mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-23 11:13:28 +01:00
b9ecf8a3ee
This commit changes the interface of the vld[1234], vld[234]lane, and vst[1234], vst[234]lane ARM neon intrinsics and associates an address space with the pointer that these intrinsics take. This changes, e.g., <2 x i32> @llvm.arm.neon.vld1.v2i32(i8*, i32) to <2 x i32> @llvm.arm.neon.vld1.v2i32.p0i8(i8*, i32) This change ensures that address spaces are fully taken into account in the ARM target during lowering of interleaved loads and stores. Differential Revision: http://reviews.llvm.org/D12985 llvm-svn: 248887
141 lines
4.7 KiB
LLVM
141 lines
4.7 KiB
LLVM
; RUN: llc -mtriple=arm-eabi -mattr=+neon %s -o - | FileCheck %s
|
|
|
|
define void @vst1i8(i8* %A, <8 x i8>* %B) nounwind {
|
|
;CHECK-LABEL: vst1i8:
|
|
;Check the alignment value. Max for this instruction is 64 bits:
|
|
;CHECK: vst1.8 {d16}, [r0:64]
|
|
%tmp1 = load <8 x i8>, <8 x i8>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v8i8(i8* %A, <8 x i8> %tmp1, i32 16)
|
|
ret void
|
|
}
|
|
|
|
define void @vst1i16(i16* %A, <4 x i16>* %B) nounwind {
|
|
;CHECK-LABEL: vst1i16:
|
|
;CHECK: vst1.16
|
|
%tmp0 = bitcast i16* %A to i8*
|
|
%tmp1 = load <4 x i16>, <4 x i16>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v4i16(i8* %tmp0, <4 x i16> %tmp1, i32 1)
|
|
ret void
|
|
}
|
|
|
|
define void @vst1i32(i32* %A, <2 x i32>* %B) nounwind {
|
|
;CHECK-LABEL: vst1i32:
|
|
;CHECK: vst1.32
|
|
%tmp0 = bitcast i32* %A to i8*
|
|
%tmp1 = load <2 x i32>, <2 x i32>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v2i32(i8* %tmp0, <2 x i32> %tmp1, i32 1)
|
|
ret void
|
|
}
|
|
|
|
define void @vst1f(float* %A, <2 x float>* %B) nounwind {
|
|
;CHECK-LABEL: vst1f:
|
|
;CHECK: vst1.32
|
|
%tmp0 = bitcast float* %A to i8*
|
|
%tmp1 = load <2 x float>, <2 x float>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v2f32(i8* %tmp0, <2 x float> %tmp1, i32 1)
|
|
ret void
|
|
}
|
|
|
|
;Check for a post-increment updating store.
|
|
define void @vst1f_update(float** %ptr, <2 x float>* %B) nounwind {
|
|
;CHECK-LABEL: vst1f_update:
|
|
;CHECK: vst1.32 {d16}, [r1]!
|
|
%A = load float*, float** %ptr
|
|
%tmp0 = bitcast float* %A to i8*
|
|
%tmp1 = load <2 x float>, <2 x float>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v2f32(i8* %tmp0, <2 x float> %tmp1, i32 1)
|
|
%tmp2 = getelementptr float, float* %A, i32 2
|
|
store float* %tmp2, float** %ptr
|
|
ret void
|
|
}
|
|
|
|
define void @vst1i64(i64* %A, <1 x i64>* %B) nounwind {
|
|
;CHECK-LABEL: vst1i64:
|
|
;CHECK: vst1.64
|
|
%tmp0 = bitcast i64* %A to i8*
|
|
%tmp1 = load <1 x i64>, <1 x i64>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v1i64(i8* %tmp0, <1 x i64> %tmp1, i32 1)
|
|
ret void
|
|
}
|
|
|
|
define void @vst1Qi8(i8* %A, <16 x i8>* %B) nounwind {
|
|
;CHECK-LABEL: vst1Qi8:
|
|
;Check the alignment value. Max for this instruction is 128 bits:
|
|
;CHECK: vst1.8 {d16, d17}, [r0:64]
|
|
%tmp1 = load <16 x i8>, <16 x i8>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v16i8(i8* %A, <16 x i8> %tmp1, i32 8)
|
|
ret void
|
|
}
|
|
|
|
define void @vst1Qi16(i16* %A, <8 x i16>* %B) nounwind {
|
|
;CHECK-LABEL: vst1Qi16:
|
|
;Check the alignment value. Max for this instruction is 128 bits:
|
|
;CHECK: vst1.16 {d16, d17}, [r0:128]
|
|
%tmp0 = bitcast i16* %A to i8*
|
|
%tmp1 = load <8 x i16>, <8 x i16>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v8i16(i8* %tmp0, <8 x i16> %tmp1, i32 32)
|
|
ret void
|
|
}
|
|
|
|
;Check for a post-increment updating store with register increment.
|
|
define void @vst1Qi16_update(i16** %ptr, <8 x i16>* %B, i32 %inc) nounwind {
|
|
;CHECK-LABEL: vst1Qi16_update:
|
|
;CHECK: vst1.16 {d16, d17}, [r1:64], r2
|
|
%A = load i16*, i16** %ptr
|
|
%tmp0 = bitcast i16* %A to i8*
|
|
%tmp1 = load <8 x i16>, <8 x i16>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v8i16(i8* %tmp0, <8 x i16> %tmp1, i32 8)
|
|
%tmp2 = getelementptr i16, i16* %A, i32 %inc
|
|
store i16* %tmp2, i16** %ptr
|
|
ret void
|
|
}
|
|
|
|
define void @vst1Qi32(i32* %A, <4 x i32>* %B) nounwind {
|
|
;CHECK-LABEL: vst1Qi32:
|
|
;CHECK: vst1.32
|
|
%tmp0 = bitcast i32* %A to i8*
|
|
%tmp1 = load <4 x i32>, <4 x i32>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v4i32(i8* %tmp0, <4 x i32> %tmp1, i32 1)
|
|
ret void
|
|
}
|
|
|
|
define void @vst1Qf(float* %A, <4 x float>* %B) nounwind {
|
|
;CHECK-LABEL: vst1Qf:
|
|
;CHECK: vst1.32
|
|
%tmp0 = bitcast float* %A to i8*
|
|
%tmp1 = load <4 x float>, <4 x float>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v4f32(i8* %tmp0, <4 x float> %tmp1, i32 1)
|
|
ret void
|
|
}
|
|
|
|
define void @vst1Qi64(i64* %A, <2 x i64>* %B) nounwind {
|
|
;CHECK-LABEL: vst1Qi64:
|
|
;CHECK: vst1.64
|
|
%tmp0 = bitcast i64* %A to i8*
|
|
%tmp1 = load <2 x i64>, <2 x i64>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v2i64(i8* %tmp0, <2 x i64> %tmp1, i32 1)
|
|
ret void
|
|
}
|
|
|
|
define void @vst1Qf64(double* %A, <2 x double>* %B) nounwind {
|
|
;CHECK-LABEL: vst1Qf64:
|
|
;CHECK: vst1.64
|
|
%tmp0 = bitcast double* %A to i8*
|
|
%tmp1 = load <2 x double>, <2 x double>* %B
|
|
call void @llvm.arm.neon.vst1.p0i8.v2f64(i8* %tmp0, <2 x double> %tmp1, i32 1)
|
|
ret void
|
|
}
|
|
|
|
declare void @llvm.arm.neon.vst1.p0i8.v8i8(i8*, <8 x i8>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v4i16(i8*, <4 x i16>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v2i32(i8*, <2 x i32>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v2f32(i8*, <2 x float>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v1i64(i8*, <1 x i64>, i32) nounwind
|
|
|
|
declare void @llvm.arm.neon.vst1.p0i8.v16i8(i8*, <16 x i8>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v8i16(i8*, <8 x i16>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v4i32(i8*, <4 x i32>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v4f32(i8*, <4 x float>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v2i64(i8*, <2 x i64>, i32) nounwind
|
|
declare void @llvm.arm.neon.vst1.p0i8.v2f64(i8*, <2 x double>, i32) nounwind
|