[llvm] r354224 - [InstCombine] reduce more unsigned saturated add with 'not' op
Sanjay Patel via llvm-commits
llvm-commits at lists.llvm.org
Sun Feb 17 08:48:50 PST 2019
Author: spatel
Date: Sun Feb 17 08:48:50 2019
New Revision: 354224
URL: http://llvm.org/viewvc/llvm-project?rev=354224&view=rev
Log:
[InstCombine] reduce more unsigned saturated add with 'not' op
We want to use the sum in the icmp to allow matching with
m_UAddWithOverflow and eliminate the 'not'. This is discussed
in D51929 and is another step towards solving PR14613:
https://bugs.llvm.org/show_bug.cgi?id=14613
Name: not op
%notx = xor i32 %x, -1
%a = add i32 %x, %y
%c = icmp ult i32 %notx, %y
%r = select i1 %c, i32 -1, i32 %a
=>
%a = add i32 %x, %y
%c2 = icmp ult i32 %a, %y
%r = select i1 %c2, i32 -1, i32 %a
Name: not op ugt
%notx = xor i32 %x, -1
%a = add i32 %x, %y
%c = icmp ugt i32 %y, %notx
%r = select i1 %c, i32 -1, i32 %a
=>
%a = add i32 %x, %y
%c2 = icmp ult i32 %a, %y
%r = select i1 %c2, i32 -1, i32 %a
https://rise4fun.com/Alive/niom
(The matching here is still incomplete.)
Modified:
llvm/trunk/lib/Transforms/InstCombine/InstCombineSelect.cpp
llvm/trunk/test/Transforms/InstCombine/saturating-add-sub.ll
Modified: llvm/trunk/lib/Transforms/InstCombine/InstCombineSelect.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/InstCombine/InstCombineSelect.cpp?rev=354224&r1=354223&r2=354224&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/InstCombine/InstCombineSelect.cpp (original)
+++ llvm/trunk/lib/Transforms/InstCombine/InstCombineSelect.cpp Sun Feb 17 08:48:50 2019
@@ -677,13 +677,23 @@ static Value *canonicalizeSaturatedSubtr
static Value *canonicalizeSaturatedAdd(ICmpInst *Cmp, Value *TVal, Value *FVal,
InstCombiner::BuilderTy &Builder) {
- if (!Cmp->hasOneUse() || Cmp->getPredicate() != ICmpInst::ICMP_ULT)
+ if (!Cmp->hasOneUse())
return nullptr;
- // Match unsigned saturated add of 2 variables with an unnecessary 'not'.
- // TODO: There are more variations of this pattern.
+ // Canonicalize to 'ULT' to simplify matching below.
Value *Cmp0 = Cmp->getOperand(0);
Value *Cmp1 = Cmp->getOperand(1);
+ ICmpInst::Predicate Pred = Cmp->getPredicate();
+ if (Pred == ICmpInst::ICMP_UGT) {
+ Pred = ICmpInst::ICMP_ULT;
+ std::swap(Cmp0, Cmp1);
+ }
+
+ if (Pred != ICmpInst::ICMP_ULT)
+ return nullptr;
+
+ // Match unsigned saturated add of 2 variables with an unnecessary 'not'.
+ // TODO: There are more variations of this pattern.
Value *X, *Y;
if (match(TVal, m_AllOnes()) && match(Cmp0, m_Not(m_Value(X))) &&
match(FVal, m_c_Add(m_Specific(X), m_Value(Y))) && Y == Cmp1) {
Modified: llvm/trunk/test/Transforms/InstCombine/saturating-add-sub.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/InstCombine/saturating-add-sub.ll?rev=354224&r1=354223&r2=354224&view=diff
==============================================================================
--- llvm/trunk/test/Transforms/InstCombine/saturating-add-sub.ll (original)
+++ llvm/trunk/test/Transforms/InstCombine/saturating-add-sub.ll Sun Feb 17 08:48:50 2019
@@ -672,11 +672,10 @@ define i32 @uadd_sat_commute_add(i32 %xp
define i32 @uadd_sat_ugt(i32 %x, i32 %yp) {
; CHECK-LABEL: @uadd_sat_ugt(
; CHECK-NEXT: [[Y:%.*]] = sdiv i32 [[YP:%.*]], 2442
-; CHECK-NEXT: [[NOTX:%.*]] = xor i32 [[X:%.*]], -1
-; CHECK-NEXT: [[A:%.*]] = add i32 [[Y]], [[X]]
-; CHECK-NEXT: [[C:%.*]] = icmp ugt i32 [[Y]], [[NOTX]]
-; CHECK-NEXT: [[R:%.*]] = select i1 [[C]], i32 -1, i32 [[A]]
-; CHECK-NEXT: ret i32 [[R]]
+; CHECK-NEXT: [[A:%.*]] = add i32 [[Y]], [[X:%.*]]
+; CHECK-NEXT: [[TMP1:%.*]] = icmp ult i32 [[A]], [[Y]]
+; CHECK-NEXT: [[TMP2:%.*]] = select i1 [[TMP1]], i32 -1, i32 [[A]]
+; CHECK-NEXT: ret i32 [[TMP2]]
;
%y = sdiv i32 %yp, 2442 ; thwart complexity-based-canonicalization
%notx = xor i32 %x, -1
@@ -690,11 +689,10 @@ define i32 @uadd_sat_ugt_commute_add(i32
; CHECK-LABEL: @uadd_sat_ugt_commute_add(
; CHECK-NEXT: [[Y:%.*]] = sdiv i32 [[YP:%.*]], 2442
; CHECK-NEXT: [[X:%.*]] = srem i32 42, [[XP:%.*]]
-; CHECK-NEXT: [[NOTX:%.*]] = xor i32 [[X]], -1
; CHECK-NEXT: [[A:%.*]] = add i32 [[X]], [[Y]]
-; CHECK-NEXT: [[C:%.*]] = icmp ugt i32 [[Y]], [[NOTX]]
-; CHECK-NEXT: [[R:%.*]] = select i1 [[C]], i32 -1, i32 [[A]]
-; CHECK-NEXT: ret i32 [[R]]
+; CHECK-NEXT: [[TMP1:%.*]] = icmp ult i32 [[A]], [[Y]]
+; CHECK-NEXT: [[TMP2:%.*]] = select i1 [[TMP1]], i32 -1, i32 [[A]]
+; CHECK-NEXT: ret i32 [[TMP2]]
;
%y = sdiv i32 %yp, 2442 ; thwart complexity-based-canonicalization
%x = srem i32 42, %xp ; thwart complexity-based-canonicalization
More information about the llvm-commits
mailing list