diff --git a/llvm/lib/Analysis/ConstantFolding.cpp b/llvm/lib/Analysis/ConstantFolding.cpp
--- a/llvm/lib/Analysis/ConstantFolding.cpp
+++ b/llvm/lib/Analysis/ConstantFolding.cpp
@@ -1555,6 +1555,13 @@
   case Intrinsic::rint:
   // Constrained intrinsics can be folded if FP environment is known
   // to compiler.
+  case Intrinsic::experimental_constrained_fma:
+  case Intrinsic::experimental_constrained_fmuladd:
+  case Intrinsic::experimental_constrained_fadd:
+  case Intrinsic::experimental_constrained_fsub:
+  case Intrinsic::experimental_constrained_fmul:
+  case Intrinsic::experimental_constrained_fdiv:
+  case Intrinsic::experimental_constrained_frem:
   case Intrinsic::experimental_constrained_ceil:
   case Intrinsic::experimental_constrained_floor:
   case Intrinsic::experimental_constrained_round:
@@ -1816,6 +1823,44 @@
   return false;
 }
 
+/// Checks if the given intrinsic call, which evaluates to constant, is allowed
+/// to be folded.
+///
+/// \param CI Constrained intrinsic call.
+/// \param RM Rounding mode used for evaluation.
+/// \param St Exception flags raised during constant evaluation.
+static bool mayFoldConstrained(ConstrainedFPIntrinsic *CI, RoundingMode RM,
+                               APFloat::opStatus St) {
+  Optional<fp::ExceptionBehavior> EB = CI->getExceptionBehavior();
+
+  // If the operation does not change exception status flags, it is safe
+  // to fold.
+  if (St == APFloat::opStatus::opOK) {
+    // When FP exceptions are not ignored, intrinsic call will not be
+    // eliminated, because it is considered as having side effect. But we
+    // know that its evaluation does not raise exceptions, so side effect
+    // is absent. To allow of removing the call, mark it as not accessing
+    // memory.
+    if (EB && *EB != fp::ExceptionBehavior::ebIgnore)
+      CI->addAttribute(AttributeList::FunctionIndex, Attribute::ReadNone);
+    return true;
+  }
+
+  // If evaluation raised FP exception, the result can depend on rounding
+  // mode. If the latter is unknown, folding is not possible.
+  if (RM == RoundingMode::Dynamic)
+    return false;
+
+  // If FP exceptions are ignored, fold the call, even if such exception is
+  // raised.
+  if (!EB || *EB == fp::ExceptionBehavior::ebIgnore)
+    return true;
+
+  // Leave the calculation for runtime so that exception flags be correctly set
+  // in hardware.
+  return false;
+}
+
 static Constant *ConstantFoldScalarCall1(StringRef Name,
                                          Intrinsic::ID IntrinsicID,
                                          Type *Ty,
@@ -2319,7 +2364,7 @@
   }
 
   if (auto *Op1 = dyn_cast<ConstantFP>(Operands[0])) {
-    if (!Ty->isHalfTy() && !Ty->isFloatTy() && !Ty->isDoubleTy())
+    if (!Ty->isFloatingPointTy())
       return nullptr;
     APFloat Op1V = Op1->getValueAPF();
 
@@ -2328,6 +2373,49 @@
         return nullptr;
       APFloat Op2V = Op2->getValueAPF();
 
+      if (auto ConstrIntr = dyn_cast<ConstrainedFPIntrinsic>(Call)) {
+        Optional<RoundingMode> ORM = ConstrIntr->getRoundingMode();
+        if (!ORM)
+          ORM = RoundingMode::Dynamic;
+        RoundingMode RM;
+        if (*ORM == RoundingMode::Dynamic)
+          // Even if the rounding mode is unknown, try evaluating the operation.
+          // If it does not raise inexact exception, rounding was not applied
+          // so the result does not depend on rounding mode.
+          RM = RoundingMode::NearestTiesToEven;
+        else
+          RM = *ORM;
+        APFloat Res = Op1V;
+        APFloat::opStatus St;
+        switch (IntrinsicID) {
+        default:
+          return nullptr;
+        case Intrinsic::experimental_constrained_fadd:
+          St = Res.add(Op2V, RM);
+          break;
+        case Intrinsic::experimental_constrained_fsub:
+          St = Res.subtract(Op2V, RM);
+          break;
+        case Intrinsic::experimental_constrained_fmul:
+          St = Res.multiply(Op2V, RM);
+          break;
+        case Intrinsic::experimental_constrained_fdiv:
+          St = Res.divide(Op2V, RM);
+          break;
+        case Intrinsic::experimental_constrained_frem:
+          // frem does not depend on rounding mode, so change dynamic mode to
+          // something definite.
+          if (*ORM == RoundingMode::Dynamic)
+            ORM = RoundingMode::NearestTiesToEven;
+          St = Res.remainder(Op2V);
+          break;
+        }
+        if (mayFoldConstrained(const_cast<ConstrainedFPIntrinsic *>(ConstrIntr),
+                               *ORM, St))
+          return ConstantFP::get(Ty->getContext(), Res);
+        return nullptr;
+      }
+
       switch (IntrinsicID) {
       default:
         break;
@@ -2723,6 +2811,35 @@
         const APFloat &C1 = Op1->getValueAPF();
         const APFloat &C2 = Op2->getValueAPF();
         const APFloat &C3 = Op3->getValueAPF();
+
+        if (auto ConstrIntr = dyn_cast<ConstrainedFPIntrinsic>(Call)) {
+          Optional<RoundingMode> ORM = ConstrIntr->getRoundingMode();
+          if (!ORM)
+            ORM = RoundingMode::Dynamic;
+          RoundingMode RM;
+          if (*ORM == RoundingMode::Dynamic)
+            // Even if the rounding mode is unknown, try evaluating the
+            // operation. If it does not raise inexact exception, rounding was
+            // not applied so the result does not depend on rounding mode.
+            RM = RoundingMode::NearestTiesToEven;
+          else
+            RM = *ORM;
+          APFloat Res = C1;
+          APFloat::opStatus St;
+          switch (IntrinsicID) {
+          default:
+            return nullptr;
+          case Intrinsic::experimental_constrained_fma:
+          case Intrinsic::experimental_constrained_fmuladd:
+            St = Res.fusedMultiplyAdd(C2, C3, RM);
+            break;
+          }
+          if (mayFoldConstrained(
+                  const_cast<ConstrainedFPIntrinsic *>(ConstrIntr), *ORM, St))
+            return ConstantFP::get(Ty->getContext(), Res);
+          return nullptr;
+        }
+
         switch (IntrinsicID) {
         default: break;
         case Intrinsic::amdgcn_fma_legacy: {
diff --git a/llvm/lib/Transforms/Utils/Local.cpp b/llvm/lib/Transforms/Utils/Local.cpp
--- a/llvm/lib/Transforms/Utils/Local.cpp
+++ b/llvm/lib/Transforms/Utils/Local.cpp
@@ -486,6 +486,16 @@
     if (isMathLibCallNoop(Call, TLI))
       return true;
 
+  // To express possible interaction with floating point environment constrained
+  // intrinsics are described as if they access memory. So they look like having
+  // side effect but actually do not have it unless they raise floating point
+  // exception. If FP exceptions are ignored, the intrinsic may be deleted.
+  if (auto *CI = dyn_cast<ConstrainedFPIntrinsic>(I)) {
+    Optional<fp::ExceptionBehavior> EB = CI->getExceptionBehavior();
+    if (!EB || *EB == fp::ExceptionBehavior::ebIgnore)
+      return true;
+  }
+
   return false;
 }
 
diff --git a/llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll b/llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll
--- a/llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll
+++ b/llvm/test/CodeGen/PowerPC/vector-constrained-fp-intrinsics.ll
@@ -213,34 +213,14 @@
 define <1 x float> @constrained_vector_frem_v1f32() #0 {
 ; PC64LE-LABEL: constrained_vector_frem_v1f32:
 ; PC64LE:       # %bb.0: # %entry
-; PC64LE-NEXT:    mflr 0
-; PC64LE-NEXT:    std 0, 16(1)
-; PC64LE-NEXT:    stdu 1, -32(1)
 ; PC64LE-NEXT:    addis 3, 2, .LCPI5_0@toc@ha
-; PC64LE-NEXT:    addis 4, 2, .LCPI5_1@toc@ha
 ; PC64LE-NEXT:    lfs 1, .LCPI5_0@toc@l(3)
-; PC64LE-NEXT:    lfs 2, .LCPI5_1@toc@l(4)
-; PC64LE-NEXT:    bl fmodf
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addi 1, 1, 32
-; PC64LE-NEXT:    ld 0, 16(1)
-; PC64LE-NEXT:    mtlr 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_frem_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
-; PC64LE9-NEXT:    mflr 0
-; PC64LE9-NEXT:    std 0, 16(1)
-; PC64LE9-NEXT:    stdu 1, -32(1)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI5_0@toc@ha
 ; PC64LE9-NEXT:    lfs 1, .LCPI5_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI5_1@toc@ha
-; PC64LE9-NEXT:    lfs 2, .LCPI5_1@toc@l(3)
-; PC64LE9-NEXT:    bl fmodf
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    addi 1, 1, 32
-; PC64LE9-NEXT:    ld 0, 16(1)
-; PC64LE9-NEXT:    mtlr 0
 ; PC64LE9-NEXT:    blr
 entry:
   %rem = call <1 x float> @llvm.experimental.constrained.frem.v1f32(
@@ -254,62 +234,17 @@
 define <2 x double> @constrained_vector_frem_v2f64() #0 {
 ; PC64LE-LABEL: constrained_vector_frem_v2f64:
 ; PC64LE:       # %bb.0: # %entry
-; PC64LE-NEXT:    mflr 0
-; PC64LE-NEXT:    std 0, 16(1)
-; PC64LE-NEXT:    stdu 1, -80(1)
-; PC64LE-NEXT:    addis 4, 2, .LCPI6_1@toc@ha
-; PC64LE-NEXT:    stfd 31, 72(1) # 8-byte Folded Spill
-; PC64LE-NEXT:    li 3, 48
-; PC64LE-NEXT:    stxvd2x 63, 1, 3 # 16-byte Folded Spill
 ; PC64LE-NEXT:    addis 3, 2, .LCPI6_0@toc@ha
-; PC64LE-NEXT:    lfs 31, .LCPI6_1@toc@l(4)
-; PC64LE-NEXT:    lfs 1, .LCPI6_0@toc@l(3)
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addis 3, 2, .LCPI6_2@toc@ha
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    xxlor 63, 1, 1
-; PC64LE-NEXT:    lfs 1, .LCPI6_2@toc@l(3)
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    li 3, 48
-; PC64LE-NEXT:    # kill: def $f1 killed $f1 def $vsl1
-; PC64LE-NEXT:    xxmrghd 34, 1, 63
-; PC64LE-NEXT:    lfd 31, 72(1) # 8-byte Folded Reload
-; PC64LE-NEXT:    lxvd2x 63, 1, 3 # 16-byte Folded Reload
-; PC64LE-NEXT:    addi 1, 1, 80
-; PC64LE-NEXT:    ld 0, 16(1)
-; PC64LE-NEXT:    mtlr 0
+; PC64LE-NEXT:    addi 3, 3, .LCPI6_0@toc@l
+; PC64LE-NEXT:    lxvd2x 0, 0, 3
+; PC64LE-NEXT:    xxswapd 34, 0
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_frem_v2f64:
 ; PC64LE9:       # %bb.0: # %entry
-; PC64LE9-NEXT:    mflr 0
-; PC64LE9-NEXT:    std 0, 16(1)
-; PC64LE9-NEXT:    stdu 1, -64(1)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI6_0@toc@ha
-; PC64LE9-NEXT:    stfd 31, 56(1) # 8-byte Folded Spill
-; PC64LE9-NEXT:    stxv 63, 32(1) # 16-byte Folded Spill
-; PC64LE9-NEXT:    lfs 1, .LCPI6_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI6_1@toc@ha
-; PC64LE9-NEXT:    lfs 31, .LCPI6_1@toc@l(3)
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    addis 3, 2, .LCPI6_2@toc@ha
-; PC64LE9-NEXT:    xscpsgndp 63, 1, 1
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    lfs 1, .LCPI6_2@toc@l(3)
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    # kill: def $f1 killed $f1 def $vsl1
-; PC64LE9-NEXT:    xxmrghd 34, 1, 63
-; PC64LE9-NEXT:    lxv 63, 32(1) # 16-byte Folded Reload
-; PC64LE9-NEXT:    lfd 31, 56(1) # 8-byte Folded Reload
-; PC64LE9-NEXT:    addi 1, 1, 64
-; PC64LE9-NEXT:    ld 0, 16(1)
-; PC64LE9-NEXT:    mtlr 0
+; PC64LE9-NEXT:    addi 3, 3, .LCPI6_0@toc@l
+; PC64LE9-NEXT:    lxvx 34, 0, 3
 ; PC64LE9-NEXT:    blr
 entry:
   %rem = call <2 x double> @llvm.experimental.constrained.frem.v2f64(
@@ -323,88 +258,16 @@
 define <3 x float> @constrained_vector_frem_v3f32() #0 {
 ; PC64LE-LABEL: constrained_vector_frem_v3f32:
 ; PC64LE:       # %bb.0: # %entry
-; PC64LE-NEXT:    mflr 0
-; PC64LE-NEXT:    stfd 29, -24(1) # 8-byte Folded Spill
-; PC64LE-NEXT:    stfd 30, -16(1) # 8-byte Folded Spill
-; PC64LE-NEXT:    stfd 31, -8(1) # 8-byte Folded Spill
-; PC64LE-NEXT:    std 0, 16(1)
-; PC64LE-NEXT:    stdu 1, -64(1)
-; PC64LE-NEXT:    addis 4, 2, .LCPI7_1@toc@ha
 ; PC64LE-NEXT:    addis 3, 2, .LCPI7_0@toc@ha
-; PC64LE-NEXT:    lfs 31, .LCPI7_1@toc@l(4)
-; PC64LE-NEXT:    lfs 1, .LCPI7_0@toc@l(3)
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    bl fmodf
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addis 3, 2, .LCPI7_2@toc@ha
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    fmr 30, 1
-; PC64LE-NEXT:    lfs 1, .LCPI7_2@toc@l(3)
-; PC64LE-NEXT:    bl fmodf
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addis 3, 2, .LCPI7_3@toc@ha
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    fmr 29, 1
-; PC64LE-NEXT:    lfs 1, .LCPI7_3@toc@l(3)
-; PC64LE-NEXT:    bl fmodf
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xscvdpspn 34, 29
-; PC64LE-NEXT:    addis 3, 2, .LCPI7_4@toc@ha
-; PC64LE-NEXT:    xscvdpspn 35, 1
-; PC64LE-NEXT:    addi 3, 3, .LCPI7_4@toc@l
-; PC64LE-NEXT:    lvx 4, 0, 3
-; PC64LE-NEXT:    vmrghw 2, 2, 3
-; PC64LE-NEXT:    xscvdpspn 35, 30
-; PC64LE-NEXT:    vperm 2, 3, 2, 4
-; PC64LE-NEXT:    addi 1, 1, 64
-; PC64LE-NEXT:    ld 0, 16(1)
-; PC64LE-NEXT:    lfd 31, -8(1) # 8-byte Folded Reload
-; PC64LE-NEXT:    lfd 30, -16(1) # 8-byte Folded Reload
-; PC64LE-NEXT:    lfd 29, -24(1) # 8-byte Folded Reload
-; PC64LE-NEXT:    mtlr 0
+; PC64LE-NEXT:    addi 3, 3, .LCPI7_0@toc@l
+; PC64LE-NEXT:    lvx 2, 0, 3
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_frem_v3f32:
 ; PC64LE9:       # %bb.0: # %entry
-; PC64LE9-NEXT:    mflr 0
-; PC64LE9-NEXT:    stfd 29, -24(1) # 8-byte Folded Spill
-; PC64LE9-NEXT:    stfd 30, -16(1) # 8-byte Folded Spill
-; PC64LE9-NEXT:    stfd 31, -8(1) # 8-byte Folded Spill
-; PC64LE9-NEXT:    std 0, 16(1)
-; PC64LE9-NEXT:    stdu 1, -64(1)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI7_0@toc@ha
-; PC64LE9-NEXT:    lfs 1, .LCPI7_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI7_1@toc@ha
-; PC64LE9-NEXT:    lfs 31, .LCPI7_1@toc@l(3)
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    bl fmodf
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    addis 3, 2, .LCPI7_2@toc@ha
-; PC64LE9-NEXT:    fmr 30, 1
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    lfs 1, .LCPI7_2@toc@l(3)
-; PC64LE9-NEXT:    bl fmodf
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    addis 3, 2, .LCPI7_3@toc@ha
-; PC64LE9-NEXT:    fmr 29, 1
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    lfs 1, .LCPI7_3@toc@l(3)
-; PC64LE9-NEXT:    bl fmodf
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    xscvdpspn 34, 1
-; PC64LE9-NEXT:    xscvdpspn 35, 29
-; PC64LE9-NEXT:    addis 3, 2, .LCPI7_4@toc@ha
-; PC64LE9-NEXT:    addi 3, 3, .LCPI7_4@toc@l
-; PC64LE9-NEXT:    lxvx 36, 0, 3
-; PC64LE9-NEXT:    vmrghw 2, 3, 2
-; PC64LE9-NEXT:    xscvdpspn 35, 30
-; PC64LE9-NEXT:    vperm 2, 3, 2, 4
-; PC64LE9-NEXT:    addi 1, 1, 64
-; PC64LE9-NEXT:    ld 0, 16(1)
-; PC64LE9-NEXT:    lfd 31, -8(1) # 8-byte Folded Reload
-; PC64LE9-NEXT:    lfd 30, -16(1) # 8-byte Folded Reload
-; PC64LE9-NEXT:    mtlr 0
-; PC64LE9-NEXT:    lfd 29, -24(1) # 8-byte Folded Reload
+; PC64LE9-NEXT:    addi 3, 3, .LCPI7_0@toc@l
+; PC64LE9-NEXT:    lxvx 34, 0, 3
 ; PC64LE9-NEXT:    blr
 entry:
   %rem = call <3 x float> @llvm.experimental.constrained.frem.v3f32(
@@ -418,80 +281,22 @@
 define <3 x double> @constrained_vector_frem_v3f64() #0 {
 ; PC64LE-LABEL: constrained_vector_frem_v3f64:
 ; PC64LE:       # %bb.0: # %entry
-; PC64LE-NEXT:    mflr 0
-; PC64LE-NEXT:    std 0, 16(1)
-; PC64LE-NEXT:    stdu 1, -80(1)
-; PC64LE-NEXT:    addis 4, 2, .LCPI8_1@toc@ha
-; PC64LE-NEXT:    stfd 31, 72(1) # 8-byte Folded Spill
-; PC64LE-NEXT:    li 3, 48
-; PC64LE-NEXT:    stxvd2x 63, 1, 3 # 16-byte Folded Spill
 ; PC64LE-NEXT:    addis 3, 2, .LCPI8_0@toc@ha
-; PC64LE-NEXT:    lfs 31, .LCPI8_1@toc@l(4)
+; PC64LE-NEXT:    addis 4, 2, .LCPI8_1@toc@ha
+; PC64LE-NEXT:    addis 5, 2, .LCPI8_2@toc@ha
 ; PC64LE-NEXT:    lfs 1, .LCPI8_0@toc@l(3)
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addis 3, 2, .LCPI8_2@toc@ha
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    xxlor 63, 1, 1
-; PC64LE-NEXT:    lfs 1, .LCPI8_2@toc@l(3)
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addis 3, 2, .LCPI8_3@toc@ha
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    # kill: def $f1 killed $f1 def $vsl1
-; PC64LE-NEXT:    xxmrghd 63, 1, 63
-; PC64LE-NEXT:    lfs 1, .LCPI8_3@toc@l(3)
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    xxswapd 0, 63
-; PC64LE-NEXT:    li 3, 48
-; PC64LE-NEXT:    xxlor 2, 63, 63
-; PC64LE-NEXT:    lfd 31, 72(1) # 8-byte Folded Reload
-; PC64LE-NEXT:    lxvd2x 63, 1, 3 # 16-byte Folded Reload
-; PC64LE-NEXT:    fmr 3, 1
-; PC64LE-NEXT:    fmr 1, 0
-; PC64LE-NEXT:    addi 1, 1, 80
-; PC64LE-NEXT:    ld 0, 16(1)
-; PC64LE-NEXT:    mtlr 0
+; PC64LE-NEXT:    lfs 2, .LCPI8_1@toc@l(4)
+; PC64LE-NEXT:    lfs 3, .LCPI8_2@toc@l(5)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_frem_v3f64:
 ; PC64LE9:       # %bb.0: # %entry
-; PC64LE9-NEXT:    mflr 0
-; PC64LE9-NEXT:    std 0, 16(1)
-; PC64LE9-NEXT:    stdu 1, -64(1)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI8_0@toc@ha
-; PC64LE9-NEXT:    stfd 31, 56(1) # 8-byte Folded Spill
-; PC64LE9-NEXT:    stxv 63, 32(1) # 16-byte Folded Spill
 ; PC64LE9-NEXT:    lfs 1, .LCPI8_0@toc@l(3)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI8_1@toc@ha
-; PC64LE9-NEXT:    lfs 31, .LCPI8_1@toc@l(3)
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
+; PC64LE9-NEXT:    lfs 2, .LCPI8_1@toc@l(3)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI8_2@toc@ha
-; PC64LE9-NEXT:    xscpsgndp 63, 1, 1
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    lfs 1, .LCPI8_2@toc@l(3)
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    addis 3, 2, .LCPI8_3@toc@ha
-; PC64LE9-NEXT:    # kill: def $f1 killed $f1 def $vsl1
-; PC64LE9-NEXT:    xxmrghd 63, 1, 63
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    lfs 1, .LCPI8_3@toc@l(3)
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    fmr 3, 1
-; PC64LE9-NEXT:    xxswapd 1, 63
-; PC64LE9-NEXT:    xscpsgndp 2, 63, 63
-; PC64LE9-NEXT:    lxv 63, 32(1) # 16-byte Folded Reload
-; PC64LE9-NEXT:    lfd 31, 56(1) # 8-byte Folded Reload
-; PC64LE9-NEXT:    # kill: def $f1 killed $f1 killed $vsl1
-; PC64LE9-NEXT:    addi 1, 1, 64
-; PC64LE9-NEXT:    ld 0, 16(1)
-; PC64LE9-NEXT:    mtlr 0
+; PC64LE9-NEXT:    lfs 3, .LCPI8_2@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %rem = call <3 x double> @llvm.experimental.constrained.frem.v3f64(
@@ -505,96 +310,24 @@
 define <4 x double> @constrained_vector_frem_v4f64() #0 {
 ; PC64LE-LABEL: constrained_vector_frem_v4f64:
 ; PC64LE:       # %bb.0:
-; PC64LE-NEXT:    mflr 0
-; PC64LE-NEXT:    std 0, 16(1)
-; PC64LE-NEXT:    stdu 1, -96(1)
-; PC64LE-NEXT:    addis 4, 2, .LCPI9_1@toc@ha
-; PC64LE-NEXT:    stfd 31, 88(1) # 8-byte Folded Spill
-; PC64LE-NEXT:    li 3, 48
-; PC64LE-NEXT:    stxvd2x 62, 1, 3 # 16-byte Folded Spill
-; PC64LE-NEXT:    li 3, 64
-; PC64LE-NEXT:    lfs 31, .LCPI9_1@toc@l(4)
-; PC64LE-NEXT:    stxvd2x 63, 1, 3 # 16-byte Folded Spill
 ; PC64LE-NEXT:    addis 3, 2, .LCPI9_0@toc@ha
-; PC64LE-NEXT:    lfs 1, .LCPI9_0@toc@l(3)
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addis 3, 2, .LCPI9_2@toc@ha
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    xxlor 63, 1, 1
-; PC64LE-NEXT:    lfs 1, .LCPI9_2@toc@l(3)
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addis 3, 2, .LCPI9_3@toc@ha
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    # kill: def $f1 killed $f1 def $vsl1
-; PC64LE-NEXT:    xxmrghd 63, 1, 63
-; PC64LE-NEXT:    lfs 1, .LCPI9_3@toc@l(3)
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    addis 3, 2, .LCPI9_4@toc@ha
-; PC64LE-NEXT:    fmr 2, 31
-; PC64LE-NEXT:    xxlor 62, 1, 1
-; PC64LE-NEXT:    lfs 1, .LCPI9_4@toc@l(3)
-; PC64LE-NEXT:    bl fmod
-; PC64LE-NEXT:    nop
-; PC64LE-NEXT:    li 3, 64
-; PC64LE-NEXT:    vmr 2, 31
-; PC64LE-NEXT:    # kill: def $f1 killed $f1 def $vsl1
-; PC64LE-NEXT:    xxmrghd 35, 1, 62
-; PC64LE-NEXT:    lfd 31, 88(1) # 8-byte Folded Reload
-; PC64LE-NEXT:    lxvd2x 63, 1, 3 # 16-byte Folded Reload
-; PC64LE-NEXT:    li 3, 48
-; PC64LE-NEXT:    lxvd2x 62, 1, 3 # 16-byte Folded Reload
-; PC64LE-NEXT:    addi 1, 1, 96
-; PC64LE-NEXT:    ld 0, 16(1)
-; PC64LE-NEXT:    mtlr 0
+; PC64LE-NEXT:    addis 4, 2, .LCPI9_1@toc@ha
+; PC64LE-NEXT:    addi 3, 3, .LCPI9_0@toc@l
+; PC64LE-NEXT:    addi 4, 4, .LCPI9_1@toc@l
+; PC64LE-NEXT:    lxvd2x 0, 0, 3
+; PC64LE-NEXT:    lxvd2x 1, 0, 4
+; PC64LE-NEXT:    xxswapd 34, 0
+; PC64LE-NEXT:    xxswapd 35, 1
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_frem_v4f64:
 ; PC64LE9:       # %bb.0:
-; PC64LE9-NEXT:    mflr 0
-; PC64LE9-NEXT:    std 0, 16(1)
-; PC64LE9-NEXT:    stdu 1, -80(1)
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI9_0@toc@ha
-; PC64LE9-NEXT:    stfd 31, 72(1) # 8-byte Folded Spill
-; PC64LE9-NEXT:    stxv 62, 32(1) # 16-byte Folded Spill
-; PC64LE9-NEXT:    lfs 1, .LCPI9_0@toc@l(3)
+; PC64LE9-NEXT:    addi 3, 3, .LCPI9_0@toc@l
+; PC64LE9-NEXT:    lxvx 34, 0, 3
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI9_1@toc@ha
-; PC64LE9-NEXT:    stxv 63, 48(1) # 16-byte Folded Spill
-; PC64LE9-NEXT:    lfs 31, .LCPI9_1@toc@l(3)
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    addis 3, 2, .LCPI9_2@toc@ha
-; PC64LE9-NEXT:    xscpsgndp 63, 1, 1
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    lfs 1, .LCPI9_2@toc@l(3)
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    addis 3, 2, .LCPI9_3@toc@ha
-; PC64LE9-NEXT:    # kill: def $f1 killed $f1 def $vsl1
-; PC64LE9-NEXT:    xxmrghd 63, 1, 63
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    lfs 1, .LCPI9_3@toc@l(3)
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    addis 3, 2, .LCPI9_4@toc@ha
-; PC64LE9-NEXT:    xscpsgndp 62, 1, 1
-; PC64LE9-NEXT:    fmr 2, 31
-; PC64LE9-NEXT:    lfs 1, .LCPI9_4@toc@l(3)
-; PC64LE9-NEXT:    bl fmod
-; PC64LE9-NEXT:    nop
-; PC64LE9-NEXT:    # kill: def $f1 killed $f1 def $vsl1
-; PC64LE9-NEXT:    xxmrghd 35, 1, 62
-; PC64LE9-NEXT:    vmr 2, 31
-; PC64LE9-NEXT:    lxv 63, 48(1) # 16-byte Folded Reload
-; PC64LE9-NEXT:    lxv 62, 32(1) # 16-byte Folded Reload
-; PC64LE9-NEXT:    lfd 31, 72(1) # 8-byte Folded Reload
-; PC64LE9-NEXT:    addi 1, 1, 80
-; PC64LE9-NEXT:    ld 0, 16(1)
-; PC64LE9-NEXT:    mtlr 0
+; PC64LE9-NEXT:    addi 3, 3, .LCPI9_1@toc@l
+; PC64LE9-NEXT:    lxvx 35, 0, 3
 ; PC64LE9-NEXT:    blr
   %rem = call <4 x double> @llvm.experimental.constrained.frem.v4f64(
            <4 x double> <double 1.000000e+00, double 2.000000e+00,
@@ -610,19 +343,13 @@
 ; PC64LE-LABEL: constrained_vector_fmul_v1f32:
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI10_0@toc@ha
-; PC64LE-NEXT:    addis 4, 2, .LCPI10_1@toc@ha
-; PC64LE-NEXT:    lfs 0, .LCPI10_0@toc@l(3)
-; PC64LE-NEXT:    lfs 1, .LCPI10_1@toc@l(4)
-; PC64LE-NEXT:    xsmulsp 1, 1, 0
+; PC64LE-NEXT:    lfs 1, .LCPI10_0@toc@l(3)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fmul_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI10_0@toc@ha
-; PC64LE9-NEXT:    lfs 0, .LCPI10_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI10_1@toc@ha
-; PC64LE9-NEXT:    lfs 1, .LCPI10_1@toc@l(3)
-; PC64LE9-NEXT:    xsmulsp 1, 1, 0
+; PC64LE9-NEXT:    lfs 1, .LCPI10_0@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %mul = call <1 x float> @llvm.experimental.constrained.fmul.v1f32(
@@ -669,48 +396,16 @@
 define <3 x float> @constrained_vector_fmul_v3f32() #0 {
 ; PC64LE-LABEL: constrained_vector_fmul_v3f32:
 ; PC64LE:       # %bb.0: # %entry
-; PC64LE-NEXT:    addis 3, 2, .LCPI12_1@toc@ha
-; PC64LE-NEXT:    addis 4, 2, .LCPI12_3@toc@ha
-; PC64LE-NEXT:    addis 5, 2, .LCPI12_2@toc@ha
-; PC64LE-NEXT:    lfs 0, .LCPI12_1@toc@l(3)
-; PC64LE-NEXT:    lfs 1, .LCPI12_3@toc@l(4)
-; PC64LE-NEXT:    lfs 2, .LCPI12_2@toc@l(5)
 ; PC64LE-NEXT:    addis 3, 2, .LCPI12_0@toc@ha
-; PC64LE-NEXT:    xsmulsp 1, 0, 1
-; PC64LE-NEXT:    lfs 3, .LCPI12_0@toc@l(3)
-; PC64LE-NEXT:    addis 3, 2, .LCPI12_4@toc@ha
-; PC64LE-NEXT:    xsmulsp 2, 0, 2
-; PC64LE-NEXT:    addi 3, 3, .LCPI12_4@toc@l
-; PC64LE-NEXT:    lvx 4, 0, 3
-; PC64LE-NEXT:    xsmulsp 0, 0, 3
-; PC64LE-NEXT:    xscvdpspn 34, 1
-; PC64LE-NEXT:    xscvdpspn 35, 2
-; PC64LE-NEXT:    vmrghw 2, 3, 2
-; PC64LE-NEXT:    xscvdpspn 35, 0
-; PC64LE-NEXT:    vperm 2, 3, 2, 4
+; PC64LE-NEXT:    addi 3, 3, .LCPI12_0@toc@l
+; PC64LE-NEXT:    lvx 2, 0, 3
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fmul_v3f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI12_0@toc@ha
-; PC64LE9-NEXT:    lfs 0, .LCPI12_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI12_1@toc@ha
-; PC64LE9-NEXT:    lfs 1, .LCPI12_1@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI12_2@toc@ha
-; PC64LE9-NEXT:    xsmulsp 0, 1, 0
-; PC64LE9-NEXT:    lfs 2, .LCPI12_2@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI12_3@toc@ha
-; PC64LE9-NEXT:    lfs 3, .LCPI12_3@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI12_4@toc@ha
-; PC64LE9-NEXT:    addi 3, 3, .LCPI12_4@toc@l
-; PC64LE9-NEXT:    lxvx 36, 0, 3
-; PC64LE9-NEXT:    xsmulsp 2, 1, 2
-; PC64LE9-NEXT:    xsmulsp 1, 1, 3
-; PC64LE9-NEXT:    xscvdpspn 35, 2
-; PC64LE9-NEXT:    xscvdpspn 34, 1
-; PC64LE9-NEXT:    vmrghw 2, 3, 2
-; PC64LE9-NEXT:    xscvdpspn 35, 0
-; PC64LE9-NEXT:    vperm 2, 3, 2, 4
+; PC64LE9-NEXT:    addi 3, 3, .LCPI12_0@toc@l
+; PC64LE9-NEXT:    lxvx 34, 0, 3
 ; PC64LE9-NEXT:    blr
 entry:
   %mul = call <3 x float> @llvm.experimental.constrained.fmul.v3f32(
@@ -820,19 +515,13 @@
 ; PC64LE-LABEL: constrained_vector_fadd_v1f32:
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI15_0@toc@ha
-; PC64LE-NEXT:    addis 4, 2, .LCPI15_1@toc@ha
-; PC64LE-NEXT:    lfs 0, .LCPI15_0@toc@l(3)
-; PC64LE-NEXT:    lfs 1, .LCPI15_1@toc@l(4)
-; PC64LE-NEXT:    xsaddsp 1, 1, 0
+; PC64LE-NEXT:    lfs 1, .LCPI15_0@toc@l(3)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fadd_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI15_0@toc@ha
-; PC64LE9-NEXT:    lfs 0, .LCPI15_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI15_1@toc@ha
-; PC64LE9-NEXT:    lfs 1, .LCPI15_1@toc@l(3)
-; PC64LE9-NEXT:    xsaddsp 1, 1, 0
+; PC64LE9-NEXT:    lfs 1, .LCPI15_0@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %add = call <1 x float> @llvm.experimental.constrained.fadd.v1f32(
@@ -879,46 +568,12 @@
 define <3 x float> @constrained_vector_fadd_v3f32() #0 {
 ; PC64LE-LABEL: constrained_vector_fadd_v3f32:
 ; PC64LE:       # %bb.0: # %entry
-; PC64LE-NEXT:    addis 3, 2, .LCPI17_0@toc@ha
-; PC64LE-NEXT:    addis 4, 2, .LCPI17_2@toc@ha
-; PC64LE-NEXT:    addis 5, 2, .LCPI17_1@toc@ha
-; PC64LE-NEXT:    xxlxor 3, 3, 3
-; PC64LE-NEXT:    lfs 0, .LCPI17_0@toc@l(3)
-; PC64LE-NEXT:    lfs 1, .LCPI17_2@toc@l(4)
-; PC64LE-NEXT:    lfs 2, .LCPI17_1@toc@l(5)
-; PC64LE-NEXT:    addis 3, 2, .LCPI17_3@toc@ha
-; PC64LE-NEXT:    addi 3, 3, .LCPI17_3@toc@l
-; PC64LE-NEXT:    xsaddsp 1, 0, 1
-; PC64LE-NEXT:    lvx 4, 0, 3
-; PC64LE-NEXT:    xsaddsp 2, 0, 2
-; PC64LE-NEXT:    xsaddsp 0, 0, 3
-; PC64LE-NEXT:    xscvdpspn 34, 1
-; PC64LE-NEXT:    xscvdpspn 35, 2
-; PC64LE-NEXT:    vmrghw 2, 3, 2
-; PC64LE-NEXT:    xscvdpspn 35, 0
-; PC64LE-NEXT:    vperm 2, 3, 2, 4
+; PC64LE-NEXT:    xxleqv 34, 34, 34
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fadd_v3f32:
 ; PC64LE9:       # %bb.0: # %entry
-; PC64LE9-NEXT:    addis 3, 2, .LCPI17_0@toc@ha
-; PC64LE9-NEXT:    xxlxor 1, 1, 1
-; PC64LE9-NEXT:    lfs 0, .LCPI17_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI17_1@toc@ha
-; PC64LE9-NEXT:    lfs 2, .LCPI17_1@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI17_2@toc@ha
-; PC64LE9-NEXT:    lfs 3, .LCPI17_2@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI17_3@toc@ha
-; PC64LE9-NEXT:    addi 3, 3, .LCPI17_3@toc@l
-; PC64LE9-NEXT:    xsaddsp 1, 0, 1
-; PC64LE9-NEXT:    lxvx 36, 0, 3
-; PC64LE9-NEXT:    xsaddsp 2, 0, 2
-; PC64LE9-NEXT:    xsaddsp 0, 0, 3
-; PC64LE9-NEXT:    xscvdpspn 35, 2
-; PC64LE9-NEXT:    xscvdpspn 34, 0
-; PC64LE9-NEXT:    vmrghw 2, 3, 2
-; PC64LE9-NEXT:    xscvdpspn 35, 1
-; PC64LE9-NEXT:    vperm 2, 3, 2, 4
+; PC64LE9-NEXT:    xxleqv 34, 34, 34
 ; PC64LE9-NEXT:    blr
 entry:
   %add = call <3 x float> @llvm.experimental.constrained.fadd.v3f32(
@@ -1026,19 +681,13 @@
 ; PC64LE-LABEL: constrained_vector_fsub_v1f32:
 ; PC64LE:       # %bb.0: # %entry
 ; PC64LE-NEXT:    addis 3, 2, .LCPI20_0@toc@ha
-; PC64LE-NEXT:    addis 4, 2, .LCPI20_1@toc@ha
-; PC64LE-NEXT:    lfs 0, .LCPI20_0@toc@l(3)
-; PC64LE-NEXT:    lfs 1, .LCPI20_1@toc@l(4)
-; PC64LE-NEXT:    xssubsp 1, 1, 0
+; PC64LE-NEXT:    lfs 1, .LCPI20_0@toc@l(3)
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fsub_v1f32:
 ; PC64LE9:       # %bb.0: # %entry
 ; PC64LE9-NEXT:    addis 3, 2, .LCPI20_0@toc@ha
-; PC64LE9-NEXT:    lfs 0, .LCPI20_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI20_1@toc@ha
-; PC64LE9-NEXT:    lfs 1, .LCPI20_1@toc@l(3)
-; PC64LE9-NEXT:    xssubsp 1, 1, 0
+; PC64LE9-NEXT:    lfs 1, .LCPI20_0@toc@l(3)
 ; PC64LE9-NEXT:    blr
 entry:
   %sub = call <1 x float> @llvm.experimental.constrained.fsub.v1f32(
@@ -1085,46 +734,12 @@
 define <3 x float> @constrained_vector_fsub_v3f32() #0 {
 ; PC64LE-LABEL: constrained_vector_fsub_v3f32:
 ; PC64LE:       # %bb.0: # %entry
-; PC64LE-NEXT:    addis 3, 2, .LCPI22_0@toc@ha
-; PC64LE-NEXT:    addis 4, 2, .LCPI22_2@toc@ha
-; PC64LE-NEXT:    addis 5, 2, .LCPI22_1@toc@ha
-; PC64LE-NEXT:    xxlxor 3, 3, 3
-; PC64LE-NEXT:    lfs 0, .LCPI22_0@toc@l(3)
-; PC64LE-NEXT:    lfs 1, .LCPI22_2@toc@l(4)
-; PC64LE-NEXT:    lfs 2, .LCPI22_1@toc@l(5)
-; PC64LE-NEXT:    addis 3, 2, .LCPI22_3@toc@ha
-; PC64LE-NEXT:    addi 3, 3, .LCPI22_3@toc@l
-; PC64LE-NEXT:    xssubsp 1, 0, 1
-; PC64LE-NEXT:    lvx 4, 0, 3
-; PC64LE-NEXT:    xssubsp 2, 0, 2
-; PC64LE-NEXT:    xssubsp 0, 0, 3
-; PC64LE-NEXT:    xscvdpspn 34, 1
-; PC64LE-NEXT:    xscvdpspn 35, 2
-; PC64LE-NEXT:    vmrghw 2, 3, 2
-; PC64LE-NEXT:    xscvdpspn 35, 0
-; PC64LE-NEXT:    vperm 2, 3, 2, 4
+; PC64LE-NEXT:    xxleqv 34, 34, 34
 ; PC64LE-NEXT:    blr
 ;
 ; PC64LE9-LABEL: constrained_vector_fsub_v3f32:
 ; PC64LE9:       # %bb.0: # %entry
-; PC64LE9-NEXT:    addis 3, 2, .LCPI22_0@toc@ha
-; PC64LE9-NEXT:    xxlxor 1, 1, 1
-; PC64LE9-NEXT:    lfs 0, .LCPI22_0@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI22_1@toc@ha
-; PC64LE9-NEXT:    lfs 2, .LCPI22_1@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI22_2@toc@ha
-; PC64LE9-NEXT:    lfs 3, .LCPI22_2@toc@l(3)
-; PC64LE9-NEXT:    addis 3, 2, .LCPI22_3@toc@ha
-; PC64LE9-NEXT:    addi 3, 3, .LCPI22_3@toc@l
-; PC64LE9-NEXT:    xssubsp 1, 0, 1
-; PC64LE9-NEXT:    lxvx 36, 0, 3
-; PC64LE9-NEXT:    xssubsp 2, 0, 2
-; PC64LE9-NEXT:    xssubsp 0, 0, 3
-; PC64LE9-NEXT:    xscvdpspn 35, 2
-; PC64LE9-NEXT:    xscvdpspn 34, 0
-; PC64LE9-NEXT:    vmrghw 2, 3, 2
-; PC64LE9-NEXT:    xscvdpspn 35, 1
-; PC64LE9-NEXT:    vperm 2, 3, 2, 4
+; PC64LE9-NEXT:    xxleqv 34, 34, 34
 ; PC64LE9-NEXT:    blr
 entry:
   %sub = call <3 x float> @llvm.experimental.constrained.fsub.v3f32(
diff --git a/llvm/test/Transforms/InstSimplify/constfold-constrained.ll b/llvm/test/Transforms/InstSimplify/constfold-constrained.ll
--- a/llvm/test/Transforms/InstSimplify/constfold-constrained.ll
+++ b/llvm/test/Transforms/InstSimplify/constfold-constrained.ll
@@ -232,6 +232,219 @@
   ; CHECK: ret double 1.000000e+01
 }
 
+define float @fadd_01() #0 {
+entry:
+  %result = call float @llvm.experimental.constrained.fadd.f32(
+                                               float 1.000000e+01,
+                                               float 2.000000e+01,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.ignore") #0
+  ret float %result
+  ; CHECK-LABEL: @fadd_01
+  ; CHECK: ret float 3.000000e+01
+}
+
+; Inexact result does not prevent from folding if exceptions are ignored and
+; rounding mode is known.
+define double @fadd_02() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fadd.f64(
+                                               double 1.0,
+                                               double 0x3FF0000000000001,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fadd_02
+  ; CHECK: ret double 2.000000e+00
+}
+
+define double @fadd_03() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fadd.f64(
+                                               double 1.0,
+                                               double 0x3FF0000000000001,
+                                               metadata !"round.upward",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fadd_03
+  ; CHECK: ret double 0x4000000000000001
+}
+
+; Inexact result prevents from folding if exceptions may be checked.
+define double @fadd_04() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fadd.f64(
+                                               double 1.0,
+                                               double 0x3FF0000000000001,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.strict") #0
+  ret double %result
+  ; CHECK-LABEL: @fadd_04
+  ; CHECK: ret double %result
+}
+
+; If result is exact, folding is allowed even if exceptions may be checked.
+define double @fadd_05() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fadd.f64(
+                                               double 1.0,
+                                               double 2.0,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.strict") #0
+  ret double %result
+  ; CHECK-LABEL: @fadd_05
+  ; CHECK: ret double 3.000000e+00
+}
+
+; Dynamic rounding mode does not prevent from folding if the result is exact.
+define double @fadd_06() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fadd.f64(
+                                               double 1.0,
+                                               double 2.0,
+                                               metadata !"round.dynamic",
+                                               metadata !"fpexcept.strict") #0
+  ret double %result
+  ; CHECK-LABEL: @fadd_06
+  ; CHECK: ret double 3.000000e+00
+}
+
+; Inexact results prevents from folding if rounding mode is unknown.
+define double @fadd_07() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fadd.f64(
+                                               double 1.0,
+                                               double 0x3FF0000000000001,
+                                               metadata !"round.dynamic",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fadd_07
+  ; CHECK: ret double %result
+}
+
+; Infinite result does not prevent from folding unless exceptions are tracked.
+define double @fadd_08() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fadd.f64(
+                                               double 0x7fEFFFFFFFFFFFFF,
+                                               double 0x7fEFFFFFFFFFFFFF,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fadd_08
+  ; CHECK: ret double 0x7FF0000000000000
+}
+
+define double @fadd_09() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fadd.f64(
+                                               double 0x7fEFFFFFFFFFFFFF,
+                                               double 0x7fEFFFFFFFFFFFFF,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.strict") #0
+  ret double %result
+  ; CHECK-LABEL: @fadd_09
+  ; CHECK: ret double %result
+}
+
+define half @fadd_10() #0 {
+entry:
+  %result = call half @llvm.experimental.constrained.fadd.f16(
+                                               half 1.0,
+                                               half 2.0,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.ignore") #0
+  ret half %result
+  ; CHECK-LABEL: @fadd_10
+  ; CHECK: ret half 0xH4200
+}
+
+define bfloat @fadd_11() #0 {
+entry:
+  %result = call bfloat @llvm.experimental.constrained.fadd.bf16(
+                                               bfloat 1.0,
+                                               bfloat 2.0,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.ignore") #0
+  ret bfloat %result
+  ; CHECK-LABEL: @fadd_11
+  ; CHECK: ret bfloat 0xR4040
+}
+
+define double @fsub_01() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fsub.f64(
+                                               double 1.0,
+                                               double 2.0,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fsub_01
+  ; CHECK: ret double -1.000000e+00
+}
+
+define double @fmul_01() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fmul.f64(
+                                               double 1.0,
+                                               double 2.0,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fmul_01
+  ; CHECK: ret double 2.000000e+00
+}
+
+define double @fdiv_01() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fdiv.f64(
+                                               double 1.0,
+                                               double 2.0,
+                                               metadata !"round.tonearest",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fdiv_01
+  ; CHECK: ret double 5.000000e-01
+}
+
+define double @frem_01() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.frem.f64(
+                                               double 1.0,
+                                               double 2.0,
+                                               metadata !"round.dynamic",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @frem_01
+  ; CHECK: ret double 1.000000e+00
+}
+
+define double @fma_01() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fma.f64(
+                                               double 1.0,
+                                               double 2.0,
+                                               double 3.0,
+                                               metadata !"round.dynamic",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fma_01
+  ; CHECK: ret double 5.000000e+00
+}
+
+define double @fmuladd_01() #0 {
+entry:
+  %result = call double @llvm.experimental.constrained.fmuladd.f64(
+                                               double 1.0,
+                                               double 2.0,
+                                               double 3.0,
+                                               metadata !"round.dynamic",
+                                               metadata !"fpexcept.ignore") #0
+  ret double %result
+  ; CHECK-LABEL: @fmuladd_01
+  ; CHECK: ret double 5.000000e+00
+}
+
 
 attributes #0 = { strictfp }
 
@@ -241,4 +454,14 @@
 declare double @llvm.experimental.constrained.trunc.f64(double, metadata)
 declare double @llvm.experimental.constrained.round.f64(double, metadata)
 declare double @llvm.experimental.constrained.rint.f64(double, metadata, metadata)
+declare double @llvm.experimental.constrained.fadd.f64(double, double, metadata, metadata)
+declare float @llvm.experimental.constrained.fadd.f32(float, float, metadata, metadata)
+declare half @llvm.experimental.constrained.fadd.f16(half, half, metadata, metadata)
+declare bfloat @llvm.experimental.constrained.fadd.bf16(bfloat, bfloat, metadata, metadata)
+declare double @llvm.experimental.constrained.fsub.f64(double, double, metadata, metadata)
+declare double @llvm.experimental.constrained.fmul.f64(double, double, metadata, metadata)
+declare double @llvm.experimental.constrained.fdiv.f64(double, double, metadata, metadata)
+declare double @llvm.experimental.constrained.frem.f64(double, double, metadata, metadata)
+declare double @llvm.experimental.constrained.fma.f64(double, double, double, metadata, metadata)
+declare double @llvm.experimental.constrained.fmuladd.f64(double, double, double, metadata, metadata)