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
1518 lines
54 KiB
LLVM
1518 lines
54 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -enable-subreg-liveness -mattr=+sve -force-streaming-compatible < %s | FileCheck %s --check-prefixes=SVE
|
|
; RUN: llc -enable-subreg-liveness -mattr=+sve2 -force-streaming-compatible < %s | FileCheck %s --check-prefixes=SVE2
|
|
; RUN: llc -enable-subreg-liveness -mattr=+sme -force-streaming < %s | FileCheck %s --check-prefixes=SVE2
|
|
; RUN: llc -enable-subreg-liveness -force-streaming-compatible < %s | FileCheck %s --check-prefix=NONEON-NOSVE
|
|
|
|
target datalayout = "e-m:o-i64:64-i128:128-n32:64-S128"
|
|
|
|
target triple = "aarch64-unknown-linux-gnu"
|
|
|
|
define void @test_copysign_f16(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_f16:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr h0, [x1]
|
|
; SVE-NEXT: ldr h1, [x0]
|
|
; SVE-NEXT: and z0.h, z0.h, #0x8000
|
|
; SVE-NEXT: and z1.h, z1.h, #0x7fff
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str h0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_f16:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.h, #32767 // =0x7fff
|
|
; SVE2-NEXT: ldr h1, [x1]
|
|
; SVE2-NEXT: ldr h2, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z1.d, z0.d
|
|
; SVE2-NEXT: str h2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_f16:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #16
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 16
|
|
; NONEON-NOSVE-NEXT: ldr h0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr h1, [x1]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: str h1, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #13]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #16
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load half, ptr %ap
|
|
%b = load half, ptr %bp
|
|
%r = call half @llvm.copysign.f16(half %a, half %b)
|
|
store half %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_bf16(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_bf16:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr h0, [x1]
|
|
; SVE-NEXT: ldr h1, [x0]
|
|
; SVE-NEXT: and z0.h, z0.h, #0x8000
|
|
; SVE-NEXT: and z1.h, z1.h, #0x7fff
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str h0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_bf16:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.h, #32767 // =0x7fff
|
|
; SVE2-NEXT: ldr h1, [x1]
|
|
; SVE2-NEXT: ldr h2, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z1.d, z0.d
|
|
; SVE2-NEXT: str h2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_bf16:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #16
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 16
|
|
; NONEON-NOSVE-NEXT: ldr h0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr h1, [x1]
|
|
; NONEON-NOSVE-NEXT: fmov w8, s0
|
|
; NONEON-NOSVE-NEXT: str h1, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: lsl w8, w8, #16
|
|
; NONEON-NOSVE-NEXT: fmov s0, w8
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #13]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: fmov w8, s0
|
|
; NONEON-NOSVE-NEXT: lsr w8, w8, #16
|
|
; NONEON-NOSVE-NEXT: fmov s0, w8
|
|
; NONEON-NOSVE-NEXT: str h0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #16
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load bfloat, ptr %ap
|
|
%b = load bfloat, ptr %bp
|
|
%r = call bfloat @llvm.copysign.bf16(bfloat %a, bfloat %b)
|
|
store bfloat %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_f32(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_f32:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr s0, [x1]
|
|
; SVE-NEXT: ldr s1, [x0]
|
|
; SVE-NEXT: and z0.s, z0.s, #0x80000000
|
|
; SVE-NEXT: and z1.s, z1.s, #0x7fffffff
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str s0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_f32:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.s, #0x7fffffff
|
|
; SVE2-NEXT: ldr s1, [x1]
|
|
; SVE2-NEXT: ldr s2, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z1.d, z0.d
|
|
; SVE2-NEXT: str s2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_f32:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: ldr s0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr w8, [x1]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: str s0, [x0]
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load float, ptr %ap
|
|
%b = load float, ptr %bp
|
|
%r = call float @llvm.copysign.f32(float %a, float %b)
|
|
store float %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_f64(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_f64:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr d0, [x1]
|
|
; SVE-NEXT: ldr d1, [x0]
|
|
; SVE-NEXT: and z0.d, z0.d, #0x8000000000000000
|
|
; SVE-NEXT: and z1.d, z1.d, #0x7fffffffffffffff
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str d0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_f64:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.d, #0x7fffffffffffffff
|
|
; SVE2-NEXT: ldr d1, [x1]
|
|
; SVE2-NEXT: ldr d2, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z1.d, z0.d
|
|
; SVE2-NEXT: str d2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_f64:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: ldr d0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr x8, [x1]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d0, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: str d0, [x0]
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load double, ptr %ap
|
|
%b = load double, ptr %bp
|
|
%r = call double @llvm.copysign.f64(double %a, double %b)
|
|
store double %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ f16
|
|
|
|
define void @test_copysign_v4f16_v4f16(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v4f16_v4f16:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr d0, [x0]
|
|
; SVE-NEXT: ldr d1, [x1]
|
|
; SVE-NEXT: and z1.h, z1.h, #0x8000
|
|
; SVE-NEXT: and z0.h, z0.h, #0x7fff
|
|
; SVE-NEXT: orr z0.d, z0.d, z1.d
|
|
; SVE-NEXT: str d0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v4f16_v4f16:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.h, #32767 // =0x7fff
|
|
; SVE2-NEXT: ldr d1, [x0]
|
|
; SVE2-NEXT: ldr d2, [x1]
|
|
; SVE2-NEXT: bsl z1.d, z1.d, z2.d, z0.d
|
|
; SVE2-NEXT: str d1, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v4f16_v4f16:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 48
|
|
; NONEON-NOSVE-NEXT: ldr d0, [x1]
|
|
; NONEON-NOSVE-NEXT: ldr d1, [x0]
|
|
; NONEON-NOSVE-NEXT: stp d1, d0, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #38]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #20]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #36]
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #21]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #34]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #17]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #30]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #13]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #46]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #28]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #9]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #44]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #26]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #42]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: str d0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <4 x half>, ptr %ap
|
|
%b = load <4 x half>, ptr %bp
|
|
%r = call <4 x half> @llvm.copysign.v4f16(<4 x half> %a, <4 x half> %b)
|
|
store <4 x half> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_v8f16_v8f16(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v8f16_v8f16:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr q0, [x0]
|
|
; SVE-NEXT: ldr q1, [x1]
|
|
; SVE-NEXT: and z1.h, z1.h, #0x8000
|
|
; SVE-NEXT: and z0.h, z0.h, #0x7fff
|
|
; SVE-NEXT: orr z0.d, z0.d, z1.d
|
|
; SVE-NEXT: str q0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v8f16_v8f16:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.h, #32767 // =0x7fff
|
|
; SVE2-NEXT: ldr q1, [x0]
|
|
; SVE2-NEXT: ldr q2, [x1]
|
|
; SVE2-NEXT: bsl z1.d, z1.d, z2.d, z0.d
|
|
; SVE2-NEXT: str q1, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v8f16_v8f16:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #80
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 80
|
|
; NONEON-NOSVE-NEXT: ldr q0, [x1]
|
|
; NONEON-NOSVE-NEXT: ldr q1, [x0]
|
|
; NONEON-NOSVE-NEXT: stp q1, q0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #62]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #28]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #60]
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #29]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #58]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #25]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #20]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #56]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #54]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #52]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #50]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #4]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #46]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #21]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #78]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #44]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #17]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #76]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #42]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #13]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #74]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #9]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #72]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #38]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #5]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #70]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #36]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #1]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #68]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #34]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #66]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: str q0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #80
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <8 x half>, ptr %ap
|
|
%b = load <8 x half>, ptr %bp
|
|
%r = call <8 x half> @llvm.copysign.v8f16(<8 x half> %a, <8 x half> %b)
|
|
store <8 x half> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_v16f16_v16f16(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v16f16_v16f16:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldp q0, q3, [x1]
|
|
; SVE-NEXT: ldp q1, q2, [x0]
|
|
; SVE-NEXT: and z0.h, z0.h, #0x8000
|
|
; SVE-NEXT: and z3.h, z3.h, #0x8000
|
|
; SVE-NEXT: and z1.h, z1.h, #0x7fff
|
|
; SVE-NEXT: and z2.h, z2.h, #0x7fff
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: orr z1.d, z2.d, z3.d
|
|
; SVE-NEXT: stp q0, q1, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v16f16_v16f16:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.h, #32767 // =0x7fff
|
|
; SVE2-NEXT: ldp q1, q4, [x1]
|
|
; SVE2-NEXT: ldp q2, q3, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z1.d, z0.d
|
|
; SVE2-NEXT: bsl z3.d, z3.d, z4.d, z0.d
|
|
; SVE2-NEXT: stp q2, q3, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v16f16_v16f16:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #160
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 160
|
|
; NONEON-NOSVE-NEXT: ldp q1, q0, [x1]
|
|
; NONEON-NOSVE-NEXT: ldp q2, q3, [x0]
|
|
; NONEON-NOSVE-NEXT: stp q2, q1, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: stp q3, q0, [sp, #96]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #126]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #28]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #124]
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #29]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #122]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #25]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #20]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #120]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #118]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #116]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #114]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #4]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #112]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #94]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #60]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #92]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #56]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #90]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #52]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #88]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #86]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #44]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #84]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #82]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #36]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #80]
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #110]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #21]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #158]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #108]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #17]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #156]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #106]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #13]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #154]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #104]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #9]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #152]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #102]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #5]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #150]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #100]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #1]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #148]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #98]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #61]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #146]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #96]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #57]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #144]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #78]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #53]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #142]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #76]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #49]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #140]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #74]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #45]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #138]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #72]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #41]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #136]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #70]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #37]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #134]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #68]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: ldrb w8, [sp, #33]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #132]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #66]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #130]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #128]
|
|
; NONEON-NOSVE-NEXT: ldp q0, q1, [sp, #128]
|
|
; NONEON-NOSVE-NEXT: stp q0, q1, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #160
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <16 x half>, ptr %ap
|
|
%b = load <16 x half>, ptr %bp
|
|
%r = call <16 x half> @llvm.copysign.v16f16(<16 x half> %a, <16 x half> %b)
|
|
store <16 x half> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ f32
|
|
|
|
define void @test_copysign_v2f32_v2f32(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v2f32_v2f32:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr d0, [x0]
|
|
; SVE-NEXT: ldr d1, [x1]
|
|
; SVE-NEXT: and z1.s, z1.s, #0x80000000
|
|
; SVE-NEXT: and z0.s, z0.s, #0x7fffffff
|
|
; SVE-NEXT: orr z0.d, z0.d, z1.d
|
|
; SVE-NEXT: str d0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v2f32_v2f32:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.s, #0x7fffffff
|
|
; SVE2-NEXT: ldr d1, [x0]
|
|
; SVE2-NEXT: ldr d2, [x1]
|
|
; SVE2-NEXT: bsl z1.d, z1.d, z2.d, z0.d
|
|
; SVE2-NEXT: str d1, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v2f32_v2f32:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #32
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 32
|
|
; NONEON-NOSVE-NEXT: ldr d0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr d1, [x1]
|
|
; NONEON-NOSVE-NEXT: stp d0, d1, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: str d0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #32
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <2 x float>, ptr %ap
|
|
%b = load <2 x float>, ptr %bp
|
|
%r = call <2 x float> @llvm.copysign.v2f32(<2 x float> %a, <2 x float> %b)
|
|
store <2 x float> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_v4f32_v4f32(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v4f32_v4f32:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr q0, [x0]
|
|
; SVE-NEXT: ldr q1, [x1]
|
|
; SVE-NEXT: and z1.s, z1.s, #0x80000000
|
|
; SVE-NEXT: and z0.s, z0.s, #0x7fffffff
|
|
; SVE-NEXT: orr z0.d, z0.d, z1.d
|
|
; SVE-NEXT: str q0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v4f32_v4f32:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.s, #0x7fffffff
|
|
; SVE2-NEXT: ldr q1, [x0]
|
|
; SVE2-NEXT: ldr q2, [x1]
|
|
; SVE2-NEXT: bsl z1.d, z1.d, z2.d, z0.d
|
|
; SVE2-NEXT: str q1, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v4f32_v4f32:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: ldr q0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr q1, [x1]
|
|
; NONEON-NOSVE-NEXT: stp q0, q1, [sp, #-48]!
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 48
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #4]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: str q0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <4 x float>, ptr %ap
|
|
%b = load <4 x float>, ptr %bp
|
|
%r = call <4 x float> @llvm.copysign.v4f32(<4 x float> %a, <4 x float> %b)
|
|
store <4 x float> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_v8f32_v8f32(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v8f32_v8f32:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldp q0, q3, [x1]
|
|
; SVE-NEXT: ldp q1, q2, [x0]
|
|
; SVE-NEXT: and z0.s, z0.s, #0x80000000
|
|
; SVE-NEXT: and z3.s, z3.s, #0x80000000
|
|
; SVE-NEXT: and z1.s, z1.s, #0x7fffffff
|
|
; SVE-NEXT: and z2.s, z2.s, #0x7fffffff
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: orr z1.d, z2.d, z3.d
|
|
; SVE-NEXT: stp q0, q1, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v8f32_v8f32:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.s, #0x7fffffff
|
|
; SVE2-NEXT: ldp q1, q4, [x1]
|
|
; SVE2-NEXT: ldp q2, q3, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z1.d, z0.d
|
|
; SVE2-NEXT: bsl z3.d, z3.d, z4.d, z0.d
|
|
; SVE2-NEXT: stp q2, q3, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v8f32_v8f32:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #96
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 96
|
|
; NONEON-NOSVE-NEXT: ldp q2, q0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldp q3, q1, [x1]
|
|
; NONEON-NOSVE-NEXT: stp q2, q3, [sp]
|
|
; NONEON-NOSVE-NEXT: stp q0, q1, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #44]
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #56]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #88]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #36]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #80]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #72]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #4]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: ldp q0, q1, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: stp q0, q1, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #96
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <8 x float>, ptr %ap
|
|
%b = load <8 x float>, ptr %bp
|
|
%r = call <8 x float> @llvm.copysign.v8f32(<8 x float> %a, <8 x float> %b)
|
|
store <8 x float> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ f64
|
|
|
|
define void @test_copysign_v2f64_v2f64(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v2f64_v2f64:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldr q0, [x0]
|
|
; SVE-NEXT: ldr q1, [x1]
|
|
; SVE-NEXT: and z1.d, z1.d, #0x8000000000000000
|
|
; SVE-NEXT: and z0.d, z0.d, #0x7fffffffffffffff
|
|
; SVE-NEXT: orr z0.d, z0.d, z1.d
|
|
; SVE-NEXT: str q0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v2f64_v2f64:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.d, #0x7fffffffffffffff
|
|
; SVE2-NEXT: ldr q1, [x0]
|
|
; SVE2-NEXT: ldr q2, [x1]
|
|
; SVE2-NEXT: bsl z1.d, z1.d, z2.d, z0.d
|
|
; SVE2-NEXT: str q1, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v2f64_v2f64:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: ldr q0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr q1, [x1]
|
|
; NONEON-NOSVE-NEXT: stp q0, q1, [sp, #-48]!
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 48
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: ldp x8, x9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: tst x9, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d2, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp]
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d0, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: stp d0, d2, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: str q0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <2 x double>, ptr %ap
|
|
%b = load <2 x double>, ptr %bp
|
|
%r = call <2 x double> @llvm.copysign.v2f64(<2 x double> %a, <2 x double> %b)
|
|
store <2 x double> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_v4f64_v4f64(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v4f64_v4f64:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldp q0, q3, [x1]
|
|
; SVE-NEXT: ldp q1, q2, [x0]
|
|
; SVE-NEXT: and z0.d, z0.d, #0x8000000000000000
|
|
; SVE-NEXT: and z3.d, z3.d, #0x8000000000000000
|
|
; SVE-NEXT: and z1.d, z1.d, #0x7fffffffffffffff
|
|
; SVE-NEXT: and z2.d, z2.d, #0x7fffffffffffffff
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: orr z1.d, z2.d, z3.d
|
|
; SVE-NEXT: stp q0, q1, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v4f64_v4f64:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: mov z0.d, #0x7fffffffffffffff
|
|
; SVE2-NEXT: ldp q1, q4, [x1]
|
|
; SVE2-NEXT: ldp q2, q3, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z1.d, z0.d
|
|
; SVE2-NEXT: bsl z3.d, z3.d, z4.d, z0.d
|
|
; SVE2-NEXT: stp q2, q3, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v4f64_v4f64:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #96
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 96
|
|
; NONEON-NOSVE-NEXT: ldp q2, q0, [x0]
|
|
; NONEON-NOSVE-NEXT: ldp q3, q1, [x1]
|
|
; NONEON-NOSVE-NEXT: stp q2, q3, [sp]
|
|
; NONEON-NOSVE-NEXT: stp q0, q1, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: ldp x8, x9, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: tst x9, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d2, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: ldp x8, x9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d0, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: tst x9, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: stp d0, d2, [sp, #80]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d2, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp]
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d0, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: stp d0, d2, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: ldp q0, q1, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: stp q0, q1, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #96
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <4 x double>, ptr %ap
|
|
%b = load <4 x double>, ptr %bp
|
|
%r = call <4 x double> @llvm.copysign.v4f64(<4 x double> %a, <4 x double> %b)
|
|
store <4 x double> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ v2f32
|
|
|
|
define void @test_copysign_v2f32_v2f64(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v2f32_v2f64:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ptrue p0.d
|
|
; SVE-NEXT: ldr q0, [x1]
|
|
; SVE-NEXT: ldr d1, [x0]
|
|
; SVE-NEXT: fcvt z0.s, p0/m, z0.d
|
|
; SVE-NEXT: and z1.s, z1.s, #0x7fffffff
|
|
; SVE-NEXT: uzp1 z0.s, z0.s, z0.s
|
|
; SVE-NEXT: and z0.s, z0.s, #0x80000000
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str d0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v2f32_v2f64:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: ptrue p0.d
|
|
; SVE2-NEXT: ldr q0, [x1]
|
|
; SVE2-NEXT: mov z1.s, #0x7fffffff
|
|
; SVE2-NEXT: ldr d2, [x0]
|
|
; SVE2-NEXT: fcvt z0.s, p0/m, z0.d
|
|
; SVE2-NEXT: uzp1 z0.s, z0.s, z0.s
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z0.d, z1.d
|
|
; SVE2-NEXT: str d2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v2f32_v2f64:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 48
|
|
; NONEON-NOSVE-NEXT: ldr d1, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [x1]
|
|
; NONEON-NOSVE-NEXT: str d1, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: str q0, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: ldp x8, x9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: tst x9, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: str d0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <2 x float>, ptr %ap
|
|
%b = load <2 x double>, ptr %bp
|
|
%tmp0 = fptrunc <2 x double> %b to <2 x float>
|
|
%r = call <2 x float> @llvm.copysign.v2f32(<2 x float> %a, <2 x float> %tmp0)
|
|
store <2 x float> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ v4f32
|
|
|
|
; SplitVecOp #1
|
|
define void @test_copysign_v4f32_v4f64(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v4f32_v4f64:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldp q0, q1, [x1]
|
|
; SVE-NEXT: ptrue p0.d
|
|
; SVE-NEXT: fcvt z1.s, p0/m, z1.d
|
|
; SVE-NEXT: fcvt z0.s, p0/m, z0.d
|
|
; SVE-NEXT: ptrue p0.s, vl2
|
|
; SVE-NEXT: uzp1 z1.s, z1.s, z1.s
|
|
; SVE-NEXT: uzp1 z0.s, z0.s, z0.s
|
|
; SVE-NEXT: splice z0.s, p0, z0.s, z1.s
|
|
; SVE-NEXT: ldr q1, [x0]
|
|
; SVE-NEXT: and z1.s, z1.s, #0x7fffffff
|
|
; SVE-NEXT: and z0.s, z0.s, #0x80000000
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str q0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v4f32_v4f64:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: ldp q1, q0, [x1]
|
|
; SVE2-NEXT: ptrue p0.d
|
|
; SVE2-NEXT: fcvt z0.s, p0/m, z0.d
|
|
; SVE2-NEXT: fcvt z1.s, p0/m, z1.d
|
|
; SVE2-NEXT: ptrue p0.s, vl2
|
|
; SVE2-NEXT: uzp1 z2.s, z0.s, z0.s
|
|
; SVE2-NEXT: uzp1 z1.s, z1.s, z1.s
|
|
; SVE2-NEXT: splice z0.s, p0, { z1.s, z2.s }
|
|
; SVE2-NEXT: mov z1.s, #0x7fffffff
|
|
; SVE2-NEXT: ldr q2, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z0.d, z1.d
|
|
; SVE2-NEXT: str q2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v4f32_v4f64:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #64
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 64
|
|
; NONEON-NOSVE-NEXT: ldp q1, q0, [x1]
|
|
; NONEON-NOSVE-NEXT: ldr q2, [x0]
|
|
; NONEON-NOSVE-NEXT: str q2, [sp]
|
|
; NONEON-NOSVE-NEXT: stp q1, q0, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: ldp x8, x9, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: tst x9, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: ldp x8, x9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst x9, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #56]
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp, #4]
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s2, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: ldr s0, [sp]
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: stp s0, s2, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: str q0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #64
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <4 x float>, ptr %ap
|
|
%b = load <4 x double>, ptr %bp
|
|
%tmp0 = fptrunc <4 x double> %b to <4 x float>
|
|
%r = call <4 x float> @llvm.copysign.v4f32(<4 x float> %a, <4 x float> %tmp0)
|
|
store <4 x float> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ v2f64
|
|
|
|
define void @test_copysign_v2f64_v2f32(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v2f64_v2f32:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ptrue p0.d, vl2
|
|
; SVE-NEXT: ldr q0, [x0]
|
|
; SVE-NEXT: ld1w { z1.d }, p0/z, [x1]
|
|
; SVE-NEXT: and z0.d, z0.d, #0x7fffffffffffffff
|
|
; SVE-NEXT: fcvt z1.d, p0/m, z1.s
|
|
; SVE-NEXT: and z1.d, z1.d, #0x8000000000000000
|
|
; SVE-NEXT: orr z0.d, z0.d, z1.d
|
|
; SVE-NEXT: str q0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v2f64_v2f32:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: ptrue p0.d, vl2
|
|
; SVE2-NEXT: mov z1.d, #0x7fffffffffffffff
|
|
; SVE2-NEXT: ldr q2, [x0]
|
|
; SVE2-NEXT: ld1w { z0.d }, p0/z, [x1]
|
|
; SVE2-NEXT: fcvt z0.d, p0/m, z0.s
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z0.d, z1.d
|
|
; SVE2-NEXT: str q2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v2f64_v2f32:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 48
|
|
; NONEON-NOSVE-NEXT: ldr q1, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [x1]
|
|
; NONEON-NOSVE-NEXT: str q1, [sp]
|
|
; NONEON-NOSVE-NEXT: str d0, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d2, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d0, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: stp d0, d2, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: str q0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <2 x double>, ptr %ap
|
|
%b = load < 2 x float>, ptr %bp
|
|
%tmp0 = fpext <2 x float> %b to <2 x double>
|
|
%r = call <2 x double> @llvm.copysign.v2f64(<2 x double> %a, <2 x double> %tmp0)
|
|
store <2 x double> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ v4f64
|
|
|
|
; SplitVecRes mismatched
|
|
define void @test_copysign_v4f64_v4f32(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v4f64_v4f32:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ptrue p0.d, vl2
|
|
; SVE-NEXT: mov x8, #2 // =0x2
|
|
; SVE-NEXT: ldp q2, q3, [x0]
|
|
; SVE-NEXT: ld1w { z0.d }, p0/z, [x1]
|
|
; SVE-NEXT: ld1w { z1.d }, p0/z, [x1, x8, lsl #2]
|
|
; SVE-NEXT: and z2.d, z2.d, #0x7fffffffffffffff
|
|
; SVE-NEXT: and z3.d, z3.d, #0x7fffffffffffffff
|
|
; SVE-NEXT: fcvt z0.d, p0/m, z0.s
|
|
; SVE-NEXT: fcvt z1.d, p0/m, z1.s
|
|
; SVE-NEXT: and z0.d, z0.d, #0x8000000000000000
|
|
; SVE-NEXT: and z1.d, z1.d, #0x8000000000000000
|
|
; SVE-NEXT: orr z0.d, z2.d, z0.d
|
|
; SVE-NEXT: orr z1.d, z3.d, z1.d
|
|
; SVE-NEXT: stp q0, q1, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v4f64_v4f32:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: ptrue p0.d, vl2
|
|
; SVE2-NEXT: mov x8, #2 // =0x2
|
|
; SVE2-NEXT: mov z2.d, #0x7fffffffffffffff
|
|
; SVE2-NEXT: ldp q3, q4, [x0]
|
|
; SVE2-NEXT: ld1w { z0.d }, p0/z, [x1]
|
|
; SVE2-NEXT: ld1w { z1.d }, p0/z, [x1, x8, lsl #2]
|
|
; SVE2-NEXT: fcvt z0.d, p0/m, z0.s
|
|
; SVE2-NEXT: fcvt z1.d, p0/m, z1.s
|
|
; SVE2-NEXT: bsl z3.d, z3.d, z0.d, z2.d
|
|
; SVE2-NEXT: bsl z4.d, z4.d, z1.d, z2.d
|
|
; SVE2-NEXT: stp q3, q4, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v4f64_v4f32:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #96
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 96
|
|
; NONEON-NOSVE-NEXT: ldp q1, q2, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [x1]
|
|
; NONEON-NOSVE-NEXT: stp q0, q2, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldp d0, d2, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: str q1, [sp]
|
|
; NONEON-NOSVE-NEXT: stp d2, d0, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d2, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #56]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d0, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: stp d0, d2, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d2, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp]
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fabs d0, d0
|
|
; NONEON-NOSVE-NEXT: fneg d1, d0
|
|
; NONEON-NOSVE-NEXT: fcsel d0, d1, d0, ne
|
|
; NONEON-NOSVE-NEXT: stp d0, d2, [sp, #80]
|
|
; NONEON-NOSVE-NEXT: ldp q1, q0, [sp, #64]
|
|
; NONEON-NOSVE-NEXT: stp q0, q1, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #96
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <4 x double>, ptr %ap
|
|
%b = load <4 x float>, ptr %bp
|
|
%tmp0 = fpext <4 x float> %b to <4 x double>
|
|
%r = call <4 x double> @llvm.copysign.v4f64(<4 x double> %a, <4 x double> %tmp0)
|
|
store <4 x double> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ v4f16
|
|
|
|
define void @test_copysign_v4f16_v4f32(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v4f16_v4f32:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ptrue p0.s
|
|
; SVE-NEXT: ldr q0, [x1]
|
|
; SVE-NEXT: ldr d1, [x0]
|
|
; SVE-NEXT: fcvt z0.h, p0/m, z0.s
|
|
; SVE-NEXT: and z1.h, z1.h, #0x7fff
|
|
; SVE-NEXT: uzp1 z0.h, z0.h, z0.h
|
|
; SVE-NEXT: and z0.h, z0.h, #0x8000
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str d0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v4f16_v4f32:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: ptrue p0.s
|
|
; SVE2-NEXT: ldr q0, [x1]
|
|
; SVE2-NEXT: mov z1.h, #32767 // =0x7fff
|
|
; SVE2-NEXT: ldr d2, [x0]
|
|
; SVE2-NEXT: fcvt z0.h, p0/m, z0.s
|
|
; SVE2-NEXT: uzp1 z0.h, z0.h, z0.h
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z0.d, z1.d
|
|
; SVE2-NEXT: str d2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v4f16_v4f32:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 48
|
|
; NONEON-NOSVE-NEXT: ldr d1, [x0]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [x1]
|
|
; NONEON-NOSVE-NEXT: str d1, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: str q0, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #14]
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #46]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #44]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #10]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #42]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: str d0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #48
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <4 x half>, ptr %ap
|
|
%b = load <4 x float>, ptr %bp
|
|
%tmp0 = fptrunc <4 x float> %b to <4 x half>
|
|
%r = call <4 x half> @llvm.copysign.v4f16(<4 x half> %a, <4 x half> %tmp0)
|
|
store <4 x half> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
define void @test_copysign_v4f16_v4f64(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v4f16_v4f64:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldp q0, q1, [x1]
|
|
; SVE-NEXT: mov z2.d, z1.d[1]
|
|
; SVE-NEXT: mov z3.d, z0.d[1]
|
|
; SVE-NEXT: fcvt h1, d1
|
|
; SVE-NEXT: fcvt h0, d0
|
|
; SVE-NEXT: fcvt h2, d2
|
|
; SVE-NEXT: fcvt h3, d3
|
|
; SVE-NEXT: zip1 z1.h, z1.h, z2.h
|
|
; SVE-NEXT: zip1 z0.h, z0.h, z3.h
|
|
; SVE-NEXT: zip1 z0.s, z0.s, z1.s
|
|
; SVE-NEXT: ldr d1, [x0]
|
|
; SVE-NEXT: and z1.h, z1.h, #0x7fff
|
|
; SVE-NEXT: and z0.h, z0.h, #0x8000
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str d0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v4f16_v4f64:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: ldp q0, q1, [x1]
|
|
; SVE2-NEXT: mov z2.d, z1.d[1]
|
|
; SVE2-NEXT: mov z3.d, z0.d[1]
|
|
; SVE2-NEXT: fcvt h1, d1
|
|
; SVE2-NEXT: fcvt h0, d0
|
|
; SVE2-NEXT: fcvt h2, d2
|
|
; SVE2-NEXT: fcvt h3, d3
|
|
; SVE2-NEXT: zip1 z1.h, z1.h, z2.h
|
|
; SVE2-NEXT: zip1 z0.h, z0.h, z3.h
|
|
; SVE2-NEXT: mov z2.h, #32767 // =0x7fff
|
|
; SVE2-NEXT: zip1 z0.s, z0.s, z1.s
|
|
; SVE2-NEXT: ldr d1, [x0]
|
|
; SVE2-NEXT: bsl z1.d, z1.d, z0.d, z2.d
|
|
; SVE2-NEXT: str d1, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v4f16_v4f64:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #64
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 64
|
|
; NONEON-NOSVE-NEXT: ldp q1, q0, [x1]
|
|
; NONEON-NOSVE-NEXT: ldr d2, [x0]
|
|
; NONEON-NOSVE-NEXT: str d2, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: stp q1, q0, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #14]
|
|
; NONEON-NOSVE-NEXT: ldp x8, x9, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: tst x9, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: ldp x8, x9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #62]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst x9, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #60]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #10]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst x8, #0x8000000000000000
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #58]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #56]
|
|
; NONEON-NOSVE-NEXT: ldr d0, [sp, #56]
|
|
; NONEON-NOSVE-NEXT: str d0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #64
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <4 x half>, ptr %ap
|
|
%b = load <4 x double>, ptr %bp
|
|
%tmp0 = fptrunc <4 x double> %b to <4 x half>
|
|
%r = call <4 x half> @llvm.copysign.v4f16(<4 x half> %a, <4 x half> %tmp0)
|
|
store <4 x half> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
;============ v8f16
|
|
|
|
define void @test_copysign_v8f16_v8f32(ptr %ap, ptr %bp) {
|
|
; SVE-LABEL: test_copysign_v8f16_v8f32:
|
|
; SVE: // %bb.0:
|
|
; SVE-NEXT: ldp q0, q1, [x1]
|
|
; SVE-NEXT: ptrue p0.s
|
|
; SVE-NEXT: fcvt z1.h, p0/m, z1.s
|
|
; SVE-NEXT: fcvt z0.h, p0/m, z0.s
|
|
; SVE-NEXT: ptrue p0.h, vl4
|
|
; SVE-NEXT: uzp1 z1.h, z1.h, z1.h
|
|
; SVE-NEXT: uzp1 z0.h, z0.h, z0.h
|
|
; SVE-NEXT: splice z0.h, p0, z0.h, z1.h
|
|
; SVE-NEXT: ldr q1, [x0]
|
|
; SVE-NEXT: and z1.h, z1.h, #0x7fff
|
|
; SVE-NEXT: and z0.h, z0.h, #0x8000
|
|
; SVE-NEXT: orr z0.d, z1.d, z0.d
|
|
; SVE-NEXT: str q0, [x0]
|
|
; SVE-NEXT: ret
|
|
;
|
|
; SVE2-LABEL: test_copysign_v8f16_v8f32:
|
|
; SVE2: // %bb.0:
|
|
; SVE2-NEXT: ldp q1, q0, [x1]
|
|
; SVE2-NEXT: ptrue p0.s
|
|
; SVE2-NEXT: fcvt z0.h, p0/m, z0.s
|
|
; SVE2-NEXT: fcvt z1.h, p0/m, z1.s
|
|
; SVE2-NEXT: ptrue p0.h, vl4
|
|
; SVE2-NEXT: uzp1 z2.h, z0.h, z0.h
|
|
; SVE2-NEXT: uzp1 z1.h, z1.h, z1.h
|
|
; SVE2-NEXT: splice z0.h, p0, { z1.h, z2.h }
|
|
; SVE2-NEXT: mov z1.h, #32767 // =0x7fff
|
|
; SVE2-NEXT: ldr q2, [x0]
|
|
; SVE2-NEXT: bsl z2.d, z2.d, z0.d, z1.d
|
|
; SVE2-NEXT: str q2, [x0]
|
|
; SVE2-NEXT: ret
|
|
;
|
|
; NONEON-NOSVE-LABEL: test_copysign_v8f16_v8f32:
|
|
; NONEON-NOSVE: // %bb.0:
|
|
; NONEON-NOSVE-NEXT: sub sp, sp, #64
|
|
; NONEON-NOSVE-NEXT: .cfi_def_cfa_offset 64
|
|
; NONEON-NOSVE-NEXT: ldp q1, q0, [x1]
|
|
; NONEON-NOSVE-NEXT: ldr q2, [x0]
|
|
; NONEON-NOSVE-NEXT: str q2, [sp]
|
|
; NONEON-NOSVE-NEXT: stp q1, q0, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #14]
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #40]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #32]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #62]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #12]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #60]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #10]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #24]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #58]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #8]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #56]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #6]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: ldp w8, w9, [sp, #16]
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #54]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #4]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w9, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #52]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp, #2]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: tst w8, #0x80000000
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #50]
|
|
; NONEON-NOSVE-NEXT: ldr h0, [sp]
|
|
; NONEON-NOSVE-NEXT: fcvt s0, h0
|
|
; NONEON-NOSVE-NEXT: fabs s0, s0
|
|
; NONEON-NOSVE-NEXT: fneg s1, s0
|
|
; NONEON-NOSVE-NEXT: fcsel s0, s1, s0, ne
|
|
; NONEON-NOSVE-NEXT: fcvt h0, s0
|
|
; NONEON-NOSVE-NEXT: str h0, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: ldr q0, [sp, #48]
|
|
; NONEON-NOSVE-NEXT: str q0, [x0]
|
|
; NONEON-NOSVE-NEXT: add sp, sp, #64
|
|
; NONEON-NOSVE-NEXT: ret
|
|
%a = load <8 x half>, ptr %ap
|
|
%b = load <8 x float>, ptr %bp
|
|
%tmp0 = fptrunc <8 x float> %b to <8 x half>
|
|
%r = call <8 x half> @llvm.copysign.v8f16(<8 x half> %a, <8 x half> %tmp0)
|
|
store <8 x half> %r, ptr %ap
|
|
ret void
|
|
}
|
|
|
|
declare <4 x half> @llvm.copysign.v4f16(<4 x half> %a, <4 x half> %b) #0
|
|
declare <8 x half> @llvm.copysign.v8f16(<8 x half> %a, <8 x half> %b) #0
|
|
declare <16 x half> @llvm.copysign.v16f16(<16 x half> %a, <16 x half> %b) #0
|
|
|
|
declare <2 x float> @llvm.copysign.v2f32(<2 x float> %a, <2 x float> %b) #0
|
|
declare <4 x float> @llvm.copysign.v4f32(<4 x float> %a, <4 x float> %b) #0
|
|
declare <8 x float> @llvm.copysign.v8f32(<8 x float> %a, <8 x float> %b) #0
|
|
|
|
declare <2 x double> @llvm.copysign.v2f64(<2 x double> %a, <2 x double> %b) #0
|
|
declare <4 x double> @llvm.copysign.v4f64(<4 x double> %a, <4 x double> %b) #0
|