[llvm] a968972 - [KnownBits] Add knownbits analysis for mulhs/mulu 'multiply high' instructions

[Simon Pilgrim via llvm-commits]

Author: Simon Pilgrim
Date: 2021-03-19T08:56:06Z
New Revision: a96897219daf43a1b90e1e0c9dbf20167c0c16af

URL: https://github.com/llvm/llvm-project/commit/a96897219daf43a1b90e1e0c9dbf20167c0c16af
DIFF: https://github.com/llvm/llvm-project/commit/a96897219daf43a1b90e1e0c9dbf20167c0c16af.diff

LOG: [KnownBits] Add knownbits analysis for mulhs/mulu 'multiply high' instructions

Split off from D98857

https://reviews.llvm.org/D98866

Added: 
    

Modified: 
    llvm/include/llvm/Support/KnownBits.h
    llvm/lib/Support/KnownBits.cpp
    llvm/unittests/Support/KnownBitsTest.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/include/llvm/Support/KnownBits.h b/llvm/include/llvm/Support/KnownBits.h
index d854aadbd430..c27ddb0ce804 100644
--- a/llvm/include/llvm/Support/KnownBits.h
+++ b/llvm/include/llvm/Support/KnownBits.h
@@ -296,6 +296,12 @@ struct KnownBits {
   /// Compute known bits resulting from multiplying LHS and RHS.
   static KnownBits computeForMul(const KnownBits &LHS, const KnownBits &RHS);
 
+  /// Compute known bits from sign-extended multiply-hi.
+  static KnownBits mulhs(const KnownBits &LHS, const KnownBits &RHS);
+
+  /// Compute known bits from zero-extended multiply-hi.
+  static KnownBits mulhu(const KnownBits &LHS, const KnownBits &RHS);
+
   /// Compute known bits for udiv(LHS, RHS).
   static KnownBits udiv(const KnownBits &LHS, const KnownBits &RHS);
 

diff  --git a/llvm/lib/Support/KnownBits.cpp b/llvm/lib/Support/KnownBits.cpp
index 6c7aaad968f5..423a908eed57 100644
--- a/llvm/lib/Support/KnownBits.cpp
+++ b/llvm/lib/Support/KnownBits.cpp
@@ -489,6 +489,24 @@ KnownBits KnownBits::computeForMul(const KnownBits &LHS, const KnownBits &RHS) {
   return Res;
 }
 
+KnownBits KnownBits::mulhs(const KnownBits &LHS, const KnownBits &RHS) {
+  unsigned BitWidth = LHS.getBitWidth();
+  assert(BitWidth == RHS.getBitWidth() && !LHS.hasConflict() &&
+         !RHS.hasConflict() && "Operand mismatch");
+  KnownBits WideLHS = LHS.sext(2 * BitWidth);
+  KnownBits WideRHS = RHS.sext(2 * BitWidth);
+  return computeForMul(WideLHS, WideRHS).extractBits(BitWidth, BitWidth);
+}
+
+KnownBits KnownBits::mulhu(const KnownBits &LHS, const KnownBits &RHS) {
+  unsigned BitWidth = LHS.getBitWidth();
+  assert(BitWidth == RHS.getBitWidth() && !LHS.hasConflict() &&
+         !RHS.hasConflict() && "Operand mismatch");
+  KnownBits WideLHS = LHS.zext(2 * BitWidth);
+  KnownBits WideRHS = RHS.zext(2 * BitWidth);
+  return computeForMul(WideLHS, WideRHS).extractBits(BitWidth, BitWidth);
+}
+
 KnownBits KnownBits::udiv(const KnownBits &LHS, const KnownBits &RHS) {
   unsigned BitWidth = LHS.getBitWidth();
   assert(!LHS.hasConflict() && !RHS.hasConflict());

diff  --git a/llvm/unittests/Support/KnownBitsTest.cpp b/llvm/unittests/Support/KnownBitsTest.cpp
index 4e69df49837e..5f2133165d24 100644
--- a/llvm/unittests/Support/KnownBitsTest.cpp
+++ b/llvm/unittests/Support/KnownBitsTest.cpp
@@ -113,6 +113,8 @@ TEST(KnownBitsTest, BinaryExhaustive) {
       KnownBits KnownSMax(KnownAnd);
       KnownBits KnownSMin(KnownAnd);
       KnownBits KnownMul(KnownAnd);
+      KnownBits KnownMulHS(KnownAnd);
+      KnownBits KnownMulHU(KnownAnd);
       KnownBits KnownUDiv(KnownAnd);
       KnownBits KnownURem(KnownAnd);
       KnownBits KnownSRem(KnownAnd);
@@ -156,6 +158,14 @@ TEST(KnownBitsTest, BinaryExhaustive) {
           KnownMul.One &= Res;
           KnownMul.Zero &= ~Res;
 
+          Res = (N1.sext(2 * Bits) * N2.sext(2 * Bits)).extractBits(Bits, Bits);
+          KnownMulHS.One &= Res;
+          KnownMulHS.Zero &= ~Res;
+
+          Res = (N1.zext(2 * Bits) * N2.zext(2 * Bits)).extractBits(Bits, Bits);
+          KnownMulHU.One &= Res;
+          KnownMulHU.Zero &= ~Res;
+
           if (!N2.isNullValue()) {
             Res = N1.udiv(N2);
             KnownUDiv.One &= Res;
@@ -218,12 +228,20 @@ TEST(KnownBitsTest, BinaryExhaustive) {
       EXPECT_EQ(KnownSMin.Zero, ComputedSMin.Zero);
       EXPECT_EQ(KnownSMin.One, ComputedSMin.One);
 
-      // ComputedMul is conservatively correct, but not guaranteed to be
+      // The following are conservatively correct, but not guaranteed to be
       // precise.
       KnownBits ComputedMul = KnownBits::computeForMul(Known1, Known2);
       EXPECT_TRUE(ComputedMul.Zero.isSubsetOf(KnownMul.Zero));
       EXPECT_TRUE(ComputedMul.One.isSubsetOf(KnownMul.One));
 
+      KnownBits ComputedMulHS = KnownBits::mulhs(Known1, Known2);
+      EXPECT_TRUE(ComputedMulHS.Zero.isSubsetOf(KnownMulHS.Zero));
+      EXPECT_TRUE(ComputedMulHS.One.isSubsetOf(KnownMulHS.One));
+
+      KnownBits ComputedMulHU = KnownBits::mulhu(Known1, Known2);
+      EXPECT_TRUE(ComputedMulHU.Zero.isSubsetOf(KnownMulHU.Zero));
+      EXPECT_TRUE(ComputedMulHU.One.isSubsetOf(KnownMulHU.One));
+
       KnownBits ComputedUDiv = KnownBits::udiv(Known1, Known2);
       EXPECT_TRUE(ComputedUDiv.Zero.isSubsetOf(KnownUDiv.Zero));
       EXPECT_TRUE(ComputedUDiv.One.isSubsetOf(KnownUDiv.One));