6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2025-07-07 08:24:27 +00:00
Code Issues Projects Releases Wiki Activity

Use new EVT::vAny type to combine Neon intrinsics for VPADD.

Browse Source 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78632 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Bob Wilson
2009-08-11 01:15:26 +00:00
parent 61fc4cf7aa
commit f24bd401eb
3 changed files with 16 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
include/llvm/IntrinsicsARM.td
View File

@ -42,6 +42,9 @@ let TargetPrefix = "arm" in { // All intrinsics start with "llvm.arm.".
class Neon_2Arg_Float_Intrinsic class Neon_2Arg_Float_Intrinsic
: Intrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>, LLVMMatchType<0>], : Intrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>, LLVMMatchType<0>],
[IntrNoMem]>; [IntrNoMem]>;
class Neon_2Arg_Vector_Intrinsic
: Intrinsic<[llvm_anyvector_ty], [LLVMMatchType<0>, LLVMMatchType<0>],
[IntrNoMem]>;
class Neon_2Arg_Narrow_Intrinsic class Neon_2Arg_Narrow_Intrinsic
: Intrinsic<[llvm_anyint_ty], : Intrinsic<[llvm_anyint_ty],
[LLVMExtendedElementVectorType<0>, [LLVMExtendedElementVectorType<0>,
@ -194,8 +197,7 @@ def int_arm_neon_vabals : Neon_3Arg_Long_Intrinsic;
def int_arm_neon_vabalu : Neon_3Arg_Long_Intrinsic; def int_arm_neon_vabalu : Neon_3Arg_Long_Intrinsic;
// Vector Pairwise Add. // Vector Pairwise Add.
def int_arm_neon_vpaddi : Neon_2Arg_Intrinsic; def int_arm_neon_vpadd : Neon_2Arg_Vector_Intrinsic;
def int_arm_neon_vpaddf : Neon_2Arg_Float_Intrinsic;
// Vector Pairwise Add Long. // Vector Pairwise Add Long.
// Note: This is different than the other "long" NEON intrinsics because // Note: This is different than the other "long" NEON intrinsics because

8
lib/Target/ARM/ARMInstrNEON.td
View File

@ -1261,13 +1261,13 @@ def VMINfq : N3VQInt<0, 0, 0b10, 0b1111, 0, "vmin.f32", v4f32, v4f32,
// VPADD : Vector Pairwise Add // VPADD : Vector Pairwise Add
def VPADDi8 : N3VDInt<0, 0, 0b00, 0b1011, 1, "vpadd.i8", v8i8, v8i8, def VPADDi8 : N3VDInt<0, 0, 0b00, 0b1011, 1, "vpadd.i8", v8i8, v8i8,
int_arm_neon_vpaddi, 0>; int_arm_neon_vpadd, 0>;
def VPADDi16 : N3VDInt<0, 0, 0b01, 0b1011, 1, "vpadd.i16", v4i16, v4i16, def VPADDi16 : N3VDInt<0, 0, 0b01, 0b1011, 1, "vpadd.i16", v4i16, v4i16,
int_arm_neon_vpaddi, 0>; int_arm_neon_vpadd, 0>;
def VPADDi32 : N3VDInt<0, 0, 0b10, 0b1011, 1, "vpadd.i32", v2i32, v2i32, def VPADDi32 : N3VDInt<0, 0, 0b10, 0b1011, 1, "vpadd.i32", v2i32, v2i32,
int_arm_neon_vpaddi, 0>; int_arm_neon_vpadd, 0>;
def VPADDf : N3VDInt<1, 0, 0b00, 0b1101, 0, "vpadd.f32", v2f32, v2f32, def VPADDf : N3VDInt<1, 0, 0b00, 0b1101, 0, "vpadd.f32", v2f32, v2f32,
int_arm_neon_vpaddf, 0>; int_arm_neon_vpadd, 0>;
// VPADDL : Vector Pairwise Add Long // VPADDL : Vector Pairwise Add Long
defm VPADDLs : N2VPLInt_QHS<0b11, 0b11, 0b00, 0b00100, 0, "vpaddl.s", defm VPADDLs : N2VPLInt_QHS<0b11, 0b11, 0b00, 0b00100, 0, "vpaddl.s",

16
test/CodeGen/ARM/vpadd.ll
View File

@ -7,33 +7,33 @@
define <8 x i8> @vpaddi8(<8 x i8>* %A, <8 x i8>* %B) nounwind { define <8 x i8> @vpaddi8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
%tmp1 = load <8 x i8>* %A %tmp1 = load <8 x i8>* %A
%tmp2 = load <8 x i8>* %B %tmp2 = load <8 x i8>* %B
%tmp3 = call <8 x i8> @llvm.arm.neon.vpaddi.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2) %tmp3 = call <8 x i8> @llvm.arm.neon.vpadd.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
ret <8 x i8> %tmp3 ret <8 x i8> %tmp3
} }
define <4 x i16> @vpaddi16(<4 x i16>* %A, <4 x i16>* %B) nounwind { define <4 x i16> @vpaddi16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
%tmp1 = load <4 x i16>* %A %tmp1 = load <4 x i16>* %A
%tmp2 = load <4 x i16>* %B %tmp2 = load <4 x i16>* %B
%tmp3 = call <4 x i16> @llvm.arm.neon.vpaddi.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2) %tmp3 = call <4 x i16> @llvm.arm.neon.vpadd.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
ret <4 x i16> %tmp3 ret <4 x i16> %tmp3
} }
define <2 x i32> @vpaddi32(<2 x i32>* %A, <2 x i32>* %B) nounwind { define <2 x i32> @vpaddi32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
%tmp1 = load <2 x i32>* %A %tmp1 = load <2 x i32>* %A
%tmp2 = load <2 x i32>* %B %tmp2 = load <2 x i32>* %B
%tmp3 = call <2 x i32> @llvm.arm.neon.vpaddi.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2) %tmp3 = call <2 x i32> @llvm.arm.neon.vpadd.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
ret <2 x i32> %tmp3 ret <2 x i32> %tmp3
} }
define <2 x float> @vpaddf32(<2 x float>* %A, <2 x float>* %B) nounwind { define <2 x float> @vpaddf32(<2 x float>* %A, <2 x float>* %B) nounwind {
%tmp1 = load <2 x float>* %A %tmp1 = load <2 x float>* %A
%tmp2 = load <2 x float>* %B %tmp2 = load <2 x float>* %B
%tmp3 = call <2 x float> @llvm.arm.neon.vpaddf.v2f32(<2 x float> %tmp1, <2 x float> %tmp2) %tmp3 = call <2 x float> @llvm.arm.neon.vpadd.v2f32(<2 x float> %tmp1, <2 x float> %tmp2)
ret <2 x float> %tmp3 ret <2 x float> %tmp3
} }
declare <8 x i8> @llvm.arm.neon.vpaddi.v8i8(<8 x i8>, <8 x i8>) nounwind readnone declare <8 x i8> @llvm.arm.neon.vpadd.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpaddi.v4i16(<4 x i16>, <4 x i16>) nounwind readnone declare <4 x i16> @llvm.arm.neon.vpadd.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpaddi.v2i32(<2 x i32>, <2 x i32>) nounwind readnone declare <2 x i32> @llvm.arm.neon.vpadd.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
declare <2 x float> @llvm.arm.neon.vpaddf.v2f32(<2 x float>, <2 x float>) nounwind readnone declare <2 x float> @llvm.arm.neon.vpadd.v2f32(<2 x float>, <2 x float>) nounwind readnone