1 ; RUN: llc < %s -march=cellspu > %t1.s 2 ; RUN: grep ceq %t1.s | count 30 3 ; RUN: grep ceqb %t1.s | count 10 4 ; RUN: grep ceqhi %t1.s | count 5 5 ; RUN: grep ceqi %t1.s | count 5 6 ; RUN: grep cgt %t1.s | count 30 7 ; RUN: grep cgtb %t1.s | count 10 8 ; RUN: grep cgthi %t1.s | count 5 9 ; RUN: grep cgti %t1.s | count 5 10 target datalayout = "E-p:32:32:128-f64:64:128-f32:32:128-i64:32:128-i32:32:128-i16:16:128-i8:8:128-i1:8:128-a0:0:128-v128:128:128-s0:128:128" 11 target triple = "spu" 12 13 declare <4 x i32> @llvm.spu.si.shli(<4 x i32>, i8) 14 15 declare <4 x i32> @llvm.spu.si.ceq(<4 x i32>, <4 x i32>) 16 declare <16 x i8> @llvm.spu.si.ceqb(<16 x i8>, <16 x i8>) 17 declare <8 x i16> @llvm.spu.si.ceqh(<8 x i16>, <8 x i16>) 18 declare <4 x i32> @llvm.spu.si.ceqi(<4 x i32>, i16) 19 declare <8 x i16> @llvm.spu.si.ceqhi(<8 x i16>, i16) 20 declare <16 x i8> @llvm.spu.si.ceqbi(<16 x i8>, i8) 21 22 declare <4 x i32> @llvm.spu.si.cgt(<4 x i32>, <4 x i32>) 23 declare <16 x i8> @llvm.spu.si.cgtb(<16 x i8>, <16 x i8>) 24 declare <8 x i16> @llvm.spu.si.cgth(<8 x i16>, <8 x i16>) 25 declare <4 x i32> @llvm.spu.si.cgti(<4 x i32>, i16) 26 declare <8 x i16> @llvm.spu.si.cgthi(<8 x i16>, i16) 27 declare <16 x i8> @llvm.spu.si.cgtbi(<16 x i8>, i8) 28 29 declare <4 x i32> @llvm.spu.si.clgt(<4 x i32>, <4 x i32>) 30 declare <16 x i8> @llvm.spu.si.clgtb(<16 x i8>, <16 x i8>) 31 declare <8 x i16> @llvm.spu.si.clgth(<8 x i16>, <8 x i16>) 32 declare <4 x i32> @llvm.spu.si.clgti(<4 x i32>, i16) 33 declare <8 x i16> @llvm.spu.si.clgthi(<8 x i16>, i16) 34 declare <16 x i8> @llvm.spu.si.clgtbi(<16 x i8>, i8) 35 36 37 38 define <4 x i32> @test(<4 x i32> %A) { 39 call <4 x i32> @llvm.spu.si.shli(<4 x i32> %A, i8 3) 40 %Y = bitcast <4 x i32> %1 to <4 x i32> 41 ret <4 x i32> %Y 42 } 43 44 define <4 x i32> @ceqtest(<4 x i32> %A, <4 x i32> %B) { 45 call <4 x i32> @llvm.spu.si.ceq(<4 x i32> %A, <4 x i32> %B) 46 %Y = bitcast <4 x i32> %1 to <4 x i32> 47 ret <4 x i32> %Y 48 } 49 50 define <8 x i16> @ceqhtest(<8 x i16> %A, <8 x i16> %B) { 51 call <8 x i16> @llvm.spu.si.ceqh(<8 x i16> %A, <8 x i16> %B) 52 %Y = bitcast <8 x i16> %1 to <8 x i16> 53 ret <8 x i16> %Y 54 } 55 56 define <16 x i8> @ceqbtest(<16 x i8> %A, <16 x i8> %B) { 57 call <16 x i8> @llvm.spu.si.ceqb(<16 x i8> %A, <16 x i8> %B) 58 %Y = bitcast <16 x i8> %1 to <16 x i8> 59 ret <16 x i8> %Y 60 } 61 62 define <4 x i32> @ceqitest(<4 x i32> %A) { 63 call <4 x i32> @llvm.spu.si.ceqi(<4 x i32> %A, i16 65) 64 %Y = bitcast <4 x i32> %1 to <4 x i32> 65 ret <4 x i32> %Y 66 } 67 68 define <8 x i16> @ceqhitest(<8 x i16> %A) { 69 call <8 x i16> @llvm.spu.si.ceqhi(<8 x i16> %A, i16 65) 70 %Y = bitcast <8 x i16> %1 to <8 x i16> 71 ret <8 x i16> %Y 72 } 73 74 define <16 x i8> @ceqbitest(<16 x i8> %A) { 75 call <16 x i8> @llvm.spu.si.ceqbi(<16 x i8> %A, i8 65) 76 %Y = bitcast <16 x i8> %1 to <16 x i8> 77 ret <16 x i8> %Y 78 } 79 80 define <4 x i32> @cgttest(<4 x i32> %A, <4 x i32> %B) { 81 call <4 x i32> @llvm.spu.si.cgt(<4 x i32> %A, <4 x i32> %B) 82 %Y = bitcast <4 x i32> %1 to <4 x i32> 83 ret <4 x i32> %Y 84 } 85 86 define <8 x i16> @cgthtest(<8 x i16> %A, <8 x i16> %B) { 87 call <8 x i16> @llvm.spu.si.cgth(<8 x i16> %A, <8 x i16> %B) 88 %Y = bitcast <8 x i16> %1 to <8 x i16> 89 ret <8 x i16> %Y 90 } 91 92 define <16 x i8> @cgtbtest(<16 x i8> %A, <16 x i8> %B) { 93 call <16 x i8> @llvm.spu.si.cgtb(<16 x i8> %A, <16 x i8> %B) 94 %Y = bitcast <16 x i8> %1 to <16 x i8> 95 ret <16 x i8> %Y 96 } 97 98 define <4 x i32> @cgtitest(<4 x i32> %A) { 99 call <4 x i32> @llvm.spu.si.cgti(<4 x i32> %A, i16 65) 100 %Y = bitcast <4 x i32> %1 to <4 x i32> 101 ret <4 x i32> %Y 102 } 103 104 define <8 x i16> @cgthitest(<8 x i16> %A) { 105 call <8 x i16> @llvm.spu.si.cgthi(<8 x i16> %A, i16 65) 106 %Y = bitcast <8 x i16> %1 to <8 x i16> 107 ret <8 x i16> %Y 108 } 109 110 define <16 x i8> @cgtbitest(<16 x i8> %A) { 111 call <16 x i8> @llvm.spu.si.cgtbi(<16 x i8> %A, i8 65) 112 %Y = bitcast <16 x i8> %1 to <16 x i8> 113 ret <16 x i8> %Y 114 } 115 116 define <4 x i32> @clgttest(<4 x i32> %A, <4 x i32> %B) { 117 call <4 x i32> @llvm.spu.si.clgt(<4 x i32> %A, <4 x i32> %B) 118 %Y = bitcast <4 x i32> %1 to <4 x i32> 119 ret <4 x i32> %Y 120 } 121 122 define <8 x i16> @clgthtest(<8 x i16> %A, <8 x i16> %B) { 123 call <8 x i16> @llvm.spu.si.clgth(<8 x i16> %A, <8 x i16> %B) 124 %Y = bitcast <8 x i16> %1 to <8 x i16> 125 ret <8 x i16> %Y 126 } 127 128 define <16 x i8> @clgtbtest(<16 x i8> %A, <16 x i8> %B) { 129 call <16 x i8> @llvm.spu.si.clgtb(<16 x i8> %A, <16 x i8> %B) 130 %Y = bitcast <16 x i8> %1 to <16 x i8> 131 ret <16 x i8> %Y 132 } 133 134 define <4 x i32> @clgtitest(<4 x i32> %A) { 135 call <4 x i32> @llvm.spu.si.clgti(<4 x i32> %A, i16 65) 136 %Y = bitcast <4 x i32> %1 to <4 x i32> 137 ret <4 x i32> %Y 138 } 139 140 define <8 x i16> @clgthitest(<8 x i16> %A) { 141 call <8 x i16> @llvm.spu.si.clgthi(<8 x i16> %A, i16 65) 142 %Y = bitcast <8 x i16> %1 to <8 x i16> 143 ret <8 x i16> %Y 144 } 145 146 define <16 x i8> @clgtbitest(<16 x i8> %A) { 147 call <16 x i8> @llvm.spu.si.clgtbi(<16 x i8> %A, i8 65) 148 %Y = bitcast <16 x i8> %1 to <16 x i8> 149 ret <16 x i8> %Y 150 } 151
