[LegalizeTypes][NVPTX] Remove extra compare from fallback code for ISD::ADD in ExpandIntRes_ADDSUB.

This is the ultimate fallback code if UADDO isn't supported.

If the target uses 0/1 we used one compare, but if the target doesn't
use 0/1 we emitted two compares. Regardless of boolean constants we
should only need to check that the Result is less than one of the
original operands. So we only need one compare.

Reviewed By: spatel

Differential Revision: https://reviews.llvm.org/D133708
This commit is contained in:
Craig Topper 2022-09-13 09:07:56 -07:00
parent 393a17b5d1
commit efd5acf120
2 changed files with 10 additions and 17 deletions

View File

@ -2988,23 +2988,17 @@ void DAGTypeLegalizer::ExpandIntRes_ADDSUB(SDNode *N,
if (N->getOpcode() == ISD::ADD) {
Lo = DAG.getNode(ISD::ADD, dl, NVT, LoOps);
Hi = DAG.getNode(ISD::ADD, dl, NVT, makeArrayRef(HiOps, 2));
SDValue Cmp1 = DAG.getSetCC(dl, getSetCCResultType(NVT), Lo, LoOps[0],
ISD::SETULT);
SDValue Cmp = DAG.getSetCC(dl, getSetCCResultType(NVT), Lo, LoOps[0],
ISD::SETULT);
if (BoolType == TargetLoweringBase::ZeroOrOneBooleanContent) {
SDValue Carry = DAG.getZExtOrTrunc(Cmp1, dl, NVT);
Hi = DAG.getNode(ISD::ADD, dl, NVT, Hi, Carry);
return;
}
SDValue Carry;
if (BoolType == TargetLoweringBase::ZeroOrOneBooleanContent)
Carry = DAG.getZExtOrTrunc(Cmp, dl, NVT);
else
Carry = DAG.getSelect(dl, NVT, Cmp, DAG.getConstant(1, dl, NVT),
DAG.getConstant(0, dl, NVT));
SDValue Carry1 = DAG.getSelect(dl, NVT, Cmp1,
DAG.getConstant(1, dl, NVT),
DAG.getConstant(0, dl, NVT));
SDValue Cmp2 = DAG.getSetCC(dl, getSetCCResultType(NVT), Lo, LoOps[1],
ISD::SETULT);
SDValue Carry2 = DAG.getSelect(dl, NVT, Cmp2,
DAG.getConstant(1, dl, NVT), Carry1);
Hi = DAG.getNode(ISD::ADD, dl, NVT, Hi, Carry2);
Hi = DAG.getNode(ISD::ADD, dl, NVT, Hi, Carry);
} else {
Lo = DAG.getNode(ISD::SUB, dl, NVT, LoOps);
Hi = DAG.getNode(ISD::SUB, dl, NVT, makeArrayRef(HiOps, 2));

View File

@ -7,10 +7,9 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f3
; COMMON-LABEL: test_add
define i128 @test_add(i128 %a, i128 %b) {
; NOCARRY: add.s64
; NOCARRY-NEXT: setp.lt.u64
; NOCARRY-NEXT: add.s64
; NOCARRY-NEXT: setp.lt.u64
; NOCARRY-NEXT: selp.u64
; NOCARRY-NEXT: selp.b64
; NOCARRY-NEXT: add.s64
; CARRY: add.cc.s64