derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
203 lines
8.7 KiB
LLVM
203 lines
8.7 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc --mtriple=loongarch32 --mattr=+32s,+lasx < %s | FileCheck %s
|
|
; RUN: llc --mtriple=loongarch64 --mattr=+lasx < %s | FileCheck %s
|
|
|
|
define void @select_v32i8_imm(ptr %res, ptr %a0) nounwind {
|
|
; CHECK-LABEL: select_v32i8_imm:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvrepli.h $xr1, -256
|
|
; CHECK-NEXT: xvbitseli.b $xr1, $xr0, 1
|
|
; CHECK-NEXT: xvst $xr1, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <32 x i8>, ptr %a0
|
|
%sel = select <32 x i1> <i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true>, <32 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, <32 x i8> %v0
|
|
store <32 x i8> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v32i8(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v32i8:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: xvrepli.h $xr2, -256
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <32 x i8>, ptr %a0
|
|
%v1 = load <32 x i8>, ptr %a1
|
|
%sel = select <32 x i1> <i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true>, <32 x i8> %v0, <32 x i8> %v1
|
|
store <32 x i8> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v32i8_1(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v32i8_1:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: pcalau12i $a1, %pc_hi20(.LCPI2_0)
|
|
; CHECK-NEXT: xvld $xr2, $a1, %pc_lo12(.LCPI2_0)
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <32 x i8>, ptr %a0
|
|
%v1 = load <32 x i8>, ptr %a1
|
|
%sel = select <32 x i1> <i1 true, i1 true, i1 true, i1 true, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false>, <32 x i8> %v0, <32 x i8> %v1
|
|
store <32 x i8> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v32i8_2(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v32i8_2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: pcalau12i $a1, %pc_hi20(.LCPI3_0)
|
|
; CHECK-NEXT: xvld $xr2, $a1, %pc_lo12(.LCPI3_0)
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <32 x i8>, ptr %a0
|
|
%v1 = load <32 x i8>, ptr %a1
|
|
%sel = select <32 x i1> <i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true>, <32 x i8> %v0, <32 x i8> %v1
|
|
store <32 x i8> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v16i16(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v16i16:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: lu12i.w $a1, -16
|
|
; CHECK-NEXT: xvreplgr2vr.w $xr2, $a1
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <16 x i16>, ptr %a0
|
|
%v1 = load <16 x i16>, ptr %a1
|
|
%sel = select <16 x i1> <i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true>, <16 x i16> %v0, <16 x i16> %v1
|
|
store <16 x i16> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v16i16_1(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v16i16_1:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: pcalau12i $a1, %pc_hi20(.LCPI5_0)
|
|
; CHECK-NEXT: xvld $xr2, $a1, %pc_lo12(.LCPI5_0)
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <16 x i16>, ptr %a0
|
|
%v1 = load <16 x i16>, ptr %a1
|
|
%sel = select <16 x i1> <i1 false, i1 false, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <16 x i16> %v0, <16 x i16> %v1
|
|
store <16 x i16> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v16i16_2(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v16i16_2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: pcalau12i $a1, %pc_hi20(.LCPI6_0)
|
|
; CHECK-NEXT: xvld $xr2, $a1, %pc_lo12(.LCPI6_0)
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <16 x i16>, ptr %a0
|
|
%v1 = load <16 x i16>, ptr %a1
|
|
%sel = select <16 x i1> <i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false>, <16 x i16> %v0, <16 x i16> %v1
|
|
store <16 x i16> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v8i32(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v8i32:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: xvldi $xr2, -1552
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <8 x i32>, ptr %a0
|
|
%v1 = load <8 x i32>, ptr %a1
|
|
%sel = select <8 x i1> <i1 false, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false, i1 true>, <8 x i32> %v0, <8 x i32> %v1
|
|
store <8 x i32> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v8i32_1(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v8i32_1:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: pcalau12i $a1, %pc_hi20(.LCPI8_0)
|
|
; CHECK-NEXT: xvld $xr2, $a1, %pc_lo12(.LCPI8_0)
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <8 x i32>, ptr %a0
|
|
%v1 = load <8 x i32>, ptr %a1
|
|
%sel = select <8 x i1> <i1 true, i1 true, i1 false, i1 false, i1 false, i1 false, i1 false, i1 false>, <8 x i32> %v0, <8 x i32> %v1
|
|
store <8 x i32> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v8f32(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v8f32:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: pcalau12i $a1, %pc_hi20(.LCPI9_0)
|
|
; CHECK-NEXT: xvld $xr2, $a1, %pc_lo12(.LCPI9_0)
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <8 x float>, ptr %a0
|
|
%v1 = load <8 x float>, ptr %a1
|
|
%sel = select <8 x i1> <i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false>, <8 x float> %v0, <8 x float> %v1
|
|
store <8 x float> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v4i64(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v4i64:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: pcalau12i $a1, %pc_hi20(.LCPI10_0)
|
|
; CHECK-NEXT: xvld $xr2, $a1, %pc_lo12(.LCPI10_0)
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <4 x i64>, ptr %a0
|
|
%v1 = load <4 x i64>, ptr %a1
|
|
%sel = select <4 x i1> <i1 true, i1 false, i1 false, i1 false>, <4 x i64> %v0, <4 x i64> %v1
|
|
store <4 x i64> %sel, ptr %res
|
|
ret void
|
|
}
|
|
|
|
define void @select_v4f64(ptr %res, ptr %a0, ptr %a1) nounwind {
|
|
; CHECK-LABEL: select_v4f64:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: xvld $xr0, $a1, 0
|
|
; CHECK-NEXT: xvld $xr1, $a2, 0
|
|
; CHECK-NEXT: pcalau12i $a1, %pc_hi20(.LCPI11_0)
|
|
; CHECK-NEXT: xvld $xr2, $a1, %pc_lo12(.LCPI11_0)
|
|
; CHECK-NEXT: xvbitsel.v $xr0, $xr1, $xr0, $xr2
|
|
; CHECK-NEXT: xvst $xr0, $a0, 0
|
|
; CHECK-NEXT: ret
|
|
%v0 = load <4 x double>, ptr %a0
|
|
%v1 = load <4 x double>, ptr %a1
|
|
%sel = select <4 x i1> <i1 true, i1 false, i1 true, i1 false>, <4 x double> %v0, <4 x double> %v1
|
|
store <4 x double> %sel, ptr %res
|
|
ret void
|
|
}
|