[FPEnv] Added a special UnrollVectorOp method to deal with the chain on StrictFP...

This change creates UnrollVectorOp_StrictFP. The purpose of this is to address a failure that consistently occurs when calling StrictFP functions on vectors whose number of elements is 3 + 2n on most platforms, such as PowerPC or SystemZ. The old UnrollVectorOp method does not expect that the vector that it will unroll will have a chain, so it has an assert that prevents it from running if this is the case. This new StrictFP version of the method deals with the chain while unrolling the vector. With this new function in place during vector widending, llc can run vector-constrained-fp-intrinsics.ll for SystemZ successfully.

Submitted by:	Drew Wock <drew.wock@sas.com>
Reviewed by:	Cameron McInally, Kevin P. Neal
Approved by:	Cameron McInally
Differential Revision:	http://reviews.llvm.org/D62546

llvm-svn: 362112

diff --git a/llvm/lib/CodeGen/SelectionDAG/LegalizeTypes.h b/llvm/lib/CodeGen/SelectionDAG/LegalizeTypes.h
index a0e7c8a..8beaf14 100644 (file)
--- a/llvm/lib/CodeGen/SelectionDAG/LegalizeTypes.h
+++ b/llvm/lib/CodeGen/SelectionDAG/LegalizeTypes.h
@@ -857,6 +857,11 @@ private:
   SDValue WidenVecOp_FCOPYSIGN(SDNode *N);
   SDValue WidenVecOp_VECREDUCE(SDNode *N);
 
+  /// Helper function to generate a set of operations to perform
+  /// a vector operation for a wider type.
+  ///
+  SDValue UnrollVectorOp_StrictFP(SDNode *N, unsigned ResNE);
+
   //===--------------------------------------------------------------------===//
   // Vector Widening Utilities Support: LegalizeVectorTypes.cpp
   //===--------------------------------------------------------------------===//

diff --git a/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp b/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
index aefc2aa..072f15b 100644 (file)
--- a/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
@@ -1318,6 +1318,63 @@ void DAGTypeLegalizer::SplitVecRes_StrictFPOp(SDNode *N, SDValue &Lo,
   ReplaceValueWith(SDValue(N, 1), Chain);
 }
 
+SDValue DAGTypeLegalizer::UnrollVectorOp_StrictFP(SDNode *N, unsigned ResNE) {
+  SDValue Chain = N->getOperand(0);
+  EVT VT = N->getValueType(0);
+  unsigned NE = VT.getVectorNumElements();
+  EVT EltVT = VT.getVectorElementType();
+  SDLoc dl(N);
+
+  SmallVector<SDValue, 8> Scalars;
+  SmallVector<SDValue, 4> Operands(N->getNumOperands());
+
+  // If ResNE is 0, fully unroll the vector op.
+  if (ResNE == 0)
+    ResNE = NE;
+  else if (NE > ResNE)
+    NE = ResNE;
+
+  //The results of each unrolled operation, including the chain.
+  EVT ChainVTs[] = {EltVT, MVT::Other};
+  SmallVector<SDValue, 8> Chains;
+
+  unsigned i;
+  for (i = 0; i != NE; ++i) {
+    Operands[0] = Chain;
+    for (unsigned j = 1, e = N->getNumOperands(); j != e; ++j) {
+      SDValue Operand = N->getOperand(j);
+      EVT OperandVT = Operand.getValueType();
+      if (OperandVT.isVector()) {
+        EVT OperandEltVT = OperandVT.getVectorElementType();
+        Operands[j] =
+            DAG.getNode(ISD::EXTRACT_VECTOR_ELT, dl, OperandEltVT, Operand,
+                    DAG.getConstant(i, dl, TLI.getVectorIdxTy(
+                          DAG.getDataLayout())));
+      } else {
+        Operands[j] = Operand;
+      }
+    }
+    SDValue Scalar = DAG.getNode(N->getOpcode(), dl, ChainVTs, Operands);
+    Scalar.getNode()->setFlags(N->getFlags());
+
+    //Add in the scalar as well as its chain value to the
+    //result vectors.
+    Scalars.push_back(Scalar);
+    Chains.push_back(Scalar.getValue(1));
+  }
+
+  for (; i < ResNE; ++i)
+    Scalars.push_back(DAG.getUNDEF(EltVT));
+
+  // Build a new factor node to connect the chain back together.
+  Chain = DAG.getNode(ISD::TokenFactor, dl, MVT::Other, Chains);
+  ReplaceValueWith(SDValue(N, 1), Chain);
+
+  // Create a new BUILD_VECTOR node
+  EVT VecVT = EVT::getVectorVT(*DAG.getContext(), EltVT, ResNE);
+  return DAG.getBuildVector(VecVT, dl, Scalars);
+}
+
 void DAGTypeLegalizer::SplitVecRes_OverflowOp(SDNode *N, unsigned ResNo,
                                               SDValue &Lo, SDValue &Hi) {
   SDLoc dl(N);
@@ -2968,7 +3025,7 @@ SDValue DAGTypeLegalizer::WidenVecRes_StrictFP(SDNode *N) {
 
   // No legal vector version so unroll the vector operation and then widen.
   if (NumElts == 1)
-    return DAG.UnrollVectorOp(N, WidenVT.getVectorNumElements());
+    return UnrollVectorOp_StrictFP(N, WidenVT.getVectorNumElements());
 
   // Since the operation can trap, apply operation on the original vector.
   EVT MaxVT = VT;

diff --git a/llvm/test/CodeGen/SystemZ/vector-constrained-fp-intrinsics.ll b/llvm/test/CodeGen/SystemZ/vector-constrained-fp-intrinsics.ll
new file mode 100644 (file)
index 0000000..fd5895f
--- /dev/null
+++ b/llvm/test/CodeGen/SystemZ/vector-constrained-fp-intrinsics.ll
@@ -0,0 +1,6445 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -O3 -mtriple=s390x-linux-gnu < %s | FileCheck --check-prefix=S390X %s
+; RUN: llc -O3 -mtriple=s390x-linux-gnu -mcpu=z13 < %s | FileCheck --check-prefix=SZ13 %s
+
+define <1 x float> @constrained_vector_fdiv_v1f32() {
+; S390X-LABEL: constrained_vector_fdiv_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI0_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI0_1
+; S390X-NEXT:    deb %f0, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fdiv_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI0_0
+; SZ13-NEXT:    vgmf %v0, 2, 8
+; SZ13-NEXT:    deb %f0, 0(%r1)
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %div = call <1 x float> @llvm.experimental.constrained.fdiv.v1f32(
+           <1 x float> <float 1.000000e+00>,
+           <1 x float> <float 1.000000e+01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <1 x float> %div
+}
+
+define <2 x double> @constrained_vector_fdiv_v2f64() {
+; S390X-LABEL: constrained_vector_fdiv_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI1_0
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI1_1
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI1_2
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ddbr %f0, %f1
+; S390X-NEXT:    ddbr %f2, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fdiv_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI1_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI1_1
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    vfddb %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %div = call <2 x double> @llvm.experimental.constrained.fdiv.v2f64(
+           <2 x double> <double 1.000000e+00, double 2.000000e+00>,
+           <2 x double> <double 1.000000e+01, double 1.000000e+01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <2 x double> %div
+}
+
+define <3 x float> @constrained_vector_fdiv_v3f32() {
+; S390X-LABEL: constrained_vector_fdiv_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI2_0
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI2_1
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI2_2
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI2_3
+; S390X-NEXT:    le %f4, 0(%r1)
+; S390X-NEXT:    debr %f0, %f1
+; S390X-NEXT:    debr %f2, %f1
+; S390X-NEXT:    debr %f4, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fdiv_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI2_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI2_1
+; SZ13-NEXT:    lde %f1, 0(%r1)
+; SZ13-NEXT:    debr %f1, %f0
+; SZ13-NEXT:    vgmf %v2, 2, 8
+; SZ13-NEXT:    vgmf %v3, 1, 1
+; SZ13-NEXT:    debr %f2, %f0
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    debr %f3, %f0
+; SZ13-NEXT:    vmrhf %v0, %v2, %v3
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    br %r14
+entry:
+  %div = call <3 x float> @llvm.experimental.constrained.fdiv.v3f32(
+           <3 x float> <float 1.000000e+00, float 2.000000e+00, float 3.000000e+00>,
+           <3 x float> <float 1.000000e+01, float 1.000000e+01, float 1.000000e+01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <3 x float> %div
+}
+
+define void @constrained_vector_fdiv_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_fdiv_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI3_1
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI3_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI3_0
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ddb %f1, 16(%r2)
+; S390X-NEXT:    ddb %f0, 8(%r2)
+; S390X-NEXT:    ddb %f2, 0(%r2)
+; S390X-NEXT:    std %f1, 16(%r2)
+; S390X-NEXT:    std %f0, 8(%r2)
+; S390X-NEXT:    std %f2, 0(%r2)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fdiv_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI3_0
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    vfddb %v0, %v1, %v0
+; SZ13-NEXT:    larl %r1, .LCPI3_1
+; SZ13-NEXT:    ldeb %f1, 0(%r1)
+; SZ13-NEXT:    ddb %f1, 16(%r2)
+; SZ13-NEXT:    std %f1, 16(%r2)
+; SZ13-NEXT:    vst %v0, 0(%r2)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %div = call <3 x double> @llvm.experimental.constrained.fdiv.v3f64(
+           <3 x double> <double 1.000000e+00, double 2.000000e+00, double 3.000000e+00>,
+           <3 x double> %b,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  store <3 x double> %div, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_fdiv_v4f64() {
+; S390X-LABEL: constrained_vector_fdiv_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI4_0
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI4_1
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI4_2
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI4_3
+; S390X-NEXT:    ldeb %f4, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI4_4
+; S390X-NEXT:    ldeb %f6, 0(%r1)
+; S390X-NEXT:    ddbr %f0, %f1
+; S390X-NEXT:    ddbr %f2, %f1
+; S390X-NEXT:    ddbr %f4, %f1
+; S390X-NEXT:    ddbr %f6, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fdiv_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI4_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI4_1
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    vfddb %v24, %v1, %v0
+; SZ13-NEXT:    larl %r1, .LCPI4_2
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    vfddb %v26, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %div = call <4 x double> @llvm.experimental.constrained.fdiv.v4f64(
+           <4 x double> <double 1.000000e+00, double 2.000000e+00,
+                         double 3.000000e+00, double 4.000000e+00>,
+           <4 x double> <double 1.000000e+01, double 1.000000e+01,
+                         double 1.000000e+01, double 1.000000e+01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <4 x double> %div
+}
+
+define <1 x float> @constrained_vector_frem_v1f32() {
+; S390X-LABEL: constrained_vector_frem_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI5_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI5_1
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    brasl %r14, fmodf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_frem_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI5_0
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    vgmf %v0, 2, 8
+; SZ13-NEXT:    brasl %r14, fmodf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %rem = call <1 x float> @llvm.experimental.constrained.frem.v1f32(
+           <1 x float> <float 1.000000e+00>,
+           <1 x float> <float 1.000000e+01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <1 x float> %rem
+}
+
+define <2 x double> @constrained_vector_frem_v2f64() {
+; S390X-LABEL: constrained_vector_frem_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI6_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI6_1
+; S390X-NEXT:    ldeb %f8, 0(%r1)
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    larl %r1, .LCPI6_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_frem_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -184
+; SZ13-NEXT:    .cfi_def_cfa_offset 344
+; SZ13-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI6_0
+; SZ13-NEXT:    ldeb %f8, 0(%r1)
+; SZ13-NEXT:    vgmg %v0, 1, 1
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vgmg %v0, 2, 11
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 296(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %rem = call <2 x double> @llvm.experimental.constrained.frem.v2f64(
+           <2 x double> <double 1.000000e+00, double 2.000000e+00>,
+           <2 x double> <double 1.000000e+01, double 1.000000e+01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <2 x double> %rem
+}
+
+define <3 x float> @constrained_vector_frem_v3f32() {
+; S390X-LABEL: constrained_vector_frem_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI7_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI7_1
+; S390X-NEXT:    le %f8, 0(%r1)
+; S390X-NEXT:    ler %f2, %f8
+; S390X-NEXT:    brasl %r14, fmodf@PLT
+; S390X-NEXT:    larl %r1, .LCPI7_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    ler %f2, %f8
+; S390X-NEXT:    brasl %r14, fmodf@PLT
+; S390X-NEXT:    larl %r1, .LCPI7_3
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f10, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    ler %f2, %f8
+; S390X-NEXT:    brasl %r14, fmodf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f9
+; S390X-NEXT:    ler %f2, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_frem_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI7_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI7_1
+; SZ13-NEXT:    lde %f8, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmodf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vgmf %v0, 2, 8
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmodf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vgmf %v0, 1, 1
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmodf@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 312(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %rem = call <3 x float> @llvm.experimental.constrained.frem.v3f32(
+           <3 x float> <float 1.000000e+00, float 2.000000e+00, float 3.000000e+00>,
+           <3 x float> <float 1.000000e+01, float 1.000000e+01, float 1.000000e+01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <3 x float> %rem
+}
+
+define void @constrained_vector_frem_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_frem_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f2, 0(%r2)
+; S390X-NEXT:    larl %r1, .LCPI8_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    larl %r1, .LCPI8_1
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    larl %r1, .LCPI8_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_frem_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    vl %v2, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    vgmg %v0, 2, 11
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v2, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f2d killed $f2d killed $v2
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v2, %v0, 1
+; SZ13-NEXT:    vgmg %v0, 1, 1
+; SZ13-NEXT:    # kill: def $f2d killed $f2d killed $v2
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    larl %r1, .LCPI8_0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %rem = call <3 x double> @llvm.experimental.constrained.frem.v3f64(
+           <3 x double> <double 1.000000e+00, double 2.000000e+00, double 3.000000e+00>,
+           <3 x double> %b,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  store <3 x double> %rem, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_frem_v4f64() {
+; S390X-LABEL: constrained_vector_frem_v4f64:
+; S390X:       # %bb.0:
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -192
+; S390X-NEXT:    .cfi_def_cfa_offset 352
+; S390X-NEXT:    std %f8, 184(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f11, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    .cfi_offset %f11, -192
+; S390X-NEXT:    larl %r1, .LCPI9_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI9_1
+; S390X-NEXT:    ldeb %f8, 0(%r1)
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    larl %r1, .LCPI9_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    larl %r1, .LCPI9_3
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    larl %r1, .LCPI9_4
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f11, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, fmod@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    ldr %f2, %f10
+; S390X-NEXT:    ldr %f4, %f11
+; S390X-NEXT:    ld %f8, 184(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f11, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 304(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_frem_v4f64:
+; SZ13:       # %bb.0:
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI9_0
+; SZ13-NEXT:    ldeb %f8, 0(%r1)
+; SZ13-NEXT:    vgmg %v0, 1, 1
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vgmg %v0, 2, 11
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI9_1
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    larl %r1, .LCPI9_2
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmod@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 312(%r15)
+; SZ13-NEXT:    br %r14
+  %rem = call <4 x double> @llvm.experimental.constrained.frem.v4f64(
+           <4 x double> <double 1.000000e+00, double 2.000000e+00,
+                         double 3.000000e+00, double 4.000000e+00>,
+           <4 x double> <double 1.000000e+01, double 1.000000e+01,
+                         double 1.000000e+01, double 1.000000e+01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <4 x double> %rem
+}
+
+define <1 x float> @constrained_vector_fmul_v1f32() {
+; S390X-LABEL: constrained_vector_fmul_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI10_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI10_1
+; S390X-NEXT:    meeb %f0, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fmul_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmf %v0, 1, 1
+; SZ13-NEXT:    vgmf %v1, 1, 8
+; SZ13-NEXT:    meebr %f1, %f0
+; SZ13-NEXT:    vlr %v24, %v1
+; SZ13-NEXT:    br %r14
+entry:
+  %mul = call <1 x float> @llvm.experimental.constrained.fmul.v1f32(
+           <1 x float> <float 0x7FF0000000000000>,
+           <1 x float> <float 2.000000e+00>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <1 x float> %mul
+}
+
+define <2 x double> @constrained_vector_fmul_v2f64() {
+; S390X-LABEL: constrained_vector_fmul_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI11_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI11_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI11_2
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    mdbr %f0, %f1
+; S390X-NEXT:    mdbr %f2, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fmul_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI11_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI11_1
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    vfmdb %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %mul = call <2 x double> @llvm.experimental.constrained.fmul.v2f64(
+           <2 x double> <double 0x7FEFFFFFFFFFFFFF, double 0x7FEFFFFFFFFFFFFF>,
+           <2 x double> <double 2.000000e+00, double 3.000000e+00>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <2 x double> %mul
+}
+
+define <3 x float> @constrained_vector_fmul_v3f32() {
+; S390X-LABEL: constrained_vector_fmul_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI12_0
+; S390X-NEXT:    le %f4, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI12_1
+; S390X-NEXT:    ler %f0, %f4
+; S390X-NEXT:    meeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI12_2
+; S390X-NEXT:    ler %f2, %f4
+; S390X-NEXT:    meeb %f2, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI12_3
+; S390X-NEXT:    meeb %f4, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fmul_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmf %v0, 1, 8
+; SZ13-NEXT:    larl %r1, .LCPI12_0
+; SZ13-NEXT:    vgmf %v2, 2, 8
+; SZ13-NEXT:    vgmf %v1, 1, 8
+; SZ13-NEXT:    meeb %f1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI12_1
+; SZ13-NEXT:    meebr %f2, %f0
+; SZ13-NEXT:    meeb %f0, 0(%r1)
+; SZ13-NEXT:    vmrhf %v0, %v2, %v0
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    br %r14
+entry:
+  %mul = call <3 x float> @llvm.experimental.constrained.fmul.v3f32(
+           <3 x float> <float 0x7FF0000000000000, float 0x7FF0000000000000,
+                        float 0x7FF0000000000000>,
+           <3 x float> <float 1.000000e+00, float 1.000000e+01, float 1.000000e+02>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <3 x float> %mul
+}
+
+define void @constrained_vector_fmul_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_fmul_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI13_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    ldr %f1, %f0
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    mdb %f0, 16(%r2)
+; S390X-NEXT:    mdb %f2, 8(%r2)
+; S390X-NEXT:    mdb %f1, 0(%r2)
+; S390X-NEXT:    std %f0, 16(%r2)
+; S390X-NEXT:    std %f2, 8(%r2)
+; S390X-NEXT:    std %f1, 0(%r2)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fmul_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI13_0
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI13_1
+; SZ13-NEXT:    vfmdb %v0, %v1, %v0
+; SZ13-NEXT:    ld %f1, 0(%r1)
+; SZ13-NEXT:    mdb %f1, 16(%r2)
+; SZ13-NEXT:    std %f1, 16(%r2)
+; SZ13-NEXT:    vst %v0, 0(%r2)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %mul = call <3 x double> @llvm.experimental.constrained.fmul.v3f64(
+           <3 x double> <double 0x7FEFFFFFFFFFFFFF, double 0x7FEFFFFFFFFFFFFF,
+                        double 0x7FEFFFFFFFFFFFFF>,
+           <3 x double> %b,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  store <3 x double> %mul, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_fmul_v4f64() {
+; S390X-LABEL: constrained_vector_fmul_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI14_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI14_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI14_2
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI14_3
+; S390X-NEXT:    ldeb %f4, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI14_4
+; S390X-NEXT:    ldeb %f6, 0(%r1)
+; S390X-NEXT:    mdbr %f0, %f1
+; S390X-NEXT:    mdbr %f2, %f1
+; S390X-NEXT:    mdbr %f4, %f1
+; S390X-NEXT:    mdbr %f6, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fmul_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI14_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI14_1
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI14_2
+; SZ13-NEXT:    vfmdb %v24, %v1, %v0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfmdb %v26, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %mul = call <4 x double> @llvm.experimental.constrained.fmul.v4f64(
+           <4 x double> <double 0x7FEFFFFFFFFFFFFF, double 0x7FEFFFFFFFFFFFFF,
+                         double 0x7FEFFFFFFFFFFFFF, double 0x7FEFFFFFFFFFFFFF>,
+           <4 x double> <double 2.000000e+00, double 3.000000e+00,
+                         double 4.000000e+00, double 5.000000e+00>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <4 x double> %mul
+}
+
+define <1 x float> @constrained_vector_fadd_v1f32() {
+; S390X-LABEL: constrained_vector_fadd_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI15_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI15_1
+; S390X-NEXT:    aeb %f0, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fadd_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmf %v0, 2, 8
+; SZ13-NEXT:    vgmf %v1, 1, 8
+; SZ13-NEXT:    aebr %f1, %f0
+; SZ13-NEXT:    vlr %v24, %v1
+; SZ13-NEXT:    br %r14
+entry:
+  %add = call <1 x float> @llvm.experimental.constrained.fadd.v1f32(
+           <1 x float> <float 0x7FF0000000000000>,
+           <1 x float> <float 1.0>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <1 x float> %add
+}
+
+define <2 x double> @constrained_vector_fadd_v2f64() {
+; S390X-LABEL: constrained_vector_fadd_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI16_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI16_1
+; S390X-NEXT:    ld %f2, 0(%r1)
+; S390X-NEXT:    adbr %f0, %f2
+; S390X-NEXT:    larl %r1, .LCPI16_2
+; S390X-NEXT:    adb %f2, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fadd_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI16_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI16_1
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    vfadb %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %add = call <2 x double> @llvm.experimental.constrained.fadd.v2f64(
+           <2 x double> <double 0x7FEFFFFFFFFFFFFF, double 0x7FEFFFFFFFFFFFFF>,
+           <2 x double> <double 1.000000e+00, double 1.000000e-01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <2 x double> %add
+}
+
+define <3 x float> @constrained_vector_fadd_v3f32() {
+; S390X-LABEL: constrained_vector_fadd_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI17_0
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI17_1
+; S390X-NEXT:    ler %f2, %f1
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    aeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI17_2
+; S390X-NEXT:    aeb %f2, 0(%r1)
+; S390X-NEXT:    lzer %f4
+; S390X-NEXT:    aebr %f4, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fadd_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgbm %v0, 15
+; SZ13-NEXT:    vgmf %v2, 1, 1
+; SZ13-NEXT:    vgmf %v3, 2, 8
+; SZ13-NEXT:    lzer %f1
+; SZ13-NEXT:    aebr %f1, %f0
+; SZ13-NEXT:    aebr %f2, %f0
+; SZ13-NEXT:    aebr %f3, %f0
+; SZ13-NEXT:    vmrhf %v0, %v2, %v3
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    br %r14
+entry:
+  %add = call <3 x float> @llvm.experimental.constrained.fadd.v3f32(
+           <3 x float> <float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000,
+                        float 0xFFFFFFFFE0000000>,
+           <3 x float> <float 2.0, float 1.0, float 0.0>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <3 x float> %add
+}
+
+define void @constrained_vector_fadd_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_fadd_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI18_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    ldr %f1, %f0
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    adb %f0, 16(%r2)
+; S390X-NEXT:    adb %f2, 8(%r2)
+; S390X-NEXT:    adb %f1, 0(%r2)
+; S390X-NEXT:    std %f0, 16(%r2)
+; S390X-NEXT:    std %f2, 8(%r2)
+; S390X-NEXT:    std %f1, 0(%r2)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fadd_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI18_0
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI18_1
+; SZ13-NEXT:    vfadb %v0, %v1, %v0
+; SZ13-NEXT:    ld %f1, 0(%r1)
+; SZ13-NEXT:    adb %f1, 16(%r2)
+; SZ13-NEXT:    std %f1, 16(%r2)
+; SZ13-NEXT:    vst %v0, 0(%r2)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %add = call <3 x double> @llvm.experimental.constrained.fadd.v3f64(
+           <3 x double> <double 0x7FEFFFFFFFFFFFFF, double 0x7FEFFFFFFFFFFFFF,
+                         double 0x7FEFFFFFFFFFFFFF>,
+           <3 x double> %b,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  store <3 x double> %add, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_fadd_v4f64() {
+; S390X-LABEL: constrained_vector_fadd_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI19_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI19_1
+; S390X-NEXT:    ld %f6, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI19_3
+; S390X-NEXT:    ldeb %f4, 0(%r1)
+; S390X-NEXT:    adbr %f0, %f6
+; S390X-NEXT:    larl %r1, .LCPI19_2
+; S390X-NEXT:    ldr %f2, %f6
+; S390X-NEXT:    adb %f2, 0(%r1)
+; S390X-NEXT:    adbr %f4, %f6
+; S390X-NEXT:    larl %r1, .LCPI19_4
+; S390X-NEXT:    adb %f6, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fadd_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI19_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI19_1
+; SZ13-NEXT:    vl %v1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI19_2
+; SZ13-NEXT:    vfadb %v24, %v1, %v0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfadb %v26, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %add = call <4 x double> @llvm.experimental.constrained.fadd.v4f64(
+           <4 x double> <double 0x7FEFFFFFFFFFFFFF, double 0x7FEFFFFFFFFFFFFF,
+                         double 0x7FEFFFFFFFFFFFFF, double 0x7FEFFFFFFFFFFFFF>,
+           <4 x double> <double 1.000000e+00, double 1.000000e-01,
+                         double 2.000000e+00, double 2.000000e-01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <4 x double> %add
+}
+
+define <1 x float> @constrained_vector_fsub_v1f32() {
+; S390X-LABEL: constrained_vector_fsub_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI20_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI20_1
+; S390X-NEXT:    seb %f0, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fsub_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmf %v0, 2, 8
+; SZ13-NEXT:    vgmf %v1, 1, 8
+; SZ13-NEXT:    sebr %f1, %f0
+; SZ13-NEXT:    vlr %v24, %v1
+; SZ13-NEXT:    br %r14
+entry:
+  %sub = call <1 x float> @llvm.experimental.constrained.fsub.v1f32(
+           <1 x float> <float 0x7FF0000000000000>,
+           <1 x float> <float 1.000000e+00>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <1 x float> %sub
+}
+
+define <2 x double> @constrained_vector_fsub_v2f64() {
+; S390X-LABEL: constrained_vector_fsub_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI21_1
+; S390X-NEXT:    ld %f2, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI21_0
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f0, %f2
+; S390X-NEXT:    larl %r1, .LCPI21_2
+; S390X-NEXT:    sdb %f2, 0(%r1)
+; S390X-NEXT:    sdbr %f0, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fsub_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI21_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vgmg %v1, 12, 10
+; SZ13-NEXT:    vfsdb %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %sub = call <2 x double> @llvm.experimental.constrained.fsub.v2f64(
+           <2 x double> <double 0xFFEFFFFFFFFFFFFF, double 0xFFEFFFFFFFFFFFFF>,
+           <2 x double> <double 1.000000e+00, double 1.000000e-01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <2 x double> %sub
+}
+
+define <3 x float> @constrained_vector_fsub_v3f32() {
+; S390X-LABEL: constrained_vector_fsub_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI22_0
+; S390X-NEXT:    le %f4, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI22_1
+; S390X-NEXT:    ler %f0, %f4
+; S390X-NEXT:    seb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI22_2
+; S390X-NEXT:    ler %f2, %f4
+; S390X-NEXT:    seb %f2, 0(%r1)
+; S390X-NEXT:    lzer %f1
+; S390X-NEXT:    sebr %f4, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fsub_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgbm %v2, 15
+; SZ13-NEXT:    lzer %f1
+; SZ13-NEXT:    sebr %f2, %f1
+; SZ13-NEXT:    vgmf %v1, 1, 1
+; SZ13-NEXT:    vgbm %v3, 15
+; SZ13-NEXT:    vgbm %v0, 15
+; SZ13-NEXT:    sebr %f3, %f1
+; SZ13-NEXT:    vgmf %v1, 2, 8
+; SZ13-NEXT:    sebr %f0, %f1
+; SZ13-NEXT:    vmrhf %v0, %v3, %v0
+; SZ13-NEXT:    vrepf %v1, %v2, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    br %r14
+entry:
+  %sub = call <3 x float> @llvm.experimental.constrained.fsub.v3f32(
+           <3 x float> <float 0xFFFFFFFFE0000000, float 0xFFFFFFFFE0000000,
+                        float 0xFFFFFFFFE0000000>,
+           <3 x float> <float 2.0, float 1.0, float 0.0>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <3 x float> %sub
+}
+
+define void @constrained_vector_fsub_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_fsub_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI23_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    ldr %f1, %f0
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    sdb %f0, 16(%r2)
+; S390X-NEXT:    sdb %f2, 8(%r2)
+; S390X-NEXT:    sdb %f1, 0(%r2)
+; S390X-NEXT:    std %f0, 16(%r2)
+; S390X-NEXT:    std %f2, 8(%r2)
+; S390X-NEXT:    std %f1, 0(%r2)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fsub_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    vgmg %v1, 12, 10
+; SZ13-NEXT:    vfsdb %v0, %v1, %v0
+; SZ13-NEXT:    sdb %f1, 16(%r2)
+; SZ13-NEXT:    std %f1, 16(%r2)
+; SZ13-NEXT:    vst %v0, 0(%r2)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %sub = call <3 x double> @llvm.experimental.constrained.fsub.v3f64(
+           <3 x double> <double 0xFFEFFFFFFFFFFFFF, double 0xFFEFFFFFFFFFFFFF,
+                         double 0xFFEFFFFFFFFFFFFF>,
+           <3 x double> %b,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  store <3 x double> %sub, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_fsub_v4f64() {
+; S390X-LABEL: constrained_vector_fsub_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI24_1
+; S390X-NEXT:    ld %f6, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI24_0
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f0, %f6
+; S390X-NEXT:    larl %r1, .LCPI24_2
+; S390X-NEXT:    ldr %f2, %f6
+; S390X-NEXT:    sdb %f2, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI24_3
+; S390X-NEXT:    ldeb %f3, 0(%r1)
+; S390X-NEXT:    ldr %f4, %f6
+; S390X-NEXT:    larl %r1, .LCPI24_4
+; S390X-NEXT:    sdb %f6, 0(%r1)
+; S390X-NEXT:    sdbr %f0, %f1
+; S390X-NEXT:    sdbr %f4, %f3
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fsub_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI24_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vgmg %v1, 12, 10
+; SZ13-NEXT:    larl %r1, .LCPI24_1
+; SZ13-NEXT:    vfsdb %v24, %v1, %v0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfsdb %v26, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %sub = call <4 x double> @llvm.experimental.constrained.fsub.v4f64(
+           <4 x double> <double 0xFFEFFFFFFFFFFFFF, double 0xFFEFFFFFFFFFFFFF,
+                         double 0xFFEFFFFFFFFFFFFF, double 0xFFEFFFFFFFFFFFFF>,
+           <4 x double> <double 1.000000e+00, double 1.000000e-01,
+                         double 2.000000e+00, double 2.000000e-01>,
+           metadata !"round.dynamic",
+           metadata !"fpexcept.strict")
+  ret <4 x double> %sub
+}
+
+define <1 x float> @constrained_vector_sqrt_v1f32() {
+; S390X-LABEL: constrained_vector_sqrt_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI25_0
+; S390X-NEXT:    sqeb %f0, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sqrt_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI25_0
+; SZ13-NEXT:    sqeb %f0, 0(%r1)
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %sqrt = call <1 x float> @llvm.experimental.constrained.sqrt.v1f32(
+                              <1 x float> <float 42.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <1 x float> %sqrt
+}
+
+define <2 x double> @constrained_vector_sqrt_v2f64() {
+; S390X-LABEL: constrained_vector_sqrt_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI26_0
+; S390X-NEXT:    larl %r2, .LCPI26_1
+; S390X-NEXT:    ldeb %f0, 0(%r2)
+; S390X-NEXT:    sqdb %f2, 0(%r1)
+; S390X-NEXT:    sqdbr %f0, %f0
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sqrt_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI26_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfsqdb %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %sqrt = call <2 x double> @llvm.experimental.constrained.sqrt.v2f64(
+                              <2 x double> <double 42.0, double 42.1>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <2 x double> %sqrt
+}
+
+define <3 x float> @constrained_vector_sqrt_v3f32() {
+; S390X-LABEL: constrained_vector_sqrt_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI27_0
+; S390X-NEXT:    sqeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI27_1
+; S390X-NEXT:    sqeb %f2, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI27_2
+; S390X-NEXT:    sqeb %f4, 0(%r1)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sqrt_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI27_0
+; SZ13-NEXT:    sqeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI27_1
+; SZ13-NEXT:    vrepf %v0, %v0, 0
+; SZ13-NEXT:    sqeb %f1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI27_2
+; SZ13-NEXT:    sqeb %f2, 0(%r1)
+; SZ13-NEXT:    vmrhf %v1, %v1, %v2
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %sqrt = call <3 x float> @llvm.experimental.constrained.sqrt.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %sqrt
+}
+
+define void @constrained_vector_sqrt_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_sqrt_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    sqdb %f0, 16(%r2)
+; S390X-NEXT:    sqdb %f1, 8(%r2)
+; S390X-NEXT:    sqdb %f2, 0(%r2)
+; S390X-NEXT:    std %f0, 16(%r2)
+; S390X-NEXT:    std %f1, 8(%r2)
+; S390X-NEXT:    std %f2, 0(%r2)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sqrt_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    sqdb %f1, 16(%r2)
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    std %f1, 16(%r2)
+; SZ13-NEXT:    vfsqdb %v0, %v0
+; SZ13-NEXT:    vst %v0, 0(%r2)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %sqrt = call <3 x double> @llvm.experimental.constrained.sqrt.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %sqrt, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_sqrt_v4f64() {
+; S390X-LABEL: constrained_vector_sqrt_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI29_0
+; S390X-NEXT:    sqdb %f2, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI29_1
+; S390X-NEXT:    sqdb %f4, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI29_3
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI29_2
+; S390X-NEXT:    sqdb %f6, 0(%r1)
+; S390X-NEXT:    sqdbr %f0, %f0
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sqrt_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI29_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfsqdb %v24, %v0
+; SZ13-NEXT:    larl %r1, .LCPI29_1
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfsqdb %v26, %v0
+; SZ13-NEXT:    br %r14
+ entry:
+  %sqrt = call <4 x double> @llvm.experimental.constrained.sqrt.v4f64(
+                              <4 x double> <double 42.0, double 42.1,
+                                            double 42.2, double 42.3>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <4 x double> %sqrt
+}
+
+define <1 x float> @constrained_vector_pow_v1f32() {
+; S390X-LABEL: constrained_vector_pow_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI30_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI30_1
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    brasl %r14, powf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_pow_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI30_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI30_1
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, powf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %pow = call <1 x float> @llvm.experimental.constrained.pow.v1f32(
+                             <1 x float> <float 42.0>,
+                             <1 x float> <float 3.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %pow
+}
+
+define <2 x double> @constrained_vector_pow_v2f64() {
+; S390X-LABEL: constrained_vector_pow_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI31_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI31_1
+; S390X-NEXT:    ldeb %f8, 0(%r1)
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    larl %r1, .LCPI31_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_pow_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -184
+; SZ13-NEXT:    .cfi_def_cfa_offset 344
+; SZ13-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI31_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI31_1
+; SZ13-NEXT:    ldeb %f8, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    larl %r1, .LCPI31_2
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 296(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %pow = call <2 x double> @llvm.experimental.constrained.pow.v2f64(
+                             <2 x double> <double 42.1, double 42.2>,
+                             <2 x double> <double 3.0, double 3.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <2 x double> %pow
+}
+
+define <3 x float> @constrained_vector_pow_v3f32() {
+; S390X-LABEL: constrained_vector_pow_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI32_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI32_1
+; S390X-NEXT:    le %f8, 0(%r1)
+; S390X-NEXT:    ler %f2, %f8
+; S390X-NEXT:    brasl %r14, powf@PLT
+; S390X-NEXT:    larl %r1, .LCPI32_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    ler %f2, %f8
+; S390X-NEXT:    brasl %r14, powf@PLT
+; S390X-NEXT:    larl %r1, .LCPI32_3
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f10, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    ler %f2, %f8
+; S390X-NEXT:    brasl %r14, powf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f9
+; S390X-NEXT:    ler %f2, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_pow_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI32_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI32_1
+; SZ13-NEXT:    lde %f8, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, powf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI32_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, powf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI32_3
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, powf@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 312(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %pow = call <3 x float> @llvm.experimental.constrained.pow.v3f32(
+                             <3 x float> <float 42.0, float 43.0, float 44.0>,
+                             <3 x float> <float 3.0, float 3.0, float 3.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <3 x float> %pow
+}
+
+define void @constrained_vector_pow_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_pow_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -192
+; S390X-NEXT:    .cfi_def_cfa_offset 352
+; S390X-NEXT:    std %f8, 184(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f11, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    .cfi_offset %f11, -192
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    larl %r1, .LCPI33_0
+; S390X-NEXT:    ldeb %f9, 0(%r1)
+; S390X-NEXT:    ld %f10, 8(%r2)
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    ldr %f11, %f0
+; S390X-NEXT:    ldr %f0, %f10
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f10, 8(%r13)
+; S390X-NEXT:    std %f11, 0(%r13)
+; S390X-NEXT:    ld %f8, 184(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f11, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_pow_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -208
+; SZ13-NEXT:    .cfi_def_cfa_offset 368
+; SZ13-NEXT:    std %f8, 200(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    std %f9, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    .cfi_offset %f9, -176
+; SZ13-NEXT:    larl %r1, .LCPI33_0
+; SZ13-NEXT:    ldeb %f9, 0(%r1)
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    ldr %f2, %f9
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ldr %f2, %f9
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    ldr %f2, %f9
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 200(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    ld %f9, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 312(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %pow = call <3 x double> @llvm.experimental.constrained.pow.v3f64(
+                          <3 x double> %b,
+                          <3 x double> <double 3.0, double 3.0, double 3.0>,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %pow, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_pow_v4f64() {
+; S390X-LABEL: constrained_vector_pow_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -192
+; S390X-NEXT:    .cfi_def_cfa_offset 352
+; S390X-NEXT:    std %f8, 184(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f11, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    .cfi_offset %f11, -192
+; S390X-NEXT:    larl %r1, .LCPI34_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI34_1
+; S390X-NEXT:    ldeb %f8, 0(%r1)
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    larl %r1, .LCPI34_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    larl %r1, .LCPI34_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    larl %r1, .LCPI34_4
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f11, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    ldr %f2, %f8
+; S390X-NEXT:    brasl %r14, pow@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    ldr %f2, %f10
+; S390X-NEXT:    ldr %f4, %f11
+; S390X-NEXT:    ld %f8, 184(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f11, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 304(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_pow_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI34_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI34_1
+; SZ13-NEXT:    ldeb %f8, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    larl %r1, .LCPI34_2
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI34_3
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    larl %r1, .LCPI34_4
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, pow@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 312(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %pow = call <4 x double> @llvm.experimental.constrained.pow.v4f64(
+                             <4 x double> <double 42.1, double 42.2,
+                                           double 42.3, double 42.4>,
+                             <4 x double> <double 3.0, double 3.0,
+                                           double 3.0, double 3.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <4 x double> %pow
+}
+
+define <1 x float> @constrained_vector_powi_v1f32() {
+; S390X-LABEL: constrained_vector_powi_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI35_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    brasl %r14, __powisf2@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_powi_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI35_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powisf2@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %powi = call <1 x float> @llvm.experimental.constrained.powi.v1f32(
+                              <1 x float> <float 42.0>,
+                              i32 3,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <1 x float> %powi
+}
+
+define <2 x double> @constrained_vector_powi_v2f64() {
+; S390X-LABEL: constrained_vector_powi_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI36_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    larl %r1, .LCPI36_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_powi_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI36_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI36_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %powi = call <2 x double> @llvm.experimental.constrained.powi.v2f64(
+                              <2 x double> <double 42.1, double 42.2>,
+                              i32 3,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <2 x double> %powi
+}
+
+define <3 x float> @constrained_vector_powi_v3f32() {
+; S390X-LABEL: constrained_vector_powi_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI37_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    brasl %r14, __powisf2@PLT
+; S390X-NEXT:    larl %r1, .LCPI37_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, __powisf2@PLT
+; S390X-NEXT:    larl %r1, .LCPI37_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, __powisf2@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_powi_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI37_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powisf2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI37_1
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powisf2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI37_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powisf2@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %powi = call <3 x float> @llvm.experimental.constrained.powi.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              i32 3,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %powi
+}
+
+define void @constrained_vector_powi_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_powi_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI38_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    larl %r1, .LCPI38_1
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    larl %r1, .LCPI38_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    std %f0, 8(%r13)
+; S390X-NEXT:    std %f9, 0(%r13)
+; S390X-NEXT:    std %f8, 16(%r13)
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_powi_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI38_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI38_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI38_2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 280(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %powi = call <3 x double> @llvm.experimental.constrained.powi.v3f64(
+                          <3 x double> <double 42.0, double 42.1, double 42.2>,
+                          i32 3,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %powi, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_powi_v4f64() {
+; S390X-LABEL: constrained_vector_powi_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI39_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    larl %r1, .LCPI39_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    larl %r1, .LCPI39_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    larl %r1, .LCPI39_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    lghi %r2, 3
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, __powidf2@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_powi_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI39_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI39_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI39_2
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI39_3
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    lghi %r2, 3
+; SZ13-NEXT:    brasl %r14, __powidf2@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %powi = call <4 x double> @llvm.experimental.constrained.powi.v4f64(
+                              <4 x double> <double 42.1, double 42.2,
+                                            double 42.3, double 42.4>,
+                              i32 3,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <4 x double> %powi
+}
+
+define <1 x float> @constrained_vector_sin_v1f32() {
+; S390X-LABEL: constrained_vector_sin_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI40_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, sinf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sin_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI40_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sinf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %sin = call <1 x float> @llvm.experimental.constrained.sin.v1f32(
+                             <1 x float> <float 42.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %sin
+}
+
+define <2 x double> @constrained_vector_sin_v2f64() {
+; S390X-LABEL: constrained_vector_sin_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI41_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    larl %r1, .LCPI41_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sin_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI41_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    larl %r1, .LCPI41_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %sin = call <2 x double> @llvm.experimental.constrained.sin.v2f64(
+                             <2 x double> <double 42.0, double 42.1>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <2 x double> %sin
+}
+
+define <3 x float> @constrained_vector_sin_v3f32() {
+; S390X-LABEL: constrained_vector_sin_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI42_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, sinf@PLT
+; S390X-NEXT:    larl %r1, .LCPI42_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, sinf@PLT
+; S390X-NEXT:    larl %r1, .LCPI42_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, sinf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sin_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI42_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sinf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI42_1
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sinf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI42_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sinf@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %sin = call <3 x float> @llvm.experimental.constrained.sin.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %sin
+}
+
+define void @constrained_vector_sin_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_sin_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sin_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %sin = call <3 x double> @llvm.experimental.constrained.sin.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %sin, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_sin_v4f64() {
+; S390X-LABEL: constrained_vector_sin_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI44_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    larl %r1, .LCPI44_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    larl %r1, .LCPI44_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    larl %r1, .LCPI44_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, sin@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_sin_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI44_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    larl %r1, .LCPI44_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI44_2
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    larl %r1, .LCPI44_3
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, sin@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %sin = call <4 x double> @llvm.experimental.constrained.sin.v4f64(
+                             <4 x double> <double 42.0, double 42.1,
+                                           double 42.2, double 42.3>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <4 x double> %sin
+}
+
+define <1 x float> @constrained_vector_cos_v1f32() {
+; S390X-LABEL: constrained_vector_cos_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI45_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, cosf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_cos_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI45_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cosf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %cos = call <1 x float> @llvm.experimental.constrained.cos.v1f32(
+                             <1 x float> <float 42.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %cos
+}
+
+define <2 x double> @constrained_vector_cos_v2f64() {
+; S390X-LABEL: constrained_vector_cos_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI46_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    larl %r1, .LCPI46_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_cos_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI46_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    larl %r1, .LCPI46_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %cos = call <2 x double> @llvm.experimental.constrained.cos.v2f64(
+                             <2 x double> <double 42.0, double 42.1>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <2 x double> %cos
+}
+
+define <3 x float> @constrained_vector_cos_v3f32() {
+; S390X-LABEL: constrained_vector_cos_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI47_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, cosf@PLT
+; S390X-NEXT:    larl %r1, .LCPI47_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, cosf@PLT
+; S390X-NEXT:    larl %r1, .LCPI47_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, cosf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_cos_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI47_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cosf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI47_1
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cosf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI47_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cosf@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %cos = call <3 x float> @llvm.experimental.constrained.cos.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %cos
+}
+
+define void @constrained_vector_cos_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_cos_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_cos_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %cos = call <3 x double> @llvm.experimental.constrained.cos.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %cos, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_cos_v4f64() {
+; S390X-LABEL: constrained_vector_cos_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI49_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    larl %r1, .LCPI49_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    larl %r1, .LCPI49_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    larl %r1, .LCPI49_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, cos@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_cos_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI49_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    larl %r1, .LCPI49_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI49_2
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    larl %r1, .LCPI49_3
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, cos@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %cos = call <4 x double> @llvm.experimental.constrained.cos.v4f64(
+                             <4 x double> <double 42.0, double 42.1,
+                                           double 42.2, double 42.3>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <4 x double> %cos
+}
+
+define <1 x float> @constrained_vector_exp_v1f32() {
+; S390X-LABEL: constrained_vector_exp_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI50_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, expf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI50_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, expf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %exp = call <1 x float> @llvm.experimental.constrained.exp.v1f32(
+                             <1 x float> <float 42.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %exp
+}
+
+define <2 x double> @constrained_vector_exp_v2f64() {
+; S390X-LABEL: constrained_vector_exp_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI51_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    larl %r1, .LCPI51_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI51_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    larl %r1, .LCPI51_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %exp = call <2 x double> @llvm.experimental.constrained.exp.v2f64(
+                             <2 x double> <double 42.0, double 42.1>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <2 x double> %exp
+}
+
+define <3 x float> @constrained_vector_exp_v3f32() {
+; S390X-LABEL: constrained_vector_exp_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI52_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, expf@PLT
+; S390X-NEXT:    larl %r1, .LCPI52_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, expf@PLT
+; S390X-NEXT:    larl %r1, .LCPI52_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, expf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI52_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, expf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI52_1
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, expf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI52_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, expf@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %exp = call <3 x float> @llvm.experimental.constrained.exp.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %exp
+}
+
+define void @constrained_vector_exp_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_exp_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %exp = call <3 x double> @llvm.experimental.constrained.exp.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %exp, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_exp_v4f64() {
+; S390X-LABEL: constrained_vector_exp_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI54_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    larl %r1, .LCPI54_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    larl %r1, .LCPI54_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    larl %r1, .LCPI54_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, exp@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI54_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    larl %r1, .LCPI54_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI54_2
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    larl %r1, .LCPI54_3
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %exp = call <4 x double> @llvm.experimental.constrained.exp.v4f64(
+                             <4 x double> <double 42.0, double 42.1,
+                                           double 42.2, double 42.3>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <4 x double> %exp
+}
+
+define <1 x float> @constrained_vector_exp2_v1f32() {
+; S390X-LABEL: constrained_vector_exp2_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI55_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, exp2f@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp2_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI55_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2f@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %exp2 = call <1 x float> @llvm.experimental.constrained.exp2.v1f32(
+                             <1 x float> <float 42.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %exp2
+}
+
+define <2 x double> @constrained_vector_exp2_v2f64() {
+; S390X-LABEL: constrained_vector_exp2_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI56_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    larl %r1, .LCPI56_1
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp2_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI56_0
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI56_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %exp2 = call <2 x double> @llvm.experimental.constrained.exp2.v2f64(
+                              <2 x double> <double 42.1, double 42.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <2 x double> %exp2
+}
+
+define <3 x float> @constrained_vector_exp2_v3f32() {
+; S390X-LABEL: constrained_vector_exp2_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI57_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, exp2f@PLT
+; S390X-NEXT:    larl %r1, .LCPI57_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, exp2f@PLT
+; S390X-NEXT:    larl %r1, .LCPI57_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, exp2f@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp2_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI57_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2f@PLT
+; SZ13-NEXT:    larl %r1, .LCPI57_1
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2f@PLT
+; SZ13-NEXT:    larl %r1, .LCPI57_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2f@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %exp2 = call <3 x float> @llvm.experimental.constrained.exp2.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %exp2
+}
+
+define void @constrained_vector_exp2_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_exp2_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp2_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %exp2 = call <3 x double> @llvm.experimental.constrained.exp2.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %exp2, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_exp2_v4f64() {
+; S390X-LABEL: constrained_vector_exp2_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI59_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    larl %r1, .LCPI59_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    larl %r1, .LCPI59_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    larl %r1, .LCPI59_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, exp2@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_exp2_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI59_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI59_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI59_2
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI59_3
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, exp2@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %exp2 = call <4 x double> @llvm.experimental.constrained.exp2.v4f64(
+                              <4 x double> <double 42.1, double 42.2,
+                                            double 42.3, double 42.4>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <4 x double> %exp2
+}
+
+define <1 x float> @constrained_vector_log_v1f32() {
+; S390X-LABEL: constrained_vector_log_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI60_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, logf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI60_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, logf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log = call <1 x float> @llvm.experimental.constrained.log.v1f32(
+                             <1 x float> <float 42.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %log
+}
+
+define <2 x double> @constrained_vector_log_v2f64() {
+; S390X-LABEL: constrained_vector_log_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI61_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    larl %r1, .LCPI61_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI61_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    larl %r1, .LCPI61_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log = call <2 x double> @llvm.experimental.constrained.log.v2f64(
+                             <2 x double> <double 42.0, double 42.1>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <2 x double> %log
+}
+
+define <3 x float> @constrained_vector_log_v3f32() {
+; S390X-LABEL: constrained_vector_log_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI62_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, logf@PLT
+; S390X-NEXT:    larl %r1, .LCPI62_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, logf@PLT
+; S390X-NEXT:    larl %r1, .LCPI62_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, logf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI62_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, logf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI62_1
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, logf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI62_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, logf@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log = call <3 x float> @llvm.experimental.constrained.log.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %log
+}
+
+define void @constrained_vector_log_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_log_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %log = call <3 x double> @llvm.experimental.constrained.log.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %log, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_log_v4f64() {
+; S390X-LABEL: constrained_vector_log_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI64_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    larl %r1, .LCPI64_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    larl %r1, .LCPI64_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    larl %r1, .LCPI64_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI64_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    larl %r1, .LCPI64_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI64_2
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    larl %r1, .LCPI64_3
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log = call <4 x double> @llvm.experimental.constrained.log.v4f64(
+                             <4 x double> <double 42.0, double 42.1,
+                                           double 42.2, double 42.3>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <4 x double> %log
+}
+
+define <1 x float> @constrained_vector_log10_v1f32() {
+; S390X-LABEL: constrained_vector_log10_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI65_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log10f@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log10_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI65_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10f@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log10 = call <1 x float> @llvm.experimental.constrained.log10.v1f32(
+                             <1 x float> <float 42.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %log10
+}
+
+define <2 x double> @constrained_vector_log10_v2f64() {
+; S390X-LABEL: constrained_vector_log10_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI66_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    larl %r1, .LCPI66_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log10_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI66_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    larl %r1, .LCPI66_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log10 = call <2 x double> @llvm.experimental.constrained.log10.v2f64(
+                               <2 x double> <double 42.0, double 42.1>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <2 x double> %log10
+}
+
+define <3 x float> @constrained_vector_log10_v3f32() {
+; S390X-LABEL: constrained_vector_log10_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI67_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log10f@PLT
+; S390X-NEXT:    larl %r1, .LCPI67_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, log10f@PLT
+; S390X-NEXT:    larl %r1, .LCPI67_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, log10f@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log10_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI67_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10f@PLT
+; SZ13-NEXT:    larl %r1, .LCPI67_1
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10f@PLT
+; SZ13-NEXT:    larl %r1, .LCPI67_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10f@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log10 = call <3 x float> @llvm.experimental.constrained.log10.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %log10
+}
+
+define void @constrained_vector_log10_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_log10_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log10_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %log10 = call <3 x double> @llvm.experimental.constrained.log10.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %log10, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_log10_v4f64() {
+; S390X-LABEL: constrained_vector_log10_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI69_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    larl %r1, .LCPI69_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    larl %r1, .LCPI69_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    larl %r1, .LCPI69_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log10@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log10_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI69_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    larl %r1, .LCPI69_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI69_2
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    larl %r1, .LCPI69_3
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log10@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log10 = call <4 x double> @llvm.experimental.constrained.log10.v4f64(
+                               <4 x double> <double 42.0, double 42.1,
+                                             double 42.2, double 42.3>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <4 x double> %log10
+}
+
+define <1 x float> @constrained_vector_log2_v1f32() {
+; S390X-LABEL: constrained_vector_log2_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI70_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log2f@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log2_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI70_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2f@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log2 = call <1 x float> @llvm.experimental.constrained.log2.v1f32(
+                             <1 x float> <float 42.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %log2
+}
+
+define <2 x double> @constrained_vector_log2_v2f64() {
+; S390X-LABEL: constrained_vector_log2_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI71_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    larl %r1, .LCPI71_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log2_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI71_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI71_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log2 = call <2 x double> @llvm.experimental.constrained.log2.v2f64(
+                              <2 x double> <double 42.0, double 42.1>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <2 x double> %log2
+}
+
+define <3 x float> @constrained_vector_log2_v3f32() {
+; S390X-LABEL: constrained_vector_log2_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI72_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log2f@PLT
+; S390X-NEXT:    larl %r1, .LCPI72_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, log2f@PLT
+; S390X-NEXT:    larl %r1, .LCPI72_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, log2f@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log2_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI72_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2f@PLT
+; SZ13-NEXT:    larl %r1, .LCPI72_1
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2f@PLT
+; SZ13-NEXT:    larl %r1, .LCPI72_2
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2f@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log2 = call <3 x float> @llvm.experimental.constrained.log2.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %log2
+}
+
+define void @constrained_vector_log2_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_log2_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log2_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %log2 = call <3 x double> @llvm.experimental.constrained.log2.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %log2, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_log2_v4f64() {
+; S390X-LABEL: constrained_vector_log2_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI74_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    larl %r1, .LCPI74_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    larl %r1, .LCPI74_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    larl %r1, .LCPI74_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, log2@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log2_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI74_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI74_1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI74_2
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    larl %r1, .LCPI74_3
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    brasl %r14, log2@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %log2 = call <4 x double> @llvm.experimental.constrained.log2.v4f64(
+                              <4 x double> <double 42.0, double 42.1,
+                                            double 42.2, double 42.3>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <4 x double> %log2
+}
+
+define <1 x float> @constrained_vector_rint_v1f32() {
+; S390X-LABEL: constrained_vector_rint_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI75_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    fiebr %f0, 0, %f0
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_rint_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI75_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    fiebr %f0, 0, %f0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %rint = call <1 x float> @llvm.experimental.constrained.rint.v1f32(
+                             <1 x float> <float 42.0>,
+                             metadata !"round.dynamic",
+                             metadata !"fpexcept.strict")
+  ret <1 x float> %rint
+}
+
+define <2 x double> @constrained_vector_rint_v2f64() {
+; S390X-LABEL: constrained_vector_rint_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI76_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI76_1
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    fidbr %f0, 0, %f0
+; S390X-NEXT:    fidbr %f2, 0, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_rint_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI76_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v24, %v0, 0, 0
+; SZ13-NEXT:    br %r14
+entry:
+  %rint = call <2 x double> @llvm.experimental.constrained.rint.v2f64(
+                        <2 x double> <double 42.1, double 42.0>,
+                        metadata !"round.dynamic",
+                        metadata !"fpexcept.strict")
+  ret <2 x double> %rint
+}
+
+define <3 x float> @constrained_vector_rint_v3f32() {
+; S390X-LABEL: constrained_vector_rint_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI77_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI77_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI77_2
+; S390X-NEXT:    le %f3, 0(%r1)
+; S390X-NEXT:    fiebr %f0, 0, %f0
+; S390X-NEXT:    fiebr %f2, 0, %f1
+; S390X-NEXT:    fiebr %f4, 0, %f3
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_rint_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI77_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI77_1
+; SZ13-NEXT:    lde %f1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI77_2
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    fiebr %f0, 0, %f0
+; SZ13-NEXT:    fiebr %f1, 0, %f1
+; SZ13-NEXT:    fiebr %f2, 0, %f2
+; SZ13-NEXT:    vmrhf %v1, %v1, %v2
+; SZ13-NEXT:    vrepf %v0, %v0, 0
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+ entry:
+  %rint = call <3 x float> @llvm.experimental.constrained.rint.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %rint
+}
+
+define void @constrained_vector_rint_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_rint_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f1, 8(%r2)
+; S390X-NEXT:    ld %f2, 16(%r2)
+; S390X-NEXT:    fidbr %f0, 0, %f0
+; S390X-NEXT:    fidbr %f1, 0, %f1
+; S390X-NEXT:    fidbr %f2, 0, %f2
+; S390X-NEXT:    std %f2, 16(%r2)
+; S390X-NEXT:    std %f1, 8(%r2)
+; S390X-NEXT:    std %f0, 0(%r2)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_rint_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vl %v1, 0(%r2)
+; SZ13-NEXT:    ld %f0, 16(%r2)
+; SZ13-NEXT:    vfidb %v1, %v1, 0, 0
+; SZ13-NEXT:    fidbra %f0, 0, %f0, 0
+; SZ13-NEXT:    std %f0, 16(%r2)
+; SZ13-NEXT:    vst %v1, 0(%r2)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %rint = call <3 x double> @llvm.experimental.constrained.rint.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %rint, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_rint_v4f64() {
+; S390X-LABEL: constrained_vector_rint_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI79_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI79_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI79_2
+; S390X-NEXT:    ld %f3, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI79_3
+; S390X-NEXT:    ld %f5, 0(%r1)
+; S390X-NEXT:    fidbr %f0, 0, %f0
+; S390X-NEXT:    fidbr %f2, 0, %f1
+; S390X-NEXT:    fidbr %f4, 0, %f3
+; S390X-NEXT:    fidbr %f6, 0, %f5
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_rint_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI79_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI79_1
+; SZ13-NEXT:    vfidb %v24, %v0, 0, 0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v26, %v0, 0, 0
+; SZ13-NEXT:    br %r14
+entry:
+  %rint = call <4 x double> @llvm.experimental.constrained.rint.v4f64(
+                        <4 x double> <double 42.1, double 42.2,
+                                      double 42.3, double 42.4>,
+                        metadata !"round.dynamic",
+                        metadata !"fpexcept.strict")
+  ret <4 x double> %rint
+}
+
+define <1 x float> @constrained_vector_nearbyint_v1f32() {
+; S390X-LABEL: constrained_vector_nearbyint_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI80_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, nearbyintf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_nearbyint_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI80_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    fiebra %f0, 0, %f0, 4
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %nearby = call <1 x float> @llvm.experimental.constrained.nearbyint.v1f32(
+                               <1 x float> <float 42.0>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <1 x float> %nearby
+}
+
+define <2 x double> @constrained_vector_nearbyint_v2f64() {
+; S390X-LABEL: constrained_vector_nearbyint_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI81_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    larl %r1, .LCPI81_1
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_nearbyint_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI81_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v24, %v0, 4, 0
+; SZ13-NEXT:    br %r14
+entry:
+  %nearby = call <2 x double> @llvm.experimental.constrained.nearbyint.v2f64(
+                                <2 x double> <double 42.1, double 42.0>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <2 x double> %nearby
+}
+
+define <3 x float> @constrained_vector_nearbyint_v3f32() {
+; S390X-LABEL: constrained_vector_nearbyint_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI82_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, nearbyintf@PLT
+; S390X-NEXT:    larl %r1, .LCPI82_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, nearbyintf@PLT
+; S390X-NEXT:    larl %r1, .LCPI82_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, nearbyintf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_nearbyint_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI82_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI82_1
+; SZ13-NEXT:    lde %f1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI82_2
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    fiebra %f0, 0, %f0, 4
+; SZ13-NEXT:    fiebra %f1, 0, %f1, 4
+; SZ13-NEXT:    fiebra %f2, 0, %f2, 4
+; SZ13-NEXT:    vmrhf %v1, %v1, %v2
+; SZ13-NEXT:    vrepf %v0, %v0, 0
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %nearby = call <3 x float> @llvm.experimental.constrained.nearbyint.v3f32(
+                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %nearby
+}
+
+define void @constrained_vector_nearbyint_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_nearbyint_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_nearbyint_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vl %v1, 0(%r2)
+; SZ13-NEXT:    ld %f0, 16(%r2)
+; SZ13-NEXT:    vfidb %v1, %v1, 4, 0
+; SZ13-NEXT:    fidbra %f0, 0, %f0, 4
+; SZ13-NEXT:    std %f0, 16(%r2)
+; SZ13-NEXT:    vst %v1, 0(%r2)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %nearby = call <3 x double> @llvm.experimental.constrained.nearbyint.v3f64(
+                          <3 x double> %b,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %nearby, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_nearbyint_v4f64() {
+; S390X-LABEL: constrained_vector_nearbyint_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI84_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    larl %r1, .LCPI84_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    larl %r1, .LCPI84_2
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    larl %r1, .LCPI84_3
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, nearbyint@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_nearbyint_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI84_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI84_1
+; SZ13-NEXT:    vfidb %v24, %v0, 4, 0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v26, %v0, 4, 0
+; SZ13-NEXT:    br %r14
+entry:
+  %nearby = call <4 x double> @llvm.experimental.constrained.nearbyint.v4f64(
+                                <4 x double> <double 42.1, double 42.2,
+                                              double 42.3, double 42.4>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <4 x double> %nearby
+}
+
+define <1 x float> @constrained_vector_maxnum_v1f32() {
+; S390X-LABEL: constrained_vector_maxnum_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI85_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI85_1
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    brasl %r14, fmaxf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_maxnum_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI85_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI85_1
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmaxf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %max = call <1 x float> @llvm.experimental.constrained.maxnum.v1f32(
+                               <1 x float> <float 42.0>, <1 x float> <float 41.0>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <1 x float> %max
+}
+
+define <2 x double> @constrained_vector_maxnum_v2f64() {
+; S390X-LABEL: constrained_vector_maxnum_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI86_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI86_1
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    larl %r1, .LCPI86_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI86_3
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_maxnum_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI86_0
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI86_1
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    larl %r1, .LCPI86_2
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI86_3
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %max = call <2 x double> @llvm.experimental.constrained.maxnum.v2f64(
+                                <2 x double> <double 43.0, double 42.0>,
+                                <2 x double> <double 41.0, double 40.0>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <2 x double> %max
+}
+
+define <3 x float> @constrained_vector_maxnum_v3f32() {
+; S390X-LABEL: constrained_vector_maxnum_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI87_0
+; S390X-NEXT:    le %f8, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI87_1
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    brasl %r14, fmaxf@PLT
+; S390X-NEXT:    larl %r1, .LCPI87_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI87_3
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, fmaxf@PLT
+; S390X-NEXT:    larl %r1, .LCPI87_4
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f10, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    ler %f2, %f8
+; S390X-NEXT:    brasl %r14, fmaxf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f9
+; S390X-NEXT:    ler %f2, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_maxnum_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI87_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI87_1
+; SZ13-NEXT:    lde %f8, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fmaxf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI87_2
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, fmaxf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI87_3
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI87_4
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmaxf@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 312(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %max = call <3 x float> @llvm.experimental.constrained.maxnum.v3f32(
+                              <3 x float> <float 43.0, float 44.0, float 45.0>,
+                              <3 x float> <float 41.0, float 42.0, float 43.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %max
+}
+
+define void @constrained_vector_log10_maxnum_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_log10_maxnum_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    larl %r1, .LCPI88_0
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ld %f9, 8(%r2)
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    larl %r1, .LCPI88_1
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f9
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    larl %r1, .LCPI88_2
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f9, 8(%r13)
+; S390X-NEXT:    std %f10, 0(%r13)
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_log10_maxnum_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI88_0
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    larl %r1, .LCPI88_1
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    larl %r1, .LCPI88_2
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %a
+  %max = call <3 x double> @llvm.experimental.constrained.maxnum.v3f64(
+                          <3 x double> %b,
+                          <3 x double> <double 40.0, double 41.0, double 42.0>,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %max, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_maxnum_v4f64() {
+; S390X-LABEL: constrained_vector_maxnum_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI89_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI89_1
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    larl %r1, .LCPI89_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI89_3
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    larl %r1, .LCPI89_4
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI89_5
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    larl %r1, .LCPI89_6
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI89_7
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, fmax@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_maxnum_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI89_0
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI89_1
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    larl %r1, .LCPI89_2
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI89_3
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI89_4
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI89_5
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    larl %r1, .LCPI89_6
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI89_7
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmax@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %max = call <4 x double> @llvm.experimental.constrained.maxnum.v4f64(
+                                <4 x double> <double 44.0, double 45.0,
+                                              double 46.0, double 47.0>,
+                                <4 x double> <double 40.0, double 41.0,
+                                              double 42.0, double 43.0>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <4 x double> %max
+}
+
+define <1 x float> @constrained_vector_minnum_v1f32() {
+; S390X-LABEL: constrained_vector_minnum_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI90_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI90_1
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    brasl %r14, fminf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_minnum_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -160
+; SZ13-NEXT:    .cfi_def_cfa_offset 320
+; SZ13-NEXT:    larl %r1, .LCPI90_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI90_1
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fminf@PLT
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    lmg %r14, %r15, 272(%r15)
+; SZ13-NEXT:    br %r14
+ entry:
+  %min = call <1 x float> @llvm.experimental.constrained.minnum.v1f32(
+                               <1 x float> <float 42.0>, <1 x float> <float 41.0>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <1 x float> %min
+}
+
+define <2 x double> @constrained_vector_minnum_v2f64() {
+; S390X-LABEL: constrained_vector_minnum_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI91_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI91_1
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    larl %r1, .LCPI91_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI91_3
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_minnum_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -176
+; SZ13-NEXT:    .cfi_def_cfa_offset 336
+; SZ13-NEXT:    larl %r1, .LCPI91_0
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI91_1
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    larl %r1, .LCPI91_2
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI91_3
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 288(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %min = call <2 x double> @llvm.experimental.constrained.minnum.v2f64(
+                                <2 x double> <double 43.0, double 42.0>,
+                                <2 x double> <double 41.0, double 40.0>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <2 x double> %min
+}
+
+define <3 x float> @constrained_vector_minnum_v3f32() {
+; S390X-LABEL: constrained_vector_minnum_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI92_0
+; S390X-NEXT:    le %f8, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI92_1
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    brasl %r14, fminf@PLT
+; S390X-NEXT:    larl %r1, .LCPI92_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI92_3
+; S390X-NEXT:    le %f2, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, fminf@PLT
+; S390X-NEXT:    larl %r1, .LCPI92_4
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f10, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    ler %f2, %f8
+; S390X-NEXT:    brasl %r14, fminf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f9
+; S390X-NEXT:    ler %f2, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_minnum_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -200
+; SZ13-NEXT:    .cfi_def_cfa_offset 360
+; SZ13-NEXT:    std %f8, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    larl %r1, .LCPI92_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI92_1
+; SZ13-NEXT:    lde %f8, 0(%r1)
+; SZ13-NEXT:    ldr %f2, %f8
+; SZ13-NEXT:    brasl %r14, fminf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI92_2
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    brasl %r14, fminf@PLT
+; SZ13-NEXT:    larl %r1, .LCPI92_3
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI92_4
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fminf@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0s killed $f0s def $v0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vrepf %v1, %v1, 0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 312(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %min = call <3 x float> @llvm.experimental.constrained.minnum.v3f32(
+                              <3 x float> <float 43.0, float 44.0, float 45.0>,
+                              <3 x float> <float 41.0, float 42.0, float 43.0>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %min
+}
+
+define void @constrained_vector_minnum_v3f64(<3 x double>* %a) {
+; S390X-LABEL: constrained_vector_minnum_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r13, %r15, 104(%r15)
+; S390X-NEXT:    .cfi_offset %r13, -56
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -192
+; S390X-NEXT:    .cfi_def_cfa_offset 352
+; S390X-NEXT:    std %f8, 184(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f11, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    .cfi_offset %f11, -192
+; S390X-NEXT:    lgr %r13, %r2
+; S390X-NEXT:    ld %f8, 16(%r2)
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    larl %r1, .LCPI93_0
+; S390X-NEXT:    ldeb %f9, 0(%r1)
+; S390X-NEXT:    ld %f10, 8(%r2)
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    ldr %f11, %f0
+; S390X-NEXT:    ldr %f0, %f10
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    std %f0, 16(%r13)
+; S390X-NEXT:    std %f10, 8(%r13)
+; S390X-NEXT:    std %f11, 0(%r13)
+; S390X-NEXT:    ld %f8, 184(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f11, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r13, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_minnum_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r13, %r15, 104(%r15)
+; SZ13-NEXT:    .cfi_offset %r13, -56
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -208
+; SZ13-NEXT:    .cfi_def_cfa_offset 368
+; SZ13-NEXT:    std %f8, 200(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    std %f9, 192(%r15) # 8-byte Folded Spill
+; SZ13-NEXT:    .cfi_offset %f8, -168
+; SZ13-NEXT:    .cfi_offset %f9, -176
+; SZ13-NEXT:    larl %r1, .LCPI93_0
+; SZ13-NEXT:    ldeb %f9, 0(%r1)
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    ld %f8, 16(%r2)
+; SZ13-NEXT:    ldr %f2, %f9
+; SZ13-NEXT:    lgr %r13, %r2
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    vl %v0, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ldr %f2, %f9
+; SZ13-NEXT:    vrepg %v0, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v1, %v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldr %f0, %f8
+; SZ13-NEXT:    ldr %f2, %f9
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    std %f0, 16(%r13)
+; SZ13-NEXT:    vl %v0, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    ld %f8, 200(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    ld %f9, 192(%r15) # 8-byte Folded Reload
+; SZ13-NEXT:    vst %v0, 0(%r13)
+; SZ13-NEXT:    lmg %r13, %r15, 312(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+ %b = load <3 x double>, <3 x double>* %a
+ %min = call <3 x double> @llvm.experimental.constrained.minnum.v3f64(
+                          <3 x double> %b,
+                          <3 x double> <double 3.0, double 3.0, double 3.0>,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  store <3 x double> %min, <3 x double>* %a
+  ret void
+}
+
+define <4 x double> @constrained_vector_minnum_v4f64() {
+; S390X-LABEL: constrained_vector_minnum_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -184
+; S390X-NEXT:    .cfi_def_cfa_offset 344
+; S390X-NEXT:    std %f8, 176(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f10, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    .cfi_offset %f10, -184
+; S390X-NEXT:    larl %r1, .LCPI94_0
+; S390X-NEXT:    ldeb %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI94_1
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    larl %r1, .LCPI94_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI94_3
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    larl %r1, .LCPI94_4
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI94_5
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    larl %r1, .LCPI94_6
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI94_7
+; S390X-NEXT:    ldeb %f2, 0(%r1)
+; S390X-NEXT:    ldr %f10, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, fmin@PLT
+; S390X-NEXT:    ldr %f6, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ldr %f4, %f10
+; S390X-NEXT:    ld %f8, 176(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f10, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 296(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_minnum_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    stmg %r14, %r15, 112(%r15)
+; SZ13-NEXT:    .cfi_offset %r14, -48
+; SZ13-NEXT:    .cfi_offset %r15, -40
+; SZ13-NEXT:    aghi %r15, -192
+; SZ13-NEXT:    .cfi_def_cfa_offset 352
+; SZ13-NEXT:    larl %r1, .LCPI94_0
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI94_1
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    larl %r1, .LCPI94_2
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI94_3
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    vl %v1, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v0, %v0, %v1
+; SZ13-NEXT:    larl %r1, .LCPI94_4
+; SZ13-NEXT:    vst %v0, 176(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI94_5
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    larl %r1, .LCPI94_6
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vst %v0, 160(%r15) # 16-byte Folded Spill
+; SZ13-NEXT:    ldeb %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI94_7
+; SZ13-NEXT:    ldeb %f2, 0(%r1)
+; SZ13-NEXT:    brasl %r14, fmin@PLT
+; SZ13-NEXT:    vl %v1, 160(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    vl %v24, 176(%r15) # 16-byte Folded Reload
+; SZ13-NEXT:    # kill: def $f0d killed $f0d def $v0
+; SZ13-NEXT:    vmrhg %v26, %v0, %v1
+; SZ13-NEXT:    lmg %r14, %r15, 304(%r15)
+; SZ13-NEXT:    br %r14
+entry:
+  %min = call <4 x double> @llvm.experimental.constrained.minnum.v4f64(
+                                <4 x double> <double 44.0, double 45.0,
+                                              double 46.0, double 47.0>,
+                                <4 x double> <double 40.0, double 41.0,
+                                              double 42.0, double 43.0>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <4 x double> %min
+}
+
+define <1 x float> @constrained_vector_fptrunc_v1f64() {
+; S390X-LABEL: constrained_vector_fptrunc_v1f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI95_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    ledbr %f0, %f0
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fptrunc_v1f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI95_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    wledb %v24, %f0, 0, 0
+; SZ13-NEXT:    br %r14
+entry:
+  %result = call <1 x float> @llvm.experimental.constrained.fptrunc.v1f32.v1f64(
+                                <1 x double><double 42.1>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <1 x float> %result
+}
+
+define <2 x float> @constrained_vector_fptrunc_v2f64() {
+; S390X-LABEL: constrained_vector_fptrunc_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI96_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI96_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ledbr %f0, %f0
+; S390X-NEXT:    ledbr %f2, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fptrunc_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI96_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI96_1
+; SZ13-NEXT:    ld %f1, 0(%r1)
+; SZ13-NEXT:    ledbra %f0, 0, %f0, 0
+; SZ13-NEXT:    ledbra %f1, 0, %f1, 0
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    vmrhg %v24, %v0, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %result = call <2 x float> @llvm.experimental.constrained.fptrunc.v2f32.v2f64(
+                                <2 x double><double 42.1, double 42.2>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <2 x float> %result
+}
+
+define void @constrained_vector_fptrunc_v3f64(<3 x double>* %src, <3 x float>* %dest) {
+; S390X-LABEL: constrained_vector_fptrunc_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    ld %f0, 0(%r2)
+; S390X-NEXT:    ld %f1, 16(%r2)
+; S390X-NEXT:    ld %f2, 8(%r2)
+; S390X-NEXT:    ledbr %f0, %f0
+; S390X-NEXT:    lgdr %r0, %f0
+; S390X-NEXT:    nilf %r0, 0
+; S390X-NEXT:    ledbr %f0, %f2
+; S390X-NEXT:    lgdr %r1, %f0
+; S390X-NEXT:    srlg %r1, %r1, 32
+; S390X-NEXT:    lr %r0, %r1
+; S390X-NEXT:    ledbr %f0, %f1
+; S390X-NEXT:    ste %f0, 8(%r3)
+; S390X-NEXT:    stg %r0, 0(%r3)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fptrunc_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vl %v1, 0(%r2)
+; SZ13-NEXT:    ledbra %f2, 0, %f1, 0
+; SZ13-NEXT:    vrepg %v1, %v1, 1
+; SZ13-NEXT:    ld %f0, 16(%r2)
+; SZ13-NEXT:    ledbra %f1, 0, %f1, 0
+; SZ13-NEXT:    ledbra %f0, 0, %f0, 0
+; SZ13-NEXT:    vmrhf %v1, %v2, %v1
+; SZ13-NEXT:    vmrhg %v1, %v1, %v1
+; SZ13-NEXT:    ste %f0, 8(%r3)
+; SZ13-NEXT:    vsteg %v1, 0(%r3), 0
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x double>, <3 x double>* %src
+  %result = call <3 x float> @llvm.experimental.constrained.fptrunc.v3f32.v3f64(
+                                <3 x double> %b,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  store <3 x float> %result, <3 x float>* %dest
+  ret void
+}
+
+define <4 x float> @constrained_vector_fptrunc_v4f64() {
+; S390X-LABEL: constrained_vector_fptrunc_v4f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI98_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI98_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI98_2
+; S390X-NEXT:    ld %f3, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI98_3
+; S390X-NEXT:    ld %f5, 0(%r1)
+; S390X-NEXT:    ledbr %f0, %f0
+; S390X-NEXT:    ledbr %f2, %f1
+; S390X-NEXT:    ledbr %f4, %f3
+; S390X-NEXT:    ledbr %f6, %f5
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fptrunc_v4f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI98_0
+; SZ13-NEXT:    ld %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI98_1
+; SZ13-NEXT:    ld %f1, 0(%r1)
+; SZ13-NEXT:    ledbra %f0, 0, %f0, 0
+; SZ13-NEXT:    ledbra %f1, 0, %f1, 0
+; SZ13-NEXT:    larl %r1, .LCPI98_2
+; SZ13-NEXT:    vmrhf %v0, %v1, %v0
+; SZ13-NEXT:    ld %f1, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI98_3
+; SZ13-NEXT:    ld %f2, 0(%r1)
+; SZ13-NEXT:    ledbra %f1, 0, %f1, 0
+; SZ13-NEXT:    ledbra %f2, 0, %f2, 0
+; SZ13-NEXT:    vmrhf %v1, %v2, %v1
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %result = call <4 x float> @llvm.experimental.constrained.fptrunc.v4f32.v4f64(
+                                <4 x double><double 42.1, double 42.2,
+                                             double 42.3, double 42.4>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <4 x float> %result
+}
+
+define <1 x double> @constrained_vector_fpext_v1f32() {
+; S390X-LABEL: constrained_vector_fpext_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI99_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    ldebr %f0, %f0
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fpext_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI99_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    wldeb %v24, %f0
+; SZ13-NEXT:    br %r14
+entry:
+  %result = call <1 x double> @llvm.experimental.constrained.fpext.v1f64.v1f32(
+                                <1 x float><float 42.0>,
+                                metadata !"fpexcept.strict")
+  ret <1 x double> %result
+}
+
+define <2 x double> @constrained_vector_fpext_v2f32() {
+; S390X-LABEL: constrained_vector_fpext_v2f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI100_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI100_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ldebr %f0, %f0
+; S390X-NEXT:    ldebr %f2, %f1
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fpext_v2f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI100_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI100_1
+; SZ13-NEXT:    lde %f1, 0(%r1)
+; SZ13-NEXT:    ldebr %f0, %f0
+; SZ13-NEXT:    ldebr %f1, %f1
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %result = call <2 x double> @llvm.experimental.constrained.fpext.v2f64.v2f32(
+                                <2 x float><float 42.0, float 43.0>,
+                                metadata !"fpexcept.strict")
+  ret <2 x double> %result
+}
+
+define void @constrained_vector_fpext_v3f64(<3 x float>* %src, <3 x double>* %dest) {
+; S390X-LABEL: constrained_vector_fpext_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    lg %r0, 0(%r2)
+; S390X-NEXT:    le %f0, 8(%r2)
+; S390X-NEXT:    sllg %r1, %r0, 32
+; S390X-NEXT:    ldgr %f1, %r1
+; S390X-NEXT:    nilf %r0, 0
+; S390X-NEXT:    ldgr %f2, %r0
+; S390X-NEXT:    ldebr %f2, %f2
+; S390X-NEXT:    ldebr %f1, %f1
+; S390X-NEXT:    ldebr %f0, %f0
+; S390X-NEXT:    std %f0, 16(%r3)
+; S390X-NEXT:    std %f1, 8(%r3)
+; S390X-NEXT:    std %f2, 0(%r3)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fpext_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vl %v0, 0(%r2)
+; SZ13-NEXT:    vrepf %v2, %v0, 1
+; SZ13-NEXT:    ldebr %f1, %f0
+; SZ13-NEXT:    ldebr %f2, %f2
+; SZ13-NEXT:    vrepf %v0, %v0, 2
+; SZ13-NEXT:    ldebr %f0, %f0
+; SZ13-NEXT:    vmrhg %v1, %v1, %v2
+; SZ13-NEXT:    std %f0, 16(%r3)
+; SZ13-NEXT:    vst %v1, 0(%r3)
+; SZ13-NEXT:    br %r14
+entry:
+  %b = load <3 x float>, <3 x float>* %src
+  %result = call <3 x double> @llvm.experimental.constrained.fpext.v3f64.v3f32(
+                              <3 x float> %b,
+                              metadata !"fpexcept.strict")
+  store <3 x double> %result, <3 x double>* %dest
+  ret void
+}
+
+define <4 x double> @constrained_vector_fpext_v4f32() {
+; S390X-LABEL: constrained_vector_fpext_v4f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    larl %r1, .LCPI102_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI102_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI102_2
+; S390X-NEXT:    le %f3, 0(%r1)
+; S390X-NEXT:    larl %r1, .LCPI102_3
+; S390X-NEXT:    le %f5, 0(%r1)
+; S390X-NEXT:    ldebr %f0, %f0
+; S390X-NEXT:    ldebr %f2, %f1
+; S390X-NEXT:    ldebr %f4, %f3
+; S390X-NEXT:    ldebr %f6, %f5
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_fpext_v4f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI102_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI102_1
+; SZ13-NEXT:    lde %f1, 0(%r1)
+; SZ13-NEXT:    ldebr %f0, %f0
+; SZ13-NEXT:    ldebr %f1, %f1
+; SZ13-NEXT:    larl %r1, .LCPI102_2
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI102_3
+; SZ13-NEXT:    lde %f1, 0(%r1)
+; SZ13-NEXT:    ldebr %f0, %f0
+; SZ13-NEXT:    ldebr %f1, %f1
+; SZ13-NEXT:    vmrhg %v26, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %result = call <4 x double> @llvm.experimental.constrained.fpext.v4f64.v4f32(
+                                <4 x float><float 42.0, float 43.0,
+                                            float 44.0, float 45.0>,
+                                metadata !"fpexcept.strict")
+  ret <4 x double> %result
+}
+
+define <1 x float> @constrained_vector_ceil_v1f32() {
+; S390X-LABEL: constrained_vector_ceil_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI103_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, ceilf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_ceil_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmf %v0, 2, 9
+; SZ13-NEXT:    fiebra %f0, 6, %f0, 4
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %ceil = call <1 x float> @llvm.experimental.constrained.ceil.v1f32(
+                               <1 x float> <float 1.5>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <1 x float> %ceil
+}
+
+define <2 x double> @constrained_vector_ceil_v2f64() {
+; S390X-LABEL: constrained_vector_ceil_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI104_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, ceil@PLT
+; S390X-NEXT:    larl %r1, .LCPI104_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, ceil@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_ceil_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI104_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v24, %v0, 4, 6
+; SZ13-NEXT:    br %r14
+entry:
+  %ceil = call <2 x double> @llvm.experimental.constrained.ceil.v2f64(
+                                <2 x double> <double 1.1, double 1.9>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <2 x double> %ceil
+}
+
+define <3 x float> @constrained_vector_ceil_v3f32() {
+; S390X-LABEL: constrained_vector_ceil_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI105_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, ceilf@PLT
+; S390X-NEXT:    larl %r1, .LCPI105_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, ceilf@PLT
+; S390X-NEXT:    larl %r1, .LCPI105_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, ceilf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_ceil_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI105_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI105_1
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    vgmf %v1, 2, 9
+; SZ13-NEXT:    fiebra %f0, 6, %f0, 4
+; SZ13-NEXT:    fiebra %f1, 6, %f1, 4
+; SZ13-NEXT:    fiebra %f2, 6, %f2, 4
+; SZ13-NEXT:    vmrhf %v1, %v1, %v2
+; SZ13-NEXT:    vrepf %v0, %v0, 0
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %ceil = call <3 x float> @llvm.experimental.constrained.ceil.v3f32(
+                              <3 x float> <float 1.5, float 2.5, float 3.5>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %ceil
+}
+
+define <3 x double> @constrained_vector_ceil_v3f64() {
+; S390X-LABEL: constrained_vector_ceil_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI106_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, ceil@PLT
+; S390X-NEXT:    larl %r1, .LCPI106_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, ceil@PLT
+; S390X-NEXT:    larl %r1, .LCPI106_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, ceil@PLT
+; S390X-NEXT:    ldr %f4, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_ceil_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmg %v0, 2, 12
+; SZ13-NEXT:    larl %r1, .LCPI106_0
+; SZ13-NEXT:    fidbra %f4, 6, %f0, 4
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v0, %v0, 4, 6
+; SZ13-NEXT:    vrepg %v2, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    # kill: def $f2d killed $f2d killed $v2
+; SZ13-NEXT:    br %r14
+entry:
+  %ceil = call <3 x double> @llvm.experimental.constrained.ceil.v3f64(
+                          <3 x double> <double 1.1, double 1.9, double 1.5>,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  ret <3 x double> %ceil
+}
+
+define <1 x float> @constrained_vector_floor_v1f32() {
+; S390X-LABEL: constrained_vector_floor_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI107_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, floorf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_floor_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmf %v0, 2, 9
+; SZ13-NEXT:    fiebra %f0, 7, %f0, 4
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %floor = call <1 x float> @llvm.experimental.constrained.floor.v1f32(
+                               <1 x float> <float 1.5>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <1 x float> %floor
+}
+
+
+define <2 x double> @constrained_vector_floor_v2f64() {
+; S390X-LABEL: constrained_vector_floor_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI108_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, floor@PLT
+; S390X-NEXT:    larl %r1, .LCPI108_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, floor@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_floor_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI108_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v24, %v0, 4, 7
+; SZ13-NEXT:    br %r14
+entry:
+  %floor = call <2 x double> @llvm.experimental.constrained.floor.v2f64(
+                                <2 x double> <double 1.1, double 1.9>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <2 x double> %floor
+}
+
+define <3 x float> @constrained_vector_floor_v3f32() {
+; S390X-LABEL: constrained_vector_floor_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI109_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, floorf@PLT
+; S390X-NEXT:    larl %r1, .LCPI109_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, floorf@PLT
+; S390X-NEXT:    larl %r1, .LCPI109_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, floorf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_floor_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI109_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI109_1
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    vgmf %v1, 2, 9
+; SZ13-NEXT:    fiebra %f0, 7, %f0, 4
+; SZ13-NEXT:    fiebra %f1, 7, %f1, 4
+; SZ13-NEXT:    fiebra %f2, 7, %f2, 4
+; SZ13-NEXT:    vmrhf %v1, %v1, %v2
+; SZ13-NEXT:    vrepf %v0, %v0, 0
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %floor = call <3 x float> @llvm.experimental.constrained.floor.v3f32(
+                              <3 x float> <float 1.5, float 2.5, float 3.5>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %floor
+}
+
+define <3 x double> @constrained_vector_floor_v3f64() {
+; S390X-LABEL: constrained_vector_floor_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI110_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, floor@PLT
+; S390X-NEXT:    larl %r1, .LCPI110_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, floor@PLT
+; S390X-NEXT:    larl %r1, .LCPI110_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, floor@PLT
+; S390X-NEXT:    ldr %f4, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_floor_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmg %v0, 2, 12
+; SZ13-NEXT:    larl %r1, .LCPI110_0
+; SZ13-NEXT:    fidbra %f4, 7, %f0, 4
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v0, %v0, 4, 7
+; SZ13-NEXT:    vrepg %v2, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    # kill: def $f2d killed $f2d killed $v2
+; SZ13-NEXT:    br %r14
+entry:
+  %floor = call <3 x double> @llvm.experimental.constrained.floor.v3f64(
+                          <3 x double> <double 1.1, double 1.9, double 1.5>,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  ret <3 x double> %floor
+}
+
+define <1 x float> @constrained_vector_round_v1f32() {
+; S390X-LABEL: constrained_vector_round_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI111_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, roundf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_round_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmf %v0, 2, 9
+; SZ13-NEXT:    fiebra %f0, 1, %f0, 4
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %round = call <1 x float> @llvm.experimental.constrained.round.v1f32(
+                               <1 x float> <float 1.5>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <1 x float> %round
+}
+
+define <2 x double> @constrained_vector_round_v2f64() {
+; S390X-LABEL: constrained_vector_round_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI112_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, round@PLT
+; S390X-NEXT:    larl %r1, .LCPI112_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, round@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_round_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI112_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v24, %v0, 4, 1
+; SZ13-NEXT:    br %r14
+entry:
+  %round = call <2 x double> @llvm.experimental.constrained.round.v2f64(
+                                <2 x double> <double 1.1, double 1.9>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <2 x double> %round
+}
+
+define <3 x float> @constrained_vector_round_v3f32() {
+; S390X-LABEL: constrained_vector_round_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI113_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, roundf@PLT
+; S390X-NEXT:    larl %r1, .LCPI113_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, roundf@PLT
+; S390X-NEXT:    larl %r1, .LCPI113_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, roundf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_round_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI113_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI113_1
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    vgmf %v1, 2, 9
+; SZ13-NEXT:    fiebra %f0, 1, %f0, 4
+; SZ13-NEXT:    fiebra %f1, 1, %f1, 4
+; SZ13-NEXT:    fiebra %f2, 1, %f2, 4
+; SZ13-NEXT:    vmrhf %v1, %v1, %v2
+; SZ13-NEXT:    vrepf %v0, %v0, 0
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %round = call <3 x float> @llvm.experimental.constrained.round.v3f32(
+                              <3 x float> <float 1.5, float 2.5, float 3.5>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %round
+}
+
+
+define <3 x double> @constrained_vector_round_v3f64() {
+; S390X-LABEL: constrained_vector_round_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI114_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, round@PLT
+; S390X-NEXT:    larl %r1, .LCPI114_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, round@PLT
+; S390X-NEXT:    larl %r1, .LCPI114_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, round@PLT
+; S390X-NEXT:    ldr %f4, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_round_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmg %v0, 2, 12
+; SZ13-NEXT:    larl %r1, .LCPI114_0
+; SZ13-NEXT:    fidbra %f4, 1, %f0, 4
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v0, %v0, 4, 1
+; SZ13-NEXT:    vrepg %v2, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    # kill: def $f2d killed $f2d killed $v2
+; SZ13-NEXT:    br %r14
+entry:
+  %round = call <3 x double> @llvm.experimental.constrained.round.v3f64(
+                          <3 x double> <double 1.1, double 1.9, double 1.5>,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  ret <3 x double> %round
+}
+
+define <1 x float> @constrained_vector_trunc_v1f32() {
+; S390X-LABEL: constrained_vector_trunc_v1f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -160
+; S390X-NEXT:    .cfi_def_cfa_offset 320
+; S390X-NEXT:    larl %r1, .LCPI115_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, truncf@PLT
+; S390X-NEXT:    lmg %r14, %r15, 272(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_trunc_v1f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmf %v0, 2, 9
+; SZ13-NEXT:    fiebra %f0, 5, %f0, 4
+; SZ13-NEXT:    vlr %v24, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %trunc = call <1 x float> @llvm.experimental.constrained.trunc.v1f32(
+                               <1 x float> <float 1.5>,
+                               metadata !"round.dynamic",
+                               metadata !"fpexcept.strict")
+  ret <1 x float> %trunc
+}
+
+define <2 x double> @constrained_vector_trunc_v2f64() {
+; S390X-LABEL: constrained_vector_trunc_v2f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -168
+; S390X-NEXT:    .cfi_def_cfa_offset 328
+; S390X-NEXT:    std %f8, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    larl %r1, .LCPI116_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, trunc@PLT
+; S390X-NEXT:    larl %r1, .LCPI116_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, trunc@PLT
+; S390X-NEXT:    ldr %f2, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ld %f8, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 280(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_trunc_v2f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI116_0
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v24, %v0, 4, 5
+; SZ13-NEXT:    br %r14
+entry:
+  %trunc = call <2 x double> @llvm.experimental.constrained.trunc.v2f64(
+                                <2 x double> <double 1.1, double 1.9>,
+                                metadata !"round.dynamic",
+                                metadata !"fpexcept.strict")
+  ret <2 x double> %trunc
+}
+
+define <3 x float> @constrained_vector_trunc_v3f32() {
+; S390X-LABEL: constrained_vector_trunc_v3f32:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI117_0
+; S390X-NEXT:    le %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, truncf@PLT
+; S390X-NEXT:    larl %r1, .LCPI117_1
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f8, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, truncf@PLT
+; S390X-NEXT:    larl %r1, .LCPI117_2
+; S390X-NEXT:    le %f1, 0(%r1)
+; S390X-NEXT:    ler %f9, %f0
+; S390X-NEXT:    ler %f0, %f1
+; S390X-NEXT:    brasl %r14, truncf@PLT
+; S390X-NEXT:    ler %f4, %f0
+; S390X-NEXT:    ler %f0, %f8
+; S390X-NEXT:    ler %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_trunc_v3f32:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    larl %r1, .LCPI117_0
+; SZ13-NEXT:    lde %f0, 0(%r1)
+; SZ13-NEXT:    larl %r1, .LCPI117_1
+; SZ13-NEXT:    lde %f2, 0(%r1)
+; SZ13-NEXT:    vgmf %v1, 2, 9
+; SZ13-NEXT:    fiebra %f0, 5, %f0, 4
+; SZ13-NEXT:    fiebra %f1, 5, %f1, 4
+; SZ13-NEXT:    fiebra %f2, 5, %f2, 4
+; SZ13-NEXT:    vmrhf %v1, %v1, %v2
+; SZ13-NEXT:    vrepf %v0, %v0, 0
+; SZ13-NEXT:    vmrhg %v24, %v1, %v0
+; SZ13-NEXT:    br %r14
+entry:
+  %trunc = call <3 x float> @llvm.experimental.constrained.trunc.v3f32(
+                              <3 x float> <float 1.5, float 2.5, float 3.5>,
+                              metadata !"round.dynamic",
+                              metadata !"fpexcept.strict")
+  ret <3 x float> %trunc
+}
+
+define <3 x double> @constrained_vector_trunc_v3f64() {
+; S390X-LABEL: constrained_vector_trunc_v3f64:
+; S390X:       # %bb.0: # %entry
+; S390X-NEXT:    stmg %r14, %r15, 112(%r15)
+; S390X-NEXT:    .cfi_offset %r14, -48
+; S390X-NEXT:    .cfi_offset %r15, -40
+; S390X-NEXT:    aghi %r15, -176
+; S390X-NEXT:    .cfi_def_cfa_offset 336
+; S390X-NEXT:    std %f8, 168(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    std %f9, 160(%r15) # 8-byte Folded Spill
+; S390X-NEXT:    .cfi_offset %f8, -168
+; S390X-NEXT:    .cfi_offset %f9, -176
+; S390X-NEXT:    larl %r1, .LCPI118_0
+; S390X-NEXT:    ld %f0, 0(%r1)
+; S390X-NEXT:    brasl %r14, trunc@PLT
+; S390X-NEXT:    larl %r1, .LCPI118_1
+; S390X-NEXT:    ld %f1, 0(%r1)
+; S390X-NEXT:    ldr %f8, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, trunc@PLT
+; S390X-NEXT:    larl %r1, .LCPI118_2
+; S390X-NEXT:    ldeb %f1, 0(%r1)
+; S390X-NEXT:    ldr %f9, %f0
+; S390X-NEXT:    ldr %f0, %f1
+; S390X-NEXT:    brasl %r14, trunc@PLT
+; S390X-NEXT:    ldr %f4, %f0
+; S390X-NEXT:    ldr %f0, %f8
+; S390X-NEXT:    ldr %f2, %f9
+; S390X-NEXT:    ld %f8, 168(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    ld %f9, 160(%r15) # 8-byte Folded Reload
+; S390X-NEXT:    lmg %r14, %r15, 288(%r15)
+; S390X-NEXT:    br %r14
+;
+; SZ13-LABEL: constrained_vector_trunc_v3f64:
+; SZ13:       # %bb.0: # %entry
+; SZ13-NEXT:    vgmg %v0, 2, 12
+; SZ13-NEXT:    larl %r1, .LCPI118_0
+; SZ13-NEXT:    fidbra %f4, 5, %f0, 4
+; SZ13-NEXT:    vl %v0, 0(%r1)
+; SZ13-NEXT:    vfidb %v0, %v0, 4, 5
+; SZ13-NEXT:    vrepg %v2, %v0, 1
+; SZ13-NEXT:    # kill: def $f0d killed $f0d killed $v0
+; SZ13-NEXT:    # kill: def $f2d killed $f2d killed $v2
+; SZ13-NEXT:    br %r14
+entry:
+  %trunc = call <3 x double> @llvm.experimental.constrained.trunc.v3f64(
+                          <3 x double> <double 1.1, double 1.9, double 1.5>,
+                          metadata !"round.dynamic",
+                          metadata !"fpexcept.strict")
+  ret <3 x double> %trunc
+}
+
+; Single width declarations
+declare <2 x double> @llvm.experimental.constrained.fadd.v2f64(<2 x double>, <2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.fsub.v2f64(<2 x double>, <2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.fmul.v2f64(<2 x double>, <2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.fdiv.v2f64(<2 x double>, <2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.frem.v2f64(<2 x double>, <2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.sqrt.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.pow.v2f64(<2 x double>, <2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.powi.v2f64(<2 x double>, i32, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.sin.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.cos.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.exp.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.exp2.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.log.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.log10.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.log2.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.rint.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.nearbyint.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.maxnum.v2f64(<2 x double>, <2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.minnum.v2f64(<2 x double>, <2 x double>, metadata, metadata)
+declare <2 x float> @llvm.experimental.constrained.fptrunc.v2f32.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.fpext.v2f64.v2f32(<2 x float>, metadata)
+declare <2 x double> @llvm.experimental.constrained.ceil.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.floor.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.round.v2f64(<2 x double>, metadata, metadata)
+declare <2 x double> @llvm.experimental.constrained.trunc.v2f64(<2 x double>, metadata, metadata)
+
+; Scalar width declarations
+declare <1 x float> @llvm.experimental.constrained.fadd.v1f32(<1 x float>, <1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.fsub.v1f32(<1 x float>, <1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.fmul.v1f32(<1 x float>, <1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.fdiv.v1f32(<1 x float>, <1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.frem.v1f32(<1 x float>, <1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.sqrt.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.pow.v1f32(<1 x float>, <1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.powi.v1f32(<1 x float>, i32, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.sin.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.cos.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.exp.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.exp2.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.log.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.log10.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.log2.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.rint.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.nearbyint.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.maxnum.v1f32(<1 x float>, <1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.minnum.v1f32(<1 x float>, <1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.fptrunc.v1f32.v1f64(<1 x double>, metadata, metadata)
+declare <1 x double> @llvm.experimental.constrained.fpext.v1f64.v1f32(<1 x float>, metadata)
+declare <1 x float> @llvm.experimental.constrained.ceil.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.floor.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.round.v1f32(<1 x float>, metadata, metadata)
+declare <1 x float> @llvm.experimental.constrained.trunc.v1f32(<1 x float>, metadata, metadata)
+
+; Illegal width declarations
+declare <3 x float> @llvm.experimental.constrained.fadd.v3f32(<3 x float>, <3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.fadd.v3f64(<3 x double>, <3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.fsub.v3f32(<3 x float>, <3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.fsub.v3f64(<3 x double>, <3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.fmul.v3f32(<3 x float>, <3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.fmul.v3f64(<3 x double>, <3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.fdiv.v3f32(<3 x float>, <3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.fdiv.v3f64(<3 x double>, <3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.frem.v3f32(<3 x float>, <3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.frem.v3f64(<3 x double>, <3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.sqrt.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.sqrt.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.pow.v3f32(<3 x float>, <3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.pow.v3f64(<3 x double>, <3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.powi.v3f32(<3 x float>, i32, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.powi.v3f64(<3 x double>, i32, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.sin.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.sin.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.cos.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.cos.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.exp.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.exp.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.exp2.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.exp2.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.log.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.log.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.log10.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.log10.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.log2.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.log2.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.rint.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.rint.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.nearbyint.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.nearbyint.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.maxnum.v3f32(<3 x float>, <3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.maxnum.v3f64(<3 x double>, <3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.minnum.v3f32(<3 x float>, <3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.minnum.v3f64(<3 x double>, <3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.fptrunc.v3f32.v3f64(<3 x double>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.fpext.v3f64.v3f32(<3 x float>, metadata)
+declare <3 x float> @llvm.experimental.constrained.ceil.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.ceil.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.floor.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.floor.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.round.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.round.v3f64(<3 x double>, metadata, metadata)
+declare <3 x float> @llvm.experimental.constrained.trunc.v3f32(<3 x float>, metadata, metadata)
+declare <3 x double> @llvm.experimental.constrained.trunc.v3f64(<3 x double>, metadata, metadata)
+
+; Double width declarations
+declare <4 x double> @llvm.experimental.constrained.fadd.v4f64(<4 x double>, <4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.fsub.v4f64(<4 x double>, <4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.fmul.v4f64(<4 x double>, <4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.fdiv.v4f64(<4 x double>, <4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.frem.v4f64(<4 x double>, <4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.sqrt.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.pow.v4f64(<4 x double>, <4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.powi.v4f64(<4 x double>, i32, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.sin.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.cos.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.exp.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.exp2.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.log.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.log10.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.log2.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.rint.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.nearbyint.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.maxnum.v4f64(<4 x double>, <4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.minnum.v4f64(<4 x double>, <4 x double>, metadata, metadata)
+declare <4 x float> @llvm.experimental.constrained.fptrunc.v4f32.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.fpext.v4f64.v4f32(<4 x float>, metadata)
+declare <4 x double> @llvm.experimental.constrained.ceil.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.floor.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.round.v4f64(<4 x double>, metadata, metadata)
+declare <4 x double> @llvm.experimental.constrained.trunc.v4f64(<4 x double>, metadata, metadata)