86 lines · plain
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 22; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve2p1,+bf16 < %s | FileCheck %s3 4define <vscale x 16 x i8> @test_zipq1_i8(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm) {5; CHECK-LABEL: test_zipq1_i8:6; CHECK: // %bb.0:7; CHECK-NEXT: zipq1 z0.b, z0.b, z1.b8; CHECK-NEXT: ret9 %res = call <vscale x 16 x i8> @llvm.aarch64.sve.zipq1.nxv16i8(<vscale x 16 x i8> %zn, <vscale x 16 x i8> %zm)10 ret <vscale x 16 x i8> %res11}12 13define <vscale x 8 x i16> @test_zipq1_i16(<vscale x 8 x i16> %zn, <vscale x 8 x i16> %zm) {14; CHECK-LABEL: test_zipq1_i16:15; CHECK: // %bb.0:16; CHECK-NEXT: zipq1 z0.h, z0.h, z1.h17; CHECK-NEXT: ret18 %res = call <vscale x 8 x i16> @llvm.aarch64.sve.zipq1.nxv8i16(<vscale x 8 x i16> %zn, <vscale x 8 x i16> %zm)19 ret <vscale x 8 x i16> %res20}21 22define <vscale x 4 x i32> @test_zipq1_i32(<vscale x 4 x i32> %zn, <vscale x 4 x i32> %zm) {23; CHECK-LABEL: test_zipq1_i32:24; CHECK: // %bb.0:25; CHECK-NEXT: zipq1 z0.s, z0.s, z1.s26; CHECK-NEXT: ret27 %res = call <vscale x 4 x i32> @llvm.aarch64.sve.zipq1.nxv4i32(<vscale x 4 x i32> %zn, <vscale x 4 x i32> %zm)28 ret <vscale x 4 x i32> %res29}30 31define <vscale x 2 x i64> @test_zipq1_i64(<vscale x 2 x i64> %zn, <vscale x 2 x i64> %zm) {32; CHECK-LABEL: test_zipq1_i64:33; CHECK: // %bb.0:34; CHECK-NEXT: zipq1 z0.d, z0.d, z1.d35; CHECK-NEXT: ret36 %res = call <vscale x 2 x i64> @llvm.aarch64.sve.zipq1.nxv2i64(<vscale x 2 x i64> %zn, <vscale x 2 x i64> %zm)37 ret <vscale x 2 x i64> %res38}39 40define <vscale x 8 x half> @test_zipq1_f16(<vscale x 8 x half> %zn, <vscale x 8 x half> %zm) {41; CHECK-LABEL: test_zipq1_f16:42; CHECK: // %bb.0:43; CHECK-NEXT: zipq1 z0.h, z0.h, z1.h44; CHECK-NEXT: ret45 %res = call <vscale x 8 x half> @llvm.aarch64.sve.zipq1.nxv8f16(<vscale x 8 x half> %zn, <vscale x 8 x half> %zm)46 ret <vscale x 8 x half> %res47}48 49define <vscale x 4 x float> @test_zipq1_f32(<vscale x 4 x float> %zn, <vscale x 4 x float> %zm) {50; CHECK-LABEL: test_zipq1_f32:51; CHECK: // %bb.0:52; CHECK-NEXT: zipq1 z0.s, z0.s, z1.s53; CHECK-NEXT: ret54 %res = call <vscale x 4 x float> @llvm.aarch64.sve.zipq1.nxv4f32(<vscale x 4 x float> %zn, <vscale x 4 x float> %zm)55 ret <vscale x 4 x float> %res56}57 58define <vscale x 2 x double> @test_zipq1_f64(<vscale x 2 x double> %zn, <vscale x 2 x double> %zm) {59; CHECK-LABEL: test_zipq1_f64:60; CHECK: // %bb.0:61; CHECK-NEXT: zipq1 z0.d, z0.d, z1.d62; CHECK-NEXT: ret63 %res = call <vscale x 2 x double> @llvm.aarch64.sve.zipq1.nxv2f64(<vscale x 2 x double> %zn, <vscale x 2 x double> %zm)64 ret <vscale x 2 x double> %res65}66 67define <vscale x 8 x bfloat> @test_zipq1_bf16(<vscale x 8 x bfloat> %zn, <vscale x 8 x bfloat> %zm) {68; CHECK-LABEL: test_zipq1_bf16:69; CHECK: // %bb.0:70; CHECK-NEXT: zipq1 z0.h, z0.h, z1.h71; CHECK-NEXT: ret72 %res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.zipq1.nxv8bf16(<vscale x 8 x bfloat> %zn, <vscale x 8 x bfloat> %zm)73 ret <vscale x 8 x bfloat> %res74}75 76 77declare <vscale x 16 x i8> @llvm.aarch64.sve.zipq1.nxv16i8(<vscale x 16 x i8>, <vscale x 16 x i8>)78declare <vscale x 8 x i16> @llvm.aarch64.sve.zipq1.nxv8i16(<vscale x 8 x i16>, <vscale x 8 x i16>)79declare <vscale x 4 x i32> @llvm.aarch64.sve.zipq1.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>)80declare <vscale x 2 x i64> @llvm.aarch64.sve.zipq1.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>)81 82declare <vscale x 8 x half> @llvm.aarch64.sve.zipq1.nxv8f16(<vscale x 8 x half>, <vscale x 8 x half>)83declare <vscale x 4 x float> @llvm.aarch64.sve.zipq1.nxv4f32(<vscale x 4 x float>, <vscale x 4 x float>)84declare <vscale x 2 x double> @llvm.aarch64.sve.zipq1.nxv2f64(<vscale x 2 x double>, <vscale x 2 x double>)85declare <vscale x 8 x bfloat> @llvm.aarch64.sve.zipq1.nxv8bf16(<vscale x 8 x bfloat>, <vscale x 8 x bfloat>)86