maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/Transforms/InstCombine/zext-bool-add-sub.ll

17 lines
432 B
LLVM
Raw Normal View History

Remove a instcombine transform that (no longer?) makes sense: // C - zext(bool) -> bool ? C - 1 : C if (ZExtInst *ZI = dyn_cast<ZExtInst>(Op1)) if (ZI->getSrcTy()->isIntegerTy(1)) return SelectInst::Create(ZI->getOperand(0), SubOne(C), C); This ends up forming sext i1 instructions that codegen to terrible code. e.g. int blah(_Bool x, _Bool y) { return (x - y) + 1; } => movzbl %dil, %eax movzbl %sil, %ecx shll $31, %ecx sarl $31, %ecx leal 1(%rax,%rcx), %eax ret Without the rule, llvm now generates: movzbl %sil, %ecx movzbl %dil, %eax incl %eax subl %ecx, %eax ret It also helps with ARM (and pretty much any target that doesn't have a sext i1 :-). The transformation was done as part of Eli's r75531. He has given the ok to remove it. rdar://11748024 llvm-svn: 159230
2012-06-27 06:03:13 +08:00
; RUN: opt < %s -instcombine -S | FileCheck %s
; rdar://11748024
Canonicalize boolean +/- a constant to a select. (I think it's reasonably clear that we want to have a canonical form for constructs like this; if anyone thinks that a select is not the best canonical form, please tell me.) llvm-svn: 75531
2009-07-14 06:27:52 +08:00
Remove a instcombine transform that (no longer?) makes sense: // C - zext(bool) -> bool ? C - 1 : C if (ZExtInst *ZI = dyn_cast<ZExtInst>(Op1)) if (ZI->getSrcTy()->isIntegerTy(1)) return SelectInst::Create(ZI->getOperand(0), SubOne(C), C); This ends up forming sext i1 instructions that codegen to terrible code. e.g. int blah(_Bool x, _Bool y) { return (x - y) + 1; } => movzbl %dil, %eax movzbl %sil, %ecx shll $31, %ecx sarl $31, %ecx leal 1(%rax,%rcx), %eax ret Without the rule, llvm now generates: movzbl %sil, %ecx movzbl %dil, %eax incl %eax subl %ecx, %eax ret It also helps with ARM (and pretty much any target that doesn't have a sext i1 :-). The transformation was done as part of Eli's r75531. He has given the ok to remove it. rdar://11748024 llvm-svn: 159230
2012-06-27 06:03:13 +08:00
define i32 @a(i1 zeroext %x, i1 zeroext %y) {
Canonicalize boolean +/- a constant to a select. (I think it's reasonably clear that we want to have a canonical form for constructs like this; if anyone thinks that a select is not the best canonical form, please tell me.) llvm-svn: 75531
2009-07-14 06:27:52 +08:00
entry:
Remove a instcombine transform that (no longer?) makes sense: // C - zext(bool) -> bool ? C - 1 : C if (ZExtInst *ZI = dyn_cast<ZExtInst>(Op1)) if (ZI->getSrcTy()->isIntegerTy(1)) return SelectInst::Create(ZI->getOperand(0), SubOne(C), C); This ends up forming sext i1 instructions that codegen to terrible code. e.g. int blah(_Bool x, _Bool y) { return (x - y) + 1; } => movzbl %dil, %eax movzbl %sil, %ecx shll $31, %ecx sarl $31, %ecx leal 1(%rax,%rcx), %eax ret Without the rule, llvm now generates: movzbl %sil, %ecx movzbl %dil, %eax incl %eax subl %ecx, %eax ret It also helps with ARM (and pretty much any target that doesn't have a sext i1 :-). The transformation was done as part of Eli's r75531. He has given the ok to remove it. rdar://11748024 llvm-svn: 159230
2012-06-27 06:03:13 +08:00
; CHECK: @a
Transform (sub 0, (zext bool to A)) to (sext bool to A) and (sub 0, (sext bool to A)) to (zext bool to A). Patch by Muhammad Ahmad Reviewed by Duncan Sands llvm-svn: 173093
2013-01-22 05:57:20 +08:00
; CHECK: [[TMP1:%.*]] = sext i1 %y to i32
Remove a instcombine transform that (no longer?) makes sense: // C - zext(bool) -> bool ? C - 1 : C if (ZExtInst *ZI = dyn_cast<ZExtInst>(Op1)) if (ZI->getSrcTy()->isIntegerTy(1)) return SelectInst::Create(ZI->getOperand(0), SubOne(C), C); This ends up forming sext i1 instructions that codegen to terrible code. e.g. int blah(_Bool x, _Bool y) { return (x - y) + 1; } => movzbl %dil, %eax movzbl %sil, %ecx shll $31, %ecx sarl $31, %ecx leal 1(%rax,%rcx), %eax ret Without the rule, llvm now generates: movzbl %sil, %ecx movzbl %dil, %eax incl %eax subl %ecx, %eax ret It also helps with ARM (and pretty much any target that doesn't have a sext i1 :-). The transformation was done as part of Eli's r75531. He has given the ok to remove it. rdar://11748024 llvm-svn: 159230
2012-06-27 06:03:13 +08:00
; CHECK: [[TMP2:%.*]] = select i1 %x, i32 2, i32 1
Transform (sub 0, (zext bool to A)) to (sext bool to A) and (sub 0, (sext bool to A)) to (zext bool to A). Patch by Muhammad Ahmad Reviewed by Duncan Sands llvm-svn: 173093
2013-01-22 05:57:20 +08:00
; CHECK-NEXT: add i32 [[TMP2]], [[TMP1]]
Remove a instcombine transform that (no longer?) makes sense: // C - zext(bool) -> bool ? C - 1 : C if (ZExtInst *ZI = dyn_cast<ZExtInst>(Op1)) if (ZI->getSrcTy()->isIntegerTy(1)) return SelectInst::Create(ZI->getOperand(0), SubOne(C), C); This ends up forming sext i1 instructions that codegen to terrible code. e.g. int blah(_Bool x, _Bool y) { return (x - y) + 1; } => movzbl %dil, %eax movzbl %sil, %ecx shll $31, %ecx sarl $31, %ecx leal 1(%rax,%rcx), %eax ret Without the rule, llvm now generates: movzbl %sil, %ecx movzbl %dil, %eax incl %eax subl %ecx, %eax ret It also helps with ARM (and pretty much any target that doesn't have a sext i1 :-). The transformation was done as part of Eli's r75531. He has given the ok to remove it. rdar://11748024 llvm-svn: 159230
2012-06-27 06:03:13 +08:00
%conv = zext i1 %x to i32
%conv3 = zext i1 %y to i32
%conv3.neg = sub i32 0, %conv3
%sub = add i32 %conv, 1
%add = add i32 %sub, %conv3.neg
ret i32 %add
Canonicalize boolean +/- a constant to a select. (I think it's reasonably clear that we want to have a canonical form for constructs like this; if anyone thinks that a select is not the best canonical form, please tell me.) llvm-svn: 75531
2009-07-14 06:27:52 +08:00
}