forked from OSchip/llvm-project
[InstCombine] Make (X|C1)^C2 -> X^(C1^C2) iff X&~C1 == 0 work for splat vectors
This also corrects the description to match what was actually implemented. The old comment said X^(C1|C2), but it implemented X^((C1|C2)&~(C1&C2)). I believe ((C1|C2)&~(C1&C2)) is equivalent to (C1^C2). Differential Revision: https://reviews.llvm.org/D36505 llvm-svn: 310658
This commit is contained in:
parent
699ae0c173
commit
9a6110b2d3
|
@ -2365,53 +2365,48 @@ Instruction *InstCombiner::visitXor(BinaryOperator &I) {
|
|||
{
|
||||
const APInt *RHSC;
|
||||
if (match(Op1, m_APInt(RHSC))) {
|
||||
Value *V;
|
||||
Value *X;
|
||||
const APInt *C;
|
||||
if (match(Op0, m_Sub(m_APInt(C), m_Value(V)))) {
|
||||
if (match(Op0, m_Sub(m_APInt(C), m_Value(X)))) {
|
||||
// ~(c-X) == X-c-1 == X+(-c-1)
|
||||
if (RHSC->isAllOnesValue()) {
|
||||
Constant *NewC = ConstantInt::get(I.getType(), -(*C) - 1);
|
||||
return BinaryOperator::CreateAdd(V, NewC);
|
||||
return BinaryOperator::CreateAdd(X, NewC);
|
||||
}
|
||||
if (RHSC->isSignMask()) {
|
||||
// (C - X) ^ signmask -> (C + signmask - X)
|
||||
Constant *NewC = ConstantInt::get(I.getType(), *C + *RHSC);
|
||||
return BinaryOperator::CreateSub(NewC, V);
|
||||
return BinaryOperator::CreateSub(NewC, X);
|
||||
}
|
||||
} else if (match(Op0, m_Add(m_Value(V), m_APInt(C)))) {
|
||||
} else if (match(Op0, m_Add(m_Value(X), m_APInt(C)))) {
|
||||
// ~(X-c) --> (-c-1)-X
|
||||
if (RHSC->isAllOnesValue()) {
|
||||
Constant *NewC = ConstantInt::get(I.getType(), -(*C) - 1);
|
||||
return BinaryOperator::CreateSub(NewC, V);
|
||||
return BinaryOperator::CreateSub(NewC, X);
|
||||
}
|
||||
if (RHSC->isSignMask()) {
|
||||
// (X + C) ^ signmask -> (X + C + signmask)
|
||||
Constant *NewC = ConstantInt::get(I.getType(), *C + *RHSC);
|
||||
return BinaryOperator::CreateAdd(V, NewC);
|
||||
return BinaryOperator::CreateAdd(X, NewC);
|
||||
}
|
||||
}
|
||||
|
||||
// (X|C1)^C2 -> X^(C1^C2) iff X&~C1 == 0
|
||||
if (match(Op0, m_Or(m_Value(X), m_APInt(C))) &&
|
||||
MaskedValueIsZero(X, *C, 0, &I)) {
|
||||
Constant *NewC = ConstantInt::get(I.getType(), *C ^ *RHSC);
|
||||
Worklist.Add(cast<Instruction>(Op0));
|
||||
I.setOperand(0, X);
|
||||
I.setOperand(1, NewC);
|
||||
return &I;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (ConstantInt *RHSC = dyn_cast<ConstantInt>(Op1)) {
|
||||
if (BinaryOperator *Op0I = dyn_cast<BinaryOperator>(Op0)) {
|
||||
if (ConstantInt *Op0CI = dyn_cast<ConstantInt>(Op0I->getOperand(1))) {
|
||||
if (Op0I->getOpcode() == Instruction::Or) {
|
||||
// (X|C1)^C2 -> X^(C1|C2) iff X&~C1 == 0
|
||||
if (MaskedValueIsZero(Op0I->getOperand(0), Op0CI->getValue(),
|
||||
0, &I)) {
|
||||
Constant *NewRHS = ConstantExpr::getOr(Op0CI, RHSC);
|
||||
// Anything in both C1 and C2 is known to be zero, remove it from
|
||||
// NewRHS.
|
||||
Constant *CommonBits = ConstantExpr::getAnd(Op0CI, RHSC);
|
||||
NewRHS = ConstantExpr::getAnd(NewRHS,
|
||||
ConstantExpr::getNot(CommonBits));
|
||||
Worklist.Add(Op0I);
|
||||
I.setOperand(0, Op0I->getOperand(0));
|
||||
I.setOperand(1, NewRHS);
|
||||
return &I;
|
||||
}
|
||||
} else if (Op0I->getOpcode() == Instruction::LShr) {
|
||||
if (Op0I->getOpcode() == Instruction::LShr) {
|
||||
// ((X^C1) >> C2) ^ C3 -> (X>>C2) ^ ((C1>>C2)^C3)
|
||||
// E1 = "X ^ C1"
|
||||
BinaryOperator *E1;
|
||||
|
|
|
@ -143,6 +143,19 @@ define i16 @test23(i16 %A) {
|
|||
ret i16 %D
|
||||
}
|
||||
|
||||
define <2 x i16> @test23vec(<2 x i16> %A) {
|
||||
; CHECK-LABEL: @test23vec(
|
||||
; CHECK-NEXT: [[B:%.*]] = lshr <2 x i16> [[A:%.*]], <i16 1, i16 1>
|
||||
; CHECK-NEXT: [[D:%.*]] = xor <2 x i16> [[B]], <i16 -24575, i16 -24575>
|
||||
; CHECK-NEXT: ret <2 x i16> [[D]]
|
||||
;
|
||||
%B = lshr <2 x i16> %A, <i16 1, i16 1>
|
||||
;; fold or into xor
|
||||
%C = or <2 x i16> %B, <i16 -32768, i16 -32768>
|
||||
%D = xor <2 x i16> %C, <i16 8193, i16 8193>
|
||||
ret <2 x i16> %D
|
||||
}
|
||||
|
||||
; PR1738
|
||||
define i1 @test24(double %X, double %Y) {
|
||||
; CHECK-LABEL: @test24(
|
||||
|
|
|
@ -273,9 +273,8 @@ define <2 x i32> @test65vec(<2 x i64> %x) {
|
|||
; CHECK-NEXT: [[TMP1:%.*]] = lshr <2 x i64> [[X:%.*]], <i64 3, i64 3>
|
||||
; CHECK-NEXT: [[TMP2:%.*]] = trunc <2 x i64> [[TMP1]] to <2 x i32>
|
||||
; CHECK-NEXT: [[TMP3:%.*]] = and <2 x i32> [[TMP2]], <i32 2, i32 2>
|
||||
; CHECK-NEXT: [[TMP4:%.*]] = or <2 x i32> [[TMP3]], <i32 40, i32 40>
|
||||
; CHECK-NEXT: [[TMP5:%.*]] = xor <2 x i32> [[TMP4]], <i32 2, i32 2>
|
||||
; CHECK-NEXT: ret <2 x i32> [[TMP5]]
|
||||
; CHECK-NEXT: [[TMP4:%.*]] = xor <2 x i32> [[TMP3]], <i32 42, i32 42>
|
||||
; CHECK-NEXT: ret <2 x i32> [[TMP4]]
|
||||
;
|
||||
%1 = and <2 x i64> %x, <i64 16, i64 16>
|
||||
%2 = icmp ne <2 x i64> %1, zeroinitializer
|
||||
|
@ -302,9 +301,8 @@ define <2 x i32> @test66vec(<2 x i64> %x) {
|
|||
; CHECK-NEXT: [[TMP1:%.*]] = lshr <2 x i64> [[X:%.*]], <i64 31, i64 31>
|
||||
; CHECK-NEXT: [[TMP2:%.*]] = trunc <2 x i64> [[TMP1]] to <2 x i32>
|
||||
; CHECK-NEXT: [[TMP3:%.*]] = and <2 x i32> [[TMP2]], <i32 2, i32 2>
|
||||
; CHECK-NEXT: [[TMP4:%.*]] = or <2 x i32> [[TMP3]], <i32 40, i32 40>
|
||||
; CHECK-NEXT: [[TMP5:%.*]] = xor <2 x i32> [[TMP4]], <i32 2, i32 2>
|
||||
; CHECK-NEXT: ret <2 x i32> [[TMP5]]
|
||||
; CHECK-NEXT: [[TMP4:%.*]] = xor <2 x i32> [[TMP3]], <i32 42, i32 42>
|
||||
; CHECK-NEXT: ret <2 x i32> [[TMP4]]
|
||||
;
|
||||
%1 = and <2 x i64> %x, <i64 4294967296, i64 4294967296>
|
||||
%2 = icmp ne <2 x i64> %1, zeroinitializer
|
||||
|
@ -344,10 +342,9 @@ define <2 x i32> @test67vec(<2 x i16> %x) {
|
|||
; CHECK-LABEL: @test67vec(
|
||||
; CHECK-NEXT: [[TMP1:%.*]] = lshr <2 x i16> [[X:%.*]], <i16 1, i16 1>
|
||||
; CHECK-NEXT: [[TMP2:%.*]] = and <2 x i16> [[TMP1]], <i16 2, i16 2>
|
||||
; CHECK-NEXT: [[TMP3:%.*]] = or <2 x i16> [[TMP2]], <i16 40, i16 40>
|
||||
; CHECK-NEXT: [[TMP4:%.*]] = xor <2 x i16> [[TMP3]], <i16 2, i16 2>
|
||||
; CHECK-NEXT: [[TMP5:%.*]] = zext <2 x i16> [[TMP4]] to <2 x i32>
|
||||
; CHECK-NEXT: ret <2 x i32> [[TMP5]]
|
||||
; CHECK-NEXT: [[TMP3:%.*]] = xor <2 x i16> [[TMP2]], <i16 42, i16 42>
|
||||
; CHECK-NEXT: [[TMP4:%.*]] = zext <2 x i16> [[TMP3]] to <2 x i32>
|
||||
; CHECK-NEXT: ret <2 x i32> [[TMP4]]
|
||||
;
|
||||
%1 = and <2 x i16> %x, <i16 4, i16 4>
|
||||
%2 = icmp ne <2 x i16> %1, zeroinitializer
|
||||
|
|
Loading…
Reference in New Issue