270 lines · plain
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py2; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+xtheadvdot \3; RUN: -verify-machineinstrs | FileCheck %s4; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+xtheadvdot \5; RUN: -verify-machineinstrs | FileCheck %s6 7define <vscale x 1 x i32> @intrinsic_th_vmaqasu_vv_nxv1i32_nxv4i8_nxv4i8(<vscale x 1 x i32> %0, <vscale x 4 x i8> %1, <vscale x 4 x i8> %2, iXLen %3) nounwind {8; CHECK-LABEL: intrinsic_th_vmaqasu_vv_nxv1i32_nxv4i8_nxv4i8:9; CHECK: # %bb.0: # %entry10; CHECK-NEXT: vsetvli zero, a0, e32, mf2, tu, ma11; CHECK-NEXT: th.vmaqasu.vv v8, v9, v1012; CHECK-NEXT: ret13entry:14 %a = call <vscale x 1 x i32> @llvm.riscv.th.vmaqasu.nxv1i32.nxv4i8(15 <vscale x 1 x i32> %0,16 <vscale x 4 x i8> %1,17 <vscale x 4 x i8> %2,18 iXLen %3, iXLen 0)19 20 ret <vscale x 1 x i32> %a21}22 23define <vscale x 1 x i32> @intrinsic_th_vmaqasu_mask_vv_nxv1i32_nxv4i8_nxv4i8(<vscale x 1 x i32> %0, <vscale x 4 x i8> %1, <vscale x 4 x i8> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {24; CHECK-LABEL: intrinsic_th_vmaqasu_mask_vv_nxv1i32_nxv4i8_nxv4i8:25; CHECK: # %bb.0: # %entry26; CHECK-NEXT: vsetvli zero, a0, e32, mf2, tu, mu27; CHECK-NEXT: th.vmaqasu.vv v8, v9, v10, v0.t28; CHECK-NEXT: ret29entry:30 %a = call <vscale x 1 x i32> @llvm.riscv.th.vmaqasu.mask.nxv1i32.nxv4i8(31 <vscale x 1 x i32> %0,32 <vscale x 4 x i8> %1,33 <vscale x 4 x i8> %2,34 <vscale x 4 x i1> %3,35 iXLen %4, iXLen 0)36 37 ret <vscale x 1 x i32> %a38}39 40define <vscale x 2 x i32> @intrinsic_th_vmaqasu_vv_nxv2i32_nxv8i8_nxv8i8(<vscale x 2 x i32> %0, <vscale x 8 x i8> %1, <vscale x 8 x i8> %2, iXLen %3) nounwind {41; CHECK-LABEL: intrinsic_th_vmaqasu_vv_nxv2i32_nxv8i8_nxv8i8:42; CHECK: # %bb.0: # %entry43; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, ma44; CHECK-NEXT: th.vmaqasu.vv v8, v9, v1045; CHECK-NEXT: ret46entry:47 %a = call <vscale x 2 x i32> @llvm.riscv.th.vmaqasu.nxv2i32.nxv8i8(48 <vscale x 2 x i32> %0,49 <vscale x 8 x i8> %1,50 <vscale x 8 x i8> %2,51 iXLen %3, iXLen 0)52 53 ret <vscale x 2 x i32> %a54}55 56define <vscale x 2 x i32> @intrinsic_th_vmaqasu_mask_vv_nxv2i32_nxv8i8_nxv8i8(<vscale x 2 x i32> %0, <vscale x 8 x i8> %1, <vscale x 8 x i8> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {57; CHECK-LABEL: intrinsic_th_vmaqasu_mask_vv_nxv2i32_nxv8i8_nxv8i8:58; CHECK: # %bb.0: # %entry59; CHECK-NEXT: vsetvli zero, a0, e32, m1, tu, mu60; CHECK-NEXT: th.vmaqasu.vv v8, v9, v10, v0.t61; CHECK-NEXT: ret62entry:63 %a = call <vscale x 2 x i32> @llvm.riscv.th.vmaqasu.mask.nxv2i32.nxv8i8(64 <vscale x 2 x i32> %0,65 <vscale x 8 x i8> %1,66 <vscale x 8 x i8> %2,67 <vscale x 8 x i1> %3,68 iXLen %4, iXLen 0)69 70 ret <vscale x 2 x i32> %a71}72 73define <vscale x 4 x i32> @intrinsic_th_vmaqasu_vv_nxv4i32_nxv16i8_nxv16i8(<vscale x 4 x i32> %0, <vscale x 16 x i8> %1, <vscale x 16 x i8> %2, iXLen %3) nounwind {74; CHECK-LABEL: intrinsic_th_vmaqasu_vv_nxv4i32_nxv16i8_nxv16i8:75; CHECK: # %bb.0: # %entry76; CHECK-NEXT: vsetvli zero, a0, e32, m2, tu, ma77; CHECK-NEXT: th.vmaqasu.vv v8, v10, v1278; CHECK-NEXT: ret79entry:80 %a = call <vscale x 4 x i32> @llvm.riscv.th.vmaqasu.nxv4i32.nxv16i8(81 <vscale x 4 x i32> %0,82 <vscale x 16 x i8> %1,83 <vscale x 16 x i8> %2,84 iXLen %3, iXLen 0)85 86 ret <vscale x 4 x i32> %a87}88 89define <vscale x 4 x i32> @intrinsic_th_vmaqasu_mask_vv_nxv4i32_nxv16i8_nxv16i8(<vscale x 4 x i32> %0, <vscale x 16 x i8> %1, <vscale x 16 x i8> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {90; CHECK-LABEL: intrinsic_th_vmaqasu_mask_vv_nxv4i32_nxv16i8_nxv16i8:91; CHECK: # %bb.0: # %entry92; CHECK-NEXT: vsetvli zero, a0, e32, m2, tu, mu93; CHECK-NEXT: th.vmaqasu.vv v8, v10, v12, v0.t94; CHECK-NEXT: ret95entry:96 %a = call <vscale x 4 x i32> @llvm.riscv.th.vmaqasu.mask.nxv4i32.nxv16i8(97 <vscale x 4 x i32> %0,98 <vscale x 16 x i8> %1,99 <vscale x 16 x i8> %2,100 <vscale x 16 x i1> %3,101 iXLen %4, iXLen 0)102 103 ret <vscale x 4 x i32> %a104}105 106define <vscale x 8 x i32> @intrinsic_th_vmaqasu_vv_nxv8i32_nxv32i8_nxv32i8(<vscale x 8 x i32> %0, <vscale x 32 x i8> %1, <vscale x 32 x i8> %2, iXLen %3) nounwind {107; CHECK-LABEL: intrinsic_th_vmaqasu_vv_nxv8i32_nxv32i8_nxv32i8:108; CHECK: # %bb.0: # %entry109; CHECK-NEXT: vsetvli zero, a0, e32, m4, tu, ma110; CHECK-NEXT: th.vmaqasu.vv v8, v12, v16111; CHECK-NEXT: ret112entry:113 %a = call <vscale x 8 x i32> @llvm.riscv.th.vmaqasu.nxv8i32.nxv32i8(114 <vscale x 8 x i32> %0,115 <vscale x 32 x i8> %1,116 <vscale x 32 x i8> %2,117 iXLen %3, iXLen 0)118 119 ret <vscale x 8 x i32> %a120}121 122define <vscale x 8 x i32> @intrinsic_th_vmaqasu_mask_vv_nxv8i32_nxv32i8_nxv32i8(<vscale x 8 x i32> %0, <vscale x 32 x i8> %1, <vscale x 32 x i8> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {123; CHECK-LABEL: intrinsic_th_vmaqasu_mask_vv_nxv8i32_nxv32i8_nxv32i8:124; CHECK: # %bb.0: # %entry125; CHECK-NEXT: vsetvli zero, a0, e32, m4, tu, mu126; CHECK-NEXT: th.vmaqasu.vv v8, v12, v16, v0.t127; CHECK-NEXT: ret128entry:129 %a = call <vscale x 8 x i32> @llvm.riscv.th.vmaqasu.mask.nxv8i32.nxv32i8(130 <vscale x 8 x i32> %0,131 <vscale x 32 x i8> %1,132 <vscale x 32 x i8> %2,133 <vscale x 32 x i1> %3,134 iXLen %4, iXLen 0)135 136 ret <vscale x 8 x i32> %a137}138 139define <vscale x 1 x i32> @intrinsic_th_vmaqasu_vx_nxv1i32_i8_nxv4i8(<vscale x 1 x i32> %0, i8 %1, <vscale x 4 x i8> %2, iXLen %3) nounwind {140; CHECK-LABEL: intrinsic_th_vmaqasu_vx_nxv1i32_i8_nxv4i8:141; CHECK: # %bb.0: # %entry142; CHECK-NEXT: vsetvli zero, a1, e32, mf2, tu, ma143; CHECK-NEXT: th.vmaqasu.vx v8, a0, v9144; CHECK-NEXT: ret145entry:146 %a = call <vscale x 1 x i32> @llvm.riscv.th.vmaqasu.nxv1i32.i8(147 <vscale x 1 x i32> %0,148 i8 %1,149 <vscale x 4 x i8> %2,150 iXLen %3, iXLen 0)151 152 ret <vscale x 1 x i32> %a153}154 155define <vscale x 1 x i32> @intrinsic_th_vmaqasu_mask_vx_nxv1i32_i8_nxv4i8(<vscale x 1 x i32> %0, i8 %1, <vscale x 4 x i8> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {156; CHECK-LABEL: intrinsic_th_vmaqasu_mask_vx_nxv1i32_i8_nxv4i8:157; CHECK: # %bb.0: # %entry158; CHECK-NEXT: vsetvli zero, a1, e32, mf2, tu, mu159; CHECK-NEXT: th.vmaqasu.vx v8, a0, v9, v0.t160; CHECK-NEXT: ret161entry:162 %a = call <vscale x 1 x i32> @llvm.riscv.th.vmaqasu.mask.nxv1i32.i8(163 <vscale x 1 x i32> %0,164 i8 %1,165 <vscale x 4 x i8> %2,166 <vscale x 4 x i1> %3,167 iXLen %4, iXLen 0)168 169 ret <vscale x 1 x i32> %a170}171 172define <vscale x 2 x i32> @intrinsic_th_vmaqasu_vx_nxv2i32_i8_nxv8i8(<vscale x 2 x i32> %0, i8 %1, <vscale x 8 x i8> %2, iXLen %3) nounwind {173; CHECK-LABEL: intrinsic_th_vmaqasu_vx_nxv2i32_i8_nxv8i8:174; CHECK: # %bb.0: # %entry175; CHECK-NEXT: vsetvli zero, a1, e32, m1, tu, ma176; CHECK-NEXT: th.vmaqasu.vx v8, a0, v9177; CHECK-NEXT: ret178entry:179 %a = call <vscale x 2 x i32> @llvm.riscv.th.vmaqasu.nxv2i32.i8(180 <vscale x 2 x i32> %0,181 i8 %1,182 <vscale x 8 x i8> %2,183 iXLen %3, iXLen 0)184 185 ret <vscale x 2 x i32> %a186}187 188define <vscale x 2 x i32> @intrinsic_th_vmaqasu_mask_vx_nxv2i32_i8_nxv8i8(<vscale x 2 x i32> %0, i8 %1, <vscale x 8 x i8> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {189; CHECK-LABEL: intrinsic_th_vmaqasu_mask_vx_nxv2i32_i8_nxv8i8:190; CHECK: # %bb.0: # %entry191; CHECK-NEXT: vsetvli zero, a1, e32, m1, tu, mu192; CHECK-NEXT: th.vmaqasu.vx v8, a0, v9, v0.t193; CHECK-NEXT: ret194entry:195 %a = call <vscale x 2 x i32> @llvm.riscv.th.vmaqasu.mask.nxv2i32.i8(196 <vscale x 2 x i32> %0,197 i8 %1,198 <vscale x 8 x i8> %2,199 <vscale x 8 x i1> %3,200 iXLen %4, iXLen 0)201 202 ret <vscale x 2 x i32> %a203}204 205define <vscale x 4 x i32> @intrinsic_th_vmaqasu_vx_nxv4i32_i8_nxv16i8(<vscale x 4 x i32> %0, i8 %1, <vscale x 16 x i8> %2, iXLen %3) nounwind {206; CHECK-LABEL: intrinsic_th_vmaqasu_vx_nxv4i32_i8_nxv16i8:207; CHECK: # %bb.0: # %entry208; CHECK-NEXT: vsetvli zero, a1, e32, m2, tu, ma209; CHECK-NEXT: th.vmaqasu.vx v8, a0, v10210; CHECK-NEXT: ret211entry:212 %a = call <vscale x 4 x i32> @llvm.riscv.th.vmaqasu.nxv4i32.i8(213 <vscale x 4 x i32> %0,214 i8 %1,215 <vscale x 16 x i8> %2,216 iXLen %3, iXLen 0)217 218 ret <vscale x 4 x i32> %a219}220 221define <vscale x 4 x i32> @intrinsic_th_vmaqasu_mask_vx_nxv4i32_i8_nxv16i8(<vscale x 4 x i32> %0, i8 %1, <vscale x 16 x i8> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {222; CHECK-LABEL: intrinsic_th_vmaqasu_mask_vx_nxv4i32_i8_nxv16i8:223; CHECK: # %bb.0: # %entry224; CHECK-NEXT: vsetvli zero, a1, e32, m2, tu, mu225; CHECK-NEXT: th.vmaqasu.vx v8, a0, v10, v0.t226; CHECK-NEXT: ret227entry:228 %a = call <vscale x 4 x i32> @llvm.riscv.th.vmaqasu.mask.nxv4i32.i8(229 <vscale x 4 x i32> %0,230 i8 %1,231 <vscale x 16 x i8> %2,232 <vscale x 16 x i1> %3,233 iXLen %4, iXLen 0)234 235 ret <vscale x 4 x i32> %a236}237 238define <vscale x 8 x i32> @intrinsic_th_vmaqasu_vx_nxv8i32_i8_nxv32i8(<vscale x 8 x i32> %0, i8 %1, <vscale x 32 x i8> %2, iXLen %3) nounwind {239; CHECK-LABEL: intrinsic_th_vmaqasu_vx_nxv8i32_i8_nxv32i8:240; CHECK: # %bb.0: # %entry241; CHECK-NEXT: vsetvli zero, a1, e32, m4, tu, ma242; CHECK-NEXT: th.vmaqasu.vx v8, a0, v12243; CHECK-NEXT: ret244entry:245 %a = call <vscale x 8 x i32> @llvm.riscv.th.vmaqasu.nxv8i32.i8(246 <vscale x 8 x i32> %0,247 i8 %1,248 <vscale x 32 x i8> %2,249 iXLen %3, iXLen 0)250 251 ret <vscale x 8 x i32> %a252}253 254define <vscale x 8 x i32> @intrinsic_th_vmaqasu_mask_vx_nxv8i32_i8_nxv32i8(<vscale x 8 x i32> %0, i8 %1, <vscale x 32 x i8> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {255; CHECK-LABEL: intrinsic_th_vmaqasu_mask_vx_nxv8i32_i8_nxv32i8:256; CHECK: # %bb.0: # %entry257; CHECK-NEXT: vsetvli zero, a1, e32, m4, tu, mu258; CHECK-NEXT: th.vmaqasu.vx v8, a0, v12, v0.t259; CHECK-NEXT: ret260entry:261 %a = call <vscale x 8 x i32> @llvm.riscv.th.vmaqasu.mask.nxv8i32.i8(262 <vscale x 8 x i32> %0,263 i8 %1,264 <vscale x 32 x i8> %2,265 <vscale x 32 x i1> %3,266 iXLen %4, iXLen 0)267 268 ret <vscale x 8 x i32> %a269}270