From 903f81ba184d45341f8f5b8b3227242cc168e535 Mon Sep 17 00:00:00 2001
From: Tim Northover <tnorthover@apple.com>
Date: Fri, 15 Apr 2016 18:17:18 +0000
Subject: [PATCH] ARM: don't try to hoist constant RHS out of a division.

Divisions by a constant can be converted into multiplies which are usually
cheaper, but this isn't possible if the constant gets separated (particularly
in loops). Fix this by telling ConstantHoisting that the immediate in a DIV is
cheap.

I considered making the check generic, but neither AArch64 (strangely) nor x86
showed any benefit on the tests I had.

llvm-svn: 266464
---
 .../lib/Target/ARM/ARMTargetTransformInfo.cpp | 15 +++++++
 llvm/lib/Target/ARM/ARMTargetTransformInfo.h  |  4 +-
 .../ConstantHoisting/ARM/bad-cases.ll         | 45 +++++++++++++++++++
 3 files changed, 61 insertions(+), 3 deletions(-)

diff --git a/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp b/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
index 932ec2d46daf..aaec3107aece 100644
--- a/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
+++ b/llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp
@@ -47,6 +47,21 @@ int ARMTTIImpl::getIntImmCost(const APInt &Imm, Type *Ty) {
   return 3;
 }
 
+int ARMTTIImpl::getIntImmCost(unsigned Opcode, unsigned Idx, const APInt &Imm,
+                              Type *Ty) {
+  // Division by a constant can be turned into multiplication, but only if we
+  // know it's constant. So it's not so much that the immediate is cheap (it's
+  // not), but that the alternative is worse.
+  // FIXME: this is probably unneeded with GlobalISel.
+  if ((Opcode == Instruction::SDiv || Opcode == Instruction::UDiv ||
+       Opcode == Instruction::SRem || Opcode == Instruction::URem) &&
+      Idx == 1)
+    return 0;
+
+  return getIntImmCost(Imm, Ty);
+}
+
+
 int ARMTTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src) {
   int ISD = TLI->InstructionOpcodeToISD(Opcode);
   assert(ISD && "Invalid opcode");
diff --git a/llvm/lib/Target/ARM/ARMTargetTransformInfo.h b/llvm/lib/Target/ARM/ARMTargetTransformInfo.h
index 0fe964f36de4..e07174eb40be 100644
--- a/llvm/lib/Target/ARM/ARMTargetTransformInfo.h
+++ b/llvm/lib/Target/ARM/ARMTargetTransformInfo.h
@@ -64,9 +64,7 @@ public:
   using BaseT::getIntImmCost;
   int getIntImmCost(const APInt &Imm, Type *Ty);
 
-  int getIntImmCost(unsigned Opcode, unsigned Idx, const APInt &Imm, Type *Ty) {
-    return getIntImmCost(Imm, Ty);
-  }
+  int getIntImmCost(unsigned Opcode, unsigned Idx, const APInt &Imm, Type *Ty);
 
   /// @}
 
diff --git a/llvm/test/Transforms/ConstantHoisting/ARM/bad-cases.ll b/llvm/test/Transforms/ConstantHoisting/ARM/bad-cases.ll
index 3602eb9f3fd2..8fa78e3d69b4 100644
--- a/llvm/test/Transforms/ConstantHoisting/ARM/bad-cases.ll
+++ b/llvm/test/Transforms/ConstantHoisting/ARM/bad-cases.ll
@@ -45,3 +45,48 @@ bb2:
 default:
   ret void
 }
+
+; We don't want to convert constant divides because the benefit from converting
+; them to a mul in the backend is larget than constant materialization savings.
+define void @signed_const_division(i32 %in1, i32 %in2, i32* %addr) {
+; CHECK-LABEL: @signed_const_division
+; CHECK: %res1 = sdiv i32 %l1, 1000000000
+; CHECK: %res2 = srem i32 %l2, 1000000000
+entry:
+  br label %loop
+
+loop:
+  %l1 = phi i32 [%res1, %loop], [%in1, %entry]
+  %l2 = phi i32 [%res2, %loop], [%in2, %entry]
+  %res1 = sdiv i32 %l1, 1000000000
+  store volatile i32 %res1, i32* %addr
+  %res2 = srem i32 %l2, 1000000000
+  store volatile i32 %res2, i32* %addr
+  %again = icmp eq i32 %res1, %res2
+  br i1 %again, label %loop, label %end
+
+end:
+  ret void
+}
+
+define void @unsigned_const_division(i32 %in1, i32 %in2, i32* %addr) {
+; CHECK-LABEL: @unsigned_const_division
+; CHECK: %res1 = udiv i32 %l1, 1000000000
+; CHECK: %res2 = urem i32 %l2, 1000000000
+
+entry:
+  br label %loop
+
+loop:
+  %l1 = phi i32 [%res1, %loop], [%in1, %entry]
+  %l2 = phi i32 [%res2, %loop], [%in2, %entry]
+  %res1 = udiv i32 %l1, 1000000000
+  store volatile i32 %res1, i32* %addr
+  %res2 = urem i32 %l2, 1000000000
+  store volatile i32 %res2, i32* %addr
+  %again = icmp eq i32 %res1, %res2
+  br i1 %again, label %loop, label %end
+
+end:
+  ret void
+}