[RISCV] Add rvv codegen support for vp.fpext.

jacquesguan · jacquesguan · commit 2509dcd58a0a · 2022-05-11T03:28:25.000Z
This patch adds rvv codegen support for vp.fpext. The lowering of fp_round, vp.fptrunc, fp_extend and vp.fpext share most code so use a common lowering function to handle these four. And this patch changes the intermediate cast from ISD::FP_EXTEND/ISD::FP_ROUND to the RVV VL version op RISCVISD::FP_EXTEND_VL and RISCVISD::FP_ROUND_VL for scalable vectors. Reviewed By: frasercrmck Differential Revision: https://reviews.llvm.org/D123975
diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
@@ -443,7 +443,8 @@ RISCVTargetLowering::RISCVTargetLowering(const TargetMachine &TM,
         ISD::VP_REDUCE_FMIN, ISD::VP_REDUCE_FMAX,
         ISD::VP_MERGE,       ISD::VP_SELECT,
         ISD::VP_SITOFP,      ISD::VP_UITOFP,
-        ISD::VP_SETCC,       ISD::VP_FP_ROUND};
+        ISD::VP_SETCC,       ISD::VP_FP_ROUND,
+        ISD::VP_FP_EXTEND};
 
     if (!Subtarget.is64Bit()) {
       // We must custom-lower certain vXi64 operations on RV32 due to the vector
@@ -2795,21 +2796,6 @@ bool RISCVTargetLowering::isShuffleMaskLegal(ArrayRef<int> M, EVT VT) const {
          isInterleaveShuffle(M, SVT, SwapSources, Subtarget);
 }
 
-static SDValue getRVVFPExtendOrRound(SDValue Op, MVT VT, MVT ContainerVT,
-                                     SDLoc DL, SelectionDAG &DAG,
-                                     const RISCVSubtarget &Subtarget) {
-  if (VT.isScalableVector())
-    return DAG.getFPExtendOrRound(Op, DL, VT);
-  assert(VT.isFixedLengthVector() &&
-         "Unexpected value type for RVV FP extend/round lowering");
-  SDValue Mask, VL;
-  std::tie(Mask, VL) = getDefaultVLOps(VT, ContainerVT, DL, DAG, Subtarget);
-  unsigned RVVOpc = ContainerVT.bitsGT(Op.getSimpleValueType())
-                        ? RISCVISD::FP_EXTEND_VL
-                        : RISCVISD::FP_ROUND_VL;
-  return DAG.getNode(RVVOpc, DL, ContainerVT, Op, Mask, VL);
-}
-
 // Lower CTLZ_ZERO_UNDEF or CTTZ_ZERO_UNDEF by converting to FP and extracting
 // the exponent.
 static SDValue lowerCTLZ_CTTZ_ZERO_UNDEF(SDValue Op, SelectionDAG &DAG) {
@@ -3126,50 +3112,11 @@ SDValue RISCVTargetLowering::LowerOperation(SDValue Op,
     }
     return SDValue();
   }
-  case ISD::FP_EXTEND: {
-    // RVV can only do fp_extend to types double the size as the source. We
-    // custom-lower f16->f64 extensions to two hops of ISD::FP_EXTEND, going
-    // via f32.
-    SDLoc DL(Op);
-    MVT VT = Op.getSimpleValueType();
-    SDValue Src = Op.getOperand(0);
-    MVT SrcVT = Src.getSimpleValueType();
-
-    // Prepare any fixed-length vector operands.
-    MVT ContainerVT = VT;
-    if (SrcVT.isFixedLengthVector()) {
-      ContainerVT = getContainerForFixedLengthVector(VT);
-      MVT SrcContainerVT =
-          ContainerVT.changeVectorElementType(SrcVT.getVectorElementType());
-      Src = convertToScalableVector(SrcContainerVT, Src, DAG, Subtarget);
-    }
-
-    if (!VT.isVector() || VT.getVectorElementType() != MVT::f64 ||
-        SrcVT.getVectorElementType() != MVT::f16) {
-      // For scalable vectors, we only need to close the gap between
-      // vXf16->vXf64.
-      if (!VT.isFixedLengthVector())
-        return Op;
-      // For fixed-length vectors, lower the FP_EXTEND to a custom "VL" version.
-      Src = getRVVFPExtendOrRound(Src, VT, ContainerVT, DL, DAG, Subtarget);
-      return convertFromScalableVector(VT, Src, DAG, Subtarget);
-    }
-
-    MVT InterVT = VT.changeVectorElementType(MVT::f32);
-    MVT InterContainerVT = ContainerVT.changeVectorElementType(MVT::f32);
-    SDValue IntermediateExtend = getRVVFPExtendOrRound(
-        Src, InterVT, InterContainerVT, DL, DAG, Subtarget);
-
-    SDValue Extend = getRVVFPExtendOrRound(IntermediateExtend, VT, ContainerVT,
-                                           DL, DAG, Subtarget);
-    if (VT.isFixedLengthVector())
-      return convertFromScalableVector(VT, Extend, DAG, Subtarget);
-    return Extend;
-  }
+  case ISD::FP_EXTEND:
   case ISD::FP_ROUND:
     if (!Op.getValueType().isVector())
       return Op;
-    return lowerVectorFPRoundLike(Op, DAG);
+    return lowerVectorFPExtendOrRoundLike(Op, DAG);
   case ISD::FP_TO_SINT:
   case ISD::FP_TO_UINT:
   case ISD::SINT_TO_FP:
@@ -3512,8 +3459,9 @@ SDValue RISCVTargetLowering::LowerOperation(SDValue Op,
                                                     : RISCVISD::VZEXT_VL);
   case ISD::VP_TRUNC:
     return lowerVectorTruncLike(Op, DAG);
+  case ISD::VP_FP_EXTEND:
   case ISD::VP_FP_ROUND:
-    return lowerVectorFPRoundLike(Op, DAG);
+    return lowerVectorFPExtendOrRoundLike(Op, DAG);
   case ISD::VP_FPTOSI:
     return lowerVPFPIntConvOp(Op, DAG, RISCVISD::FP_TO_SINT_VL);
   case ISD::VP_FPTOUI:
@@ -4281,9 +4229,13 @@ SDValue RISCVTargetLowering::lowerVectorTruncLike(SDValue Op,
   return Result;
 }
 
-SDValue RISCVTargetLowering::lowerVectorFPRoundLike(SDValue Op,
+SDValue
+RISCVTargetLowering::lowerVectorFPExtendOrRoundLike(SDValue Op,
                                                     SelectionDAG &DAG) const {
-  bool IsVPFPTrunc = Op.getOpcode() == ISD::VP_FP_ROUND;
+  bool IsVP =
+      Op.getOpcode() == ISD::VP_FP_ROUND || Op.getOpcode() == ISD::VP_FP_EXTEND;
+  bool IsExtend =
+      Op.getOpcode() == ISD::VP_FP_EXTEND || Op.getOpcode() == ISD::FP_EXTEND;
   // RVV can only do truncate fp to types half the size as the source. We
   // custom-lower f64->f16 rounds via RVV's round-to-odd float
   // conversion instruction.
@@ -4295,17 +4247,21 @@ SDValue RISCVTargetLowering::lowerVectorFPRoundLike(SDValue Op,
   SDValue Src = Op.getOperand(0);
   MVT SrcVT = Src.getSimpleValueType();
 
-  bool IsDirectConv = VT.getVectorElementType() != MVT::f16 ||
-                      SrcVT.getVectorElementType() != MVT::f64;
+  bool IsDirectExtend = IsExtend && (VT.getVectorElementType() != MVT::f64 ||
+                                     SrcVT.getVectorElementType() != MVT::f16);
+  bool IsDirectTrunc = !IsExtend && (VT.getVectorElementType() != MVT::f16 ||
+                                     SrcVT.getVectorElementType() != MVT::f64);
 
-  // For FP_ROUND of scalable vectors, leave it to the pattern. 
-  if (!VT.isFixedLengthVector() && !IsVPFPTrunc && IsDirectConv)
+  bool IsDirectConv = IsDirectExtend || IsDirectTrunc;
+
+  // For FP_ROUND/FP_EXTEND of scalable vectors, leave it to the pattern.
+  if (!VT.isFixedLengthVector() && !IsVP && IsDirectConv)
     return Op;
 
   // Prepare any fixed-length vector operands.
   MVT ContainerVT = VT;
   SDValue Mask, VL;
-  if (IsVPFPTrunc) {
+  if (IsVP) {
     Mask = Op.getOperand(1);
     VL = Op.getOperand(2);
   }
@@ -4314,31 +4270,36 @@ SDValue RISCVTargetLowering::lowerVectorFPRoundLike(SDValue Op,
     ContainerVT =
         SrcContainerVT.changeVectorElementType(VT.getVectorElementType());
     Src = convertToScalableVector(SrcContainerVT, Src, DAG, Subtarget);
-    if (IsVPFPTrunc) {
+    if (IsVP) {
       MVT MaskVT = getMaskTypeFor(ContainerVT);
       Mask = convertToScalableVector(MaskVT, Mask, DAG, Subtarget);
     }
   }
 
-  if (!IsVPFPTrunc)
+  if (!IsVP)
     std::tie(Mask, VL) =
         getDefaultVLOps(SrcVT, ContainerVT, DL, DAG, Subtarget);
 
+  unsigned ConvOpc = IsExtend ? RISCVISD::FP_EXTEND_VL : RISCVISD::FP_ROUND_VL;
+
   if (IsDirectConv) {
-    Src = DAG.getNode(RISCVISD::FP_ROUND_VL, DL, ContainerVT, Src, Mask, VL);
+    Src = DAG.getNode(ConvOpc, DL, ContainerVT, Src, Mask, VL);
     if (VT.isFixedLengthVector())
       Src = convertFromScalableVector(VT, Src, DAG, Subtarget);
     return Src;
   }
 
+  unsigned InterConvOpc =
+      IsExtend ? RISCVISD::FP_EXTEND_VL : RISCVISD::VFNCVT_ROD_VL;
+
   MVT InterVT = ContainerVT.changeVectorElementType(MVT::f32);
-  SDValue IntermediateRound =
-      DAG.getNode(RISCVISD::VFNCVT_ROD_VL, DL, InterVT, Src, Mask, VL);
-  SDValue Round = DAG.getNode(RISCVISD::FP_ROUND_VL, DL, ContainerVT,
-                              IntermediateRound, Mask, VL);
+  SDValue IntermediateConv =
+      DAG.getNode(InterConvOpc, DL, InterVT, Src, Mask, VL);
+  SDValue Result =
+      DAG.getNode(ConvOpc, DL, ContainerVT, IntermediateConv, Mask, VL);
   if (VT.isFixedLengthVector())
-    return convertFromScalableVector(VT, Round, DAG, Subtarget);
-  return Round;
+    return convertFromScalableVector(VT, Result, DAG, Subtarget);
+  return Result;
 }
 
 // Custom-legalize INSERT_VECTOR_ELT so that the value is inserted into the
diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.h b/llvm/lib/Target/RISCV/RISCVISelLowering.h
@@ -614,7 +614,7 @@ class RISCVTargetLowering : public TargetLowering {
                              int64_t ExtTrueVal) const;
   SDValue lowerVectorMaskTruncLike(SDValue Op, SelectionDAG &DAG) const;
   SDValue lowerVectorTruncLike(SDValue Op, SelectionDAG &DAG) const;
-  SDValue lowerVectorFPRoundLike(SDValue Op, SelectionDAG &DAG) const;
+  SDValue lowerVectorFPExtendOrRoundLike(SDValue Op, SelectionDAG &DAG) const;
   SDValue lowerINSERT_VECTOR_ELT(SDValue Op, SelectionDAG &DAG) const;
   SDValue lowerEXTRACT_VECTOR_ELT(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerINTRINSIC_WO_CHAIN(SDValue Op, SelectionDAG &DAG) const;
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vector-fpext-vp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vector-fpext-vp.ll
@@ -0,0 +1,77 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv32 -mattr=+d,+zfh,+experimental-zvfh,+v -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=riscv64 -mattr=+d,+zfh,+experimental-zvfh,+v -riscv-v-vector-bits-min=128 -verify-machineinstrs < %s | FileCheck %s
+
+declare <2 x float> @llvm.vp.fpext.v2f32.v2f16(<2 x half>, <2 x i1>, i32)
+
+define <2 x float> @vfpext_v2f16_v2f32(<2 x half> %a, <2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_v2f16_v2f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v9, v8, v0.t
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+  %v = call <2 x float> @llvm.vp.fpext.v2f32.v2f16(<2 x half> %a, <2 x i1> %m, i32 %vl)
+  ret <2 x float> %v
+}
+
+define <2 x float> @vfpext_v2f16_v2f32_unmasked(<2 x half> %a, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_v2f16_v2f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+  %v = call <2 x float> @llvm.vp.fpext.v2f32.v2f16(<2 x half> %a, <2 x i1> shufflevector (<2 x i1> insertelement (<2 x i1> undef, i1 true, i32 0), <2 x i1> undef, <2 x i32> zeroinitializer), i32 %vl)
+  ret <2 x float> %v
+}
+
+declare <2 x double> @llvm.vp.fpext.v2f64.v2f16(<2 x half>, <2 x i1>, i32)
+
+define <2 x double> @vfpext_v2f16_v2f64(<2 x half> %a, <2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_v2f16_v2f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v9, v8, v0.t
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+  %v = call <2 x double> @llvm.vp.fpext.v2f64.v2f16(<2 x half> %a, <2 x i1> %m, i32 %vl)
+  ret <2 x double> %v
+}
+
+define <2 x double> @vfpext_v2f16_v2f64_unmasked(<2 x half> %a, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_v2f16_v2f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v9, v8
+; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v8, v9
+; CHECK-NEXT:    ret
+  %v = call <2 x double> @llvm.vp.fpext.v2f64.v2f16(<2 x half> %a, <2 x i1> shufflevector (<2 x i1> insertelement (<2 x i1> undef, i1 true, i32 0), <2 x i1> undef, <2 x i32> zeroinitializer), i32 %vl)
+  ret <2 x double> %v
+}
+
+declare <2 x double> @llvm.vp.fpext.v2f64.v2f32(<2 x float>, <2 x i1>, i32)
+
+define <2 x double> @vfpext_v2f32_v2f64(<2 x float> %a, <2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_v2f32_v2f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v9, v8, v0.t
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+  %v = call <2 x double> @llvm.vp.fpext.v2f64.v2f32(<2 x float> %a, <2 x i1> %m, i32 %vl)
+  ret <2 x double> %v
+}
+
+define <2 x double> @vfpext_v2f32_v2f64_unmasked(<2 x float> %a, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_v2f32_v2f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+  %v = call <2 x double> @llvm.vp.fpext.v2f64.v2f32(<2 x float> %a, <2 x i1> shufflevector (<2 x i1> insertelement (<2 x i1> undef, i1 true, i32 0), <2 x i1> undef, <2 x i32> zeroinitializer), i32 %vl)
+  ret <2 x double> %v
+}
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfpext-vp.ll b/llvm/test/CodeGen/RISCV/rvv/vfpext-vp.ll
@@ -0,0 +1,77 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv32 -mattr=+d,+zfh,+experimental-zvfh,+v -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=riscv64 -mattr=+d,+zfh,+experimental-zvfh,+v -verify-machineinstrs < %s | FileCheck %s
+
+declare <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half>, <vscale x 2 x i1>, i32)
+
+define <vscale x 2 x float> @vfpext_nxv2f16_nxv2f32(<vscale x 2 x half> %a, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_nxv2f16_nxv2f32:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v9, v8, v0.t
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x float> %v
+}
+
+define <vscale x 2 x float> @vfpext_nxv2f16_nxv2f32_unmasked(<vscale x 2 x half> %a, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_nxv2f16_nxv2f32_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x i1> shufflevector (<vscale x 2 x i1> insertelement (<vscale x 2 x i1> undef, i1 true, i32 0), <vscale x 2 x i1> undef, <vscale x 2 x i32> zeroinitializer), i32 %vl)
+  ret <vscale x 2 x float> %v
+}
+
+declare <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f16(<vscale x 2 x half>, <vscale x 2 x i1>, i32)
+
+define <vscale x 2 x double> @vfpext_nxv2f16_nxv2f64(<vscale x 2 x half> %a, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_nxv2f16_nxv2f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v10, v8, v0.t
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v8, v10, v0.t
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x double> %v
+}
+
+define <vscale x 2 x double> @vfpext_nxv2f16_nxv2f64_unmasked(<vscale x 2 x half> %a, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_nxv2f16_nxv2f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v10, v8
+; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v8, v10
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x i1> shufflevector (<vscale x 2 x i1> insertelement (<vscale x 2 x i1> undef, i1 true, i32 0), <vscale x 2 x i1> undef, <vscale x 2 x i32> zeroinitializer), i32 %vl)
+  ret <vscale x 2 x double> %v
+}
+
+declare <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float>, <vscale x 2 x i1>, i32)
+
+define <vscale x 2 x double> @vfpext_nxv2f32_nxv2f64(<vscale x 2 x float> %a, <vscale x 2 x i1> %m, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_nxv2f32_nxv2f64:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v10, v8, v0.t
+; CHECK-NEXT:    vmv2r.v v8, v10
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %a, <vscale x 2 x i1> %m, i32 %vl)
+  ret <vscale x 2 x double> %v
+}
+
+define <vscale x 2 x double> @vfpext_nxv2f32_nxv2f64_unmasked(<vscale x 2 x float> %a, i32 zeroext %vl) {
+; CHECK-LABEL: vfpext_nxv2f32_nxv2f64_unmasked:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, mu
+; CHECK-NEXT:    vfwcvt.f.f.v v10, v8
+; CHECK-NEXT:    vmv2r.v v8, v10
+; CHECK-NEXT:    ret
+  %v = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %a, <vscale x 2 x i1> shufflevector (<vscale x 2 x i1> insertelement (<vscale x 2 x i1> undef, i1 true, i32 0), <vscale x 2 x i1> undef, <vscale x 2 x i32> zeroinitializer), i32 %vl)
+  ret <vscale x 2 x double> %v
+}
