1 // REQUIRES: arm-registered-target 2 // RUN: %clang_cc1 -triple thumbv7-apple-darwin \ 3 // RUN: -target-abi apcs-gnu \ 4 // RUN: -target-cpu cortex-a8 \ 5 // RUN: -mfloat-abi soft \ 6 // RUN: -target-feature +soft-float-abi \ 7 // RUN: -ffreestanding \ 8 // RUN: -emit-llvm -w -O1 -o - %s | FileCheck %s 9 10 #include <arm_neon.h> 11 12 // Check that the vget_low/vget_high intrinsics generate a single shuffle 13 // without any bitcasting. 14 int8x8_t low_s8(int8x16_t a) { 15 // CHECK: shufflevector <16 x i8> %a, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> 16 return vget_low_s8(a); 17 } 18 19 uint8x8_t low_u8 (uint8x16_t a) { 20 // CHECK: shufflevector <16 x i8> %a, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> 21 return vget_low_u8(a); 22 } 23 24 int16x4_t low_s16( int16x8_t a) { 25 // CHECK: shufflevector <8 x i16> %a, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3> 26 return vget_low_s16(a); 27 } 28 29 uint16x4_t low_u16(uint16x8_t a) { 30 // CHECK: shufflevector <8 x i16> %a, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3> 31 return vget_low_u16(a); 32 } 33 34 int32x2_t low_s32( int32x4_t a) { 35 // CHECK: shufflevector <4 x i32> %a, <4 x i32> undef, <2 x i32> <i32 0, i32 1> 36 return vget_low_s32(a); 37 } 38 39 uint32x2_t low_u32(uint32x4_t a) { 40 // CHECK: shufflevector <4 x i32> %a, <4 x i32> undef, <2 x i32> <i32 0, i32 1> 41 return vget_low_u32(a); 42 } 43 44 int64x1_t low_s64( int64x2_t a) { 45 // CHECK: shufflevector <2 x i64> %a, <2 x i64> undef, <1 x i32> zeroinitializer 46 return vget_low_s64(a); 47 } 48 49 uint64x1_t low_u64(uint64x2_t a) { 50 // CHECK: shufflevector <2 x i64> %a, <2 x i64> undef, <1 x i32> zeroinitializer 51 return vget_low_u64(a); 52 } 53 54 poly8x8_t low_p8 (poly8x16_t a) { 55 // CHECK: shufflevector <16 x i8> %a, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> 56 return vget_low_p8(a); 57 } 58 59 poly16x4_t low_p16(poly16x8_t a) { 60 // CHECK: shufflevector <8 x i16> %a, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3> 61 return vget_low_p16(a); 62 } 63 64 float32x2_t low_f32(float32x4_t a) { 65 // CHECK: shufflevector <4 x float> %a, <4 x float> undef, <2 x i32> <i32 0, i32 1> 66 return vget_low_f32(a); 67 } 68 69 70 int8x8_t high_s8(int8x16_t a) { 71 // CHECK: shufflevector <16 x i8> %a, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15> 72 return vget_high_s8(a); 73 } 74 75 uint8x8_t high_u8 (uint8x16_t a) { 76 // CHECK: shufflevector <16 x i8> %a, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15> 77 return vget_high_u8(a); 78 } 79 80 int16x4_t high_s16( int16x8_t a) { 81 // CHECK: shufflevector <8 x i16> %a, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7> 82 return vget_high_s16(a); 83 } 84 85 uint16x4_t high_u16(uint16x8_t a) { 86 // CHECK: shufflevector <8 x i16> %a, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7> 87 return vget_high_u16(a); 88 } 89 90 int32x2_t high_s32( int32x4_t a) { 91 // CHECK: shufflevector <4 x i32> %a, <4 x i32> undef, <2 x i32> <i32 2, i32 3> 92 return vget_high_s32(a); 93 } 94 95 uint32x2_t high_u32(uint32x4_t a) { 96 // CHECK: shufflevector <4 x i32> %a, <4 x i32> undef, <2 x i32> <i32 2, i32 3> 97 return vget_high_u32(a); 98 } 99 100 int64x1_t high_s64( int64x2_t a) { 101 // CHECK: shufflevector <2 x i64> %a, <2 x i64> undef, <1 x i32> <i32 1> 102 return vget_high_s64(a); 103 } 104 105 uint64x1_t high_u64(uint64x2_t a) { 106 // CHECK: shufflevector <2 x i64> %a, <2 x i64> undef, <1 x i32> <i32 1> 107 return vget_high_u64(a); 108 } 109 110 poly8x8_t high_p8 (poly8x16_t a) { 111 // CHECK: shufflevector <16 x i8> %a, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15> 112 return vget_high_p8(a); 113 } 114 115 poly16x4_t high_p16(poly16x8_t a) { 116 // CHECK: shufflevector <8 x i16> %a, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7> 117 return vget_high_p16(a); 118 } 119 120 float32x2_t high_f32(float32x4_t a) { 121 // CHECK: shufflevector <4 x float> %a, <4 x float> undef, <2 x i32> <i32 2, i32 3> 122 return vget_high_f32(a); 123 } 124 125