diff options
author | Tony Jiang <jtony@ca.ibm.com> | 2017-06-12 17:58:42 +0000 |
---|---|---|
committer | Tony Jiang <jtony@ca.ibm.com> | 2017-06-12 17:58:42 +0000 |
commit | 30a49d1a3d5bd996eb3d6d7403a86dadde75c57c (patch) | |
tree | 0f07780e4051240c1a6558a822d1256bdeaf1efa /llvm/lib/Target | |
parent | 4f01175ef926a40569089fc5f7310c64f508bed6 (diff) | |
download | bcm5719-llvm-30a49d1a3d5bd996eb3d6d7403a86dadde75c57c.tar.gz bcm5719-llvm-30a49d1a3d5bd996eb3d6d7403a86dadde75c57c.zip |
[Power9] Added support for the modsw, moduw, modsd, modud hardware instructions.
Note that if we need the result of both the divide and the modulo then we
compute the modulo based on the result of the divide and not using the new
hardware instruction.
Commit on behalf of STEFAN PINTILIE.
Differential Revision: https://reviews.llvm.org/D33940
llvm-svn: 305210
Diffstat (limited to 'llvm/lib/Target')
-rw-r--r-- | llvm/lib/Target/PowerPC/PPCISelLowering.cpp | 37 | ||||
-rw-r--r-- | llvm/lib/Target/PowerPC/PPCISelLowering.h | 1 | ||||
-rw-r--r-- | llvm/lib/Target/PowerPC/PPCInstr64Bit.td | 10 | ||||
-rw-r--r-- | llvm/lib/Target/PowerPC/PPCInstrInfo.td | 8 |
4 files changed, 51 insertions, 5 deletions
diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp index c78d3fa45c5..8880e9d88fb 100644 --- a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp +++ b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp @@ -204,11 +204,23 @@ PPCTargetLowering::PPCTargetLowering(const PPCTargetMachine &TM, setOperationAction(ISD::FNEARBYINT, MVT::ppcf128, Expand); setOperationAction(ISD::FREM, MVT::ppcf128, Expand); - // PowerPC has no SREM/UREM instructions - setOperationAction(ISD::SREM, MVT::i32, Expand); - setOperationAction(ISD::UREM, MVT::i32, Expand); - setOperationAction(ISD::SREM, MVT::i64, Expand); - setOperationAction(ISD::UREM, MVT::i64, Expand); + // PowerPC has no SREM/UREM instructions unless we are on P9 + // On P9 we may use a hardware instruction to compute the remainder. + // The instructions are not legalized directly because in the cases where the + // result of both the remainder and the division is required it is more + // efficient to compute the remainder from the result of the division rather + // than use the remainder instruction. + if (Subtarget.isISA3_0()) { + setOperationAction(ISD::SREM, MVT::i32, Custom); + setOperationAction(ISD::UREM, MVT::i32, Custom); + setOperationAction(ISD::SREM, MVT::i64, Custom); + setOperationAction(ISD::UREM, MVT::i64, Custom); + } else { + setOperationAction(ISD::SREM, MVT::i32, Expand); + setOperationAction(ISD::UREM, MVT::i32, Expand); + setOperationAction(ISD::SREM, MVT::i64, Expand); + setOperationAction(ISD::UREM, MVT::i64, Expand); + } // Don't use SMUL_LOHI/UMUL_LOHI or SDIVREM/UDIVREM to lower SREM/UREM. setOperationAction(ISD::UMUL_LOHI, MVT::i32, Expand); @@ -8394,6 +8406,18 @@ SDValue PPCTargetLowering::LowerINTRINSIC_VOID(SDValue Op, return SDValue(); } +SDValue PPCTargetLowering::LowerREM(SDValue Op, SelectionDAG &DAG) const { + // Check for a DIV with the same operands as this REM. + for (auto UI : Op.getOperand(1)->uses()) { + if ((Op.getOpcode() == ISD::SREM && UI->getOpcode() == ISD::SDIV) || + (Op.getOpcode() == ISD::UREM && UI->getOpcode() == ISD::UDIV)) + if (UI->getOperand(0) == Op.getOperand(0) && + UI->getOperand(1) == Op.getOperand(1)) + return SDValue(); + } + return Op; +} + SDValue PPCTargetLowering::LowerSIGN_EXTEND_INREG(SDValue Op, SelectionDAG &DAG) const { SDLoc dl(Op); @@ -8862,6 +8886,9 @@ SDValue PPCTargetLowering::LowerOperation(SDValue Op, SelectionDAG &DAG) const { case ISD::INTRINSIC_VOID: return LowerINTRINSIC_VOID(Op, DAG); + case ISD::SREM: + case ISD::UREM: + return LowerREM(Op, DAG); } } diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.h b/llvm/lib/Target/PowerPC/PPCISelLowering.h index 7982a4a9e9f..678d92898a0 100644 --- a/llvm/lib/Target/PowerPC/PPCISelLowering.h +++ b/llvm/lib/Target/PowerPC/PPCISelLowering.h @@ -918,6 +918,7 @@ namespace llvm { SDValue LowerEXTRACT_VECTOR_ELT(SDValue Op, SelectionDAG &DAG) const; SDValue LowerINTRINSIC_WO_CHAIN(SDValue Op, SelectionDAG &DAG) const; SDValue LowerINTRINSIC_VOID(SDValue Op, SelectionDAG &DAG) const; + SDValue LowerREM(SDValue Op, SelectionDAG &DAG) const; SDValue LowerSCALAR_TO_VECTOR(SDValue Op, SelectionDAG &DAG) const; SDValue LowerSIGN_EXTEND_INREG(SDValue Op, SelectionDAG &DAG) const; SDValue LowerMUL(SDValue Op, SelectionDAG &DAG) const; diff --git a/llvm/lib/Target/PowerPC/PPCInstr64Bit.td b/llvm/lib/Target/PowerPC/PPCInstr64Bit.td index 295590b2acf..70536a6039b 100644 --- a/llvm/lib/Target/PowerPC/PPCInstr64Bit.td +++ b/llvm/lib/Target/PowerPC/PPCInstr64Bit.td @@ -683,6 +683,16 @@ def DIVDE : XOForm_1<31, 425, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB), "divde $rT, $rA, $rB", IIC_IntDivD, [(set i64:$rT, (int_ppc_divde g8rc:$rA, g8rc:$rB))]>, isPPC64, Requires<[HasExtDiv]>; + +let Predicates = [IsISA3_0] in { +def MODSD : XForm_8<31, 777, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB), + "modsd $rT, $rA, $rB", IIC_IntDivW, + [(set i64:$rT, (srem i64:$rA, i64:$rB))]>; +def MODUD : XForm_8<31, 265, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB), + "modud $rT, $rA, $rB", IIC_IntDivW, + [(set i64:$rT, (urem i64:$rA, i64:$rB))]>; +} + let Defs = [CR0] in def DIVDEo : XOForm_1<31, 425, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB), "divde. $rT, $rA, $rB", IIC_IntDivD, diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.td b/llvm/lib/Target/PowerPC/PPCInstrInfo.td index 8223aa655e3..37506239c58 100644 --- a/llvm/lib/Target/PowerPC/PPCInstrInfo.td +++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.td @@ -2544,6 +2544,14 @@ let Uses = [RM] in { "mffs. $rT", IIC_IntMFFS, []>, isDOT; } +let Predicates = [IsISA3_0] in { +def MODSW : XForm_8<31, 779, (outs gprc:$rT), (ins gprc:$rA, gprc:$rB), + "modsw $rT, $rA, $rB", IIC_IntDivW, + [(set i32:$rT, (srem i32:$rA, i32:$rB))]>; +def MODUW : XForm_8<31, 267, (outs gprc:$rT), (ins gprc:$rA, gprc:$rB), + "moduw $rT, $rA, $rB", IIC_IntDivW, + [(set i32:$rT, (urem i32:$rA, i32:$rB))]>; +} let PPC970_Unit = 1, hasSideEffects = 0 in { // FXU Operations. // XO-Form instructions. Arithmetic instructions that can set overflow bit |