51 +0,0 @@
-//===- MathExtras.h - Math functions relevant to MLIR -----------*- C++ -*-===//
-//
-// 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
-//
-//===----------------------------------------------------------------------===//
-//
-// This file contains math functions relevant to MLIR.
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef MLIR_SUPPORT_MATHEXTRAS_H_
-#define MLIR_SUPPORT_MATHEXTRAS_H_
-
-#include "mlir/Support/LLVM.h"
-#include "llvm/ADT/APInt.h"
-
-namespace mlir {
-
-/// Returns the result of MLIR's ceildiv operation on constants. The RHS is
-/// expected to be non-zero.
-inline int64_t ceilDiv(int64_t lhs, int64_t rhs) {
-  assert(rhs != 0);
-  // C/C++'s integer division rounds towards 0.
-  int64_t x = (rhs > 0) ? -1 : 1;
-  return ((lhs != 0) && (lhs > 0) == (rhs > 0)) ? ((lhs + x) / rhs) + 1
-                                                : -(-lhs / rhs);
-}
-
-/// Returns the result of MLIR's floordiv operation on constants. The RHS is
-/// expected to be non-zero.
-inline int64_t floorDiv(int64_t lhs, int64_t rhs) {
-  assert(rhs != 0);
-  // C/C++'s integer division rounds towards 0.
-  int64_t x = (rhs < 0) ? 1 : -1;
-  return ((lhs != 0) && ((lhs < 0) != (rhs < 0))) ? -((-lhs + x) / rhs) - 1
-                                                  : lhs / rhs;
-}
-
-/// Returns MLIR's mod operation on constants. MLIR's mod operation yields the
-/// remainder of the Euclidean division of 'lhs' by 'rhs', and is therefore not
-/// C's % operator.  The RHS is always expected to be positive, and the result
-/// is always non-negative.
-inline int64_t mod(int64_t lhs, int64_t rhs) {
-  assert(rhs >= 1);
-  return lhs % rhs < 0 ? lhs % rhs + rhs : lhs % rhs;
-}
-} // namespace mlir
-
-#endif // MLIR_SUPPORT_MATHEXTRAS_H_
diff --git a/mlir/lib/Analysis/FlatLinearValueConstraints.cpp b/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
index bf7e3121ea32a..5c4f353f310d6 100644
--- a/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
+++ b/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
@@ -16,7 +16,6 @@
 #include "mlir/IR/Builders.h"
 #include "mlir/IR/IntegerSet.h"
 #include "mlir/Support/LLVM.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp b/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
index 94b7c8d4f2fdf..744236692fbb6 100644
--- a/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
+++ b/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
@@ -36,7 +36,6 @@
 #include "mlir/IR/SymbolTable.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/Support/LogicalResult.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/DialectConversion.h"
 #include "mlir/Transforms/Passes.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp b/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
index da084b89ceadc..0910a1ea076c0 100644
--- a/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
+++ b/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
@@ -12,7 +12,7 @@
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/Dialect/LLVMIR/LLVMTypes.h"
 #include "mlir/IR/Builders.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 using namespace mlir;
 
@@ -363,9 +363,9 @@ void UnrankedMemRefDescriptor::computeSizes(
   // Initialize shared constants.
   Value one = createIndexAttrConstant(builder, loc, indexType, 1);
   Value two = createIndexAttrConstant(builder, loc, indexType, 2);
-  Value indexSize =
-      createIndexAttrConstant(builder, loc, indexType,
-                              ceilDiv(typeConverter.getIndexTypeBitwidth(), 8));
+  Value indexSize = createIndexAttrConstant(
+      builder, loc, indexType,
+      llvm::ceilDiv(typeConverter.getIndexTypeBitwidth(), 8));
 
   sizes.reserve(sizes.size() + values.size());
   for (auto [desc, addressSpace] : llvm::zip(values, addressSpaces)) {
@@ -378,7 +378,7 @@ void UnrankedMemRefDescriptor::computeSizes(
     // to data layout) into the unranked descriptor.
     Value pointerSize = createIndexAttrConstant(
         builder, loc, indexType,
-        ceilDiv(typeConverter.getPointerBitwidth(addressSpace), 8));
+        llvm::ceilDiv(typeConverter.getPointerBitwidth(addressSpace), 8));
     Value doublePointerSize =
         builder.create<LLVM::MulOp>(loc, indexType, two, pointerSize);
 
diff --git a/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp b/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
index 2dc42f0a85e66..d7eab77b97d0c 100644
--- a/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
+++ b/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
@@ -25,8 +25,8 @@
 #include "mlir/IR/BuiltinTypes.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/Pass/Pass.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/SmallBitVector.h"
+#include "llvm/Support/MathExtras.h"
 #include <optional>
 
 namespace mlir {
@@ -971,8 +971,8 @@ struct MemorySpaceCastOpLowering
                                resultUnderlyingDesc, resultElemPtrType);
 
       int64_t bytesToSkip =
-          2 *
-          ceilDiv(getTypeConverter()->getPointerBitwidth(resultAddrSpace), 8);
+          2 * llvm::ceilDiv(
+                  getTypeConverter()->getPointerBitwidth(resultAddrSpace), 8);
       Value bytesToSkipConst = rewriter.create<LLVM::ConstantOp>(
           loc, getIndexType(), rewriter.getIndexAttr(bytesToSkip));
       Value copySize = rewriter.create<LLVM::SubOp>(
diff --git a/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp b/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
index 24427c25d79ff..034484af83b79 100644
--- a/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
+++ b/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
@@ -21,7 +21,6 @@
 #include "mlir/IR/AffineExprVisitor.h"
 #include "mlir/IR/IntegerSet.h"
 #include "mlir/Support/LLVM.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp b/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
index 1c28d6b00b3c8..11a9f1485514a 100644
--- a/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
+++ b/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
@@ -18,7 +18,7 @@
 #include "mlir/Dialect/Affine/Analysis/NestedMatcher.h"
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Affine/IR/AffineValueMap.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 #include "llvm/ADT/DenseSet.h"
 #include "llvm/ADT/SmallPtrSet.h"
@@ -47,7 +47,8 @@ void mlir::affine::getTripCountMapAndOperands(
     loopSpan = ub - lb;
     if (loopSpan < 0)
       loopSpan = 0;
-    *tripCountMap = AffineMap::getConstantMap(ceilDiv(loopSpan, step), context);
+    *tripCountMap =
+        AffineMap::getConstantMap(llvm::ceilDiv(loopSpan, step), context);
     tripCountOperands->clear();
     return;
   }
diff --git a/mlir/lib/Dialect/Affine/IR/AffineOps.cpp b/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
index 0a58d2fdb02f5..b52bc05e908fc 100644
--- a/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
+++ b/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
@@ -18,19 +18,23 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Interfaces/ShapedOpInterfaces.h"
 #include "mlir/Interfaces/ValueBoundsOpInterface.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/ScopeExit.h"
 #include "llvm/ADT/SmallBitVector.h"
 #include "llvm/ADT/SmallVectorExtras.h"
 #include "llvm/ADT/TypeSwitch.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 #include <numeric>
 #include <optional>
 
 using namespace mlir;
 using namespace mlir::affine;
 
+using llvm::ceilDiv;
+using llvm::floorDiv;
+using llvm::mod;
+
 #define DEBUG_TYPE "affine-ops"
 
 #include "mlir/Dialect/Affine/IR/AffineOpsDialect.cpp.inc"
diff --git a/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp b/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
index 268050a30e002..c934f229bb6c4 100644
--- a/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
+++ b/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
@@ -23,7 +23,6 @@
 #include "mlir/Dialect/SCF/IR/SCF.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/IR/IntegerSet.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/GreedyPatternRewriteDriver.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
diff --git a/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp b/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
index 1320db3f9e543..ede158db911ea 100644
--- a/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
+++ b/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
@@ -23,7 +23,6 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/IR/Value.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/APFloat.h"
 #include "llvm/ADT/STLExtras.h"
diff --git a/mlir/lib/Dialect/Func/IR/FuncOps.cpp b/mlir/lib/Dialect/Func/IR/FuncOps.cpp
index 95589e8989e27..c719981769b9e 100644
--- a/mlir/lib/Dialect/Func/IR/FuncOps.cpp
+++ b/mlir/lib/Dialect/Func/IR/FuncOps.cpp
@@ -19,7 +19,6 @@
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/IR/Value.h"
 #include "mlir/Interfaces/FunctionImplementation.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/APFloat.h"
 #include "llvm/ADT/MapVector.h"
diff --git a/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp b/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
index 38abdd122d633..24763c036192f 100644
--- a/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
+++ b/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
@@ -9,11 +9,11 @@
 #include "mlir/Dialect/Linalg/TransformOps/GPUHeuristics.h"
 
 #include "mlir/Dialect/GPU/IR/GPUDialect.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 #include "llvm/Support/raw_ostream.h"
 #include <cmath>
 #include <numeric>
@@ -76,7 +76,7 @@ transform::gpu::CopyMappingInfo::CopyMappingInfo(MLIRContext *ctx,
       llvm::map_range(llvm::zip(copySizes, this->numThreads), [](auto &&pair) {
         int64_t size, numThreads;
         std::tie(size, numThreads) = pair;
-        return mlir::ceilDiv(size, numThreads);
+        return llvm::ceilDiv(size, numThreads);
       }));
   SmallVector<Attribute> allThreadMappings{linearId2(ctx), linearId1(ctx),
                                            linearId0(ctx)};
diff --git a/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp b/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
index 2392e10522dd0..4c49ebb3b4cd7 100644
--- a/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
+++ b/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
@@ -20,9 +20,9 @@
 #include "mlir/IR/BuiltinTypes.h"
 #include "mlir/IR/OpDefinition.h"
 #include "mlir/Support/LogicalResult.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/DialectConversion.h"
 #include "llvm/Support/FormatVariadic.h"
+#include "llvm/Support/MathExtras.h"
 #include <cassert>
 #include <type_traits>
 
@@ -74,7 +74,7 @@ convertCastingOp(ConversionPatternRewriter &rewriter,
 
   SmallVector<int64_t> size;
   if (sizes.size())
-    size.push_back(ceilDiv(sizes[0], elementsPerByte));
+    size.push_back(llvm::ceilDiv(sizes[0], elementsPerByte));
   offset = offset / elementsPerByte;
 
   rewriter.replaceOpWithNewOp<memref::ReinterpretCastOp>(
diff --git a/mlir/lib/Dialect/SCF/IR/SCF.cpp b/mlir/lib/Dialect/SCF/IR/SCF.cpp
index 5e94f4dc612a7..524711dbbb6fd 100644
--- a/mlir/lib/Dialect/SCF/IR/SCF.cpp
+++ b/mlir/lib/Dialect/SCF/IR/SCF.cpp
@@ -21,7 +21,6 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Interfaces/FunctionInterfaces.h"
 #include "mlir/Interfaces/ValueBoundsOpInterface.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
diff --git a/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp b/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
index 82ec95d31f525..f5b7a6e28a178 100644
--- a/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
+++ b/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
@@ -17,10 +17,10 @@
 #include "mlir/Dialect/SCF/Utils/Utils.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/IR/PatternMatch.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/RegionUtils.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 
 #define DEBUG_TYPE "scf-loop-pipelining"
 #define DBGS() (llvm::dbgs() << "[" DEBUG_TYPE "]: ")
@@ -119,7 +119,7 @@ bool LoopPipelinerInternal::initializeLoopInfo(
     int64_t ubImm = upperBoundCst.value();
     int64_t lbImm = lowerBoundCst.value();
     int64_t stepImm = stepCst.value();
-    int64_t numIteration = ceilDiv(ubImm - lbImm, stepImm);
+    int64_t numIteration = llvm::ceilDiv(ubImm - lbImm, stepImm);
     if (numIteration > maxStage) {
       dynamicLoop = false;
     } else if (!options.supportDynamicLoops) {
diff --git a/mlir/lib/Dialect/SCF/Utils/Utils.cpp b/mlir/lib/Dialect/SCF/Utils/Utils.cpp
index 6658cca03eba7....
[truncated]

llvmbot · 2024-06-11T09:12:05Z

@llvm/pr-subscribers-mlir-llvm

Author: Ramkumar Ramachandra (artagnon)

Changes

This patch is part of a project to move the Presburger library into LLVM.

Patch is 27.71 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/95087.diff

30 Files Affected:

(modified) llvm/include/llvm/Support/MathExtras.h (+38-2)
(modified) llvm/unittests/Support/MathExtrasTest.cpp (+19-3)
(modified) mlir/include/mlir/Analysis/Presburger/Fraction.h (-1)
(modified) mlir/include/mlir/Analysis/Presburger/MPInt.h (+6-4)
(modified) mlir/include/mlir/Analysis/Presburger/SlowMPInt.h (-1)
(modified) mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h (+2-2)
(removed) mlir/include/mlir/Support/MathExtras.h (-51)
(modified) mlir/lib/Analysis/FlatLinearValueConstraints.cpp (-1)
(modified) mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp (-1)
(modified) mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp (+5-5)
(modified) mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp (+3-3)
(modified) mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp (-1)
(modified) mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp (+3-2)
(modified) mlir/lib/Dialect/Affine/IR/AffineOps.cpp (+5-1)
(modified) mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp (-1)
(modified) mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp (-1)
(modified) mlir/lib/Dialect/Func/IR/FuncOps.cpp (-1)
(modified) mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp (+2-2)
(modified) mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp (+2-2)
(modified) mlir/lib/Dialect/SCF/IR/SCF.cpp (-1)
(modified) mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp (+2-2)
(modified) mlir/lib/Dialect/SCF/Utils/Utils.cpp (+2-2)
(modified) mlir/lib/Dialect/Tensor/IR/TensorOps.cpp (+5-1)
(modified) mlir/lib/Dialect/Utils/StaticValueUtils.cpp (+2-2)
(modified) mlir/lib/Dialect/Vector/Transforms/VectorUnroll.cpp (-1)
(modified) mlir/lib/Dialect/Vector/Utils/VectorUtils.cpp (-1)
(modified) mlir/lib/IR/AffineExpr.cpp (+7-3)
(modified) mlir/lib/IR/AffineMap.cpp (+5-1)
(modified) mlir/unittests/Support/CMakeLists.txt (-1)
(removed) mlir/unittests/Support/MathExtrasTest.cpp (-31)

diff --git a/llvm/include/llvm/Support/MathExtras.h b/llvm/include/llvm/Support/MathExtras.h
index f0e4ee534ece3..c5cac51130b1d 100644
--- a/llvm/include/llvm/Support/MathExtras.h
+++ b/llvm/include/llvm/Support/MathExtras.h
@@ -424,11 +424,47 @@ template <uint64_t Align> constexpr inline uint64_t alignTo(uint64_t Value) {
   return (Value + Align - 1) / Align * Align;
 }
 
-/// Returns the integer ceil(Numerator / Denominator).
-inline uint64_t divideCeil(uint64_t Numerator, uint64_t Denominator) {
+/// Returns the integer ceil(Numerator / Denominator). Unsigned integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE uint64_t divideCeil(uint64_t Numerator,
+                                                 uint64_t Denominator) {
   return alignTo(Numerator, Denominator) / Denominator;
 }
 
+/// Returns the integer ceil(Numerator / Denominator). Signed integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t ceilDiv(int64_t Numerator,
+                                             int64_t Denominator) {
+  assert(Denominator);
+  if (!Numerator)
+    return 0;
+  // C's integer division rounds towards 0.
+  int64_t X = (Denominator > 0) ? -1 : 1;
+  bool SameSign = (Numerator > 0) == (Denominator > 0);
+  return SameSign ? ((Numerator + X) / Denominator) + 1
+                  : -(-Numerator / Denominator);
+}
+
+/// Returns the integer floor(Numerator / Denominator). Signed integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t floorDiv(int64_t Numerator,
+                                              int64_t Denominator) {
+  assert(Denominator);
+  if (!Numerator)
+    return 0;
+  // C's integer division rounds towards 0.
+  int64_t X = (Denominator > 0) ? -1 : 1;
+  bool SameSign = (Numerator > 0) == (Denominator > 0);
+  return SameSign ? Numerator / Denominator
+                  : -((-Numerator + X) / Denominator) - 1;
+}
+
+/// Returns the remainder of the Euclidean division of LHS by RHS. Result is
+/// always non-negative.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t mod(int64_t Numerator,
+                                         int64_t Denominator) {
+  assert(Denominator >= 1);
+  return Numerator % Denominator < 0 ? Numerator % Denominator + Denominator
+                                     : Numerator % Denominator;
+}
+
 /// Returns the integer nearest(Numerator / Denominator).
 inline uint64_t divideNearest(uint64_t Numerator, uint64_t Denominator) {
   return (Numerator + (Denominator / 2)) / Denominator;
diff --git a/llvm/unittests/Support/MathExtrasTest.cpp b/llvm/unittests/Support/MathExtrasTest.cpp
index 67239a24c47ad..b32094e0acdfe 100644
--- a/llvm/unittests/Support/MathExtrasTest.cpp
+++ b/llvm/unittests/Support/MathExtrasTest.cpp
@@ -434,8 +434,25 @@ TEST(MathExtras, IsShiftedInt) {
   EXPECT_FALSE((isShiftedInt<6, 10>(int64_t(1) << 15)));
 }
 
-template <typename T>
-class OverflowTest : public ::testing::Test { };
+TEST(MathExtras, CeilDivTest) {
+  EXPECT_EQ(ceilDiv(14, 3), 5);
+  EXPECT_EQ(ceilDiv(14, -3), -4);
+  EXPECT_EQ(ceilDiv(-14, -3), 5);
+  EXPECT_EQ(ceilDiv(-14, 3), -4);
+  EXPECT_EQ(ceilDiv(0, 3), 0);
+  EXPECT_EQ(ceilDiv(0, -3), 0);
+}
+
+TEST(MathExtras, FloorDivTest) {
+  EXPECT_EQ(floorDiv(14, 3), 4);
+  EXPECT_EQ(floorDiv(14, -3), -5);
+  EXPECT_EQ(floorDiv(-14, -3), 4);
+  EXPECT_EQ(floorDiv(-14, 3), -5);
+  EXPECT_EQ(floorDiv(0, 3), 0);
+  EXPECT_EQ(floorDiv(0, -3), 0);
+}
+
+template <typename T> class OverflowTest : public ::testing::Test {};
 
 using OverflowTestTypes = ::testing::Types<signed char, short, int, long,
                                            long long>;
@@ -560,5 +577,4 @@ TYPED_TEST(OverflowTest, MulResultZero) {
   EXPECT_FALSE(MulOverflow<TypeParam>(0, -5, Result));
   EXPECT_EQ(Result, TypeParam(0));
 }
-
 } // namespace
diff --git a/mlir/include/mlir/Analysis/Presburger/Fraction.h b/mlir/include/mlir/Analysis/Presburger/Fraction.h
index c07bb767f50bf..f76ba3f006d07 100644
--- a/mlir/include/mlir/Analysis/Presburger/Fraction.h
+++ b/mlir/include/mlir/Analysis/Presburger/Fraction.h
@@ -15,7 +15,6 @@
 #define MLIR_ANALYSIS_PRESBURGER_FRACTION_H
 
 #include "mlir/Analysis/Presburger/MPInt.h"
-#include "mlir/Support/MathExtras.h"
 
 namespace mlir {
 namespace presburger {
diff --git a/mlir/include/mlir/Analysis/Presburger/MPInt.h b/mlir/include/mlir/Analysis/Presburger/MPInt.h
index 12ab0598d10d9..343f50251c7dc 100644
--- a/mlir/include/mlir/Analysis/Presburger/MPInt.h
+++ b/mlir/include/mlir/Analysis/Presburger/MPInt.h
@@ -17,7 +17,8 @@
 #define MLIR_ANALYSIS_PRESBURGER_MPINT_H
 
 #include "mlir/Analysis/Presburger/SlowMPInt.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/ADT/ArrayRef.h"
+#include "llvm/Support/MathExtras.h"
 #include "llvm/Support/raw_ostream.h"
 #include <numeric>
 
@@ -31,9 +32,10 @@ namespace presburger {
 /// from the mlir::presburger namespace. So to access the 64-bit overloads, an
 /// explict call to mlir::ceilDiv would be required. These using declarations
 /// allow overload resolution to transparently call the right function.
-using ::mlir::ceilDiv;
-using ::mlir::floorDiv;
-using ::mlir::mod;
+using ::llvm::ArrayRef;
+using ::llvm::ceilDiv;
+using ::llvm::floorDiv;
+using ::llvm::mod;
 
 namespace detail {
 /// If builtin intrinsics for overflow-checked arithmetic are available,
diff --git a/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h b/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
index c70306761c549..482581c573cea 100644
--- a/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
+++ b/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
@@ -18,7 +18,6 @@
 #ifndef MLIR_ANALYSIS_PRESBURGER_SLOWMPINT_H
 #define MLIR_ANALYSIS_PRESBURGER_SLOWMPINT_H
 
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/APInt.h"
 #include "llvm/ADT/Hashing.h"
 #include "llvm/Support/raw_ostream.h"
diff --git a/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h b/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
index 7a24c201a39a7..0e39928b1d6a9 100644
--- a/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
+++ b/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
@@ -17,7 +17,7 @@
 #include "mlir/IR/SymbolTable.h"
 #include "mlir/Interfaces/InferTypeOpInterface.h"
 #include "mlir/Interfaces/SideEffectInterfaces.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 namespace mlir {
 namespace mesh {
@@ -114,7 +114,7 @@ inline int64_t shardDimension(int64_t dimSize, int64_t shardCount) {
     return ShapedType::kDynamic;
 
   assert(dimSize % shardCount == 0);
-  return ceilDiv(dimSize, shardCount);
+  return llvm::ceilDiv(dimSize, shardCount);
 }
 
 // Get the size of an unsharded dimension.
diff --git a/mlir/include/mlir/Support/MathExtras.h b/mlir/include/mlir/Support/MathExtras.h
deleted file mode 100644
index 17a747393d26e..0000000000000
--- a/mlir/include/mlir/Support/MathExtras.h
+++ /dev/null
@@ -1,51 +0,0 @@
-//===- MathExtras.h - Math functions relevant to MLIR -----------*- C++ -*-===//
-//
-// 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
-//
-//===----------------------------------------------------------------------===//
-//
-// This file contains math functions relevant to MLIR.
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef MLIR_SUPPORT_MATHEXTRAS_H_
-#define MLIR_SUPPORT_MATHEXTRAS_H_
-
-#include "mlir/Support/LLVM.h"
-#include "llvm/ADT/APInt.h"
-
-namespace mlir {
-
-/// Returns the result of MLIR's ceildiv operation on constants. The RHS is
-/// expected to be non-zero.
-inline int64_t ceilDiv(int64_t lhs, int64_t rhs) {
-  assert(rhs != 0);
-  // C/C++'s integer division rounds towards 0.
-  int64_t x = (rhs > 0) ? -1 : 1;
-  return ((lhs != 0) && (lhs > 0) == (rhs > 0)) ? ((lhs + x) / rhs) + 1
-                                                : -(-lhs / rhs);
-}
-
-/// Returns the result of MLIR's floordiv operation on constants. The RHS is
-/// expected to be non-zero.
-inline int64_t floorDiv(int64_t lhs, int64_t rhs) {
-  assert(rhs != 0);
-  // C/C++'s integer division rounds towards 0.
-  int64_t x = (rhs < 0) ? 1 : -1;
-  return ((lhs != 0) && ((lhs < 0) != (rhs < 0))) ? -((-lhs + x) / rhs) - 1
-                                                  : lhs / rhs;
-}
-
-/// Returns MLIR's mod operation on constants. MLIR's mod operation yields the
-/// remainder of the Euclidean division of 'lhs' by 'rhs', and is therefore not
-/// C's % operator.  The RHS is always expected to be positive, and the result
-/// is always non-negative.
-inline int64_t mod(int64_t lhs, int64_t rhs) {
-  assert(rhs >= 1);
-  return lhs % rhs < 0 ? lhs % rhs + rhs : lhs % rhs;
-}
-} // namespace mlir
-
-#endif // MLIR_SUPPORT_MATHEXTRAS_H_
diff --git a/mlir/lib/Analysis/FlatLinearValueConstraints.cpp b/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
index bf7e3121ea32a..5c4f353f310d6 100644
--- a/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
+++ b/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
@@ -16,7 +16,6 @@
 #include "mlir/IR/Builders.h"
 #include "mlir/IR/IntegerSet.h"
 #include "mlir/Support/LLVM.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp b/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
index 94b7c8d4f2fdf..744236692fbb6 100644
--- a/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
+++ b/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
@@ -36,7 +36,6 @@
 #include "mlir/IR/SymbolTable.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/Support/LogicalResult.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/DialectConversion.h"
 #include "mlir/Transforms/Passes.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp b/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
index da084b89ceadc..0910a1ea076c0 100644
--- a/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
+++ b/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
@@ -12,7 +12,7 @@
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/Dialect/LLVMIR/LLVMTypes.h"
 #include "mlir/IR/Builders.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 using namespace mlir;
 
@@ -363,9 +363,9 @@ void UnrankedMemRefDescriptor::computeSizes(
   // Initialize shared constants.
   Value one = createIndexAttrConstant(builder, loc, indexType, 1);
   Value two = createIndexAttrConstant(builder, loc, indexType, 2);
-  Value indexSize =
-      createIndexAttrConstant(builder, loc, indexType,
-                              ceilDiv(typeConverter.getIndexTypeBitwidth(), 8));
+  Value indexSize = createIndexAttrConstant(
+      builder, loc, indexType,
+      llvm::ceilDiv(typeConverter.getIndexTypeBitwidth(), 8));
 
   sizes.reserve(sizes.size() + values.size());
   for (auto [desc, addressSpace] : llvm::zip(values, addressSpaces)) {
@@ -378,7 +378,7 @@ void UnrankedMemRefDescriptor::computeSizes(
     // to data layout) into the unranked descriptor.
     Value pointerSize = createIndexAttrConstant(
         builder, loc, indexType,
-        ceilDiv(typeConverter.getPointerBitwidth(addressSpace), 8));
+        llvm::ceilDiv(typeConverter.getPointerBitwidth(addressSpace), 8));
     Value doublePointerSize =
         builder.create<LLVM::MulOp>(loc, indexType, two, pointerSize);
 
diff --git a/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp b/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
index 2dc42f0a85e66..d7eab77b97d0c 100644
--- a/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
+++ b/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
@@ -25,8 +25,8 @@
 #include "mlir/IR/BuiltinTypes.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/Pass/Pass.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/SmallBitVector.h"
+#include "llvm/Support/MathExtras.h"
 #include <optional>
 
 namespace mlir {
@@ -971,8 +971,8 @@ struct MemorySpaceCastOpLowering
                                resultUnderlyingDesc, resultElemPtrType);
 
       int64_t bytesToSkip =
-          2 *
-          ceilDiv(getTypeConverter()->getPointerBitwidth(resultAddrSpace), 8);
+          2 * llvm::ceilDiv(
+                  getTypeConverter()->getPointerBitwidth(resultAddrSpace), 8);
       Value bytesToSkipConst = rewriter.create<LLVM::ConstantOp>(
           loc, getIndexType(), rewriter.getIndexAttr(bytesToSkip));
       Value copySize = rewriter.create<LLVM::SubOp>(
diff --git a/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp b/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
index 24427c25d79ff..034484af83b79 100644
--- a/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
+++ b/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
@@ -21,7 +21,6 @@
 #include "mlir/IR/AffineExprVisitor.h"
 #include "mlir/IR/IntegerSet.h"
 #include "mlir/Support/LLVM.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp b/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
index 1c28d6b00b3c8..11a9f1485514a 100644
--- a/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
+++ b/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
@@ -18,7 +18,7 @@
 #include "mlir/Dialect/Affine/Analysis/NestedMatcher.h"
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Affine/IR/AffineValueMap.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 #include "llvm/ADT/DenseSet.h"
 #include "llvm/ADT/SmallPtrSet.h"
@@ -47,7 +47,8 @@ void mlir::affine::getTripCountMapAndOperands(
     loopSpan = ub - lb;
     if (loopSpan < 0)
       loopSpan = 0;
-    *tripCountMap = AffineMap::getConstantMap(ceilDiv(loopSpan, step), context);
+    *tripCountMap =
+        AffineMap::getConstantMap(llvm::ceilDiv(loopSpan, step), context);
     tripCountOperands->clear();
     return;
   }
diff --git a/mlir/lib/Dialect/Affine/IR/AffineOps.cpp b/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
index 0a58d2fdb02f5..b52bc05e908fc 100644
--- a/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
+++ b/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
@@ -18,19 +18,23 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Interfaces/ShapedOpInterfaces.h"
 #include "mlir/Interfaces/ValueBoundsOpInterface.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/ScopeExit.h"
 #include "llvm/ADT/SmallBitVector.h"
 #include "llvm/ADT/SmallVectorExtras.h"
 #include "llvm/ADT/TypeSwitch.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 #include <numeric>
 #include <optional>
 
 using namespace mlir;
 using namespace mlir::affine;
 
+using llvm::ceilDiv;
+using llvm::floorDiv;
+using llvm::mod;
+
 #define DEBUG_TYPE "affine-ops"
 
 #include "mlir/Dialect/Affine/IR/AffineOpsDialect.cpp.inc"
diff --git a/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp b/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
index 268050a30e002..c934f229bb6c4 100644
--- a/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
+++ b/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
@@ -23,7 +23,6 @@
 #include "mlir/Dialect/SCF/IR/SCF.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/IR/IntegerSet.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/GreedyPatternRewriteDriver.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
diff --git a/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp b/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
index 1320db3f9e543..ede158db911ea 100644
--- a/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
+++ b/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
@@ -23,7 +23,6 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/IR/Value.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/APFloat.h"
 #include "llvm/ADT/STLExtras.h"
diff --git a/mlir/lib/Dialect/Func/IR/FuncOps.cpp b/mlir/lib/Dialect/Func/IR/FuncOps.cpp
index 95589e8989e27..c719981769b9e 100644
--- a/mlir/lib/Dialect/Func/IR/FuncOps.cpp
+++ b/mlir/lib/Dialect/Func/IR/FuncOps.cpp
@@ -19,7 +19,6 @@
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/IR/Value.h"
 #include "mlir/Interfaces/FunctionImplementation.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/APFloat.h"
 #include "llvm/ADT/MapVector.h"
diff --git a/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp b/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
index 38abdd122d633..24763c036192f 100644
--- a/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
+++ b/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
@@ -9,11 +9,11 @@
 #include "mlir/Dialect/Linalg/TransformOps/GPUHeuristics.h"
 
 #include "mlir/Dialect/GPU/IR/GPUDialect.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 #include "llvm/Support/raw_ostream.h"
 #include <cmath>
 #include <numeric>
@@ -76,7 +76,7 @@ transform::gpu::CopyMappingInfo::CopyMappingInfo(MLIRContext *ctx,
       llvm::map_range(llvm::zip(copySizes, this->numThreads), [](auto &&pair) {
         int64_t size, numThreads;
         std::tie(size, numThreads) = pair;
-        return mlir::ceilDiv(size, numThreads);
+        return llvm::ceilDiv(size, numThreads);
       }));
   SmallVector<Attribute> allThreadMappings{linearId2(ctx), linearId1(ctx),
                                            linearId0(ctx)};
diff --git a/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp b/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
index 2392e10522dd0..4c49ebb3b4cd7 100644
--- a/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
+++ b/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
@@ -20,9 +20,9 @@
 #include "mlir/IR/BuiltinTypes.h"
 #include "mlir/IR/OpDefinition.h"
 #include "mlir/Support/LogicalResult.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/DialectConversion.h"
 #include "llvm/Support/FormatVariadic.h"
+#include "llvm/Support/MathExtras.h"
 #include <cassert>
 #include <type_traits>
 
@@ -74,7 +74,7 @@ convertCastingOp(ConversionPatternRewriter &rewriter,
 
   SmallVector<int64_t> size;
   if (sizes.size())
-    size.push_back(ceilDiv(sizes[0], elementsPerByte));
+    size.push_back(llvm::ceilDiv(sizes[0], elementsPerByte));
   offset = offset / elementsPerByte;
 
   rewriter.replaceOpWithNewOp<memref::ReinterpretCastOp>(
diff --git a/mlir/lib/Dialect/SCF/IR/SCF.cpp b/mlir/lib/Dialect/SCF/IR/SCF.cpp
index 5e94f4dc612a7..524711dbbb6fd 100644
--- a/mlir/lib/Dialect/SCF/IR/SCF.cpp
+++ b/mlir/lib/Dialect/SCF/IR/SCF.cpp
@@ -21,7 +21,6 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Interfaces/FunctionInterfaces.h"
 #include "mlir/Interfaces/ValueBoundsOpInterface.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
diff --git a/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp b/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
index 82ec95d31f525..f5b7a6e28a178 100644
--- a/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
+++ b/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
@@ -17,10 +17,10 @@
 #include "mlir/Dialect/SCF/Utils/Utils.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/IR/PatternMatch.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/RegionUtils.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 
 #define DEBUG_TYPE "scf-loop-pipelining"
 #define DBGS() (llvm::dbgs() << "[" DEBUG_TYPE "]: ")
@@ -119,7 +119,7 @@ bool LoopPipelinerInternal::initializeLoopInfo(
     int64_t ubImm = upperBoundCst.value();
     int64_t lbImm = lowerBoundCst.value();
     int64_t stepImm = stepCst.value();
-    int64_t numIteration = ceilDiv(ubImm - lbImm, stepImm);
+    int64_t numIteration = llvm::ceilDiv(ubImm - lbImm, stepImm);
     if (numIteration > maxStage) {
       dynamicLoop = false;
     } else if (!options.supportDynamicLoops) {
diff --git a/mlir/lib/Dialect/SCF/Utils/Utils.cpp b/mlir/lib/Dialect/SCF/Utils/Utils.cpp
index 6658cca03eba7....
[truncated]

llvmbot · 2024-06-11T09:12:05Z

@llvm/pr-subscribers-llvm-support

Author: Ramkumar Ramachandra (artagnon)

Changes

This patch is part of a project to move the Presburger library into LLVM.

Patch is 27.71 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/95087.diff

30 Files Affected:

(modified) llvm/include/llvm/Support/MathExtras.h (+38-2)
(modified) llvm/unittests/Support/MathExtrasTest.cpp (+19-3)
(modified) mlir/include/mlir/Analysis/Presburger/Fraction.h (-1)
(modified) mlir/include/mlir/Analysis/Presburger/MPInt.h (+6-4)
(modified) mlir/include/mlir/Analysis/Presburger/SlowMPInt.h (-1)
(modified) mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h (+2-2)
(removed) mlir/include/mlir/Support/MathExtras.h (-51)
(modified) mlir/lib/Analysis/FlatLinearValueConstraints.cpp (-1)
(modified) mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp (-1)
(modified) mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp (+5-5)
(modified) mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp (+3-3)
(modified) mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp (-1)
(modified) mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp (+3-2)
(modified) mlir/lib/Dialect/Affine/IR/AffineOps.cpp (+5-1)
(modified) mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp (-1)
(modified) mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp (-1)
(modified) mlir/lib/Dialect/Func/IR/FuncOps.cpp (-1)
(modified) mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp (+2-2)
(modified) mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp (+2-2)
(modified) mlir/lib/Dialect/SCF/IR/SCF.cpp (-1)
(modified) mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp (+2-2)
(modified) mlir/lib/Dialect/SCF/Utils/Utils.cpp (+2-2)
(modified) mlir/lib/Dialect/Tensor/IR/TensorOps.cpp (+5-1)
(modified) mlir/lib/Dialect/Utils/StaticValueUtils.cpp (+2-2)
(modified) mlir/lib/Dialect/Vector/Transforms/VectorUnroll.cpp (-1)
(modified) mlir/lib/Dialect/Vector/Utils/VectorUtils.cpp (-1)
(modified) mlir/lib/IR/AffineExpr.cpp (+7-3)
(modified) mlir/lib/IR/AffineMap.cpp (+5-1)
(modified) mlir/unittests/Support/CMakeLists.txt (-1)
(removed) mlir/unittests/Support/MathExtrasTest.cpp (-31)

diff --git a/llvm/include/llvm/Support/MathExtras.h b/llvm/include/llvm/Support/MathExtras.h
index f0e4ee534ece3..c5cac51130b1d 100644
--- a/llvm/include/llvm/Support/MathExtras.h
+++ b/llvm/include/llvm/Support/MathExtras.h
@@ -424,11 +424,47 @@ template <uint64_t Align> constexpr inline uint64_t alignTo(uint64_t Value) {
   return (Value + Align - 1) / Align * Align;
 }
 
-/// Returns the integer ceil(Numerator / Denominator).
-inline uint64_t divideCeil(uint64_t Numerator, uint64_t Denominator) {
+/// Returns the integer ceil(Numerator / Denominator). Unsigned integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE uint64_t divideCeil(uint64_t Numerator,
+                                                 uint64_t Denominator) {
   return alignTo(Numerator, Denominator) / Denominator;
 }
 
+/// Returns the integer ceil(Numerator / Denominator). Signed integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t ceilDiv(int64_t Numerator,
+                                             int64_t Denominator) {
+  assert(Denominator);
+  if (!Numerator)
+    return 0;
+  // C's integer division rounds towards 0.
+  int64_t X = (Denominator > 0) ? -1 : 1;
+  bool SameSign = (Numerator > 0) == (Denominator > 0);
+  return SameSign ? ((Numerator + X) / Denominator) + 1
+                  : -(-Numerator / Denominator);
+}
+
+/// Returns the integer floor(Numerator / Denominator). Signed integer version.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t floorDiv(int64_t Numerator,
+                                              int64_t Denominator) {
+  assert(Denominator);
+  if (!Numerator)
+    return 0;
+  // C's integer division rounds towards 0.
+  int64_t X = (Denominator > 0) ? -1 : 1;
+  bool SameSign = (Numerator > 0) == (Denominator > 0);
+  return SameSign ? Numerator / Denominator
+                  : -((-Numerator + X) / Denominator) - 1;
+}
+
+/// Returns the remainder of the Euclidean division of LHS by RHS. Result is
+/// always non-negative.
+LLVM_ATTRIBUTE_ALWAYS_INLINE int64_t mod(int64_t Numerator,
+                                         int64_t Denominator) {
+  assert(Denominator >= 1);
+  return Numerator % Denominator < 0 ? Numerator % Denominator + Denominator
+                                     : Numerator % Denominator;
+}
+
 /// Returns the integer nearest(Numerator / Denominator).
 inline uint64_t divideNearest(uint64_t Numerator, uint64_t Denominator) {
   return (Numerator + (Denominator / 2)) / Denominator;
diff --git a/llvm/unittests/Support/MathExtrasTest.cpp b/llvm/unittests/Support/MathExtrasTest.cpp
index 67239a24c47ad..b32094e0acdfe 100644
--- a/llvm/unittests/Support/MathExtrasTest.cpp
+++ b/llvm/unittests/Support/MathExtrasTest.cpp
@@ -434,8 +434,25 @@ TEST(MathExtras, IsShiftedInt) {
   EXPECT_FALSE((isShiftedInt<6, 10>(int64_t(1) << 15)));
 }
 
-template <typename T>
-class OverflowTest : public ::testing::Test { };
+TEST(MathExtras, CeilDivTest) {
+  EXPECT_EQ(ceilDiv(14, 3), 5);
+  EXPECT_EQ(ceilDiv(14, -3), -4);
+  EXPECT_EQ(ceilDiv(-14, -3), 5);
+  EXPECT_EQ(ceilDiv(-14, 3), -4);
+  EXPECT_EQ(ceilDiv(0, 3), 0);
+  EXPECT_EQ(ceilDiv(0, -3), 0);
+}
+
+TEST(MathExtras, FloorDivTest) {
+  EXPECT_EQ(floorDiv(14, 3), 4);
+  EXPECT_EQ(floorDiv(14, -3), -5);
+  EXPECT_EQ(floorDiv(-14, -3), 4);
+  EXPECT_EQ(floorDiv(-14, 3), -5);
+  EXPECT_EQ(floorDiv(0, 3), 0);
+  EXPECT_EQ(floorDiv(0, -3), 0);
+}
+
+template <typename T> class OverflowTest : public ::testing::Test {};
 
 using OverflowTestTypes = ::testing::Types<signed char, short, int, long,
                                            long long>;
@@ -560,5 +577,4 @@ TYPED_TEST(OverflowTest, MulResultZero) {
   EXPECT_FALSE(MulOverflow<TypeParam>(0, -5, Result));
   EXPECT_EQ(Result, TypeParam(0));
 }
-
 } // namespace
diff --git a/mlir/include/mlir/Analysis/Presburger/Fraction.h b/mlir/include/mlir/Analysis/Presburger/Fraction.h
index c07bb767f50bf..f76ba3f006d07 100644
--- a/mlir/include/mlir/Analysis/Presburger/Fraction.h
+++ b/mlir/include/mlir/Analysis/Presburger/Fraction.h
@@ -15,7 +15,6 @@
 #define MLIR_ANALYSIS_PRESBURGER_FRACTION_H
 
 #include "mlir/Analysis/Presburger/MPInt.h"
-#include "mlir/Support/MathExtras.h"
 
 namespace mlir {
 namespace presburger {
diff --git a/mlir/include/mlir/Analysis/Presburger/MPInt.h b/mlir/include/mlir/Analysis/Presburger/MPInt.h
index 12ab0598d10d9..343f50251c7dc 100644
--- a/mlir/include/mlir/Analysis/Presburger/MPInt.h
+++ b/mlir/include/mlir/Analysis/Presburger/MPInt.h
@@ -17,7 +17,8 @@
 #define MLIR_ANALYSIS_PRESBURGER_MPINT_H
 
 #include "mlir/Analysis/Presburger/SlowMPInt.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/ADT/ArrayRef.h"
+#include "llvm/Support/MathExtras.h"
 #include "llvm/Support/raw_ostream.h"
 #include <numeric>
 
@@ -31,9 +32,10 @@ namespace presburger {
 /// from the mlir::presburger namespace. So to access the 64-bit overloads, an
 /// explict call to mlir::ceilDiv would be required. These using declarations
 /// allow overload resolution to transparently call the right function.
-using ::mlir::ceilDiv;
-using ::mlir::floorDiv;
-using ::mlir::mod;
+using ::llvm::ArrayRef;
+using ::llvm::ceilDiv;
+using ::llvm::floorDiv;
+using ::llvm::mod;
 
 namespace detail {
 /// If builtin intrinsics for overflow-checked arithmetic are available,
diff --git a/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h b/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
index c70306761c549..482581c573cea 100644
--- a/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
+++ b/mlir/include/mlir/Analysis/Presburger/SlowMPInt.h
@@ -18,7 +18,6 @@
 #ifndef MLIR_ANALYSIS_PRESBURGER_SLOWMPINT_H
 #define MLIR_ANALYSIS_PRESBURGER_SLOWMPINT_H
 
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/APInt.h"
 #include "llvm/ADT/Hashing.h"
 #include "llvm/Support/raw_ostream.h"
diff --git a/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h b/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
index 7a24c201a39a7..0e39928b1d6a9 100644
--- a/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
+++ b/mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h
@@ -17,7 +17,7 @@
 #include "mlir/IR/SymbolTable.h"
 #include "mlir/Interfaces/InferTypeOpInterface.h"
 #include "mlir/Interfaces/SideEffectInterfaces.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 namespace mlir {
 namespace mesh {
@@ -114,7 +114,7 @@ inline int64_t shardDimension(int64_t dimSize, int64_t shardCount) {
     return ShapedType::kDynamic;
 
   assert(dimSize % shardCount == 0);
-  return ceilDiv(dimSize, shardCount);
+  return llvm::ceilDiv(dimSize, shardCount);
 }
 
 // Get the size of an unsharded dimension.
diff --git a/mlir/include/mlir/Support/MathExtras.h b/mlir/include/mlir/Support/MathExtras.h
deleted file mode 100644
index 17a747393d26e..0000000000000
--- a/mlir/include/mlir/Support/MathExtras.h
+++ /dev/null
@@ -1,51 +0,0 @@
-//===- MathExtras.h - Math functions relevant to MLIR -----------*- C++ -*-===//
-//
-// 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
-//
-//===----------------------------------------------------------------------===//
-//
-// This file contains math functions relevant to MLIR.
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef MLIR_SUPPORT_MATHEXTRAS_H_
-#define MLIR_SUPPORT_MATHEXTRAS_H_
-
-#include "mlir/Support/LLVM.h"
-#include "llvm/ADT/APInt.h"
-
-namespace mlir {
-
-/// Returns the result of MLIR's ceildiv operation on constants. The RHS is
-/// expected to be non-zero.
-inline int64_t ceilDiv(int64_t lhs, int64_t rhs) {
-  assert(rhs != 0);
-  // C/C++'s integer division rounds towards 0.
-  int64_t x = (rhs > 0) ? -1 : 1;
-  return ((lhs != 0) && (lhs > 0) == (rhs > 0)) ? ((lhs + x) / rhs) + 1
-                                                : -(-lhs / rhs);
-}
-
-/// Returns the result of MLIR's floordiv operation on constants. The RHS is
-/// expected to be non-zero.
-inline int64_t floorDiv(int64_t lhs, int64_t rhs) {
-  assert(rhs != 0);
-  // C/C++'s integer division rounds towards 0.
-  int64_t x = (rhs < 0) ? 1 : -1;
-  return ((lhs != 0) && ((lhs < 0) != (rhs < 0))) ? -((-lhs + x) / rhs) - 1
-                                                  : lhs / rhs;
-}
-
-/// Returns MLIR's mod operation on constants. MLIR's mod operation yields the
-/// remainder of the Euclidean division of 'lhs' by 'rhs', and is therefore not
-/// C's % operator.  The RHS is always expected to be positive, and the result
-/// is always non-negative.
-inline int64_t mod(int64_t lhs, int64_t rhs) {
-  assert(rhs >= 1);
-  return lhs % rhs < 0 ? lhs % rhs + rhs : lhs % rhs;
-}
-} // namespace mlir
-
-#endif // MLIR_SUPPORT_MATHEXTRAS_H_
diff --git a/mlir/lib/Analysis/FlatLinearValueConstraints.cpp b/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
index bf7e3121ea32a..5c4f353f310d6 100644
--- a/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
+++ b/mlir/lib/Analysis/FlatLinearValueConstraints.cpp
@@ -16,7 +16,6 @@
 #include "mlir/IR/Builders.h"
 #include "mlir/IR/IntegerSet.h"
 #include "mlir/Support/LLVM.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp b/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
index 94b7c8d4f2fdf..744236692fbb6 100644
--- a/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
+++ b/mlir/lib/Conversion/FuncToLLVM/FuncToLLVM.cpp
@@ -36,7 +36,6 @@
 #include "mlir/IR/SymbolTable.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/Support/LogicalResult.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/DialectConversion.h"
 #include "mlir/Transforms/Passes.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp b/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
index da084b89ceadc..0910a1ea076c0 100644
--- a/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
+++ b/mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp
@@ -12,7 +12,7 @@
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/Dialect/LLVMIR/LLVMTypes.h"
 #include "mlir/IR/Builders.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 using namespace mlir;
 
@@ -363,9 +363,9 @@ void UnrankedMemRefDescriptor::computeSizes(
   // Initialize shared constants.
   Value one = createIndexAttrConstant(builder, loc, indexType, 1);
   Value two = createIndexAttrConstant(builder, loc, indexType, 2);
-  Value indexSize =
-      createIndexAttrConstant(builder, loc, indexType,
-                              ceilDiv(typeConverter.getIndexTypeBitwidth(), 8));
+  Value indexSize = createIndexAttrConstant(
+      builder, loc, indexType,
+      llvm::ceilDiv(typeConverter.getIndexTypeBitwidth(), 8));
 
   sizes.reserve(sizes.size() + values.size());
   for (auto [desc, addressSpace] : llvm::zip(values, addressSpaces)) {
@@ -378,7 +378,7 @@ void UnrankedMemRefDescriptor::computeSizes(
     // to data layout) into the unranked descriptor.
     Value pointerSize = createIndexAttrConstant(
         builder, loc, indexType,
-        ceilDiv(typeConverter.getPointerBitwidth(addressSpace), 8));
+        llvm::ceilDiv(typeConverter.getPointerBitwidth(addressSpace), 8));
     Value doublePointerSize =
         builder.create<LLVM::MulOp>(loc, indexType, two, pointerSize);
 
diff --git a/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp b/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
index 2dc42f0a85e66..d7eab77b97d0c 100644
--- a/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
+++ b/mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp
@@ -25,8 +25,8 @@
 #include "mlir/IR/BuiltinTypes.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/Pass/Pass.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/SmallBitVector.h"
+#include "llvm/Support/MathExtras.h"
 #include <optional>
 
 namespace mlir {
@@ -971,8 +971,8 @@ struct MemorySpaceCastOpLowering
                                resultUnderlyingDesc, resultElemPtrType);
 
       int64_t bytesToSkip =
-          2 *
-          ceilDiv(getTypeConverter()->getPointerBitwidth(resultAddrSpace), 8);
+          2 * llvm::ceilDiv(
+                  getTypeConverter()->getPointerBitwidth(resultAddrSpace), 8);
       Value bytesToSkipConst = rewriter.create<LLVM::ConstantOp>(
           loc, getIndexType(), rewriter.getIndexAttr(bytesToSkip));
       Value copySize = rewriter.create<LLVM::SubOp>(
diff --git a/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp b/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
index 24427c25d79ff..034484af83b79 100644
--- a/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
+++ b/mlir/lib/Dialect/Affine/Analysis/AffineStructures.cpp
@@ -21,7 +21,6 @@
 #include "mlir/IR/AffineExprVisitor.h"
 #include "mlir/IR/IntegerSet.h"
 #include "mlir/Support/LLVM.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallVector.h"
diff --git a/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp b/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
index 1c28d6b00b3c8..11a9f1485514a 100644
--- a/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
+++ b/mlir/lib/Dialect/Affine/Analysis/LoopAnalysis.cpp
@@ -18,7 +18,7 @@
 #include "mlir/Dialect/Affine/Analysis/NestedMatcher.h"
 #include "mlir/Dialect/Affine/IR/AffineOps.h"
 #include "mlir/Dialect/Affine/IR/AffineValueMap.h"
-#include "mlir/Support/MathExtras.h"
+#include "llvm/Support/MathExtras.h"
 
 #include "llvm/ADT/DenseSet.h"
 #include "llvm/ADT/SmallPtrSet.h"
@@ -47,7 +47,8 @@ void mlir::affine::getTripCountMapAndOperands(
     loopSpan = ub - lb;
     if (loopSpan < 0)
       loopSpan = 0;
-    *tripCountMap = AffineMap::getConstantMap(ceilDiv(loopSpan, step), context);
+    *tripCountMap =
+        AffineMap::getConstantMap(llvm::ceilDiv(loopSpan, step), context);
     tripCountOperands->clear();
     return;
   }
diff --git a/mlir/lib/Dialect/Affine/IR/AffineOps.cpp b/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
index 0a58d2fdb02f5..b52bc05e908fc 100644
--- a/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
+++ b/mlir/lib/Dialect/Affine/IR/AffineOps.cpp
@@ -18,19 +18,23 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Interfaces/ShapedOpInterfaces.h"
 #include "mlir/Interfaces/ValueBoundsOpInterface.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/ScopeExit.h"
 #include "llvm/ADT/SmallBitVector.h"
 #include "llvm/ADT/SmallVectorExtras.h"
 #include "llvm/ADT/TypeSwitch.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 #include <numeric>
 #include <optional>
 
 using namespace mlir;
 using namespace mlir::affine;
 
+using llvm::ceilDiv;
+using llvm::floorDiv;
+using llvm::mod;
+
 #define DEBUG_TYPE "affine-ops"
 
 #include "mlir/Dialect/Affine/IR/AffineOpsDialect.cpp.inc"
diff --git a/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp b/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
index 268050a30e002..c934f229bb6c4 100644
--- a/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
+++ b/mlir/lib/Dialect/Affine/Utils/LoopUtils.cpp
@@ -23,7 +23,6 @@
 #include "mlir/Dialect/SCF/IR/SCF.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/IR/IntegerSet.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/GreedyPatternRewriteDriver.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
diff --git a/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp b/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
index 1320db3f9e543..ede158db911ea 100644
--- a/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
+++ b/mlir/lib/Dialect/ControlFlow/IR/ControlFlowOps.cpp
@@ -23,7 +23,6 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/IR/Value.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/APFloat.h"
 #include "llvm/ADT/STLExtras.h"
diff --git a/mlir/lib/Dialect/Func/IR/FuncOps.cpp b/mlir/lib/Dialect/Func/IR/FuncOps.cpp
index 95589e8989e27..c719981769b9e 100644
--- a/mlir/lib/Dialect/Func/IR/FuncOps.cpp
+++ b/mlir/lib/Dialect/Func/IR/FuncOps.cpp
@@ -19,7 +19,6 @@
 #include "mlir/IR/TypeUtilities.h"
 #include "mlir/IR/Value.h"
 #include "mlir/Interfaces/FunctionImplementation.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/APFloat.h"
 #include "llvm/ADT/MapVector.h"
diff --git a/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp b/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
index 38abdd122d633..24763c036192f 100644
--- a/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
+++ b/mlir/lib/Dialect/Linalg/TransformOps/GPUHeuristics.cpp
@@ -9,11 +9,11 @@
 #include "mlir/Dialect/Linalg/TransformOps/GPUHeuristics.h"
 
 #include "mlir/Dialect/GPU/IR/GPUDialect.h"
-#include "mlir/Support/MathExtras.h"
 #include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 #include "llvm/Support/raw_ostream.h"
 #include <cmath>
 #include <numeric>
@@ -76,7 +76,7 @@ transform::gpu::CopyMappingInfo::CopyMappingInfo(MLIRContext *ctx,
       llvm::map_range(llvm::zip(copySizes, this->numThreads), [](auto &&pair) {
         int64_t size, numThreads;
         std::tie(size, numThreads) = pair;
-        return mlir::ceilDiv(size, numThreads);
+        return llvm::ceilDiv(size, numThreads);
       }));
   SmallVector<Attribute> allThreadMappings{linearId2(ctx), linearId1(ctx),
                                            linearId0(ctx)};
diff --git a/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp b/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
index 2392e10522dd0..4c49ebb3b4cd7 100644
--- a/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
+++ b/mlir/lib/Dialect/MemRef/Transforms/EmulateNarrowType.cpp
@@ -20,9 +20,9 @@
 #include "mlir/IR/BuiltinTypes.h"
 #include "mlir/IR/OpDefinition.h"
 #include "mlir/Support/LogicalResult.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/DialectConversion.h"
 #include "llvm/Support/FormatVariadic.h"
+#include "llvm/Support/MathExtras.h"
 #include <cassert>
 #include <type_traits>
 
@@ -74,7 +74,7 @@ convertCastingOp(ConversionPatternRewriter &rewriter,
 
   SmallVector<int64_t> size;
   if (sizes.size())
-    size.push_back(ceilDiv(sizes[0], elementsPerByte));
+    size.push_back(llvm::ceilDiv(sizes[0], elementsPerByte));
   offset = offset / elementsPerByte;
 
   rewriter.replaceOpWithNewOp<memref::ReinterpretCastOp>(
diff --git a/mlir/lib/Dialect/SCF/IR/SCF.cpp b/mlir/lib/Dialect/SCF/IR/SCF.cpp
index 5e94f4dc612a7..524711dbbb6fd 100644
--- a/mlir/lib/Dialect/SCF/IR/SCF.cpp
+++ b/mlir/lib/Dialect/SCF/IR/SCF.cpp
@@ -21,7 +21,6 @@
 #include "mlir/IR/PatternMatch.h"
 #include "mlir/Interfaces/FunctionInterfaces.h"
 #include "mlir/Interfaces/ValueBoundsOpInterface.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/InliningUtils.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
diff --git a/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp b/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
index 82ec95d31f525..f5b7a6e28a178 100644
--- a/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
+++ b/mlir/lib/Dialect/SCF/Transforms/LoopPipelining.cpp
@@ -17,10 +17,10 @@
 #include "mlir/Dialect/SCF/Utils/Utils.h"
 #include "mlir/IR/IRMapping.h"
 #include "mlir/IR/PatternMatch.h"
-#include "mlir/Support/MathExtras.h"
 #include "mlir/Transforms/RegionUtils.h"
 #include "llvm/ADT/MapVector.h"
 #include "llvm/Support/Debug.h"
+#include "llvm/Support/MathExtras.h"
 
 #define DEBUG_TYPE "scf-loop-pipelining"
 #define DBGS() (llvm::dbgs() << "[" DEBUG_TYPE "]: ")
@@ -119,7 +119,7 @@ bool LoopPipelinerInternal::initializeLoopInfo(
     int64_t ubImm = upperBoundCst.value();
     int64_t lbImm = lowerBoundCst.value();
     int64_t stepImm = stepCst.value();
-    int64_t numIteration = ceilDiv(ubImm - lbImm, stepImm);
+    int64_t numIteration = llvm::ceilDiv(ubImm - lbImm, stepImm);
     if (numIteration > maxStage) {
       dynamicLoop = false;
     } else if (!options.supportDynamicLoops) {
diff --git a/mlir/lib/Dialect/SCF/Utils/Utils.cpp b/mlir/lib/Dialect/SCF/Utils/Utils.cpp
index 6658cca03eba7....
[truncated]

llvm/include/llvm/Support/MathExtras.h

llvm/unittests/Support/MathExtrasTest.cpp

mlir/include/mlir/Analysis/Presburger/MPInt.h

nikic

Implementation looks fine. Some of the usages look inappropriate, though it's a pre-existing problem...

nikic · 2024-06-11T19:38:47Z

mlir/include/mlir/Dialect/Mesh/IR/MeshOps.h

@@ -114,7 +114,7 @@ inline int64_t shardDimension(int64_t dimSize, int64_t shardCount) {
    return ShapedType::kDynamic;

  assert(dimSize % shardCount == 0);
-  return ceilDiv(dimSize, shardCount);
+  return llvm::divideCeilSigned(dimSize, shardCount);


Using divideCeil here doesn't make sense because the preceding assert ensures that the division is exact, so this is equivalent to a normal division.

nikic · 2024-06-11T19:39:59Z

mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp

-                              ceilDiv(typeConverter.getIndexTypeBitwidth(), 8));
+  Value indexSize = createIndexAttrConstant(
+      builder, loc, indexType,
+      llvm::divideCeilSigned(typeConverter.getIndexTypeBitwidth(), 8));


It looks like this used signed division by accident -- both quantities here are unsigned.

nikic · 2024-06-11T19:40:04Z

mlir/lib/Conversion/LLVMCommon/MemRefBuilder.cpp

@@ -378,7 +378,8 @@ void UnrankedMemRefDescriptor::computeSizes(
    // to data layout) into the unranked descriptor.
    Value pointerSize = createIndexAttrConstant(
        builder, loc, indexType,
-        ceilDiv(typeConverter.getPointerBitwidth(addressSpace), 8));
+        llvm::divideCeilSigned(typeConverter.getPointerBitwidth(addressSpace),


nikic · 2024-06-11T19:41:25Z

llvm/include/llvm/Support/MathExtras.h

@@ -424,11 +424,43 @@ template <uint64_t Align> constexpr inline uint64_t alignTo(uint64_t Value) {
  return (Value + Align - 1) / Align * Align;
 }

-/// Returns the integer ceil(Numerator / Denominator).
+/// Returns the integer ceil(Numerator / Denominator). Unsigned integer version.


Orthogonal to this change, but I think it would be good to template these functions in a follow-up. I think we have many places where we're upcasting unsigned to uint64_t to make use of these, and especially for divisions performing it in a larger width can be much more expensive.

Yes, I was planning to do that in a follow-up.

nikic · 2024-06-11T19:41:43Z

mlir/lib/Conversion/MemRefToLLVM/MemRefToLLVM.cpp

-          2 *
-          ceilDiv(getTypeConverter()->getPointerBitwidth(resultAddrSpace), 8);
+          2 * llvm::divideCeilSigned(
+                  getTypeConverter()->getPointerBitwidth(resultAddrSpace), 8);


Here as well, this should use the unsigned variant.

joker-eph

LGTM on the principle, @nikic find some issues to fix of course.
This is NFC right?

artagnon · 2024-06-11T21:59:38Z

This is NFC right?

I normally reserve the NFC tag for refactoring changes with no test additions, but yes, this patch doesn't change functionality.

artagnon · 2024-06-11T22:02:17Z

@nikic For a follow-up, I think we can template over signed/unsigned and different bitwidths. Then there won't be a separate unsigned/signed version, and we don't have to worry about each usage instance.

Several multi-argument functions unnecessarily widen types beyond the argument types. Template'ize the functions, and use std::common_type_t to avoid this, hence optimizing the functions. While at it, address usage issues raised in llvm#95087. One of the requirements of this patch is to add overflow checks, and one caller in LoopVectorize is manually widened (marked as TODO).

Several multi-argument functions unnecessarily widen types beyond the argument types. Template'ize the functions, and use std::common_type_t to avoid this, hence optimizing the functions. While at it, address usage issues raised in llvm#95087. One of the requirements of this patch is to add overflow checks, and one caller in LoopVectorize and one in AMDGPUBaseInfo is manually widened.

Follow up on llvm#95087 to fix incorrect usage instances of divideCeilSigned.

Follow up on #95087 to fix incorrect usage instances of divideCeilSigned.

Changes in updated submodules that require adaptation in iree-amd-aie: - `NpuWriteBdOp` lost its ddr_id argument - AIRPipeline.cpp was removed - `mlir::ceilDiv` and the rest of `mlir/Support/MathExtras.h` was removed in favor of the LLVM equivalents (llvm/llvm-project#95087) --------- Signed-off-by: daveliddell <dave.liddell@amd.com>

mlir/MathExtras: consolidate with llvm/MathExtras

3fdbdf6

This patch is part of a project to move the Presburger library into LLVM.

artagnon requested review from nikic, RKSimon and joker-eph June 11, 2024 09:11

artagnon requested review from ftynse, nicolasvasilache, Groverkss, Superty, dcaballe and hanhanW as code owners June 11, 2024 09:11

llvmbot added mlir:core MLIR Core Infrastructure mlir:llvm mlir:linalg mlir:affine mlir:vectorops mlir mlir:tensor mlir:presburger llvm:support mlir:memref mlir:scf mlir:cf mlir:func mlir:vector labels Jun 11, 2024

nikic reviewed Jun 11, 2024

View reviewed changes

llvm/include/llvm/Support/MathExtras.h Outdated Show resolved Hide resolved

nikic reviewed Jun 11, 2024

View reviewed changes

llvm/include/llvm/Support/MathExtras.h Outdated Show resolved Hide resolved

llvm/MathExtras: address review

10c979a

nikic reviewed Jun 11, 2024

View reviewed changes

llvm/MathExtras: fix nits, increase test coverage

3006125

Superty reviewed Jun 11, 2024

View reviewed changes

mlir/include/mlir/Analysis/Presburger/MPInt.h Outdated Show resolved Hide resolved

MPInt: fix nit in comment

19960b7

artagnon mentioned this pull request Jun 11, 2024

mlir/Presburger/MPInt: move into llvm/ADT #94953

Merged

MPInt: strip bad comment

a70c51d

nikic reviewed Jun 11, 2024

View reviewed changes

joker-eph approved these changes Jun 11, 2024

View reviewed changes

artagnon merged commit 0fb216f into llvm:main Jun 11, 2024
5 of 6 checks passed

artagnon deleted the mathextras-consolidate branch June 11, 2024 22:00

artagnon mentioned this pull request Jun 13, 2024

MathExtras: avoid unnecessarily widening types #95426

Merged

artagnon added a commit to artagnon/llvm-project that referenced this pull request Jun 15, 2024

mlir: fix incorrect usages of divideCeilSigned

d91b489

Follow up on llvm#95087 to fix incorrect usage instances of divideCeilSigned.

artagnon mentioned this pull request Jun 15, 2024

mlir: fix incorrect usages of divideCeilSigned #95680

Merged

artagnon added a commit that referenced this pull request Jun 17, 2024

mlir: fix incorrect usages of divideCeilSigned (#95680)

e843f02

Follow up on #95087 to fix incorrect usage instances of divideCeilSigned.

daveliddell mentioned this pull request Jun 26, 2024

Bump iree, mlir-air, mlir-aie nod-ai/iree-amd-aie#465

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

mlir/MathExtras: consolidate with llvm/MathExtras #95087

mlir/MathExtras: consolidate with llvm/MathExtras #95087

artagnon commented Jun 11, 2024

llvmbot commented Jun 11, 2024 •

edited

Loading

llvmbot commented Jun 11, 2024

llvmbot commented Jun 11, 2024

nikic left a comment

nikic Jun 11, 2024

nikic Jun 11, 2024

nikic Jun 11, 2024

nikic Jun 11, 2024

artagnon Jun 11, 2024

nikic Jun 11, 2024

joker-eph left a comment

artagnon commented Jun 11, 2024

artagnon commented Jun 11, 2024

mlir/MathExtras: consolidate with llvm/MathExtras #95087

mlir/MathExtras: consolidate with llvm/MathExtras #95087

Conversation

artagnon commented Jun 11, 2024

llvmbot commented Jun 11, 2024 • edited Loading

llvmbot commented Jun 11, 2024

llvmbot commented Jun 11, 2024

nikic left a comment

Choose a reason for hiding this comment

nikic Jun 11, 2024

Choose a reason for hiding this comment

nikic Jun 11, 2024

Choose a reason for hiding this comment

nikic Jun 11, 2024

Choose a reason for hiding this comment

nikic Jun 11, 2024

Choose a reason for hiding this comment

artagnon Jun 11, 2024

Choose a reason for hiding this comment

nikic Jun 11, 2024

Choose a reason for hiding this comment

joker-eph left a comment

Choose a reason for hiding this comment

artagnon commented Jun 11, 2024

artagnon commented Jun 11, 2024

llvmbot commented Jun 11, 2024 •

edited

Loading