From 12842f9865a3d7f7d636afd3959851dd6c601aa1 Mon Sep 17 00:00:00 2001 From: Bob Wilson Date: Tue, 11 Aug 2009 05:39:44 +0000 Subject: Use vAny type to get rid of Neon intrinsics that differed only in whether the overloaded vector types allowed floating-point or integer vector elements. Most of these operations actually depend on the element type, so bitcasting was not an option. If you include the vpadd intrinsics that I updated earlier, this gets rid of 20 intrinsics. llvm-svn: 78646 --- llvm/test/CodeGen/ARM/vld3.ll | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'llvm/test/CodeGen/ARM/vld3.ll') diff --git a/llvm/test/CodeGen/ARM/vld3.ll b/llvm/test/CodeGen/ARM/vld3.ll index 49665f6de33..5e528c04560 100644 --- a/llvm/test/CodeGen/ARM/vld3.ll +++ b/llvm/test/CodeGen/ARM/vld3.ll @@ -8,7 +8,7 @@ define <8 x i8> @vld3i8(i8* %A) nounwind { ;CHECK: vld3i8: ;CHECK: vld3.8 - %tmp1 = call %struct.__builtin_neon_v8qi3 @llvm.arm.neon.vld3i.v8i8(i8* %A) + %tmp1 = call %struct.__builtin_neon_v8qi3 @llvm.arm.neon.vld3.v8i8(i8* %A) %tmp2 = extractvalue %struct.__builtin_neon_v8qi3 %tmp1, 0 %tmp3 = extractvalue %struct.__builtin_neon_v8qi3 %tmp1, 2 %tmp4 = add <8 x i8> %tmp2, %tmp3 @@ -18,7 +18,7 @@ define <8 x i8> @vld3i8(i8* %A) nounwind { define <4 x i16> @vld3i16(i16* %A) nounwind { ;CHECK: vld3i16: ;CHECK: vld3.16 - %tmp1 = call %struct.__builtin_neon_v4hi3 @llvm.arm.neon.vld3i.v4i16(i16* %A) + %tmp1 = call %struct.__builtin_neon_v4hi3 @llvm.arm.neon.vld3.v4i16(i16* %A) %tmp2 = extractvalue %struct.__builtin_neon_v4hi3 %tmp1, 0 %tmp3 = extractvalue %struct.__builtin_neon_v4hi3 %tmp1, 2 %tmp4 = add <4 x i16> %tmp2, %tmp3 @@ -28,7 +28,7 @@ define <4 x i16> @vld3i16(i16* %A) nounwind { define <2 x i32> @vld3i32(i32* %A) nounwind { ;CHECK: vld3i32: ;CHECK: vld3.32 - %tmp1 = call %struct.__builtin_neon_v2si3 @llvm.arm.neon.vld3i.v2i32(i32* %A) + %tmp1 = call %struct.__builtin_neon_v2si3 @llvm.arm.neon.vld3.v2i32(i32* %A) %tmp2 = extractvalue %struct.__builtin_neon_v2si3 %tmp1, 0 %tmp3 = extractvalue %struct.__builtin_neon_v2si3 %tmp1, 2 %tmp4 = add <2 x i32> %tmp2, %tmp3 @@ -38,14 +38,14 @@ define <2 x i32> @vld3i32(i32* %A) nounwind { define <2 x float> @vld3f(float* %A) nounwind { ;CHECK: vld3f: ;CHECK: vld3.32 - %tmp1 = call %struct.__builtin_neon_v2sf3 @llvm.arm.neon.vld3f.v2f32(float* %A) + %tmp1 = call %struct.__builtin_neon_v2sf3 @llvm.arm.neon.vld3.v2f32(float* %A) %tmp2 = extractvalue %struct.__builtin_neon_v2sf3 %tmp1, 0 %tmp3 = extractvalue %struct.__builtin_neon_v2sf3 %tmp1, 2 %tmp4 = add <2 x float> %tmp2, %tmp3 ret <2 x float> %tmp4 } -declare %struct.__builtin_neon_v8qi3 @llvm.arm.neon.vld3i.v8i8(i8*) nounwind readonly -declare %struct.__builtin_neon_v4hi3 @llvm.arm.neon.vld3i.v4i16(i8*) nounwind readonly -declare %struct.__builtin_neon_v2si3 @llvm.arm.neon.vld3i.v2i32(i8*) nounwind readonly -declare %struct.__builtin_neon_v2sf3 @llvm.arm.neon.vld3f.v2f32(i8*) nounwind readonly +declare %struct.__builtin_neon_v8qi3 @llvm.arm.neon.vld3.v8i8(i8*) nounwind readonly +declare %struct.__builtin_neon_v4hi3 @llvm.arm.neon.vld3.v4i16(i8*) nounwind readonly +declare %struct.__builtin_neon_v2si3 @llvm.arm.neon.vld3.v2i32(i8*) nounwind readonly +declare %struct.__builtin_neon_v2sf3 @llvm.arm.neon.vld3.v2f32(i8*) nounwind readonly -- cgit v1.2.3