Files
TheRock/compiler/amd-llvm/llvm/test/CodeGen/X86/vec-strict-inttofp-256-fp16.ll
T
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
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

176 lines
9.1 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=avx512fp16,avx512vl -O3 | FileCheck %s --check-prefixes=CHECK,X86
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=avx512fp16,avx512vl -O3 | FileCheck %s --check-prefixes=CHECK,X64
declare <16 x half> @llvm.experimental.constrained.sitofp.v16f16.v16i1(<16 x i1>, metadata, metadata)
declare <16 x half> @llvm.experimental.constrained.uitofp.v16f16.v16i1(<16 x i1>, metadata, metadata)
declare <16 x half> @llvm.experimental.constrained.sitofp.v16f16.v16i8(<16 x i8>, metadata, metadata)
declare <16 x half> @llvm.experimental.constrained.uitofp.v16f16.v16i8(<16 x i8>, metadata, metadata)
declare <16 x half> @llvm.experimental.constrained.sitofp.v16f16.v16i16(<16 x i16>, metadata, metadata)
declare <16 x half> @llvm.experimental.constrained.uitofp.v16f16.v16i16(<16 x i16>, metadata, metadata)
declare <8 x half> @llvm.experimental.constrained.sitofp.v8f16.v8i32(<8 x i32>, metadata, metadata)
declare <8 x half> @llvm.experimental.constrained.uitofp.v8f16.v8i32(<8 x i32>, metadata, metadata)
declare <4 x half> @llvm.experimental.constrained.sitofp.v4f16.v4i64(<4 x i64>, metadata, metadata)
declare <4 x half> @llvm.experimental.constrained.uitofp.v4f16.v4i64(<4 x i64>, metadata, metadata)
declare <8 x half> @llvm.experimental.constrained.sitofp.v8f16.v8i64(<8 x i64>, metadata, metadata)
declare <8 x half> @llvm.experimental.constrained.uitofp.v8f16.v8i64(<8 x i64>, metadata, metadata)
define <16 x half> @sitofp_v16i1_v16f16(<16 x i1> %x) #0 {
; CHECK-LABEL: sitofp_v16i1_v16f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; CHECK-NEXT: vpsllw $15, %ymm0, %ymm0
; CHECK-NEXT: vpsraw $15, %ymm0, %ymm0
; CHECK-NEXT: vcvtw2ph %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%result = call <16 x half> @llvm.experimental.constrained.sitofp.v16f16.v16i1(<16 x i1> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <16 x half> %result
}
define <16 x half> @uitofp_v16i1_v16f16(<16 x i1> %x) #0 {
; X86-LABEL: uitofp_v16i1_v16f16:
; X86: # %bb.0:
; X86-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}{1to4}, %xmm0, %xmm0
; X86-NEXT: vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; X86-NEXT: vcvtw2ph %ymm0, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: uitofp_v16i1_v16f16:
; X64: # %bb.0:
; X64-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; X64-NEXT: vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; X64-NEXT: vcvtw2ph %ymm0, %ymm0
; X64-NEXT: retq
%result = call <16 x half> @llvm.experimental.constrained.uitofp.v16f16.v16i1(<16 x i1> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <16 x half> %result
}
define <16 x half> @sitofp_v16i8_v16f16(<16 x i8> %x) #0 {
; CHECK-LABEL: sitofp_v16i8_v16f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovsxbw %xmm0, %ymm0
; CHECK-NEXT: vcvtw2ph %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%result = call <16 x half> @llvm.experimental.constrained.sitofp.v16f16.v16i8(<16 x i8> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <16 x half> %result
}
define <16 x half> @uitofp_v16i8_v16f16(<16 x i8> %x) #0 {
; CHECK-LABEL: uitofp_v16i8_v16f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; CHECK-NEXT: vcvtw2ph %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%result = call <16 x half> @llvm.experimental.constrained.uitofp.v16f16.v16i8(<16 x i8> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <16 x half> %result
}
define <16 x half> @sitofp_v16i16_v16f16(<16 x i16> %x) #0 {
; CHECK-LABEL: sitofp_v16i16_v16f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vcvtw2ph %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%result = call <16 x half> @llvm.experimental.constrained.sitofp.v16f16.v16i16(<16 x i16> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <16 x half> %result
}
define <16 x half> @uitofp_v16i16_v16f16(<16 x i16> %x) #0 {
; CHECK-LABEL: uitofp_v16i16_v16f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vcvtuw2ph %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%result = call <16 x half> @llvm.experimental.constrained.uitofp.v16f16.v16i16(<16 x i16> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <16 x half> %result
}
define <8 x half> @sitofp_v8i32_v8f16(<8 x i32> %x) #0 {
; CHECK-LABEL: sitofp_v8i32_v8f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vcvtdq2ph %ymm0, %xmm0
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: ret{{[l|q]}}
%result = call <8 x half> @llvm.experimental.constrained.sitofp.v8f16.v8i32(<8 x i32> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <8 x half> %result
}
define <8 x half> @uitofp_v8i32_v8f16(<8 x i32> %x) #0 {
; CHECK-LABEL: uitofp_v8i32_v8f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vcvtudq2ph %ymm0, %xmm0
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: ret{{[l|q]}}
%result = call <8 x half> @llvm.experimental.constrained.uitofp.v8f16.v8i32(<8 x i32> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <8 x half> %result
}
define <4 x half> @sitofp_v4i64_v4f16(<4 x i64> %x) #0 {
; CHECK-LABEL: sitofp_v4i64_v4f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vcvtqq2ph %ymm0, %xmm0
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: ret{{[l|q]}}
%result = call <4 x half> @llvm.experimental.constrained.sitofp.v4f16.v4i64(<4 x i64> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <4 x half> %result
}
define <4 x half> @uitofp_v4i64_v4f16(<4 x i64> %x) #0 {
; CHECK-LABEL: uitofp_v4i64_v4f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vcvtuqq2ph %ymm0, %xmm0
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: ret{{[l|q]}}
%result = call <4 x half> @llvm.experimental.constrained.uitofp.v4f16.v4i64(<4 x i64> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <4 x half> %result
}
define <8 x half> @sitofp_v8i64_v8f16(<8 x i64> %x) #1 {
; CHECK-LABEL: sitofp_v8i64_v8f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vcvtqq2ph %ymm1, %xmm1
; CHECK-NEXT: vcvtqq2ph %ymm0, %xmm0
; CHECK-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: ret{{[l|q]}}
%result = call <8 x half> @llvm.experimental.constrained.sitofp.v8f16.v8i64(<8 x i64> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <8 x half> %result
}
define <8 x half> @uitofp_v8i64_v8f16(<8 x i64> %x) #1 {
; CHECK-LABEL: uitofp_v8i64_v8f16:
; CHECK: # %bb.0:
; CHECK-NEXT: vcvtuqq2ph %ymm1, %xmm1
; CHECK-NEXT: vcvtuqq2ph %ymm0, %xmm0
; CHECK-NEXT: vmovlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: ret{{[l|q]}}
%result = call <8 x half> @llvm.experimental.constrained.uitofp.v8f16.v8i64(<8 x i64> %x,
metadata !"round.dynamic",
metadata !"fpexcept.strict") #0
ret <8 x half> %result
}
attributes #0 = { strictfp }
attributes #1 = { strictfp "min-legal-vector-width"="256" "prefer-vector-width"="256" }