replica
/
hanchenye-llvm-project
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[AArch64] Refactor the Neon vector/scalar floating-point convert intrinsics so 

that they use float/double rather than the vector equivalents when appropriate.

llvm-svn: 196931
This commit is contained in:
Chad Rosier 2013-12-10 16:11:55 +00:00
parent 7a9bba442f
commit 58f6a1fee7
1 changed files with 11 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
clang/lib/CodeGen/CGBuiltin.cpp
View File

@ -2144,38 +2144,38 @@ static Value *EmitAArch64ScalarBuiltinExpr(CodeGenFunction &CGF,
// Scalar Floating-point Converts
case AArch64::BI__builtin_neon_vcvtxd_f32_f64:
Int = Intrinsic::aarch64_neon_fcvtxn;
s = "vcvtxn"; OverloadCvtInt = true; break;
s = "vcvtxn"; OverloadCvtInt = false; break;
case AArch64::BI__builtin_neon_vcvtas_s32_f32:
case AArch64::BI__builtin_neon_vcvtad_s64_f64:
Int = Intrinsic::arm_neon_vcvtas;
Int = Intrinsic::aarch64_neon_fcvtas;
s = "vcvtas"; OverloadCvtInt = true; break;
case AArch64::BI__builtin_neon_vcvtas_u32_f32:
case AArch64::BI__builtin_neon_vcvtad_u64_f64:
Int = Intrinsic::arm_neon_vcvtau;
Int = Intrinsic::aarch64_neon_fcvtau;
s = "vcvtau"; OverloadCvtInt = true; break;
case AArch64::BI__builtin_neon_vcvtms_s32_f32:
case AArch64::BI__builtin_neon_vcvtmd_s64_f64:
Int = Intrinsic::arm_neon_vcvtms;
Int = Intrinsic::aarch64_neon_fcvtms;
s = "vcvtms"; OverloadCvtInt = true; break;
case AArch64::BI__builtin_neon_vcvtms_u32_f32:
case AArch64::BI__builtin_neon_vcvtmd_u64_f64:
Int = Intrinsic::arm_neon_vcvtmu;
Int = Intrinsic::aarch64_neon_fcvtmu;
s = "vcvtmu"; OverloadCvtInt = true; break;
case AArch64::BI__builtin_neon_vcvtns_s32_f32:
case AArch64::BI__builtin_neon_vcvtnd_s64_f64:
Int = Intrinsic::arm_neon_vcvtns;
Int = Intrinsic::aarch64_neon_fcvtns;
s = "vcvtns"; OverloadCvtInt = true; break;
case AArch64::BI__builtin_neon_vcvtns_u32_f32:
case AArch64::BI__builtin_neon_vcvtnd_u64_f64:
Int = Intrinsic::arm_neon_vcvtnu;
Int = Intrinsic::aarch64_neon_fcvtnu;
s = "vcvtnu"; OverloadCvtInt = true; break;
case AArch64::BI__builtin_neon_vcvtps_s32_f32:
case AArch64::BI__builtin_neon_vcvtpd_s64_f64:
Int = Intrinsic::arm_neon_vcvtps;
Int = Intrinsic::aarch64_neon_fcvtps;
s = "vcvtps"; OverloadCvtInt = true; break;
case AArch64::BI__builtin_neon_vcvtps_u32_f32:
case AArch64::BI__builtin_neon_vcvtpd_u64_f64:
Int = Intrinsic::arm_neon_vcvtpu;
Int = Intrinsic::aarch64_neon_fcvtpu;
s = "vcvtpu"; OverloadCvtInt = true; break;
case AArch64::BI__builtin_neon_vcvts_s32_f32:
case AArch64::BI__builtin_neon_vcvtd_s64_f64:
@ -2637,8 +2637,7 @@ static Value *EmitAArch64ScalarBuiltinExpr(CodeGenFunction &CGF,
llvm::VectorType *VTy = llvm::VectorType::get(Ty, 1);
Tys.push_back(VTy);
Ty = CGF.ConvertType(Arg->getType());
VTy = llvm::VectorType::get(Ty, 1);
Tys.push_back(VTy);
Tys.push_back(Ty);
F = CGF.CGM.getIntrinsic(Int, Tys);
} else
@ -3747,7 +3746,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID,
llvm::Type *EltTy = FloatTy;
llvm::Type *ResTy = llvm::VectorType::get(EltTy, 2);
llvm::Type *Tys[2] = { ResTy, Ty };
Int = Intrinsic::aarch64_neon_fcvtxn;
Int = Intrinsic::aarch64_neon_vcvtxn;
return EmitNeonCall(CGM.getIntrinsic(Int, Tys), Ops, "vcvtx_f32_f64");
}
case AArch64::BI__builtin_neon_vcvt_f64_f32: {