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
100 lines
3.1 KiB
LLVM
100 lines
3.1 KiB
LLVM
; RUN: llc -mtriple=hexagon -hexagon-hvx-widen=32 < %s | FileCheck %s
|
|
|
|
; v32i8 -> v32i16
|
|
; CHECK-LABEL: f0:
|
|
; CHECK: r[[R0:[0-9]+]] = #64
|
|
; CHECK: v[[V0:[0-9]+]] = vmem(r0+#0)
|
|
; CHECK: v[[V1:[0-9]+]]:[[V2:[0-9]+]].h = vunpack(v[[V0]].b)
|
|
; CHECK: q[[Q0:[0-3]]] = vsetq(r[[R0]])
|
|
; CHECK: if (q[[Q0]]) vmem(r1+#0) = v[[V2]]
|
|
define void @f0(ptr %a0, ptr %a1) #0 {
|
|
%v0 = load <32 x i8>, ptr %a0, align 128
|
|
%v1 = sext <32 x i8> %v0 to <32 x i16>
|
|
store <32 x i16> %v1, ptr %a1, align 128
|
|
ret void
|
|
}
|
|
|
|
; v32i8 -> v32i32
|
|
; CHECK-LABEL: f1:
|
|
; CHECK: v[[V0:[0-9]+]] = vmem(r0+#0)
|
|
; CHECK: v[[V1:[0-9]+]]:[[V2:[0-9]+]].h = vunpack(v[[V0]].b)
|
|
; CHECK: v[[V3:[0-9]+]]:[[V4:[0-9]+]].w = vunpack(v[[V2]].h)
|
|
; CHECK: vmem(r1+#0) = v[[V4]]
|
|
define void @f1(ptr %a0, ptr %a1) #0 {
|
|
%v0 = load <32 x i8>, ptr %a0, align 128
|
|
%v1 = sext <32 x i8> %v0 to <32 x i32>
|
|
store <32 x i32> %v1, ptr %a1, align 128
|
|
ret void
|
|
}
|
|
|
|
; v64i8 -> v64i16
|
|
; CHECK-LABEL: f2:
|
|
; CHECK: v[[V0:[0-9]+]] = vmem(r0+#0)
|
|
; CHECK: v[[V1:[0-9]+]]:[[V2:[0-9]+]].h = vunpack(v[[V0]].b)
|
|
; CHECK: vmem(r1+#0) = v[[V2]]
|
|
define void @f2(ptr %a0, ptr %a1) #0 {
|
|
%v0 = load <64 x i8>, ptr %a0, align 128
|
|
%v1 = sext <64 x i8> %v0 to <64 x i16>
|
|
store <64 x i16> %v1, ptr %a1, align 128
|
|
ret void
|
|
}
|
|
|
|
; v64i8 -> v64i32
|
|
; CHECK-LABEL: f3:
|
|
; CHECK: v[[V0:[0-9]+]] = vmem(r0+#0)
|
|
; CHECK: v[[V1:[0-9]+]]:[[V2:[0-9]+]].h = vunpack(v[[V0]].b)
|
|
; CHECK: v[[V3:[0-9]+]]:[[V4:[0-9]+]].w = vunpack(v[[V2]].h)
|
|
; CHECK-DAG: vmem(r1+#0) = v[[V4]]
|
|
; CHECK-DAG: vmem(r1+#1) = v[[V3]]
|
|
define void @f3(ptr %a0, ptr %a1) #0 {
|
|
%v0 = load <64 x i8>, ptr %a0, align 128
|
|
%v1 = sext <64 x i8> %v0 to <64 x i32>
|
|
store <64 x i32> %v1, ptr %a1, align 128
|
|
ret void
|
|
}
|
|
|
|
; v16i16 -> v16i32
|
|
; CHECK-LABEL: f4:
|
|
; CHECK: r[[R0:[0-9]+]] = #64
|
|
; CHECK: v[[V0:[0-9]+]] = vmem(r0+#0)
|
|
; CHECK: v[[V1:[0-9]+]]:[[V2:[0-9]+]].w = vunpack(v[[V0]].h)
|
|
; CHECK: q[[Q0:[0-3]]] = vsetq(r[[R0]])
|
|
; CHECK: if (q[[Q0]]) vmem(r1+#0) = v[[V2]]
|
|
define void @f4(ptr %a0, ptr %a1) #0 {
|
|
%v0 = load <16 x i16>, ptr %a0, align 128
|
|
%v1 = sext <16 x i16> %v0 to <16 x i32>
|
|
store <16 x i32> %v1, ptr %a1, align 128
|
|
ret void
|
|
}
|
|
|
|
; v32i16 -> v32i32
|
|
; CHECK-LABEL: f5:
|
|
; CHECK: v[[V0:[0-9]+]] = vmem(r0+#0)
|
|
; CHECK: v[[V1:[0-9]+]]:[[V2:[0-9]+]].w = vunpack(v[[V0]].h)
|
|
; CHECK: vmem(r1+#0) = v[[V2]]
|
|
define void @f5(ptr %a0, ptr %a1) #0 {
|
|
%v0 = load <32 x i16>, ptr %a0, align 128
|
|
%v1 = sext <32 x i16> %v0 to <32 x i32>
|
|
store <32 x i32> %v1, ptr %a1, align 128
|
|
ret void
|
|
}
|
|
|
|
; v8i8 -> v8i32
|
|
; CHECK-LABEL: f6:
|
|
; CHECK: r[[R0:[0-9]+]]:[[R1:[0-9]+]] = memd(r0+#0)
|
|
; CHECK-DAG: v[[V0:[0-9]+]].w = vinsert(r[[R0]])
|
|
; CHECK-DAG: v[[V0]].w = vinsert(r[[R1]])
|
|
; CHECK-DAG: q[[Q0:[0-3]]] = vsetq
|
|
; CHECK: v[[V1:[0-9]+]]:[[V2:[0-9]+]].h = vunpack(v[[V0]].b)
|
|
; CHECK: v[[V3:[0-9]+]]:[[V4:[0-9]+]].w = vunpack(v[[V2]].h)
|
|
; CHECK: if (q[[Q0]]) vmem(r1+#0) = v[[V4]]
|
|
define void @f6(ptr %a0, ptr %a1) #0 {
|
|
%v0 = load <8 x i8>, ptr %a0, align 128
|
|
%v1 = sext <8 x i8> %v0 to <8 x i32>
|
|
store <8 x i32> %v1, ptr %a1, align 128
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = { "target-cpu"="hexagonv65" "target-features"="+hvx,+hvx-length128b,-packets" }
|
|
|