| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5 |
| ; RUN: llc --mtriple=loongarch32 --mattr=+32s,+lsx,+lasx %s -o - | FileCheck %s --check-prefixes=CHECK,LA32 |
| ; RUN: llc --mtriple=loongarch64 --mattr=+lsx,+lasx %s -o - | FileCheck %s --check-prefixes=CHECK,LA64 |
| |
| define void @load_sext_2i8_to_2i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_2i8_to_2i64: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: ld.h $a0, $a0, 0 |
| ; CHECK-NEXT: vinsgr2vr.h $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.b $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <2 x i8>, ptr %ptr |
| %B = sext <2 x i8> %A to <2 x i64> |
| store <2 x i64> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_4i8_to_4i32(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_4i8_to_4i32: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: ld.w $a0, $a0, 0 |
| ; CHECK-NEXT: vinsgr2vr.w $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.w.b $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <4 x i8>, ptr %ptr |
| %B = sext <4 x i8> %A to <4 x i32> |
| store <4 x i32> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_8i8_to_8i16(ptr %ptr, ptr %dst) { |
| ; LA32-LABEL: load_sext_8i8_to_8i16: |
| ; LA32: # %bb.0: # %entry |
| ; LA32-NEXT: ld.w $a2, $a0, 0 |
| ; LA32-NEXT: ld.w $a0, $a0, 4 |
| ; LA32-NEXT: vinsgr2vr.w $vr0, $a2, 0 |
| ; LA32-NEXT: vinsgr2vr.w $vr0, $a0, 1 |
| ; LA32-NEXT: vext2xv.h.b $xr0, $xr0 |
| ; LA32-NEXT: vst $vr0, $a1, 0 |
| ; LA32-NEXT: ret |
| ; |
| ; LA64-LABEL: load_sext_8i8_to_8i16: |
| ; LA64: # %bb.0: # %entry |
| ; LA64-NEXT: ld.d $a0, $a0, 0 |
| ; LA64-NEXT: vinsgr2vr.d $vr0, $a0, 0 |
| ; LA64-NEXT: vext2xv.h.b $xr0, $xr0 |
| ; LA64-NEXT: vst $vr0, $a1, 0 |
| ; LA64-NEXT: ret |
| entry: |
| %A = load <8 x i8>, ptr %ptr |
| %B = sext <8 x i8> %A to <8 x i16> |
| store <8 x i16> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_2i16_to_2i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_2i16_to_2i64: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: ld.w $a0, $a0, 0 |
| ; CHECK-NEXT: vinsgr2vr.w $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.h $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <2 x i16>, ptr %ptr |
| %B = sext <2 x i16> %A to <2 x i64> |
| store <2 x i64> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_4i16_to_4i32(ptr %ptr, ptr %dst) { |
| ; LA32-LABEL: load_sext_4i16_to_4i32: |
| ; LA32: # %bb.0: # %entry |
| ; LA32-NEXT: ld.w $a2, $a0, 0 |
| ; LA32-NEXT: ld.w $a0, $a0, 4 |
| ; LA32-NEXT: vinsgr2vr.w $vr0, $a2, 0 |
| ; LA32-NEXT: vinsgr2vr.w $vr0, $a0, 1 |
| ; LA32-NEXT: vext2xv.w.h $xr0, $xr0 |
| ; LA32-NEXT: vst $vr0, $a1, 0 |
| ; LA32-NEXT: ret |
| ; |
| ; LA64-LABEL: load_sext_4i16_to_4i32: |
| ; LA64: # %bb.0: # %entry |
| ; LA64-NEXT: ld.d $a0, $a0, 0 |
| ; LA64-NEXT: vinsgr2vr.d $vr0, $a0, 0 |
| ; LA64-NEXT: vext2xv.w.h $xr0, $xr0 |
| ; LA64-NEXT: vst $vr0, $a1, 0 |
| ; LA64-NEXT: ret |
| entry: |
| %A = load <4 x i16>, ptr %ptr |
| %B = sext <4 x i16> %A to <4 x i32> |
| store <4 x i32> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_2i32_to_2i64(ptr %ptr, ptr %dst) { |
| ; LA32-LABEL: load_sext_2i32_to_2i64: |
| ; LA32: # %bb.0: # %entry |
| ; LA32-NEXT: ld.w $a2, $a0, 0 |
| ; LA32-NEXT: ld.w $a0, $a0, 4 |
| ; LA32-NEXT: vinsgr2vr.w $vr0, $a2, 0 |
| ; LA32-NEXT: vinsgr2vr.w $vr0, $a0, 1 |
| ; LA32-NEXT: vext2xv.d.w $xr0, $xr0 |
| ; LA32-NEXT: vst $vr0, $a1, 0 |
| ; LA32-NEXT: ret |
| ; |
| ; LA64-LABEL: load_sext_2i32_to_2i64: |
| ; LA64: # %bb.0: # %entry |
| ; LA64-NEXT: ld.d $a0, $a0, 0 |
| ; LA64-NEXT: vinsgr2vr.d $vr0, $a0, 0 |
| ; LA64-NEXT: vext2xv.d.w $xr0, $xr0 |
| ; LA64-NEXT: vst $vr0, $a1, 0 |
| ; LA64-NEXT: ret |
| entry: |
| %A = load <2 x i32>, ptr %ptr |
| %B = sext <2 x i32> %A to <2 x i64> |
| store <2 x i64> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_16i8_to_16i16(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_16i8_to_16i16: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.h.b $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <16 x i8>, ptr %ptr |
| %B = sext <16 x i8> %A to <16 x i16> |
| store <16 x i16> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_16i8_to_16i32(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_16i8_to_16i32: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.h.b $xr0, $xr0 |
| ; CHECK-NEXT: xvpermi.q $xr1, $xr0, 1 |
| ; CHECK-NEXT: vext2xv.w.h $xr1, $xr1 |
| ; CHECK-NEXT: vext2xv.w.h $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: xvst $xr1, $a1, 32 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <16 x i8>, ptr %ptr |
| %B = sext <16 x i8> %A to <16 x i32> |
| store <16 x i32> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_16i8_to_16i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_16i8_to_16i64: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.h.b $xr0, $xr0 |
| ; CHECK-NEXT: xvpermi.q $xr1, $xr0, 1 |
| ; CHECK-NEXT: vext2xv.w.h $xr1, $xr1 |
| ; CHECK-NEXT: xvpermi.q $xr2, $xr1, 1 |
| ; CHECK-NEXT: vext2xv.d.w $xr2, $xr2 |
| ; CHECK-NEXT: vext2xv.w.h $xr0, $xr0 |
| ; CHECK-NEXT: xvpermi.q $xr3, $xr0, 1 |
| ; CHECK-NEXT: vext2xv.d.w $xr3, $xr3 |
| ; CHECK-NEXT: vext2xv.d.w $xr1, $xr1 |
| ; CHECK-NEXT: vext2xv.d.w $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: xvst $xr1, $a1, 64 |
| ; CHECK-NEXT: xvst $xr3, $a1, 32 |
| ; CHECK-NEXT: xvst $xr2, $a1, 96 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <16 x i8>, ptr %ptr |
| %B = sext <16 x i8> %A to <16 x i64> |
| store <16 x i64> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_8i16_to_8i32(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_8i16_to_8i32: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.w.h $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <8 x i16>, ptr %ptr |
| %B = sext <8 x i16> %A to <8 x i32> |
| store <8 x i32> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_8i16_to_8i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_8i16_to_8i64: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.w.h $xr0, $xr0 |
| ; CHECK-NEXT: xvpermi.q $xr1, $xr0, 1 |
| ; CHECK-NEXT: vext2xv.d.w $xr1, $xr1 |
| ; CHECK-NEXT: vext2xv.d.w $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: xvst $xr1, $a1, 32 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <8 x i16>, ptr %ptr |
| %B = sext <8 x i16> %A to <8 x i64> |
| store <8 x i64> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_4i32_to_4i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_4i32_to_4i64: |
| ; CHECK: # %bb.0: # %entry |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.w $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| entry: |
| %A = load <4 x i32>, ptr %ptr |
| %B = sext <4 x i32> %A to <4 x i64> |
| store <4 x i64> %B, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v16i8_to_v2i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v16i8_to_v2i64: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.b $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <16 x i8>, ptr %ptr |
| %B = shufflevector <16 x i8> %A, <16 x i8> poison, |
| <2 x i32> <i32 0, i32 1> |
| %C = sext <2 x i8> %B to <2 x i64> |
| store <2 x i64> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v8i16_to_v2i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v8i16_to_v2i64: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.h $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <8 x i16>, ptr %ptr |
| %B = shufflevector <8 x i16> %A, <8 x i16> poison, |
| <2 x i32> <i32 0, i32 1> |
| %C = sext <2 x i16> %B to <2 x i64> |
| store <2 x i64> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v4i32_to_v2i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v4i32_to_v2i64: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.w $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <4 x i32>, ptr %ptr |
| %B = shufflevector <4 x i32> %A, <4 x i32> poison, |
| <2 x i32> <i32 0, i32 1> |
| %C = sext <2 x i32> %B to <2 x i64> |
| store <2 x i64> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v16i8_to_v4i32(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v16i8_to_v4i32: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.w.b $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <16 x i8>, ptr %ptr |
| %B = shufflevector <16 x i8> %A, <16 x i8> poison, |
| <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| %C = sext <4 x i8> %B to <4 x i32> |
| store <4 x i32> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v8i16_to_v4i32(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v8i16_to_v4i32: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.w.h $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <8 x i16>, ptr %ptr |
| %B = shufflevector <8 x i16> %A, <8 x i16> poison, |
| <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| %C = sext <4 x i16> %B to <4 x i32> |
| store <4 x i32> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v32i8_to_v4i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v32i8_to_v4i64: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: xvld $xr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.b $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <32 x i8>, ptr %ptr |
| %B = shufflevector <32 x i8> %A, <32 x i8> poison, |
| <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| %C = sext <4 x i8> %B to <4 x i64> |
| store <4 x i64> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v16i16_to_v4i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v16i16_to_v4i64: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: xvld $xr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.h $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <16 x i16>, ptr %ptr |
| %B = shufflevector <16 x i16> %A, <16 x i16> poison, |
| <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| %C = sext <4 x i16> %B to <4 x i64> |
| store <4 x i64> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v8i32_to_v4i64(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v8i32_to_v4i64: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.d.w $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <8 x i32>, ptr %ptr |
| %B = shufflevector <8 x i32> %A, <8 x i32> poison, |
| <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| %C = sext <4 x i32> %B to <4 x i64> |
| store <4 x i64> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v16i8_to_v8i16(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v16i8_to_v8i16: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.h.b $xr0, $xr0 |
| ; CHECK-NEXT: vst $vr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <16 x i8>, ptr %ptr |
| %B = shufflevector <16 x i8> %A, <16 x i8> poison, |
| <8 x i32> <i32 0, i32 1, i32 2, i32 3, |
| i32 4, i32 5, i32 6, i32 7> |
| %C = sext <8 x i8> %B to <8 x i16> |
| store <8 x i16> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v32i8_to_v16i16(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v32i8_to_v16i16: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.h.b $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <32 x i8>, ptr %ptr |
| %B = shufflevector <32 x i8> %A, <32 x i8> poison, |
| <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15> |
| %C = sext <16 x i8> %B to <16 x i16> |
| store <16 x i16> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v32i8_to_v8i32(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v32i8_to_v8i32: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: xvld $xr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.w.b $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <32 x i8>, ptr %ptr |
| %B = shufflevector <32 x i8> %A, <32 x i8> poison, |
| <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> |
| %C = sext <8 x i8> %B to <8 x i32> |
| store <8 x i32> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v16i16_to_v8i32(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v16i16_to_v8i32: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.w.h $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <16 x i16>, ptr %ptr |
| %B = shufflevector <16 x i16> %A, <16 x i16> poison, |
| <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> |
| %C = sext <8 x i16> %B to <8 x i32> |
| store <8 x i32> %C, ptr %dst |
| ret void |
| } |
| |
| define void @load_sext_invec_v16i8_to_v8i32(ptr %ptr, ptr %dst) { |
| ; CHECK-LABEL: load_sext_invec_v16i8_to_v8i32: |
| ; CHECK: # %bb.0: |
| ; CHECK-NEXT: vld $vr0, $a0, 0 |
| ; CHECK-NEXT: vext2xv.w.b $xr0, $xr0 |
| ; CHECK-NEXT: xvst $xr0, $a1, 0 |
| ; CHECK-NEXT: ret |
| %A = load <16 x i8>, ptr %ptr |
| %B = shufflevector <16 x i8> %A, <16 x i8> poison, |
| <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> |
| %C = sext <8 x i8> %B to <8 x i32> |
| store <8 x i32> %C, ptr %dst |
| ret void |
| } |