r200524 - ARM & AArch64: merge another NEON block completely.
Tim Northover
tnorthover at apple.com
Fri Jan 31 02:46:37 PST 2014
Author: tnorthover
Date: Fri Jan 31 04:46:36 2014
New Revision: 200524
URL: http://llvm.org/viewvc/llvm-project?rev=200524&view=rev
Log:
ARM & AArch64: merge another NEON block completely.
Modified:
cfe/trunk/lib/CodeGen/CGBuiltin.cpp
Modified: cfe/trunk/lib/CodeGen/CGBuiltin.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CGBuiltin.cpp?rev=200524&r1=200523&r2=200524&view=diff
==============================================================================
--- cfe/trunk/lib/CodeGen/CGBuiltin.cpp (original)
+++ cfe/trunk/lib/CodeGen/CGBuiltin.cpp Fri Jan 31 04:46:36 2014
@@ -1778,6 +1778,24 @@ Value *CodeGenFunction::EmitCommonNeonBu
case NEON::BI__builtin_neon_vabdq_v:
Int = Usgn ? Intrinsic::arm_neon_vabdu : Intrinsic::arm_neon_vabds;
return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vabd");
+ case NEON::BI__builtin_neon_vaddhn_v: {
+ llvm::VectorType *SrcTy =
+ llvm::VectorType::getExtendedElementVectorType(VTy);
+
+ // %sum = add <4 x i32> %lhs, %rhs
+ Ops[0] = Builder.CreateBitCast(Ops[0], SrcTy);
+ Ops[1] = Builder.CreateBitCast(Ops[1], SrcTy);
+ Ops[0] = Builder.CreateAdd(Ops[0], Ops[1], "vaddhn");
+
+ // %high = lshr <4 x i32> %sum, <i32 16, i32 16, i32 16, i32 16>
+ Constant *ShiftAmt = ConstantInt::get(SrcTy->getElementType(),
+ SrcTy->getScalarSizeInBits() / 2);
+ ShiftAmt = ConstantVector::getSplat(VTy->getNumElements(), ShiftAmt);
+ Ops[0] = Builder.CreateLShr(Ops[0], ShiftAmt, "vaddhn");
+
+ // %res = trunc <4 x i32> %high to <4 x i16>
+ return Builder.CreateTrunc(Ops[0], VTy, "vaddhn");
+ }
case NEON::BI__builtin_neon_vbsl_v:
case NEON::BI__builtin_neon_vbslq_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vbsl, Ty),
@@ -1804,19 +1822,146 @@ Value *CodeGenFunction::EmitCommonNeonBu
// NEON intrinsic puts accumulator first, unlike the LLVM fma.
return Builder.CreateCall3(F, Ops[1], Ops[2], Ops[0]);
}
+ case NEON::BI__builtin_neon_vhadd_v:
+ case NEON::BI__builtin_neon_vhaddq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vhaddu : Intrinsic::arm_neon_vhadds;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vhadd");
+ case NEON::BI__builtin_neon_vhsub_v:
+ case NEON::BI__builtin_neon_vhsubq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vhsubu : Intrinsic::arm_neon_vhsubs;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vhsub");
+ case NEON::BI__builtin_neon_vqadd_v:
+ case NEON::BI__builtin_neon_vqaddq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vqaddu : Intrinsic::arm_neon_vqadds;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqadd");
+ case NEON::BI__builtin_neon_vqshl_n_v:
+ case NEON::BI__builtin_neon_vqshlq_n_v:
+ Int = Usgn ? Intrinsic::arm_neon_vqshiftu : Intrinsic::arm_neon_vqshifts;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqshl_n",
+ 1, false);
+ case NEON::BI__builtin_neon_vqsub_v:
+ case NEON::BI__builtin_neon_vqsubq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vqsubu : Intrinsic::arm_neon_vqsubs;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqsub");
+ case NEON::BI__builtin_neon_vrhadd_v:
+ case NEON::BI__builtin_neon_vrhaddq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vrhaddu : Intrinsic::arm_neon_vrhadds;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vrhadd");
+ case NEON::BI__builtin_neon_vmax_v:
+ case NEON::BI__builtin_neon_vmaxq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vmaxu : Intrinsic::arm_neon_vmaxs;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vmax");
+ case NEON::BI__builtin_neon_vmin_v:
+ case NEON::BI__builtin_neon_vminq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vminu : Intrinsic::arm_neon_vmins;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vmin");
case NEON::BI__builtin_neon_vmul_v:
case NEON::BI__builtin_neon_vmulq_v:
assert(Type.isPoly() && "vmul builtin only supported for polynomial types");
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vmulp, Ty),
Ops, "vmul");
+ case NEON::BI__builtin_neon_vmull_v:
+ // FIXME: the integer vmull operations could be emitted in terms of pure
+ // LLVM IR (2 exts followed by a mul). Unfortunately LLVM has a habit of
+ // hoisting the exts outside loops. Until global ISel comes along that can
+ // see through such movement this leads to bad CodeGen. So we need an
+ // intrinsic for now.
+ Int = Usgn ? Intrinsic::arm_neon_vmullu : Intrinsic::arm_neon_vmulls;
+ Int = Type.isPoly() ? (unsigned)Intrinsic::arm_neon_vmullp : Int;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vmull");
+ case NEON::BI__builtin_neon_vpadd_v:
+ return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vpadd, Ty),
+ Ops, "vpadd");
+ case NEON::BI__builtin_neon_vpmax_v:
+ Int = Usgn ? Intrinsic::arm_neon_vpmaxu : Intrinsic::arm_neon_vpmaxs;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vpmax");
+ case NEON::BI__builtin_neon_vpmin_v:
+ Int = Usgn ? Intrinsic::arm_neon_vpminu : Intrinsic::arm_neon_vpmins;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vpmin");
+ case NEON::BI__builtin_neon_vqdmlal_v: {
+ SmallVector<Value *, 2> MulOps(Ops.begin() + 1, Ops.end());
+ Value *Mul = EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqdmull, Ty),
+ MulOps, "vqdmlal");
+
+ SmallVector<Value *, 2> AddOps;
+ AddOps.push_back(Ops[0]);
+ AddOps.push_back(Mul);
+ return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqadds, Ty),
+ AddOps, "vqdmlal");
+ }
+ case NEON::BI__builtin_neon_vqdmlsl_v: {
+ SmallVector<Value *, 2> MulOps(Ops.begin() + 1, Ops.end());
+ Value *Mul = EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqdmull, Ty),
+ MulOps, "vqdmlsl");
+
+ SmallVector<Value *, 2> SubOps;
+ SubOps.push_back(Ops[0]);
+ SubOps.push_back(Mul);
+ return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqsubs, Ty),
+ SubOps, "vqdmlsl");
+ }
+ case NEON::BI__builtin_neon_vqdmulh_v:
+ case NEON::BI__builtin_neon_vqdmulhq_v:
+ return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqdmulh, Ty),
+ Ops, "vqdmulh");
+ case NEON::BI__builtin_neon_vqdmull_v:
+ return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqdmull, Ty),
+ Ops, "vqdmull");
+ case NEON::BI__builtin_neon_vqrdmulh_v:
+ case NEON::BI__builtin_neon_vqrdmulhq_v:
+ return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqrdmulh, Ty),
+ Ops, "vqrdmulh");
+ case NEON::BI__builtin_neon_vqrshl_v:
+ case NEON::BI__builtin_neon_vqrshlq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vqrshiftu : Intrinsic::arm_neon_vqrshifts;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqrshl");
+ case NEON::BI__builtin_neon_vqshl_v:
+ case NEON::BI__builtin_neon_vqshlq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vqshiftu : Intrinsic::arm_neon_vqshifts;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqshl");
+ case NEON::BI__builtin_neon_vraddhn_v:
+ return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vraddhn, Ty),
+ Ops, "vraddhn");
case NEON::BI__builtin_neon_vrecps_v:
case NEON::BI__builtin_neon_vrecpsq_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vrecps, Ty),
Ops, "vrecps");
+ case NEON::BI__builtin_neon_vrshl_v:
+ case NEON::BI__builtin_neon_vrshlq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vrshiftu : Intrinsic::arm_neon_vrshifts;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vrshl");
case NEON::BI__builtin_neon_vrsqrts_v:
case NEON::BI__builtin_neon_vrsqrtsq_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vrsqrts, Ty),
Ops, "vrsqrts");
+ case NEON::BI__builtin_neon_vrsubhn_v:
+ return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vrsubhn, Ty),
+ Ops, "vrsubhn");
+ case NEON::BI__builtin_neon_vshl_v:
+ case NEON::BI__builtin_neon_vshlq_v:
+ Int = Usgn ? Intrinsic::arm_neon_vshiftu : Intrinsic::arm_neon_vshifts;
+ return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vshl");
+ case NEON::BI__builtin_neon_vshr_n_v:
+ case NEON::BI__builtin_neon_vshrq_n_v:
+ return EmitNeonRShiftImm(Ops[0], Ops[1], Ty, Usgn, "vshr_n");
+ case NEON::BI__builtin_neon_vsubhn_v: {
+ llvm::VectorType *SrcTy =
+ llvm::VectorType::getExtendedElementVectorType(VTy);
+
+ // %sum = add <4 x i32> %lhs, %rhs
+ Ops[0] = Builder.CreateBitCast(Ops[0], SrcTy);
+ Ops[1] = Builder.CreateBitCast(Ops[1], SrcTy);
+ Ops[0] = Builder.CreateSub(Ops[0], Ops[1], "vsubhn");
+
+ // %high = lshr <4 x i32> %sum, <i32 16, i32 16, i32 16, i32 16>
+ Constant *ShiftAmt = ConstantInt::get(SrcTy->getElementType(),
+ SrcTy->getScalarSizeInBits() / 2);
+ ShiftAmt = ConstantVector::getSplat(VTy->getNumElements(), ShiftAmt);
+ Ops[0] = Builder.CreateLShr(Ops[0], ShiftAmt, "vsubhn");
+
+ // %res = trunc <4 x i32> %high to <4 x i16>
+ return Builder.CreateTrunc(Ops[0], VTy, "vsubhn");
+ }
case NEON::BI__builtin_neon_vtrn_v:
case NEON::BI__builtin_neon_vtrnq_v: {
Ops[0] = Builder.CreateBitCast(Ops[0], llvm::PointerType::getUnqual(Ty));
@@ -1837,6 +1982,15 @@ Value *CodeGenFunction::EmitCommonNeonBu
}
return SV;
}
+ case NEON::BI__builtin_neon_vtst_v:
+ case NEON::BI__builtin_neon_vtstq_v: {
+ Ops[0] = Builder.CreateBitCast(Ops[0], Ty);
+ Ops[1] = Builder.CreateBitCast(Ops[1], Ty);
+ Ops[0] = Builder.CreateAnd(Ops[0], Ops[1]);
+ Ops[0] = Builder.CreateICmp(ICmpInst::ICMP_NE, Ops[0],
+ ConstantAggregateZero::get(Ty));
+ return Builder.CreateSExt(Ops[0], Ty, "vtst");
+ }
case NEON::BI__builtin_neon_vuzp_v:
case NEON::BI__builtin_neon_vuzpq_v: {
Ops[0] = Builder.CreateBitCast(Ops[0], llvm::PointerType::getUnqual(Ty));
@@ -3177,90 +3331,8 @@ Value *CodeGenFunction::EmitAArch64Built
F = CGM.getIntrinsic(Intrinsic::arm_neon_vacgtq);
return EmitNeonCall(F, Ops, "vcagt");
}
- case NEON::BI__builtin_neon_vtst_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vtst_v, E);
- case NEON::BI__builtin_neon_vtstq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vtstq_v, E);
- case NEON::BI__builtin_neon_vhadd_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vhadd_v, E);
- case NEON::BI__builtin_neon_vhaddq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vhaddq_v, E);
- case NEON::BI__builtin_neon_vhsub_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vhsub_v, E);
- case NEON::BI__builtin_neon_vhsubq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vhsubq_v, E);
- case NEON::BI__builtin_neon_vrhadd_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vrhadd_v, E);
- case NEON::BI__builtin_neon_vrhaddq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vrhaddq_v, E);
- case NEON::BI__builtin_neon_vqadd_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqadd_v, E);
- case NEON::BI__builtin_neon_vqaddq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqaddq_v, E);
- case NEON::BI__builtin_neon_vqsub_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqsub_v, E);
- case NEON::BI__builtin_neon_vqsubq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqsubq_v, E);
- case NEON::BI__builtin_neon_vshl_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vshl_v, E);
- case NEON::BI__builtin_neon_vshlq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vshlq_v, E);
- case NEON::BI__builtin_neon_vqshl_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqshl_v, E);
- case NEON::BI__builtin_neon_vqshlq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqshlq_v, E);
- case NEON::BI__builtin_neon_vrshl_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vrshl_v, E);
- case NEON::BI__builtin_neon_vrshlq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vrshlq_v, E);
- case NEON::BI__builtin_neon_vqrshl_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqrshl_v, E);
- case NEON::BI__builtin_neon_vqrshlq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqrshlq_v, E);
- case NEON::BI__builtin_neon_vaddhn_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vaddhn_v, E);
- case NEON::BI__builtin_neon_vraddhn_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vraddhn_v, E);
- case NEON::BI__builtin_neon_vsubhn_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vsubhn_v, E);
- case NEON::BI__builtin_neon_vrsubhn_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vrsubhn_v, E);
- case NEON::BI__builtin_neon_vmull_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vmull_v, E);
- case NEON::BI__builtin_neon_vqdmull_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqdmull_v, E);
- case NEON::BI__builtin_neon_vqdmlal_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqdmlal_v, E);
- case NEON::BI__builtin_neon_vqdmlsl_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqdmlsl_v, E);
- case NEON::BI__builtin_neon_vmax_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vmax_v, E);
- case NEON::BI__builtin_neon_vmaxq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vmaxq_v, E);
- case NEON::BI__builtin_neon_vmin_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vmin_v, E);
- case NEON::BI__builtin_neon_vminq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vminq_v, E);
- case NEON::BI__builtin_neon_vpmax_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vpmax_v, E);
- case NEON::BI__builtin_neon_vpmin_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vpmin_v, E);
- case NEON::BI__builtin_neon_vpadd_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vpadd_v, E);
- case NEON::BI__builtin_neon_vqdmulh_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqdmulh_v, E);
- case NEON::BI__builtin_neon_vqdmulhq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqdmulhq_v, E);
- case NEON::BI__builtin_neon_vqrdmulh_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqrdmulh_v, E);
- case NEON::BI__builtin_neon_vqrdmulhq_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vqrdmulhq_v, E);
// Shift by immediate
- case NEON::BI__builtin_neon_vshr_n_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vshr_n_v, E);
- case NEON::BI__builtin_neon_vshrq_n_v:
- return EmitARMBuiltinExpr(NEON::BI__builtin_neon_vshrq_n_v, E);
case NEON::BI__builtin_neon_vrshr_n_v:
case NEON::BI__builtin_neon_vrshrq_n_v:
Int = usgn ? Intrinsic::aarch64_neon_vurshr
@@ -4317,24 +4389,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
case NEON::BI__builtin_neon_vabsq_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vabs, Ty),
Ops, "vabs");
- case NEON::BI__builtin_neon_vaddhn_v: {
- llvm::VectorType *SrcTy =
- llvm::VectorType::getExtendedElementVectorType(VTy);
-
- // %sum = add <4 x i32> %lhs, %rhs
- Ops[0] = Builder.CreateBitCast(Ops[0], SrcTy);
- Ops[1] = Builder.CreateBitCast(Ops[1], SrcTy);
- Ops[0] = Builder.CreateAdd(Ops[0], Ops[1], "vaddhn");
-
- // %high = lshr <4 x i32> %sum, <i32 16, i32 16, i32 16, i32 16>
- Constant *ShiftAmt = ConstantInt::get(SrcTy->getElementType(),
- SrcTy->getScalarSizeInBits() / 2);
- ShiftAmt = ConstantVector::getSplat(VTy->getNumElements(), ShiftAmt);
- Ops[0] = Builder.CreateLShr(Ops[0], ShiftAmt, "vaddhn");
-
- // %res = trunc <4 x i32> %high to <4 x i16>
- return Builder.CreateTrunc(Ops[0], VTy, "vaddhn");
- }
case NEON::BI__builtin_neon_vcale_v:
std::swap(Ops[0], Ops[1]);
case NEON::BI__builtin_neon_vcage_v: {
@@ -4428,14 +4482,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
Function *F = CGM.getIntrinsic(Int, Tys);
return EmitNeonCall(F, Ops, "vcvt_n");
}
- case NEON::BI__builtin_neon_vhadd_v:
- case NEON::BI__builtin_neon_vhaddq_v:
- Int = usgn ? Intrinsic::arm_neon_vhaddu : Intrinsic::arm_neon_vhadds;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vhadd");
- case NEON::BI__builtin_neon_vhsub_v:
- case NEON::BI__builtin_neon_vhsubq_v:
- Int = usgn ? Intrinsic::arm_neon_vhsubu : Intrinsic::arm_neon_vhsubs;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vhsub");
case NEON::BI__builtin_neon_vld1_v:
case NEON::BI__builtin_neon_vld1q_v:
Ops.push_back(Align);
@@ -4600,14 +4646,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
Ops[0] = Builder.CreateBitCast(Ops[0], Ty);
return Builder.CreateStore(Ops[1], Ops[0]);
}
- case NEON::BI__builtin_neon_vmax_v:
- case NEON::BI__builtin_neon_vmaxq_v:
- Int = usgn ? Intrinsic::arm_neon_vmaxu : Intrinsic::arm_neon_vmaxs;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vmax");
- case NEON::BI__builtin_neon_vmin_v:
- case NEON::BI__builtin_neon_vminq_v:
- Int = usgn ? Intrinsic::arm_neon_vminu : Intrinsic::arm_neon_vmins;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vmin");
case NEON::BI__builtin_neon_vmovl_v: {
llvm::Type *DTy =llvm::VectorType::getTruncatedElementVectorType(VTy);
Ops[0] = Builder.CreateBitCast(Ops[0], DTy);
@@ -4620,15 +4658,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
Ops[0] = Builder.CreateBitCast(Ops[0], QTy);
return Builder.CreateTrunc(Ops[0], Ty, "vmovn");
}
- case NEON::BI__builtin_neon_vmull_v:
- // FIXME: the integer vmull operations could be emitted in terms of pure
- // LLVM IR (2 exts followed by a mul). Unfortunately LLVM has a habit of
- // hoisting the exts outside loops. Until global ISel comes along that can
- // see through such movement this leads to bad CodeGen. So we need an
- // intrinsic for now.
- Int = usgn ? Intrinsic::arm_neon_vmullu : Intrinsic::arm_neon_vmulls;
- Int = Type.isPoly() ? (unsigned)Intrinsic::arm_neon_vmullp : Int;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vmull");
case NEON::BI__builtin_neon_vpadal_v:
case NEON::BI__builtin_neon_vpadalq_v: {
Int = usgn ? Intrinsic::arm_neon_vpadalu : Intrinsic::arm_neon_vpadals;
@@ -4641,9 +4670,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
llvm::Type *Tys[2] = { Ty, NarrowTy };
return EmitNeonCall(CGM.getIntrinsic(Int, Tys), Ops, "vpadal");
}
- case NEON::BI__builtin_neon_vpadd_v:
- return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vpadd, Ty),
- Ops, "vpadd");
case NEON::BI__builtin_neon_vpaddl_v:
case NEON::BI__builtin_neon_vpaddlq_v: {
Int = usgn ? Intrinsic::arm_neon_vpaddlu : Intrinsic::arm_neon_vpaddls;
@@ -4655,49 +4681,10 @@ Value *CodeGenFunction::EmitARMBuiltinEx
llvm::Type *Tys[2] = { Ty, NarrowTy };
return EmitNeonCall(CGM.getIntrinsic(Int, Tys), Ops, "vpaddl");
}
- case NEON::BI__builtin_neon_vpmax_v:
- Int = usgn ? Intrinsic::arm_neon_vpmaxu : Intrinsic::arm_neon_vpmaxs;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vpmax");
- case NEON::BI__builtin_neon_vpmin_v:
- Int = usgn ? Intrinsic::arm_neon_vpminu : Intrinsic::arm_neon_vpmins;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vpmin");
case NEON::BI__builtin_neon_vqabs_v:
case NEON::BI__builtin_neon_vqabsq_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqabs, Ty),
Ops, "vqabs");
- case NEON::BI__builtin_neon_vqadd_v:
- case NEON::BI__builtin_neon_vqaddq_v:
- Int = usgn ? Intrinsic::arm_neon_vqaddu : Intrinsic::arm_neon_vqadds;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqadd");
- case NEON::BI__builtin_neon_vqdmlal_v: {
- SmallVector<Value *, 2> MulOps(Ops.begin() + 1, Ops.end());
- Value *Mul = EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqdmull, Ty),
- MulOps, "vqdmlal");
-
- SmallVector<Value *, 2> AddOps;
- AddOps.push_back(Ops[0]);
- AddOps.push_back(Mul);
- return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqadds, Ty),
- AddOps, "vqdmlal");
- }
- case NEON::BI__builtin_neon_vqdmlsl_v: {
- SmallVector<Value *, 2> MulOps(Ops.begin() + 1, Ops.end());
- Value *Mul = EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqdmull, Ty),
- MulOps, "vqdmlsl");
-
- SmallVector<Value *, 2> SubOps;
- SubOps.push_back(Ops[0]);
- SubOps.push_back(Mul);
- return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqsubs, Ty),
- SubOps, "vqdmlsl");
- }
- case NEON::BI__builtin_neon_vqdmulh_v:
- case NEON::BI__builtin_neon_vqdmulhq_v:
- return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqdmulh, Ty),
- Ops, "vqdmulh");
- case NEON::BI__builtin_neon_vqdmull_v:
- return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqdmull, Ty),
- Ops, "vqdmull");
case NEON::BI__builtin_neon_vqmovn_v:
Int = usgn ? Intrinsic::arm_neon_vqmovnu : Intrinsic::arm_neon_vqmovns;
return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqmovn");
@@ -4708,14 +4695,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
case NEON::BI__builtin_neon_vqnegq_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqneg, Ty),
Ops, "vqneg");
- case NEON::BI__builtin_neon_vqrdmulh_v:
- case NEON::BI__builtin_neon_vqrdmulhq_v:
- return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqrdmulh, Ty),
- Ops, "vqrdmulh");
- case NEON::BI__builtin_neon_vqrshl_v:
- case NEON::BI__builtin_neon_vqrshlq_v:
- Int = usgn ? Intrinsic::arm_neon_vqrshiftu : Intrinsic::arm_neon_vqrshifts;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqrshl");
case NEON::BI__builtin_neon_vqrshrn_n_v:
Int =
usgn ? Intrinsic::arm_neon_vqrshiftnu : Intrinsic::arm_neon_vqrshiftns;
@@ -4724,15 +4703,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
case NEON::BI__builtin_neon_vqrshrun_n_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqrshiftnsu, Ty),
Ops, "vqrshrun_n", 1, true);
- case NEON::BI__builtin_neon_vqshl_v:
- case NEON::BI__builtin_neon_vqshlq_v:
- Int = usgn ? Intrinsic::arm_neon_vqshiftu : Intrinsic::arm_neon_vqshifts;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqshl");
- case NEON::BI__builtin_neon_vqshl_n_v:
- case NEON::BI__builtin_neon_vqshlq_n_v:
- Int = usgn ? Intrinsic::arm_neon_vqshiftu : Intrinsic::arm_neon_vqshifts;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqshl_n",
- 1, false);
case NEON::BI__builtin_neon_vqshlu_n_v:
case NEON::BI__builtin_neon_vqshluq_n_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqshiftsu, Ty),
@@ -4744,25 +4714,10 @@ Value *CodeGenFunction::EmitARMBuiltinEx
case NEON::BI__builtin_neon_vqshrun_n_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vqshiftnsu, Ty),
Ops, "vqshrun_n", 1, true);
- case NEON::BI__builtin_neon_vqsub_v:
- case NEON::BI__builtin_neon_vqsubq_v:
- Int = usgn ? Intrinsic::arm_neon_vqsubu : Intrinsic::arm_neon_vqsubs;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vqsub");
- case NEON::BI__builtin_neon_vraddhn_v:
- return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vraddhn, Ty),
- Ops, "vraddhn");
case NEON::BI__builtin_neon_vrecpe_v:
case NEON::BI__builtin_neon_vrecpeq_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vrecpe, Ty),
Ops, "vrecpe");
- case NEON::BI__builtin_neon_vrhadd_v:
- case NEON::BI__builtin_neon_vrhaddq_v:
- Int = usgn ? Intrinsic::arm_neon_vrhaddu : Intrinsic::arm_neon_vrhadds;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vrhadd");
- case NEON::BI__builtin_neon_vrshl_v:
- case NEON::BI__builtin_neon_vrshlq_v:
- Int = usgn ? Intrinsic::arm_neon_vrshiftu : Intrinsic::arm_neon_vrshifts;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vrshl");
case NEON::BI__builtin_neon_vrshrn_n_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vrshiftn, Ty),
Ops, "vrshrn_n", 1, true);
@@ -4782,13 +4737,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
Int = usgn ? Intrinsic::arm_neon_vrshiftu : Intrinsic::arm_neon_vrshifts;
Ops[1] = Builder.CreateCall2(CGM.getIntrinsic(Int, Ty), Ops[1], Ops[2]);
return Builder.CreateAdd(Ops[0], Ops[1], "vrsra_n");
- case NEON::BI__builtin_neon_vrsubhn_v:
- return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vrsubhn, Ty),
- Ops, "vrsubhn");
- case NEON::BI__builtin_neon_vshl_v:
- case NEON::BI__builtin_neon_vshlq_v:
- Int = usgn ? Intrinsic::arm_neon_vshiftu : Intrinsic::arm_neon_vshifts;
- return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vshl");
case NEON::BI__builtin_neon_vshll_n_v:
Int = usgn ? Intrinsic::arm_neon_vshiftlu : Intrinsic::arm_neon_vshiftls;
return EmitNeonCall(CGM.getIntrinsic(Int, Ty), Ops, "vshll", 1);
@@ -4800,9 +4748,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
case NEON::BI__builtin_neon_vshrn_n_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vshiftn, Ty),
Ops, "vshrn_n", 1, true);
- case NEON::BI__builtin_neon_vshr_n_v:
- case NEON::BI__builtin_neon_vshrq_n_v:
- return EmitNeonRShiftImm(Ops[0], Ops[1], Ty, usgn, "vshr_n");
case NEON::BI__builtin_neon_vsri_n_v:
case NEON::BI__builtin_neon_vsriq_n_v:
rightShift = true;
@@ -4872,24 +4817,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
Ops.push_back(Align);
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vst4lane, Ty),
Ops, "");
- case NEON::BI__builtin_neon_vsubhn_v: {
- llvm::VectorType *SrcTy =
- llvm::VectorType::getExtendedElementVectorType(VTy);
-
- // %sum = add <4 x i32> %lhs, %rhs
- Ops[0] = Builder.CreateBitCast(Ops[0], SrcTy);
- Ops[1] = Builder.CreateBitCast(Ops[1], SrcTy);
- Ops[0] = Builder.CreateSub(Ops[0], Ops[1], "vsubhn");
-
- // %high = lshr <4 x i32> %sum, <i32 16, i32 16, i32 16, i32 16>
- Constant *ShiftAmt = ConstantInt::get(SrcTy->getElementType(),
- SrcTy->getScalarSizeInBits() / 2);
- ShiftAmt = ConstantVector::getSplat(VTy->getNumElements(), ShiftAmt);
- Ops[0] = Builder.CreateLShr(Ops[0], ShiftAmt, "vsubhn");
-
- // %res = trunc <4 x i32> %high to <4 x i16>
- return Builder.CreateTrunc(Ops[0], VTy, "vsubhn");
- }
case NEON::BI__builtin_neon_vtbl1_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vtbl1),
Ops, "vtbl1");
@@ -4914,15 +4841,6 @@ Value *CodeGenFunction::EmitARMBuiltinEx
case NEON::BI__builtin_neon_vtbx4_v:
return EmitNeonCall(CGM.getIntrinsic(Intrinsic::arm_neon_vtbx4),
Ops, "vtbx4");
- case NEON::BI__builtin_neon_vtst_v:
- case NEON::BI__builtin_neon_vtstq_v: {
- Ops[0] = Builder.CreateBitCast(Ops[0], Ty);
- Ops[1] = Builder.CreateBitCast(Ops[1], Ty);
- Ops[0] = Builder.CreateAnd(Ops[0], Ops[1]);
- Ops[0] = Builder.CreateICmp(ICmpInst::ICMP_NE, Ops[0],
- ConstantAggregateZero::get(Ty));
- return Builder.CreateSExt(Ops[0], Ty, "vtst");
- }
}
}
More information about the cfe-commits
mailing list