| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 6 | |
| ; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve < %s | FileCheck %s | |
| ; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve -lower-interleaved-accesses=false < %s | FileCheck %s | |
| define void @store2_twice(ptr %dst0, ptr %dst1, <4 x i32> %v0, <4 x i32> %v1) { | |
| ; CHECK-LABEL: store2_twice: | |
| ; CHECK: // %bb.0: | |
| ; CHECK-NEXT: zip2 v2.4s, v0.4s, v1.4s | |
| ; CHECK-NEXT: zip1 v0.4s, v0.4s, v1.4s | |
| ; CHECK-NEXT: stp q0, q2, [x0] | |
| ; CHECK-NEXT: stp q0, q2, [x1] | |
| ; CHECK-NEXT: ret | |
| %interleaved = call <8 x i32> @llvm.vector.interleave2.v8i32(<4 x i32> %v0, <4 x i32> %v1) | |
| store <8 x i32> %interleaved, ptr %dst0, align 4 | |
| store <8 x i32> %interleaved, ptr %dst1, align 4 | |
| ret void | |
| } | |
| define void @store2_twice_optsize(ptr %dst0, ptr %dst1, <4 x i32> %v0, <4 x i32> %v1) optsize { | |
| ; CHECK-LABEL: store2_twice_optsize: | |
| ; CHECK: // %bb.0: | |
| ; CHECK-NEXT: // kill: def $q1 killed $q1 killed $q0_q1 def $q0_q1 | |
| ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $q0_q1 def $q0_q1 | |
| ; CHECK-NEXT: st2 { v0.4s, v1.4s }, [x0] | |
| ; CHECK-NEXT: st2 { v0.4s, v1.4s }, [x1] | |
| ; CHECK-NEXT: ret | |
| %interleaved = call <8 x i32> @llvm.vector.interleave2.v8i32(<4 x i32> %v0, <4 x i32> %v1) | |
| store <8 x i32> %interleaved, ptr %dst0, align 4 | |
| store <8 x i32> %interleaved, ptr %dst1, align 4 | |
| ret void | |
| } | |
| define void @store3_twice(ptr %dst0, ptr %dst1, <4 x i32> %v0, <4 x i32> %v1, <4 x i32> %v2) { | |
| ; CHECK-LABEL: store3_twice: | |
| ; CHECK: // %bb.0: | |
| ; CHECK-NEXT: // kill: def $q2 killed $q2 killed $q0_q1_q2 def $q0_q1_q2 | |
| ; CHECK-NEXT: // kill: def $q1 killed $q1 killed $q0_q1_q2 def $q0_q1_q2 | |
| ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $q0_q1_q2 def $q0_q1_q2 | |
| ; CHECK-NEXT: st3 { v0.4s, v1.4s, v2.4s }, [x0] | |
| ; CHECK-NEXT: st3 { v0.4s, v1.4s, v2.4s }, [x1] | |
| ; CHECK-NEXT: ret | |
| %interleaved = call <12 x i32> @llvm.vector.interleave3.v12i32(<4 x i32> %v0, <4 x i32> %v1, <4 x i32> %v2) | |
| store <12 x i32> %interleaved, ptr %dst0, align 4 | |
| store <12 x i32> %interleaved, ptr %dst1, align 4 | |
| ret void | |
| } | |
| define void @store4_twice(ptr %dst0, ptr %dst1, <4 x i32> %v0, <4 x i32> %v1, <4 x i32> %v2, <4 x i32> %v3) { | |
| ; CHECK-LABEL: store4_twice: | |
| ; CHECK: // %bb.0: | |
| ; CHECK-NEXT: // kill: def $q3 killed $q3 killed $q0_q1_q2_q3 def $q0_q1_q2_q3 | |
| ; CHECK-NEXT: // kill: def $q2 killed $q2 killed $q0_q1_q2_q3 def $q0_q1_q2_q3 | |
| ; CHECK-NEXT: // kill: def $q1 killed $q1 killed $q0_q1_q2_q3 def $q0_q1_q2_q3 | |
| ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $q0_q1_q2_q3 def $q0_q1_q2_q3 | |
| ; CHECK-NEXT: st4 { v0.4s, v1.4s, v2.4s, v3.4s }, [x0] | |
| ; CHECK-NEXT: st4 { v0.4s, v1.4s, v2.4s, v3.4s }, [x1] | |
| ; CHECK-NEXT: ret | |
| %interleaved = call <16 x i32> @llvm.vector.interleave4.v16i32(<4 x i32> %v0, <4 x i32> %v1, <4 x i32> %v2, <4 x i32> %v3) | |
| store <16 x i32> %interleaved, ptr %dst0, align 4 | |
| store <16 x i32> %interleaved, ptr %dst1, align 4 | |
| ret void | |
| } | |
| define void @store2_twice_scalable(ptr %dst0, ptr %dst1, <vscale x 4 x i32> %v0, <vscale x 4 x i32> %v1) { | |
| ; CHECK-LABEL: store2_twice_scalable: | |
| ; CHECK: // %bb.0: | |
| ; CHECK-NEXT: ptrue p0.s | |
| ; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1 | |
| ; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1 | |
| ; CHECK-NEXT: st2w { z0.s, z1.s }, p0, [x0] | |
| ; CHECK-NEXT: st2w { z0.s, z1.s }, p0, [x1] | |
| ; CHECK-NEXT: ret | |
| %interleaved = call <vscale x 8 x i32> @llvm.vector.interleave2.nxv8i32(<vscale x 4 x i32> %v0, <vscale x 4 x i32> %v1) | |
| store <vscale x 8 x i32> %interleaved, ptr %dst0, align 4 | |
| store <vscale x 8 x i32> %interleaved, ptr %dst1, align 4 | |
| ret void | |
| } | |
| define void @store3_twice_scalable(ptr %dst0, ptr %dst1, <vscale x 4 x i32> %v0, <vscale x 4 x i32> %v1, <vscale x 4 x i32> %v2) { | |
| ; CHECK-LABEL: store3_twice_scalable: | |
| ; CHECK: // %bb.0: | |
| ; CHECK-NEXT: // kill: def $z2 killed $z2 killed $z0_z1_z2 def $z0_z1_z2 | |
| ; CHECK-NEXT: ptrue p0.s | |
| ; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1_z2 def $z0_z1_z2 | |
| ; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1_z2 def $z0_z1_z2 | |
| ; CHECK-NEXT: st3w { z0.s - z2.s }, p0, [x0] | |
| ; CHECK-NEXT: st3w { z0.s - z2.s }, p0, [x1] | |
| ; CHECK-NEXT: ret | |
| %interleaved = call <vscale x 12 x i32> @llvm.vector.interleave3.nxv12i32(<vscale x 4 x i32> %v0, <vscale x 4 x i32> %v1, <vscale x 4 x i32> %v2) | |
| store <vscale x 12 x i32> %interleaved, ptr %dst0, align 4 | |
| store <vscale x 12 x i32> %interleaved, ptr %dst1, align 4 | |
| ret void | |
| } | |
| define void @store4_twice_scalable(ptr %dst0, ptr %dst1, <vscale x 4 x i32> %v0, <vscale x 4 x i32> %v1, <vscale x 4 x i32> %v2, <vscale x 4 x i32> %v3) { | |
| ; CHECK-LABEL: store4_twice_scalable: | |
| ; CHECK: // %bb.0: | |
| ; CHECK-NEXT: // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3 | |
| ; CHECK-NEXT: ptrue p0.s | |
| ; CHECK-NEXT: // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3 | |
| ; CHECK-NEXT: // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3 | |
| ; CHECK-NEXT: // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3 | |
| ; CHECK-NEXT: st4w { z0.s - z3.s }, p0, [x0] | |
| ; CHECK-NEXT: st4w { z0.s - z3.s }, p0, [x1] | |
| ; CHECK-NEXT: ret | |
| %interleaved = call <vscale x 16 x i32> @llvm.vector.interleave4.nxv16i32(<vscale x 4 x i32> %v0, <vscale x 4 x i32> %v1, <vscale x 4 x i32> %v2, <vscale x 4 x i32> %v3) | |
| store <vscale x 16 x i32> %interleaved, ptr %dst0, align 4 | |
| store <vscale x 16 x i32> %interleaved, ptr %dst1, align 4 | |
| ret void | |
| } |