[llvm] r285459 - Don't claim the udiv created in BypassSlowDivision is exact.

[Justin Lebar via llvm-commits]

Author: jlebar
Date: Fri Oct 28 16:43:51 2016
New Revision: 285459

URL: http://llvm.org/viewvc/llvm-project?rev=285459&view=rev
Log:
Don't claim the udiv created in BypassSlowDivision is exact.

Summary:
In BypassSlowDivision's short-dividend path, we would create e.g.

  udiv exact i32 %a, %b

"exact" here means that we are asserting that %a is a multiple of %b.
But we have no reason to believe this must be true -- this is just a
bug, as far as I can tell.

Reviewers: tra

Subscribers: jholewinski, llvm-commits

Differential Revision: https://reviews.llvm.org/D26097

Added:
    llvm/trunk/test/Transforms/CodeGenPrepare/NVPTX/
    llvm/trunk/test/Transforms/CodeGenPrepare/NVPTX/bypass-slow-div-not-exact.ll
Modified:
    llvm/trunk/lib/Transforms/Utils/BypassSlowDivision.cpp

Modified: llvm/trunk/lib/Transforms/Utils/BypassSlowDivision.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Utils/BypassSlowDivision.cpp?rev=285459&r1=285458&r2=285459&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/Utils/BypassSlowDivision.cpp (original)
+++ llvm/trunk/lib/Transforms/Utils/BypassSlowDivision.cpp Fri Oct 28 16:43:51 2016
@@ -120,8 +120,7 @@ static bool insertFastDiv(Instruction *I
                                                  BypassType);
 
   // udiv/urem because optimization only handles positive numbers
-  Value *ShortQuotientV = FastBuilder.CreateExactUDiv(ShortDividendV,
-                                                      ShortDivisorV);
+  Value *ShortQuotientV = FastBuilder.CreateUDiv(ShortDividendV, ShortDivisorV);
   Value *ShortRemainderV = FastBuilder.CreateURem(ShortDividendV,
                                                   ShortDivisorV);
   Value *FastQuotientV = FastBuilder.CreateCast(Instruction::ZExt,

Added: llvm/trunk/test/Transforms/CodeGenPrepare/NVPTX/bypass-slow-div-not-exact.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/CodeGenPrepare/NVPTX/bypass-slow-div-not-exact.ll?rev=285459&view=auto
==============================================================================
--- llvm/trunk/test/Transforms/CodeGenPrepare/NVPTX/bypass-slow-div-not-exact.ll (added)
+++ llvm/trunk/test/Transforms/CodeGenPrepare/NVPTX/bypass-slow-div-not-exact.ll Fri Oct 28 16:43:51 2016
@@ -0,0 +1,16 @@
+; RUN: opt -S -codegenprepare < %s | FileCheck %s
+
+target datalayout = "e-i64:64-v16:16-v32:32-n16:32:64"
+target triple = "nvptx64-nvidia-cuda"
+
+; Check that the smaller-width division that the BypassSlowDivision pass
+; creates is not marked as "exact" (that is, it doesn't claim that the
+; numerator is a multiple of the denominator).
+;
+; CHECK-LABEL: @test
+define void @test(i64 %a, i64 %b, i64* %retptr) {
+  ; CHECK: udiv i32
+  %d = sdiv i64 %a, %b
+  store i64 %d, i64* %retptr
+  ret void
+}