[Power9] Added support for the modsw, moduw, modsd, modud hardware instructions that are new to P9.

[stefanp]

Summary

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.

[stefanp]

I did not set the
let Predicates = [IsISA3_0] in {
correctly.
Fixed in this update.

[jtony]

lib/Target/PowerPC/PPCISelLowering.cpp

207–223

SREM/UREM are actually not instructions, they are SDNode or operations, I think.

test/CodeGen/PowerPC/ppc64-P9-mod.ll

2

Do we need to test Big-endian also?

[nemanjai]

Aside from a few minor inline nits, LGTM. Please address the inline comments on the commit.

lib/Target/PowerPC/PPCISelLowering.cpp

207–223

Sure. Technically. But I think the comment is fine.

8405

Well, this isn't completely accurate. Probably more accurate to say something like
// Check for a DIV with the same operands as this REM.

8406

The temporary isn't really needed here. The loop can just be
for (auto UI : Op.getOperand(1)->uses())

test/CodeGen/PowerPC/ppc64-P9-mod.ll

2

Yes, I agree. The codegen (and therefore CHECK directives) don't need to change, but please add a RUN line with triple -mtriple=powerpc64-unknown-linux-gnu.

21

For the actual MOD instructions you're checking for where the operands are parameters, please add the input registers to the CHECK directives since they're mandated by the ABI. They should look something like
; CHECK: modsw {{[0-9]+}}, 3, 4

I don't think we care to do that for the checks of existing code though (i.e. the dif/mul/sub).

[inouehrs]

lib/Target/PowerPC/PPCISelLowering.cpp

208

nit: harware -> hardware

test/CodeGen/PowerPC/ppc64-P9-mod.ll

2

Also, it is better to add -verify-machineinstrs

[stefanp]

Fixed various comments from reviewers.

lib/Target/PowerPC/PPCISelLowering.h

private:
     SDValue LowerSRL_PARTS(SDValue Op, SelectionDAG &DAG) const;
     SDValue LowerSRA_PARTS(SDValue Op, SelectionDAG &DAG) const;
     SDValue LowerBUILD_VECTOR(SDValue Op, SelectionDAG &DAG) const;
     SDValue LowerVECTOR_SHUFFLE(SDValue Op, SelectionDAG &DAG) const;
     SDValue LowerINSERT_VECTOR_ELT(SDValue Op, SelectionDAG &DAG) const;
     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;
 
     SDValue LowerVectorLoad(SDValue Op, SelectionDAG &DAG) const;
     SDValue LowerVectorStore(SDValue Op, SelectionDAG &DAG) const;
 
     SDValue LowerCallResult(SDValue Chain, SDValue InFlag,

lib/Target/PowerPC/PPCISelLowering.cpp

PPCTargetLowering::PPCTargetLowering(const PPCTargetMachine &TM,
   // We do not currently implement these libm ops for PowerPC.
   setOperationAction(ISD::FFLOOR, MVT::ppcf128, Expand);
   setOperationAction(ISD::FCEIL,  MVT::ppcf128, Expand);
   setOperationAction(ISD::FTRUNC, MVT::ppcf128, Expand);
   setOperationAction(ISD::FRINT,  MVT::ppcf128, Expand);
   setOperationAction(ISD::FNEARBYINT, MVT::ppcf128, Expand);
   setOperationAction(ISD::FREM, MVT::ppcf128, Expand);
 
-  // PowerPC has no SREM/UREM instructions
+  // PowerPC has no SREM/UREM instructions unless we are on P9
+  // On P9 we may use a harware 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);
   setOperationAction(ISD::SMUL_LOHI, MVT::i32, Expand);
   setOperationAction(ISD::UMUL_LOHI, MVT::i64, Expand);
   setOperationAction(ISD::SMUL_LOHI, MVT::i64, Expand);
   setOperationAction(ISD::UDIVREM, MVT::i32, Expand);
   setOperationAction(ISD::SDIVREM, MVT::i32, Expand);
return SDValue(DAG.getMachineNode(PPC::CFENCE8, DL, MVT::Other,
                    0);
   }
   default:
     break;
   }
   return SDValue();
 }
 
+SDValue PPCTargetLowering::LowerREM(SDValue Op, SelectionDAG &DAG) const {
+  // Check for a DIV that requires the same result as this REM.
+  SDNode* Denominator = Op.getOperand(1).getNode();
+  for (auto UI : Denominator->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);
   // For v2i64 (VSX), we can pattern patch the v2i32 case (using fp <-> int
   // instructions), but for smaller types, we need to first extend up to v2i32
   // before doing going farther.
   if (Op.getValueType() == MVT::v2i64) {
     EVT ExtVT = cast<VTSDNode>(Op.getOperand(1))->getVT();
SDValue PPCTargetLowering::LowerOperation(SDValue Op, SelectionDAG &DAG) const {
   case ISD::INTRINSIC_W_CHAIN:  return SDValue();
 
   // Frame & Return address.
   case ISD::RETURNADDR:         return LowerRETURNADDR(Op, DAG);
   case ISD::FRAMEADDR:          return LowerFRAMEADDR(Op, DAG);
 
   case ISD::INTRINSIC_VOID:
     return LowerINTRINSIC_VOID(Op, DAG);
+  case ISD::SREM:
+  case ISD::UREM:
+    return LowerREM(Op, DAG);
   }
 }
 
 void PPCTargetLowering::ReplaceNodeResults(SDNode *N,
                                            SmallVectorImpl<SDValue>&Results,
                                            SelectionDAG &DAG) const {
   SDLoc dl(N);
   switch (N->getOpcode()) {

lib/Target/PowerPC/PPCInstr64Bit.td

defm DIVD  : XOForm_1rcr<31, 489, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
                           [(set i64:$rT, (sdiv i64:$rA, i64:$rB))]>, isPPC64;
 defm DIVDU : XOForm_1rcr<31, 457, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
                           "divdu", "$rT, $rA, $rB", IIC_IntDivD,
                           [(set i64:$rT, (udiv i64:$rA, i64:$rB))]>, isPPC64;
 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]>;
+
+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, $rB", IIC_IntDivD,
                       []>, isDOT, PPC970_DGroup_Cracked, PPC970_DGroup_First,
                       isPPC64, Requires<[HasExtDiv]>;
 def DIVDEU : XOForm_1<31, 393, 0, (outs g8rc:$rT), (ins g8rc:$rA, g8rc:$rB),
                       "divdeu $rT, $rA, $rB", IIC_IntDivD,
                       [(set i64:$rT, (int_ppc_divdeu g8rc:$rA, g8rc:$rB))]>,

lib/Target/PowerPC/PPCInstrInfo.td

defm DIVW  : XOForm_1rcr<31, 491, 0, (outs gprc:$rT), (ins gprc:$rA, gprc:$rB),
                           [(set i32:$rT, (sdiv i32:$rA, i32:$rB))]>;
 defm DIVWU : XOForm_1rcr<31, 459, 0, (outs gprc:$rT), (ins gprc:$rA, gprc:$rB),
                           "divwu", "$rT, $rA, $rB", IIC_IntDivW,
                           [(set i32:$rT, (udiv i32:$rA, i32:$rB))]>;
 def DIVWE : XOForm_1<31, 427, 0, (outs gprc:$rT), (ins gprc:$rA, gprc:$rB),
                      "divwe $rT, $rA, $rB", IIC_IntDivW,
                      [(set i32:$rT, (int_ppc_divwe gprc:$rA, gprc:$rB))]>,
                      Requires<[HasExtDiv]>;
+
+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 Defs = [CR0] in
 def DIVWEo : XOForm_1<31, 427, 0, (outs gprc:$rT), (ins gprc:$rA, gprc:$rB),
                       "divwe. $rT, $rA, $rB", IIC_IntDivW,
                       []>, isDOT, PPC970_DGroup_Cracked, PPC970_DGroup_First,
                       Requires<[HasExtDiv]>;
 def DIVWEU : XOForm_1<31, 395, 0, (outs gprc:$rT), (ins gprc:$rA, gprc:$rB),
                       "divweu $rT, $rA, $rB", IIC_IntDivW,
                       [(set i32:$rT, (int_ppc_divweu gprc:$rA, gprc:$rB))]>,

test/CodeGen/PowerPC/ppc64-P9-mod.ll

+; RUN: llc < %s -mtriple=powerpc64le-unknown-linux-gnu -mcpu=pwr9 | FileCheck %s
+; RUN: llc < %s -mtriple=powerpc64le-unknown-linux-gnu -mcpu=pwr8 | FileCheck %s -check-prefix=CHECK-PWR8 -implicit-check-not mod[us][wd]
+
+@mod_resultsw = common local_unnamed_addr global i32 0, align 4
+@mod_resultud = common local_unnamed_addr global i64 0, align 8
+@div_resultsw = common local_unnamed_addr global i32 0, align 4
+@mod_resultuw = common local_unnamed_addr global i32 0, align 4
+@div_resultuw = common local_unnamed_addr global i32 0, align 4
+@div_resultsd = common local_unnamed_addr global i64 0, align 8
+@mod_resultsd = common local_unnamed_addr global i64 0, align 8
+@div_resultud = common local_unnamed_addr global i64 0, align 8
+
+; Function Attrs: norecurse nounwind
+define void @modulo_sw(i32 signext %a, i32 signext %b) local_unnamed_addr {
+entry:
+  %rem = srem i32 %a, %b
+  store i32 %rem, i32* @mod_resultsw, align 4
+  ret void
+; CHECK-LABEL: modulo_sw
+; CHECK: modsw
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_sw
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind readnone
+define zeroext i32 @modulo_uw(i32 zeroext %a, i32 zeroext %b) local_unnamed_addr {
+entry:
+  %rem = urem i32 %a, %b
+  ret i32 %rem
+; CHECK-LABEL: modulo_uw
+; CHECK: moduw
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_uw
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind readnone
+define i64 @modulo_sd(i64 %a, i64 %b) local_unnamed_addr {
+entry:
+  %rem = srem i64 %a, %b
+  ret i64 %rem
+; CHECK-LABEL: modulo_sd
+; CHECK: modsd
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_sd
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind
+define void @modulo_ud(i64 %a, i64 %b) local_unnamed_addr {
+entry:
+  %rem = urem i64 %a, %b
+  store i64 %rem, i64* @mod_resultud, align 8
+  ret void
+; CHECK-LABEL: modulo_ud
+; CHECK: modud
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_ud
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind
+define void @modulo_div_sw(i32 signext %a, i32 signext %b) local_unnamed_addr {
+entry:
+  %rem = srem i32 %a, %b
+  store i32 %rem, i32* @mod_resultsw, align 4
+  %div = sdiv i32 %a, %b
+  store i32 %div, i32* @div_resultsw, align 4
+  ret void
+; CHECK-LABEL: modulo_div_sw
+; CHECK-NOT: modsw
+; CHECK: div
+; CHECK-NOT: modsw
+; CHECK: mull
+; CHECK-NOT: modsw
+; CHECK: sub
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_div_sw
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind
+define void @modulo_div_abc_sw(i32 signext %a, i32 signext %b, i32 signext %c) local_unnamed_addr {
+entry:
+  %rem = srem i32 %a, %c
+  store i32 %rem, i32* @mod_resultsw, align 4
+  %div = sdiv i32 %b, %c
+  store i32 %div, i32* @div_resultsw, align 4
+  ret void
+; CHECK-LABEL: modulo_div_abc_sw
+; CHECK: modsw
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_div_abc_sw
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind
+define void @modulo_div_uw(i32 zeroext %a, i32 zeroext %b) local_unnamed_addr {
+entry:
+  %rem = urem i32 %a, %b
+  store i32 %rem, i32* @mod_resultuw, align 4
+  %div = udiv i32 %a, %b
+  store i32 %div, i32* @div_resultuw, align 4
+  ret void
+; CHECK-LABEL: modulo_div_uw
+; CHECK-NOT: modsw
+; CHECK: div
+; CHECK-NOT: modsw
+; CHECK: mull
+; CHECK-NOT: modsw
+; CHECK: sub
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_div_uw
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind
+define void @modulo_div_swuw(i32 signext %a, i32 signext %b) local_unnamed_addr {
+entry:
+  %rem = srem i32 %a, %b
+  store i32 %rem, i32* @mod_resultsw, align 4
+  %div = udiv i32 %a, %b
+  store i32 %div, i32* @div_resultsw, align 4
+  ret void
+; CHECK-LABEL: modulo_div_swuw
+; CHECK: modsw
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_div_swuw
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind
+define void @modulo_div_udsd(i64 %a, i64 %b) local_unnamed_addr {
+entry:
+  %rem = urem i64 %a, %b
+  store i64 %rem, i64* @mod_resultud, align 8
+  %div = sdiv i64 %a, %b
+  store i64 %div, i64* @div_resultsd, align 8
+  ret void
+; CHECK-LABEL: modulo_div_udsd
+; CHECK: modud
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_div_udsd
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind
+define void @modulo_const32_sw(i32 signext %a) local_unnamed_addr {
+entry:
+  %rem = srem i32 %a, 32
+  store i32 %rem, i32* @mod_resultsw, align 4
+  ret void
+; CHECK-LABEL: modulo_const32_sw
+; CHECK-NOT: modsw
+; CHECK: srawi
+; CHECK-NOT: modsw
+; CHECK: addze
+; CHECK-NOT: modsw
+; CHECK: slwi
+; CHECK-NOT: modsw
+; CHECK: subf
+; CHECK-NOT: modsw
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_const32_sw
+; CHECK-PWR8: srawi
+; CHECK-PWR8: addze
+; CHECK-PWR8: slwi
+; CHECK-PWR8: subf
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind readnone
+define signext i32 @modulo_const3_sw(i32 signext %a) local_unnamed_addr {
+entry:
+  %rem = srem i32 %a, 3
+  ret i32 %rem
+; CHECK-LABEL: modulo_const3_sw
+; CHECK-NOT: modsw
+; CHECK: mull
+; CHECK-NOT: modsw
+; CHECK: sub
+; CHECK-NOT: modsw
+; CHECK: blr
+; CHECK-PWR8-LABEL: modulo_const3_sw
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind readnone
+define signext i32 @const2_modulo_sw(i32 signext %a) local_unnamed_addr {
+entry:
+  %rem = srem i32 2, %a
+  ret i32 %rem
+; CHECK-LABEL: const2_modulo_sw
+; CHECK: modsw
+; CHECK: blr
+; CHECK-PWR8-LABEL: const2_modulo_sw
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+; Function Attrs: norecurse nounwind
+; FIXME On power 9 this test will still produce modsw because the divide is in
+; a different block than the remainder. Due to the nature of the SDAG we cannot
+; see the div in the other block.
+define void @blocks_modulo_div_sw(i32 signext %a, i32 signext %b, i32 signext %c) local_unnamed_addr {
+entry:
+  %div = sdiv i32 %a, %b
+  store i32 %div, i32* @div_resultsw, align 4
+  %cmp = icmp sgt i32 %c, 0
+  br i1 %cmp, label %if.then, label %if.end
+
+if.then:                                          ; preds = %entry
+  %rem = srem i32 %a, %b
+  store i32 %rem, i32* @mod_resultsw, align 4
+  br label %if.end
+
+if.end:                                           ; preds = %if.then, %entry
+  ret void
+; CHECK-LABEL: blocks_modulo_div_sw
+; CHECK: div
+; CHECK: modsw
+; CHECK: blr
+; CHECK-PWR8-LABEL: blocks_modulo_div_sw
+; CHECK-PWR8: div
+; CHECK-PWR8: mull
+; CHECK-PWR8: sub
+; CHECK-PWR8: blr
+}
+
+

test/MC/Disassembler/PowerPC/ppc64-encoding.txt

 0x7c 0x43 0x23 0x57
 
 # CHECK: divweu 2, 3, 4
 0x7c 0x43 0x23 0x16
 
 # CHECK: divweu. 2, 3, 4
 0x7c 0x43 0x23 0x17
 
+# CHECK: modsw 2, 3, 4
+0x7c 0x43 0x26 0x16
+
+# CHECK: moduw 2, 3, 4
+0x7c 0x43 0x22 0x16
+
+# CHECK: modsd 2, 3, 4
+0x7c 0x43 0x26 0x12
+
+# CHECK: modud 2, 3, 4
+0x7c 0x43 0x22 0x12
+
 # CHECK: mulld 2, 3, 4                   
 0x7c 0x43 0x21 0xd2
 
 # CHECK: mulld. 2, 3, 4                  
 0x7c 0x43 0x21 0xd3
 
 # CHECK: mulhd 2, 3, 4                   
 0x7c 0x43 0x20 0x92

test/MC/Disassembler/PowerPC/ppc64le-encoding.txt

 0x57 0x23 0x43 0x7c
 
 # CHECK: divweu 2, 3, 4
 0x16 0x23 0x43 0x7c
 
 # CHECK: divweu. 2, 3, 4
 0x17 0x23 0x43 0x7c
 
+# CHECK: modsw 2, 3, 4
+0x16 0x26 0x43 0x7c
+
+# CHECK: moduw 2, 3, 4
+0x16 0x22 0x43 0x7c
+
+# CHECK: modsd 2, 3, 4
+0x12 0x26 0x43 0x7c
+
+# CHECK: modud 2, 3, 4
+0x12 0x22 0x43 0x7c
+
 # CHECK: mulld 2, 3, 4
 0xd2 0x21 0x43 0x7c
 
 # CHECK: mulld. 2, 3, 4
 0xd3 0x21 0x43 0x7c
 
 # CHECK: mulhd 2, 3, 4
 0x92 0x20 0x43 0x7c

test/MC/PowerPC/ppc64-encoding.s

 # CHECK-LE: divweu 2, 3, 4                  # encoding: [0x16,0x23,0x43,0x7c]
             divweu 2, 3, 4
 # CHECK-BE: divweu. 2, 3, 4                 # encoding: [0x7c,0x43,0x23,0x17]
 # CHECK-LE: divweu. 2, 3, 4                 # encoding: [0x17,0x23,0x43,0x7c]
             divweu. 2, 3, 4
 # FIXME:    divweuo 2, 3, 4
 # FIXME:    divweuo. 2, 3, 4
 
+# CHECK-BE: modsw 2, 3, 4                   # encoding: [0x7c,0x43,0x26,0x16]
+# CHECK-LE: modsw 2, 3, 4                   # encoding: [0x16,0x26,0x43,0x7c]
+            modsw 2, 3, 4
+# CHECK-BE: moduw 2, 3, 4                   # encoding: [0x7c,0x43,0x22,0x16]
+# CHECK-LE: moduw 2, 3, 4                   # encoding: [0x16,0x22,0x43,0x7c]
+            moduw 2, 3, 4
+# CHECK-BE: modsd 2, 3, 4                   # encoding: [0x7c,0x43,0x26,0x12]
+# CHECK-LE: modsd 2, 3, 4                   # encoding: [0x12,0x26,0x43,0x7c]
+            modsd 2, 3, 4
+# CHECK-BE: modud 2, 3, 4                   # encoding: [0x7c,0x43,0x22,0x12]
+# CHECK-LE: modud 2, 3, 4                   # encoding: [0x12,0x22,0x43,0x7c]
+            modud 2, 3, 4
+
 # CHECK-BE: mulld 2, 3, 4                   # encoding: [0x7c,0x43,0x21,0xd2]
 # CHECK-LE: mulld 2, 3, 4                   # encoding: [0xd2,0x21,0x43,0x7c]
             mulld 2, 3, 4
 # CHECK-BE: mulld. 2, 3, 4                  # encoding: [0x7c,0x43,0x21,0xd3]
 # CHECK-LE: mulld. 2, 3, 4                  # encoding: [0xd3,0x21,0x43,0x7c]
             mulld. 2, 3, 4
 # FIXME:    mulldo 2, 3, 4
 # FIXME:    mulldo. 2, 3, 4