1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206
|
; Test the MSA intrinsics that are encoded with the 3RF instruction format and
; use the result as a third operand and perform fixed-point operations.
; RUN: llc -march=mips -mattr=+msa,+fp64,+mips32r2 < %s | FileCheck %s
; RUN: llc -march=mipsel -mattr=+msa,+fp64,+mips32r2 < %s | FileCheck %s
@llvm_mips_madd_q_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
@llvm_mips_madd_q_h_ARG2 = global <8 x i16> <i16 8, i16 9, i16 10, i16 11, i16 12, i16 13, i16 14, i16 15>, align 16
@llvm_mips_madd_q_h_ARG3 = global <8 x i16> <i16 16, i16 17, i16 18, i16 19, i16 20, i16 21, i16 22, i16 23>, align 16
@llvm_mips_madd_q_h_RES = global <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, align 16
define void @llvm_mips_madd_q_h_test() nounwind {
entry:
%0 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG1
%1 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG2
%2 = load <8 x i16>, <8 x i16>* @llvm_mips_madd_q_h_ARG3
%3 = tail call <8 x i16> @llvm.mips.madd.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
store <8 x i16> %3, <8 x i16>* @llvm_mips_madd_q_h_RES
ret void
}
declare <8 x i16> @llvm.mips.madd.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
; CHECK: llvm_mips_madd_q_h_test:
; CHECK: ld.h
; CHECK: ld.h
; CHECK: ld.h
; CHECK: madd_q.h
; CHECK: st.h
; CHECK: .size llvm_mips_madd_q_h_test
;
@llvm_mips_madd_q_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
@llvm_mips_madd_q_w_ARG2 = global <4 x i32> <i32 4, i32 5, i32 6, i32 7>, align 16
@llvm_mips_madd_q_w_ARG3 = global <4 x i32> <i32 8, i32 9, i32 10, i32 11>, align 16
@llvm_mips_madd_q_w_RES = global <4 x i32> <i32 0, i32 0, i32 0, i32 0>, align 16
define void @llvm_mips_madd_q_w_test() nounwind {
entry:
%0 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG1
%1 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG2
%2 = load <4 x i32>, <4 x i32>* @llvm_mips_madd_q_w_ARG3
%3 = tail call <4 x i32> @llvm.mips.madd.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
store <4 x i32> %3, <4 x i32>* @llvm_mips_madd_q_w_RES
ret void
}
declare <4 x i32> @llvm.mips.madd.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
; CHECK: llvm_mips_madd_q_w_test:
; CHECK: ld.w
; CHECK: ld.w
; CHECK: ld.w
; CHECK: madd_q.w
; CHECK: st.w
; CHECK: .size llvm_mips_madd_q_w_test
;
@llvm_mips_maddr_q_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
@llvm_mips_maddr_q_h_ARG2 = global <8 x i16> <i16 8, i16 9, i16 10, i16 11, i16 12, i16 13, i16 14, i16 15>, align 16
@llvm_mips_maddr_q_h_ARG3 = global <8 x i16> <i16 16, i16 17, i16 18, i16 19, i16 20, i16 21, i16 22, i16 23>, align 16
@llvm_mips_maddr_q_h_RES = global <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, align 16
define void @llvm_mips_maddr_q_h_test() nounwind {
entry:
%0 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG1
%1 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG2
%2 = load <8 x i16>, <8 x i16>* @llvm_mips_maddr_q_h_ARG3
%3 = tail call <8 x i16> @llvm.mips.maddr.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
store <8 x i16> %3, <8 x i16>* @llvm_mips_maddr_q_h_RES
ret void
}
declare <8 x i16> @llvm.mips.maddr.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
; CHECK: llvm_mips_maddr_q_h_test:
; CHECK: ld.h
; CHECK: ld.h
; CHECK: ld.h
; CHECK: maddr_q.h
; CHECK: st.h
; CHECK: .size llvm_mips_maddr_q_h_test
;
@llvm_mips_maddr_q_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
@llvm_mips_maddr_q_w_ARG2 = global <4 x i32> <i32 4, i32 5, i32 6, i32 7>, align 16
@llvm_mips_maddr_q_w_ARG3 = global <4 x i32> <i32 8, i32 9, i32 10, i32 11>, align 16
@llvm_mips_maddr_q_w_RES = global <4 x i32> <i32 0, i32 0, i32 0, i32 0>, align 16
define void @llvm_mips_maddr_q_w_test() nounwind {
entry:
%0 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG1
%1 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG2
%2 = load <4 x i32>, <4 x i32>* @llvm_mips_maddr_q_w_ARG3
%3 = tail call <4 x i32> @llvm.mips.maddr.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
store <4 x i32> %3, <4 x i32>* @llvm_mips_maddr_q_w_RES
ret void
}
declare <4 x i32> @llvm.mips.maddr.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
; CHECK: llvm_mips_maddr_q_w_test:
; CHECK: ld.w
; CHECK: ld.w
; CHECK: ld.w
; CHECK: maddr_q.w
; CHECK: st.w
; CHECK: .size llvm_mips_maddr_q_w_test
;
@llvm_mips_msub_q_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
@llvm_mips_msub_q_h_ARG2 = global <8 x i16> <i16 8, i16 9, i16 10, i16 11, i16 12, i16 13, i16 14, i16 15>, align 16
@llvm_mips_msub_q_h_ARG3 = global <8 x i16> <i16 16, i16 17, i16 18, i16 19, i16 20, i16 21, i16 22, i16 23>, align 16
@llvm_mips_msub_q_h_RES = global <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, align 16
define void @llvm_mips_msub_q_h_test() nounwind {
entry:
%0 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG1
%1 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG2
%2 = load <8 x i16>, <8 x i16>* @llvm_mips_msub_q_h_ARG3
%3 = tail call <8 x i16> @llvm.mips.msub.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
store <8 x i16> %3, <8 x i16>* @llvm_mips_msub_q_h_RES
ret void
}
declare <8 x i16> @llvm.mips.msub.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
; CHECK: llvm_mips_msub_q_h_test:
; CHECK: ld.h
; CHECK: ld.h
; CHECK: ld.h
; CHECK: msub_q.h
; CHECK: st.h
; CHECK: .size llvm_mips_msub_q_h_test
;
@llvm_mips_msub_q_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
@llvm_mips_msub_q_w_ARG2 = global <4 x i32> <i32 4, i32 5, i32 6, i32 7>, align 16
@llvm_mips_msub_q_w_ARG3 = global <4 x i32> <i32 8, i32 9, i32 10, i32 11>, align 16
@llvm_mips_msub_q_w_RES = global <4 x i32> <i32 0, i32 0, i32 0, i32 0>, align 16
define void @llvm_mips_msub_q_w_test() nounwind {
entry:
%0 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG1
%1 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG2
%2 = load <4 x i32>, <4 x i32>* @llvm_mips_msub_q_w_ARG3
%3 = tail call <4 x i32> @llvm.mips.msub.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
store <4 x i32> %3, <4 x i32>* @llvm_mips_msub_q_w_RES
ret void
}
declare <4 x i32> @llvm.mips.msub.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
; CHECK: llvm_mips_msub_q_w_test:
; CHECK: ld.w
; CHECK: ld.w
; CHECK: ld.w
; CHECK: msub_q.w
; CHECK: st.w
; CHECK: .size llvm_mips_msub_q_w_test
;
@llvm_mips_msubr_q_h_ARG1 = global <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>, align 16
@llvm_mips_msubr_q_h_ARG2 = global <8 x i16> <i16 8, i16 9, i16 10, i16 11, i16 12, i16 13, i16 14, i16 15>, align 16
@llvm_mips_msubr_q_h_ARG3 = global <8 x i16> <i16 16, i16 17, i16 18, i16 19, i16 20, i16 21, i16 22, i16 23>, align 16
@llvm_mips_msubr_q_h_RES = global <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0, i16 0>, align 16
define void @llvm_mips_msubr_q_h_test() nounwind {
entry:
%0 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG1
%1 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG2
%2 = load <8 x i16>, <8 x i16>* @llvm_mips_msubr_q_h_ARG3
%3 = tail call <8 x i16> @llvm.mips.msubr.q.h(<8 x i16> %0, <8 x i16> %1, <8 x i16> %2)
store <8 x i16> %3, <8 x i16>* @llvm_mips_msubr_q_h_RES
ret void
}
declare <8 x i16> @llvm.mips.msubr.q.h(<8 x i16>, <8 x i16>, <8 x i16>) nounwind
; CHECK: llvm_mips_msubr_q_h_test:
; CHECK: ld.h
; CHECK: ld.h
; CHECK: ld.h
; CHECK: msubr_q.h
; CHECK: st.h
; CHECK: .size llvm_mips_msubr_q_h_test
;
@llvm_mips_msubr_q_w_ARG1 = global <4 x i32> <i32 0, i32 1, i32 2, i32 3>, align 16
@llvm_mips_msubr_q_w_ARG2 = global <4 x i32> <i32 4, i32 5, i32 6, i32 7>, align 16
@llvm_mips_msubr_q_w_ARG3 = global <4 x i32> <i32 8, i32 9, i32 10, i32 11>, align 16
@llvm_mips_msubr_q_w_RES = global <4 x i32> <i32 0, i32 0, i32 0, i32 0>, align 16
define void @llvm_mips_msubr_q_w_test() nounwind {
entry:
%0 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG1
%1 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG2
%2 = load <4 x i32>, <4 x i32>* @llvm_mips_msubr_q_w_ARG3
%3 = tail call <4 x i32> @llvm.mips.msubr.q.w(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
store <4 x i32> %3, <4 x i32>* @llvm_mips_msubr_q_w_RES
ret void
}
declare <4 x i32> @llvm.mips.msubr.q.w(<4 x i32>, <4 x i32>, <4 x i32>) nounwind
; CHECK: llvm_mips_msubr_q_w_test:
; CHECK: ld.w
; CHECK: ld.w
; CHECK: ld.w
; CHECK: msubr_q.w
; CHECK: st.w
; CHECK: .size llvm_mips_msubr_q_w_test
;
|