[VP][RISCV] Add vp.copysign and RISC-V support.

[fakepaper56]

ISD::FCOPYSIGN could have different types of magnitude operand and sign operand. Should ISD::VP_FOPYSIGN have same property? My implements does not loose the constraint like other vp intrinsics.

[craig.topper]

In D134935#3825988, @fakepaper56 wrote:

ISD::FCOPYSIGN could have different types of magnitude operand and sign operand. Should ISD::VP_FOPYSIGN have same property? My implements does not loose the constraint like other vp intrinsics.

I don't think we should support differing types for ISD::VP_COPYSIGN. It was disabled for vector ISD::FCOPYSIGN in February 2021 D96037 Though it look likes a command line flag was added to DAGCombiner.cpp in August 2022 to allow it to be re-enabled. Using that flag would very likely cause isel failures on RISC-V.

[craig.topper]

LGTM

llvm/docs/LangRef.rst

 
 Semantics:
 """"""""""
 If either operand is a NaN, returns NaN. Otherwise returns the greater
 of the two arguments. -0.0 is considered to be less than +0.0 for this
 intrinsic. Note that these are the semantics specified in the draft of
 IEEE 754-2018.
 
+.. _int_copysign:
+
 '``llvm.copysign.*``' Intrinsic
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 
 This is an overloaded intrinsic. You can use ``llvm.copysign`` on any
 floating-point or vector of floating-point type. Not all targets support
.. code-block:: llvm
 
       %r = call <4 x i32> @llvm.vp.xor.v4i32(<4 x i32> %a, <4 x i32> %b, <4 x i1> %mask, i32 %evl)
       ;; For all lanes below %evl, %r is lane-wise equivalent to %also.r
 
       %t = xor <4 x i32> %a, %b
       %also.r = select <4 x i1> %mask, <4 x i32> %t, <4 x i32> poison
 
 
+.. _int_vp_copysign:
+
+'``llvm.vp.copysign.*``' Intrinsics
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+This is an overloaded intrinsic.
+
+::
+
+      declare <16 x float>  @llvm.vp.copysign.v16f32 (<16 x float> <mag_op>, <16 x float> <sign_op>, <16 x i1> <mask>, i32 <vector_length>)
+      declare <vscale x 4 x float>  @llvm.vp.copysign.nxv4f32 (<vscale x 4 x float> <mag_op>, <vscale x 4 x float> <sign_op>, <vscale x 4 x i1> <mask>, i32 <vector_length>)
+      declare <256 x double>  @llvm.vp.copysign.v256f64 (<256 x double> <mag_op>, <256 x double> <sign_op>, <256 x i1> <mask>, i32 <vector_length>)
+
+Overview:
+"""""""""
+
+Predicated floating-point copysign of two vectors of floating-point values.
+
+
+Arguments:
+""""""""""
+
+The first two operands and the result have the same vector of floating-point type. The
+third operand is the vector mask and has the same number of elements as the
+result vector type. The fourth operand is the explicit vector length of the
+operation.
+
+Semantics:
+""""""""""
+
+The '``llvm.vp.copysign``' intrinsic performs floating-point copysign (:ref:`copysign <int_copysign>`)
+of the first and second vector operand on each enabled lane.  The result on
+disabled lanes is a :ref:`poison value <poisonvalues>`.  The operation is
+performed in the default floating-point environment.
+
+Examples:
+"""""""""
+
+.. code-block:: llvm
+
+      %r = call <4 x float> @llvm.vp.copysign.v4f32(<4 x float> %mag, <4 x float> %sign, <4 x i1> %mask, i32 %evl)
+      ;; For all lanes below %evl, %r is lane-wise equivalent to %also.r
+
+      %t = call <4 x float> @llvm.copysign.v4f32(<4 x float> %mag, <4 x float> %sign)
+      %also.r = select <4 x i1> %mask, <4 x float> %t, <4 x float> poison
+
+
 .. _int_vp_minnum:
 
 '``llvm.vp.minnum.*``' Intrinsics
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Syntax:
 """""""
 This is an overloaded intrinsic.

llvm/include/llvm/IR/Intrinsics.td

def int_vp_minnum : DefaultAttrsIntrinsic<[ llvm_anyvector_ty ],
                                LLVMMatchType<0>,
                                LLVMScalarOrSameVectorWidth<0, llvm_i1_ty>,
                                llvm_i32_ty]>;
   def int_vp_maxnum : DefaultAttrsIntrinsic<[ llvm_anyvector_ty ],
                              [ LLVMMatchType<0>,
                                LLVMMatchType<0>,
                                LLVMScalarOrSameVectorWidth<0, llvm_i1_ty>,
                                llvm_i32_ty]>;
+  def int_vp_copysign : DefaultAttrsIntrinsic<[ llvm_anyvector_ty ],
+                             [ LLVMMatchType<0>,
+                               LLVMMatchType<0>,
+                               LLVMScalarOrSameVectorWidth<0, llvm_i1_ty>,
+                               llvm_i32_ty]>;
   def int_vp_ceil : DefaultAttrsIntrinsic<[ llvm_anyvector_ty ],
                              [ LLVMMatchType<0>,
                                LLVMScalarOrSameVectorWidth<0, llvm_i1_ty>,
                                llvm_i32_ty]>;
   def int_vp_floor : DefaultAttrsIntrinsic<[ llvm_anyvector_ty ],
                              [ LLVMMatchType<0>,
                                LLVMScalarOrSameVectorWidth<0, llvm_i1_ty>,
                                llvm_i32_ty]>;

llvm/include/llvm/IR/VPIntrinsics.def

 VP_PROPERTY_CONSTRAINEDFP(1, 1, experimental_constrained_fma)
 END_REGISTER_VP(vp_fma, VP_FMA)
 
 // llvm.vp.fmuladd(x,y,z,mask,vlen)
 BEGIN_REGISTER_VP(vp_fmuladd, 3, 4, VP_FMULADD, -1)
 VP_PROPERTY_CONSTRAINEDFP(1, 1, experimental_constrained_fmuladd)
 END_REGISTER_VP(vp_fmuladd, VP_FMULADD)
 
+// llvm.vp.copysign(x,y,mask,vlen)
+BEGIN_REGISTER_VP(vp_copysign, 2, 3, VP_FCOPYSIGN, -1)
+END_REGISTER_VP(vp_copysign, VP_FCOPYSIGN)
+
 // llvm.vp.minnum(x, y, mask,vlen)
 BEGIN_REGISTER_VP(vp_minnum, 2, 3, VP_FMINNUM, -1)
 END_REGISTER_VP(vp_minnum, VP_FMINNUM)
 
 // llvm.vp.maxnum(x, y, mask,vlen)
 BEGIN_REGISTER_VP(vp_maxnum, 2, 3, VP_FMAXNUM, -1)
 END_REGISTER_VP(vp_maxnum, VP_FMAXNUM)
 

llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp

#endif
   case ISD::SADDSAT:
   case ISD::UADDSAT:
   case ISD::SSUBSAT:
   case ISD::USUBSAT:
   case ISD::SSHLSAT:
   case ISD::USHLSAT:
   case ISD::ROTL:
   case ISD::ROTR:
+  case ISD::VP_FCOPYSIGN:
     SplitVecRes_BinOp(N, Lo, Hi);
     break;
   case ISD::FMA: case ISD::VP_FMA:
   case ISD::FSHL:
   case ISD::FSHR:
     SplitVecRes_TernaryOp(N, Lo, Hi);
     break;
 
#endif
   case ISD::VP_UDIV:
   case ISD::VP_SREM:
   case ISD::VP_UREM:
   case ISD::VP_FADD:
   case ISD::VP_FSUB:
   case ISD::VP_FMUL:
   case ISD::VP_FDIV:
   case ISD::VP_FREM:
+  case ISD::VP_FCOPYSIGN:
     Res = WidenVecRes_Binary(N);
     break;
 
   case ISD::FPOW:
   case ISD::FREM:
     if (unrollExpandedOp())
       break;
     // If the target has custom/legal support for the scalar FP intrinsic ops

llvm/lib/Target/RISCV/RISCVISelLowering.cpp

static const unsigned FloatingPointVPOps[] = {
         ISD::VP_FADD,        ISD::VP_FSUB,        ISD::VP_FMUL,
         ISD::VP_FDIV,        ISD::VP_FNEG,        ISD::VP_FABS,
         ISD::VP_FMA,         ISD::VP_REDUCE_FADD, ISD::VP_REDUCE_SEQ_FADD,
         ISD::VP_REDUCE_FMIN, ISD::VP_REDUCE_FMAX, ISD::VP_MERGE,
         ISD::VP_SELECT,      ISD::VP_SINT_TO_FP,  ISD::VP_UINT_TO_FP,
         ISD::VP_SETCC,       ISD::VP_FP_ROUND,    ISD::VP_FP_EXTEND,
         ISD::VP_SQRT,        ISD::VP_FMINNUM,     ISD::VP_FMAXNUM,
         ISD::VP_FCEIL,       ISD::VP_FFLOOR,      ISD::VP_FROUND,
-        ISD::VP_FROUNDEVEN};
+        ISD::VP_FROUNDEVEN,  ISD::VP_FCOPYSIGN};
 
     static const unsigned IntegerVecReduceOps[] = {
         ISD::VECREDUCE_ADD,  ISD::VECREDUCE_AND,  ISD::VECREDUCE_OR,
         ISD::VECREDUCE_XOR,  ISD::VECREDUCE_SMAX, ISD::VECREDUCE_SMIN,
         ISD::VECREDUCE_UMAX, ISD::VECREDUCE_UMIN};
 
     static const unsigned FloatingPointVecReduceOps[] = {
         ISD::VECREDUCE_FADD, ISD::VECREDUCE_SEQ_FADD, ISD::VECREDUCE_FMIN,
SDValue RISCVTargetLowering::LowerOperation(SDValue Op,
   case ISD::VP_SQRT:
     return lowerVPOp(Op, DAG, RISCVISD::FSQRT_VL);
   case ISD::VP_FMA:
     return lowerVPOp(Op, DAG, RISCVISD::VFMADD_VL);
   case ISD::VP_FMINNUM:
     return lowerVPOp(Op, DAG, RISCVISD::FMINNUM_VL, /*HasMergeOp*/ true);
   case ISD::VP_FMAXNUM:
     return lowerVPOp(Op, DAG, RISCVISD::FMAXNUM_VL, /*HasMergeOp*/ true);
+  case ISD::VP_FCOPYSIGN:
+    return lowerVPOp(Op, DAG, RISCVISD::FCOPYSIGN_VL, /*HasMergeOp*/ true);
   case ISD::VP_SIGN_EXTEND:
   case ISD::VP_ZERO_EXTEND:
     if (Op.getOperand(0).getSimpleValueType().getVectorElementType() == MVT::i1)
       return lowerVPExtMaskOp(Op, DAG);
     return lowerVPOp(Op, DAG,
                      Op.getOpcode() == ISD::VP_SIGN_EXTEND
                          ? RISCVISD::VSEXT_VL
                          : RISCVISD::VZEXT_VL);

llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vcopysign-vp.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv32 -mattr=+m,+d,+zfh,+experimental-zvfh,+v -target-abi=ilp32d -riscv-v-vector-bits-min=128 \
+; RUN:   -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=riscv64 -mattr=+m,+d,+zfh,+experimental-zvfh,+v -target-abi=lp64d -riscv-v-vector-bits-min=128 \
+; RUN:   -verify-machineinstrs < %s | FileCheck %s
+
+declare <2 x half> @llvm.vp.copysign.v2f16(<2 x half>, <2 x half>, <2 x i1>, i32)
+
+define <2 x half> @vfsgnj_vv_v2f16(<2 x half> %va, <2 x half> %vb, <2 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v2f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <2 x half> @llvm.vp.copysign.v2f16(<2 x half> %va, <2 x half> %vb, <2 x i1> %m, i32 %evl)
+  ret <2 x half> %v
+}
+
+define <2 x half> @vfsgnj_vv_v2f16_unmasked(<2 x half> %va, <2 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v2f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <2 x i1> poison, i1 true, i32 0
+  %m = shufflevector <2 x i1> %head, <2 x i1> poison, <2 x i32> zeroinitializer
+  %v = call <2 x half> @llvm.vp.copysign.v2f16(<2 x half> %va, <2 x half> %vb, <2 x i1> %m, i32 %evl)
+  ret <2 x half> %v
+}
+
+declare <4 x half> @llvm.vp.copysign.v4f16(<4 x half>, <4 x half>, <4 x i1>, i32)
+
+define <4 x half> @vfsgnj_vv_v4f16(<4 x half> %va, <4 x half> %vb, <4 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v4f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <4 x half> @llvm.vp.copysign.v4f16(<4 x half> %va, <4 x half> %vb, <4 x i1> %m, i32 %evl)
+  ret <4 x half> %v
+}
+
+define <4 x half> @vfsgnj_vv_v4f16_unmasked(<4 x half> %va, <4 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v4f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <4 x i1> poison, i1 true, i32 0
+  %m = shufflevector <4 x i1> %head, <4 x i1> poison, <4 x i32> zeroinitializer
+  %v = call <4 x half> @llvm.vp.copysign.v4f16(<4 x half> %va, <4 x half> %vb, <4 x i1> %m, i32 %evl)
+  ret <4 x half> %v
+}
+
+declare <8 x half> @llvm.vp.copysign.v8f16(<8 x half>, <8 x half>, <8 x i1>, i32)
+
+define <8 x half> @vfsgnj_vv_v8f16(<8 x half> %va, <8 x half> %vb, <8 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v8f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <8 x half> @llvm.vp.copysign.v8f16(<8 x half> %va, <8 x half> %vb, <8 x i1> %m, i32 %evl)
+  ret <8 x half> %v
+}
+
+define <8 x half> @vfsgnj_vv_v8f16_unmasked(<8 x half> %va, <8 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v8f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <8 x i1> poison, i1 true, i32 0
+  %m = shufflevector <8 x i1> %head, <8 x i1> poison, <8 x i32> zeroinitializer
+  %v = call <8 x half> @llvm.vp.copysign.v8f16(<8 x half> %va, <8 x half> %vb, <8 x i1> %m, i32 %evl)
+  ret <8 x half> %v
+}
+
+declare <16 x half> @llvm.vp.copysign.v16f16(<16 x half>, <16 x half>, <16 x i1>, i32)
+
+define <16 x half> @vfsgnj_vv_v16f16(<16 x half> %va, <16 x half> %vb, <16 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v16f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10, v0.t
+; CHECK-NEXT:    ret
+  %v = call <16 x half> @llvm.vp.copysign.v16f16(<16 x half> %va, <16 x half> %vb, <16 x i1> %m, i32 %evl)
+  ret <16 x half> %v
+}
+
+define <16 x half> @vfsgnj_vv_v16f16_unmasked(<16 x half> %va, <16 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v16f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10
+; CHECK-NEXT:    ret
+  %head = insertelement <16 x i1> poison, i1 true, i32 0
+  %m = shufflevector <16 x i1> %head, <16 x i1> poison, <16 x i32> zeroinitializer
+  %v = call <16 x half> @llvm.vp.copysign.v16f16(<16 x half> %va, <16 x half> %vb, <16 x i1> %m, i32 %evl)
+  ret <16 x half> %v
+}
+
+declare <2 x float> @llvm.vp.copysign.v2f32(<2 x float>, <2 x float>, <2 x i1>, i32)
+
+define <2 x float> @vfsgnj_vv_v2f32(<2 x float> %va, <2 x float> %vb, <2 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v2f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <2 x float> @llvm.vp.copysign.v2f32(<2 x float> %va, <2 x float> %vb, <2 x i1> %m, i32 %evl)
+  ret <2 x float> %v
+}
+
+define <2 x float> @vfsgnj_vv_v2f32_unmasked(<2 x float> %va, <2 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v2f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <2 x i1> poison, i1 true, i32 0
+  %m = shufflevector <2 x i1> %head, <2 x i1> poison, <2 x i32> zeroinitializer
+  %v = call <2 x float> @llvm.vp.copysign.v2f32(<2 x float> %va, <2 x float> %vb, <2 x i1> %m, i32 %evl)
+  ret <2 x float> %v
+}
+
+declare <4 x float> @llvm.vp.copysign.v4f32(<4 x float>, <4 x float>, <4 x i1>, i32)
+
+define <4 x float> @vfsgnj_vv_v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v4f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <4 x float> @llvm.vp.copysign.v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> %m, i32 %evl)
+  ret <4 x float> %v
+}
+
+define <4 x float> @vfsgnj_vv_v4f32_unmasked(<4 x float> %va, <4 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v4f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <4 x i1> poison, i1 true, i32 0
+  %m = shufflevector <4 x i1> %head, <4 x i1> poison, <4 x i32> zeroinitializer
+  %v = call <4 x float> @llvm.vp.copysign.v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> %m, i32 %evl)
+  ret <4 x float> %v
+}
+
+declare <8 x float> @llvm.vp.copysign.v8f32(<8 x float>, <8 x float>, <8 x i1>, i32)
+
+define <8 x float> @vfsgnj_vv_v8f32(<8 x float> %va, <8 x float> %vb, <8 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v8f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10, v0.t
+; CHECK-NEXT:    ret
+  %v = call <8 x float> @llvm.vp.copysign.v8f32(<8 x float> %va, <8 x float> %vb, <8 x i1> %m, i32 %evl)
+  ret <8 x float> %v
+}
+
+define <8 x float> @vfsgnj_vv_v8f32_unmasked(<8 x float> %va, <8 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v8f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10
+; CHECK-NEXT:    ret
+  %head = insertelement <8 x i1> poison, i1 true, i32 0
+  %m = shufflevector <8 x i1> %head, <8 x i1> poison, <8 x i32> zeroinitializer
+  %v = call <8 x float> @llvm.vp.copysign.v8f32(<8 x float> %va, <8 x float> %vb, <8 x i1> %m, i32 %evl)
+  ret <8 x float> %v
+}
+
+declare <16 x float> @llvm.vp.copysign.v16f32(<16 x float>, <16 x float>, <16 x i1>, i32)
+
+define <16 x float> @vfsgnj_vv_v16f32(<16 x float> %va, <16 x float> %vb, <16 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v16f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12, v0.t
+; CHECK-NEXT:    ret
+  %v = call <16 x float> @llvm.vp.copysign.v16f32(<16 x float> %va, <16 x float> %vb, <16 x i1> %m, i32 %evl)
+  ret <16 x float> %v
+}
+
+define <16 x float> @vfsgnj_vv_v16f32_unmasked(<16 x float> %va, <16 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v16f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12
+; CHECK-NEXT:    ret
+  %head = insertelement <16 x i1> poison, i1 true, i32 0
+  %m = shufflevector <16 x i1> %head, <16 x i1> poison, <16 x i32> zeroinitializer
+  %v = call <16 x float> @llvm.vp.copysign.v16f32(<16 x float> %va, <16 x float> %vb, <16 x i1> %m, i32 %evl)
+  ret <16 x float> %v
+}
+
+declare <2 x double> @llvm.vp.copysign.v2f64(<2 x double>, <2 x double>, <2 x i1>, i32)
+
+define <2 x double> @vfsgnj_vv_v2f64(<2 x double> %va, <2 x double> %vb, <2 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v2f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <2 x double> @llvm.vp.copysign.v2f64(<2 x double> %va, <2 x double> %vb, <2 x i1> %m, i32 %evl)
+  ret <2 x double> %v
+}
+
+define <2 x double> @vfsgnj_vv_v2f64_unmasked(<2 x double> %va, <2 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v2f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <2 x i1> poison, i1 true, i32 0
+  %m = shufflevector <2 x i1> %head, <2 x i1> poison, <2 x i32> zeroinitializer
+  %v = call <2 x double> @llvm.vp.copysign.v2f64(<2 x double> %va, <2 x double> %vb, <2 x i1> %m, i32 %evl)
+  ret <2 x double> %v
+}
+
+declare <4 x double> @llvm.vp.copysign.v4f64(<4 x double>, <4 x double>, <4 x i1>, i32)
+
+define <4 x double> @vfsgnj_vv_v4f64(<4 x double> %va, <4 x double> %vb, <4 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v4f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10, v0.t
+; CHECK-NEXT:    ret
+  %v = call <4 x double> @llvm.vp.copysign.v4f64(<4 x double> %va, <4 x double> %vb, <4 x i1> %m, i32 %evl)
+  ret <4 x double> %v
+}
+
+define <4 x double> @vfsgnj_vv_v4f64_unmasked(<4 x double> %va, <4 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v4f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10
+; CHECK-NEXT:    ret
+  %head = insertelement <4 x i1> poison, i1 true, i32 0
+  %m = shufflevector <4 x i1> %head, <4 x i1> poison, <4 x i32> zeroinitializer
+  %v = call <4 x double> @llvm.vp.copysign.v4f64(<4 x double> %va, <4 x double> %vb, <4 x i1> %m, i32 %evl)
+  ret <4 x double> %v
+}
+
+declare <8 x double> @llvm.vp.copysign.v8f64(<8 x double>, <8 x double>, <8 x i1>, i32)
+
+define <8 x double> @vfsgnj_vv_v8f64(<8 x double> %va, <8 x double> %vb, <8 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v8f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12, v0.t
+; CHECK-NEXT:    ret
+  %v = call <8 x double> @llvm.vp.copysign.v8f64(<8 x double> %va, <8 x double> %vb, <8 x i1> %m, i32 %evl)
+  ret <8 x double> %v
+}
+
+define <8 x double> @vfsgnj_vv_v8f64_unmasked(<8 x double> %va, <8 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v8f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12
+; CHECK-NEXT:    ret
+  %head = insertelement <8 x i1> poison, i1 true, i32 0
+  %m = shufflevector <8 x i1> %head, <8 x i1> poison, <8 x i32> zeroinitializer
+  %v = call <8 x double> @llvm.vp.copysign.v8f64(<8 x double> %va, <8 x double> %vb, <8 x i1> %m, i32 %evl)
+  ret <8 x double> %v
+}
+
+declare <15 x double> @llvm.vp.copysign.v15f64(<15 x double>, <15 x double>, <15 x i1>, i32)
+
+define <15 x double> @vfsgnj_vv_v15f64(<15 x double> %va, <15 x double> %vb, <15 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v15f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16, v0.t
+; CHECK-NEXT:    ret
+  %v = call <15 x double> @llvm.vp.copysign.v15f64(<15 x double> %va, <15 x double> %vb, <15 x i1> %m, i32 %evl)
+  ret <15 x double> %v
+}
+
+define <15 x double> @vfsgnj_vv_v15f64_unmasked(<15 x double> %va, <15 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v15f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16
+; CHECK-NEXT:    ret
+  %head = insertelement <15 x i1> poison, i1 true, i32 0
+  %m = shufflevector <15 x i1> %head, <15 x i1> poison, <15 x i32> zeroinitializer
+  %v = call <15 x double> @llvm.vp.copysign.v15f64(<15 x double> %va, <15 x double> %vb, <15 x i1> %m, i32 %evl)
+  ret <15 x double> %v
+}
+
+declare <16 x double> @llvm.vp.copysign.v16f64(<16 x double>, <16 x double>, <16 x i1>, i32)
+
+define <16 x double> @vfsgnj_vv_v16f64(<16 x double> %va, <16 x double> %vb, <16 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v16f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16, v0.t
+; CHECK-NEXT:    ret
+  %v = call <16 x double> @llvm.vp.copysign.v16f64(<16 x double> %va, <16 x double> %vb, <16 x i1> %m, i32 %evl)
+  ret <16 x double> %v
+}
+
+define <16 x double> @vfsgnj_vv_v16f64_unmasked(<16 x double> %va, <16 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v16f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16
+; CHECK-NEXT:    ret
+  %head = insertelement <16 x i1> poison, i1 true, i32 0
+  %m = shufflevector <16 x i1> %head, <16 x i1> poison, <16 x i32> zeroinitializer
+  %v = call <16 x double> @llvm.vp.copysign.v16f64(<16 x double> %va, <16 x double> %vb, <16 x i1> %m, i32 %evl)
+  ret <16 x double> %v
+}
+
+declare <32 x double> @llvm.vp.copysign.v32f64(<32 x double>, <32 x double>, <32 x i1>, i32)
+
+define <32 x double> @vfsgnj_vv_v32f64(<32 x double> %va, <32 x double> %vb, <32 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v32f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    addi sp, sp, -16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    csrr a1, vlenb
+; CHECK-NEXT:    li a3, 24
+; CHECK-NEXT:    mul a1, a1, a3
+; CHECK-NEXT:    sub sp, sp, a1
+; CHECK-NEXT:    vmv1r.v v1, v0
+; CHECK-NEXT:    vsetivli zero, 2, e8, mf4, ta, mu
+; CHECK-NEXT:    vslidedown.vi v0, v0, 2
+; CHECK-NEXT:    addi a1, a0, 128
+; CHECK-NEXT:    vsetivli zero, 16, e64, m8, ta, mu
+; CHECK-NEXT:    vle64.v v24, (a1)
+; CHECK-NEXT:    csrr a1, vlenb
+; CHECK-NEXT:    slli a1, a1, 3
+; CHECK-NEXT:    add a1, sp, a1
+; CHECK-NEXT:    addi a1, a1, 16
+; CHECK-NEXT:    vs8r.v v24, (a1) # Unknown-size Folded Spill
+; CHECK-NEXT:    addi a3, a2, -16
+; CHECK-NEXT:    csrr a1, vlenb
+; CHECK-NEXT:    slli a1, a1, 4
+; CHECK-NEXT:    add a1, sp, a1
+; CHECK-NEXT:    addi a1, a1, 16
+; CHECK-NEXT:    vs8r.v v8, (a1) # Unknown-size Folded Spill
+; CHECK-NEXT:    li a1, 0
+; CHECK-NEXT:    bltu a2, a3, .LBB26_2
+; CHECK-NEXT:  # %bb.1:
+; CHECK-NEXT:    mv a1, a3
+; CHECK-NEXT:  .LBB26_2:
+; CHECK-NEXT:    vle64.v v8, (a0)
+; CHECK-NEXT:    addi a0, sp, 16
+; CHECK-NEXT:    vs8r.v v8, (a0) # Unknown-size Folded Spill
+; CHECK-NEXT:    vsetvli zero, a1, e64, m8, ta, mu
+; CHECK-NEXT:    li a0, 16
+; CHECK-NEXT:    csrr a1, vlenb
+; CHECK-NEXT:    slli a1, a1, 3
+; CHECK-NEXT:    add a1, sp, a1
+; CHECK-NEXT:    addi a1, a1, 16
+; CHECK-NEXT:    vl8re8.v v8, (a1) # Unknown-size Folded Reload
+; CHECK-NEXT:    vfsgnj.vv v16, v16, v8, v0.t
+; CHECK-NEXT:    bltu a2, a0, .LBB26_4
+; CHECK-NEXT:  # %bb.3:
+; CHECK-NEXT:    li a2, 16
+; CHECK-NEXT:  .LBB26_4:
+; CHECK-NEXT:    vsetvli zero, a2, e64, m8, ta, mu
+; CHECK-NEXT:    vmv1r.v v0, v1
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    slli a0, a0, 4
+; CHECK-NEXT:    add a0, sp, a0
+; CHECK-NEXT:    addi a0, a0, 16
+; CHECK-NEXT:    vl8re8.v v24, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    addi a0, sp, 16
+; CHECK-NEXT:    vl8re8.v v8, (a0) # Unknown-size Folded Reload
+; CHECK-NEXT:    vfsgnj.vv v8, v24, v8, v0.t
+; CHECK-NEXT:    csrr a0, vlenb
+; CHECK-NEXT:    li a1, 24
+; CHECK-NEXT:    mul a0, a0, a1
+; CHECK-NEXT:    add sp, sp, a0
+; CHECK-NEXT:    addi sp, sp, 16
+; CHECK-NEXT:    ret
+  %v = call <32 x double> @llvm.vp.copysign.v32f64(<32 x double> %va, <32 x double> %vb, <32 x i1> %m, i32 %evl)
+  ret <32 x double> %v
+}
+
+define <32 x double> @vfsgnj_vv_v32f64_unmasked(<32 x double> %va, <32 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_v32f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    addi a1, a0, 128
+; CHECK-NEXT:    vsetivli zero, 16, e64, m8, ta, mu
+; CHECK-NEXT:    vle64.v v24, (a1)
+; CHECK-NEXT:    addi a3, a2, -16
+; CHECK-NEXT:    li a1, 0
+; CHECK-NEXT:    bltu a2, a3, .LBB27_2
+; CHECK-NEXT:  # %bb.1:
+; CHECK-NEXT:    mv a1, a3
+; CHECK-NEXT:  .LBB27_2:
+; CHECK-NEXT:    vle64.v v0, (a0)
+; CHECK-NEXT:    vsetvli zero, a1, e64, m8, ta, mu
+; CHECK-NEXT:    li a0, 16
+; CHECK-NEXT:    vfsgnj.vv v16, v16, v24
+; CHECK-NEXT:    bltu a2, a0, .LBB27_4
+; CHECK-NEXT:  # %bb.3:
+; CHECK-NEXT:    li a2, 16
+; CHECK-NEXT:  .LBB27_4:
+; CHECK-NEXT:    vsetvli zero, a2, e64, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v0
+; CHECK-NEXT:    ret
+  %head = insertelement <32 x i1> poison, i1 true, i32 0
+  %m = shufflevector <32 x i1> %head, <32 x i1> poison, <32 x i32> zeroinitializer
+  %v = call <32 x double> @llvm.vp.copysign.v32f64(<32 x double> %va, <32 x double> %vb, <32 x i1> %m, i32 %evl)
+  ret <32 x double> %v
+}

llvm/test/CodeGen/RISCV/rvv/vcopysign-vp.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv32 -mattr=+d,+zfh,+experimental-zvfh,+v -target-abi=ilp32d \
+; RUN:     -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=riscv64 -mattr=+d,+zfh,+experimental-zvfh,+v -target-abi=lp64d \
+; RUN:     -verify-machineinstrs < %s | FileCheck %s
+
+declare <vscale x 1 x half> @llvm.vp.copysign.nxv1f16(<vscale x 1 x half>, <vscale x 1 x half>, <vscale x 1 x i1>, i32)
+
+define <vscale x 1 x half> @vfsgnj_vv_nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x half> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv1f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 1 x half> @llvm.vp.copysign.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x half> %vb, <vscale x 1 x i1> %m, i32 %evl)
+  ret <vscale x 1 x half> %v
+}
+
+define <vscale x 1 x half> @vfsgnj_vv_nxv1f16_unmasked(<vscale x 1 x half> %va, <vscale x 1 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv1f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 1 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 1 x i1> %head, <vscale x 1 x i1> poison, <vscale x 1 x i32> zeroinitializer
+  %v = call <vscale x 1 x half> @llvm.vp.copysign.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x half> %vb, <vscale x 1 x i1> %m, i32 %evl)
+  ret <vscale x 1 x half> %v
+}
+
+declare <vscale x 2 x half> @llvm.vp.copysign.nxv2f16(<vscale x 2 x half>, <vscale x 2 x half>, <vscale x 2 x i1>, i32)
+
+define <vscale x 2 x half> @vfsgnj_vv_nxv2f16(<vscale x 2 x half> %va, <vscale x 2 x half> %vb, <vscale x 2 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv2f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x half> @llvm.vp.copysign.nxv2f16(<vscale x 2 x half> %va, <vscale x 2 x half> %vb, <vscale x 2 x i1> %m, i32 %evl)
+  ret <vscale x 2 x half> %v
+}
+
+define <vscale x 2 x half> @vfsgnj_vv_nxv2f16_unmasked(<vscale x 2 x half> %va, <vscale x 2 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv2f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 2 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 2 x i1> %head, <vscale x 2 x i1> poison, <vscale x 2 x i32> zeroinitializer
+  %v = call <vscale x 2 x half> @llvm.vp.copysign.nxv2f16(<vscale x 2 x half> %va, <vscale x 2 x half> %vb, <vscale x 2 x i1> %m, i32 %evl)
+  ret <vscale x 2 x half> %v
+}
+
+declare <vscale x 4 x half> @llvm.vp.copysign.nxv4f16(<vscale x 4 x half>, <vscale x 4 x half>, <vscale x 4 x i1>, i32)
+
+define <vscale x 4 x half> @vfsgnj_vv_nxv4f16(<vscale x 4 x half> %va, <vscale x 4 x half> %vb, <vscale x 4 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv4f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 4 x half> @llvm.vp.copysign.nxv4f16(<vscale x 4 x half> %va, <vscale x 4 x half> %vb, <vscale x 4 x i1> %m, i32 %evl)
+  ret <vscale x 4 x half> %v
+}
+
+define <vscale x 4 x half> @vfsgnj_vv_nxv4f16_unmasked(<vscale x 4 x half> %va, <vscale x 4 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv4f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 4 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 4 x i1> %head, <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer
+  %v = call <vscale x 4 x half> @llvm.vp.copysign.nxv4f16(<vscale x 4 x half> %va, <vscale x 4 x half> %vb, <vscale x 4 x i1> %m, i32 %evl)
+  ret <vscale x 4 x half> %v
+}
+
+declare <vscale x 8 x half> @llvm.vp.copysign.nxv8f16(<vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x i1>, i32)
+
+define <vscale x 8 x half> @vfsgnj_vv_nxv8f16(<vscale x 8 x half> %va, <vscale x 8 x half> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv8f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 8 x half> @llvm.vp.copysign.nxv8f16(<vscale x 8 x half> %va, <vscale x 8 x half> %vb, <vscale x 8 x i1> %m, i32 %evl)
+  ret <vscale x 8 x half> %v
+}
+
+define <vscale x 8 x half> @vfsgnj_vv_nxv8f16_unmasked(<vscale x 8 x half> %va, <vscale x 8 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv8f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 8 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 8 x i1> %head, <vscale x 8 x i1> poison, <vscale x 8 x i32> zeroinitializer
+  %v = call <vscale x 8 x half> @llvm.vp.copysign.nxv8f16(<vscale x 8 x half> %va, <vscale x 8 x half> %vb, <vscale x 8 x i1> %m, i32 %evl)
+  ret <vscale x 8 x half> %v
+}
+
+declare <vscale x 16 x half> @llvm.vp.copysign.nxv16f16(<vscale x 16 x half>, <vscale x 16 x half>, <vscale x 16 x i1>, i32)
+
+define <vscale x 16 x half> @vfsgnj_vv_nxv16f16(<vscale x 16 x half> %va, <vscale x 16 x half> %vb, <vscale x 16 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv16f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 16 x half> @llvm.vp.copysign.nxv16f16(<vscale x 16 x half> %va, <vscale x 16 x half> %vb, <vscale x 16 x i1> %m, i32 %evl)
+  ret <vscale x 16 x half> %v
+}
+
+define <vscale x 16 x half> @vfsgnj_vv_nxv16f16_unmasked(<vscale x 16 x half> %va, <vscale x 16 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv16f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 16 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 16 x i1> %head, <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer
+  %v = call <vscale x 16 x half> @llvm.vp.copysign.nxv16f16(<vscale x 16 x half> %va, <vscale x 16 x half> %vb, <vscale x 16 x i1> %m, i32 %evl)
+  ret <vscale x 16 x half> %v
+}
+
+declare <vscale x 32 x half> @llvm.vp.copysign.nxv32f16(<vscale x 32 x half>, <vscale x 32 x half>, <vscale x 32 x i1>, i32)
+
+define <vscale x 32 x half> @vfsgnj_vv_nxv32f16(<vscale x 32 x half> %va, <vscale x 32 x half> %vb, <vscale x 32 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv32f16:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 32 x half> @llvm.vp.copysign.nxv32f16(<vscale x 32 x half> %va, <vscale x 32 x half> %vb, <vscale x 32 x i1> %m, i32 %evl)
+  ret <vscale x 32 x half> %v
+}
+
+define <vscale x 32 x half> @vfsgnj_vv_nxv32f16_unmasked(<vscale x 32 x half> %va, <vscale x 32 x half> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv32f16_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 32 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 32 x i1> %head, <vscale x 32 x i1> poison, <vscale x 32 x i32> zeroinitializer
+  %v = call <vscale x 32 x half> @llvm.vp.copysign.nxv32f16(<vscale x 32 x half> %va, <vscale x 32 x half> %vb, <vscale x 32 x i1> %m, i32 %evl)
+  ret <vscale x 32 x half> %v
+}
+
+declare <vscale x 1 x float> @llvm.vp.copysign.nxv1f32(<vscale x 1 x float>, <vscale x 1 x float>, <vscale x 1 x i1>, i32)
+
+define <vscale x 1 x float> @vfsgnj_vv_nxv1f32(<vscale x 1 x float> %va, <vscale x 1 x float> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv1f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 1 x float> @llvm.vp.copysign.nxv1f32(<vscale x 1 x float> %va, <vscale x 1 x float> %vb, <vscale x 1 x i1> %m, i32 %evl)
+  ret <vscale x 1 x float> %v
+}
+
+define <vscale x 1 x float> @vfsgnj_vv_nxv1f32_unmasked(<vscale x 1 x float> %va, <vscale x 1 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv1f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 1 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 1 x i1> %head, <vscale x 1 x i1> poison, <vscale x 1 x i32> zeroinitializer
+  %v = call <vscale x 1 x float> @llvm.vp.copysign.nxv1f32(<vscale x 1 x float> %va, <vscale x 1 x float> %vb, <vscale x 1 x i1> %m, i32 %evl)
+  ret <vscale x 1 x float> %v
+}
+
+declare <vscale x 2 x float> @llvm.vp.copysign.nxv2f32(<vscale x 2 x float>, <vscale x 2 x float>, <vscale x 2 x i1>, i32)
+
+define <vscale x 2 x float> @vfsgnj_vv_nxv2f32(<vscale x 2 x float> %va, <vscale x 2 x float> %vb, <vscale x 2 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv2f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x float> @llvm.vp.copysign.nxv2f32(<vscale x 2 x float> %va, <vscale x 2 x float> %vb, <vscale x 2 x i1> %m, i32 %evl)
+  ret <vscale x 2 x float> %v
+}
+
+define <vscale x 2 x float> @vfsgnj_vv_nxv2f32_unmasked(<vscale x 2 x float> %va, <vscale x 2 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv2f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 2 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 2 x i1> %head, <vscale x 2 x i1> poison, <vscale x 2 x i32> zeroinitializer
+  %v = call <vscale x 2 x float> @llvm.vp.copysign.nxv2f32(<vscale x 2 x float> %va, <vscale x 2 x float> %vb, <vscale x 2 x i1> %m, i32 %evl)
+  ret <vscale x 2 x float> %v
+}
+
+declare <vscale x 4 x float> @llvm.vp.copysign.nxv4f32(<vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x i1>, i32)
+
+define <vscale x 4 x float> @vfsgnj_vv_nxv4f32(<vscale x 4 x float> %va, <vscale x 4 x float> %vb, <vscale x 4 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv4f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 4 x float> @llvm.vp.copysign.nxv4f32(<vscale x 4 x float> %va, <vscale x 4 x float> %vb, <vscale x 4 x i1> %m, i32 %evl)
+  ret <vscale x 4 x float> %v
+}
+
+define <vscale x 4 x float> @vfsgnj_vv_nxv4f32_unmasked(<vscale x 4 x float> %va, <vscale x 4 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv4f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 4 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 4 x i1> %head, <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer
+  %v = call <vscale x 4 x float> @llvm.vp.copysign.nxv4f32(<vscale x 4 x float> %va, <vscale x 4 x float> %vb, <vscale x 4 x i1> %m, i32 %evl)
+  ret <vscale x 4 x float> %v
+}
+
+declare <vscale x 8 x float> @llvm.vp.copysign.nxv8f32(<vscale x 8 x float>, <vscale x 8 x float>, <vscale x 8 x i1>, i32)
+
+define <vscale x 8 x float> @vfsgnj_vv_nxv8f32(<vscale x 8 x float> %va, <vscale x 8 x float> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv8f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 8 x float> @llvm.vp.copysign.nxv8f32(<vscale x 8 x float> %va, <vscale x 8 x float> %vb, <vscale x 8 x i1> %m, i32 %evl)
+  ret <vscale x 8 x float> %v
+}
+
+define <vscale x 8 x float> @vfsgnj_vv_nxv8f32_unmasked(<vscale x 8 x float> %va, <vscale x 8 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv8f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 8 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 8 x i1> %head, <vscale x 8 x i1> poison, <vscale x 8 x i32> zeroinitializer
+  %v = call <vscale x 8 x float> @llvm.vp.copysign.nxv8f32(<vscale x 8 x float> %va, <vscale x 8 x float> %vb, <vscale x 8 x i1> %m, i32 %evl)
+  ret <vscale x 8 x float> %v
+}
+
+declare <vscale x 16 x float> @llvm.vp.copysign.nxv16f32(<vscale x 16 x float>, <vscale x 16 x float>, <vscale x 16 x i1>, i32)
+
+define <vscale x 16 x float> @vfsgnj_vv_nxv16f32(<vscale x 16 x float> %va, <vscale x 16 x float> %vb, <vscale x 16 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv16f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 16 x float> @llvm.vp.copysign.nxv16f32(<vscale x 16 x float> %va, <vscale x 16 x float> %vb, <vscale x 16 x i1> %m, i32 %evl)
+  ret <vscale x 16 x float> %v
+}
+
+define <vscale x 16 x float> @vfsgnj_vv_nxv16f32_unmasked(<vscale x 16 x float> %va, <vscale x 16 x float> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv16f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 16 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 16 x i1> %head, <vscale x 16 x i1> poison, <vscale x 16 x i32> zeroinitializer
+  %v = call <vscale x 16 x float> @llvm.vp.copysign.nxv16f32(<vscale x 16 x float> %va, <vscale x 16 x float> %vb, <vscale x 16 x i1> %m, i32 %evl)
+  ret <vscale x 16 x float> %v
+}
+
+declare <vscale x 1 x double> @llvm.vp.copysign.nxv1f64(<vscale x 1 x double>, <vscale x 1 x double>, <vscale x 1 x i1>, i32)
+
+define <vscale x 1 x double> @vfsgnj_vv_nxv1f64(<vscale x 1 x double> %va, <vscale x 1 x double> %vb, <vscale x 1 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv1f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 1 x double> @llvm.vp.copysign.nxv1f64(<vscale x 1 x double> %va, <vscale x 1 x double> %vb, <vscale x 1 x i1> %m, i32 %evl)
+  ret <vscale x 1 x double> %v
+}
+
+define <vscale x 1 x double> @vfsgnj_vv_nxv1f64_unmasked(<vscale x 1 x double> %va, <vscale x 1 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv1f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 1 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 1 x i1> %head, <vscale x 1 x i1> poison, <vscale x 1 x i32> zeroinitializer
+  %v = call <vscale x 1 x double> @llvm.vp.copysign.nxv1f64(<vscale x 1 x double> %va, <vscale x 1 x double> %vb, <vscale x 1 x i1> %m, i32 %evl)
+  ret <vscale x 1 x double> %v
+}
+
+declare <vscale x 2 x double> @llvm.vp.copysign.nxv2f64(<vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x i1>, i32)
+
+define <vscale x 2 x double> @vfsgnj_vv_nxv2f64(<vscale x 2 x double> %va, <vscale x 2 x double> %vb, <vscale x 2 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv2f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x double> @llvm.vp.copysign.nxv2f64(<vscale x 2 x double> %va, <vscale x 2 x double> %vb, <vscale x 2 x i1> %m, i32 %evl)
+  ret <vscale x 2 x double> %v
+}
+
+define <vscale x 2 x double> @vfsgnj_vv_nxv2f64_unmasked(<vscale x 2 x double> %va, <vscale x 2 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv2f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 2 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 2 x i1> %head, <vscale x 2 x i1> poison, <vscale x 2 x i32> zeroinitializer
+  %v = call <vscale x 2 x double> @llvm.vp.copysign.nxv2f64(<vscale x 2 x double> %va, <vscale x 2 x double> %vb, <vscale x 2 x i1> %m, i32 %evl)
+  ret <vscale x 2 x double> %v
+}
+
+declare <vscale x 4 x double> @llvm.vp.copysign.nxv4f64(<vscale x 4 x double>, <vscale x 4 x double>, <vscale x 4 x i1>, i32)
+
+define <vscale x 4 x double> @vfsgnj_vv_nxv4f64(<vscale x 4 x double> %va, <vscale x 4 x double> %vb, <vscale x 4 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv4f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 4 x double> @llvm.vp.copysign.nxv4f64(<vscale x 4 x double> %va, <vscale x 4 x double> %vb, <vscale x 4 x i1> %m, i32 %evl)
+  ret <vscale x 4 x double> %v
+}
+
+define <vscale x 4 x double> @vfsgnj_vv_nxv4f64_unmasked(<vscale x 4 x double> %va, <vscale x 4 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv4f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 4 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 4 x i1> %head, <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer
+  %v = call <vscale x 4 x double> @llvm.vp.copysign.nxv4f64(<vscale x 4 x double> %va, <vscale x 4 x double> %vb, <vscale x 4 x i1> %m, i32 %evl)
+  ret <vscale x 4 x double> %v
+}
+
+declare <vscale x 8 x double> @llvm.vp.copysign.nxv8f64(<vscale x 8 x double>, <vscale x 8 x double>, <vscale x 8 x i1>, i32)
+
+define <vscale x 8 x double> @vfsgnj_vv_nxv8f64(<vscale x 8 x double> %va, <vscale x 8 x double> %vb, <vscale x 8 x i1> %m, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv8f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 8 x double> @llvm.vp.copysign.nxv8f64(<vscale x 8 x double> %va, <vscale x 8 x double> %vb, <vscale x 8 x i1> %m, i32 %evl)
+  ret <vscale x 8 x double> %v
+}
+
+define <vscale x 8 x double> @vfsgnj_vv_nxv8f64_unmasked(<vscale x 8 x double> %va, <vscale x 8 x double> %vb, i32 zeroext %evl) {
+; CHECK-LABEL: vfsgnj_vv_nxv8f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e64, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16
+; CHECK-NEXT:    ret
+  %head = insertelement <vscale x 8 x i1> poison, i1 true, i32 0
+  %m = shufflevector <vscale x 8 x i1> %head, <vscale x 8 x i1> poison, <vscale x 8 x i32> zeroinitializer
+  %v = call <vscale x 8 x double> @llvm.vp.copysign.nxv8f64(<vscale x 8 x double> %va, <vscale x 8 x double> %vb, <vscale x 8 x i1> %m, i32 %evl)
+  ret <vscale x 8 x double> %v
+}

llvm/unittests/IR/VPIntrinsicTest.cpp

std::unique_ptr<Module> createVPDeclarationModule() {
     const char *BinaryIntOpcodes[] = {"add",  "sub",  "mul", "sdiv", "srem",
                                       "udiv", "urem", "and", "xor",  "or",
                                       "ashr", "lshr", "shl"};
     std::stringstream Str;
     for (const char *BinaryIntOpcode : BinaryIntOpcodes)
       Str << " declare <8 x i32> @llvm.vp." << BinaryIntOpcode
           << ".v8i32(<8 x i32>, <8 x i32>, <8 x i1>, i32) ";
 
-    const char *BinaryFPOpcodes[] = {"fadd", "fsub",   "fmul",  "fdiv",
-                                     "frem", "minnum", "maxnum"};
+    const char *BinaryFPOpcodes[] = {"fadd", "fsub",   "fmul",   "fdiv",
+                                     "frem", "minnum", "maxnum", "copysign"};
     for (const char *BinaryFPOpcode : BinaryFPOpcodes)
       Str << " declare <8 x float> @llvm.vp." << BinaryFPOpcode
           << ".v8f32(<8 x float>, <8 x float>, <8 x i1>, i32) ";
 
     Str << " declare <8 x float> @llvm.vp.floor.v8f32(<8 x float>, <8 x i1>, "
            "i32)";
     Str << " declare <8 x float> @llvm.vp.round.v8f32(<8 x float>, <8 x i1>, "
            "i32)";