derek/TheRock
1
0 Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
Files
64d7f43c751a4e2cdcbadd467aaf47daf2c7a244
TheRock/compiler/amd-llvm/llvm/test/CodeGen/RISCV/combine-is_fpclass.ll
T
derek.holloway 7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
Details
CodeQL / Analyze (python) (push) Canceled after 0s
Details
pre-commit / pre-commit (push) Canceled after 0s
Details
Multi-Arch CI / setup (push) Canceled after 0s
Details
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Details
Setup (Multi-Arch) / setup (push) Canceled after 0s
Details
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Details
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Details
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Details
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Details
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Details
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
Details
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Details
Multi-Arch CI / CI Summary (push) Canceled after 0s
Details
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Details
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Details
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Details
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Details
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Details
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Details
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Details
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Details
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Details
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Details
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Details
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Details
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
Details
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

449 lines
18 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc -mtriple=riscv64 -mattr=+v,+f,+d < %s | FileCheck %s
define i1 @isneginf_constant_f32() nounwind {
; CHECK-LABEL: isneginf_constant_f32:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%f = tail call i1 @llvm.is.fpclass.f32(float 1.0, i32 4) ; 0x4 = "neginf"
ret i1 %f
}
define i8 @iszero_constant_v4f32() nounwind {
; CHECK-LABEL: iszero_constant_v4f32:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 8, e8, mf2, ta, ma
; CHECK-NEXT: vmclr.m v8
; CHECK-NEXT: vmv.x.s a0, v8
; CHECK-NEXT: ret
%f = tail call <8 x i1> @llvm.is.fpclass.v4f32(<8 x float> <float 1.0, float 2.0, float 3.0, float 4.0, float -1.0, float -2.0, float -3.0, float -4.0>, i32 96) ; 0x60 = "zero"
%r = bitcast <8 x i1> %f to i8
ret i8 %r
}
; elts 0,2 sign set, idx=0, ispos -> false
define i1 @extract_bitcast_sign_set_not_pos(<4 x i32> %bits, ptr %p) nounwind {
; CHECK-LABEL: extract_bitcast_sign_set_not_pos:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vid.v v9
; CHECK-NEXT: vsll.vi v9, v9, 31
; CHECK-NEXT: lui a1, 524288
; CHECK-NEXT: vrsub.vx v9, v9, a1
; CHECK-NEXT: vor.vv v8, v8, v9
; CHECK-NEXT: vse32.v v8, (a0)
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%masked = or <4 x i32> %bits, <i32 u0x80000000, i32 0, i32 u0x80000000, i32 0>
%fvec = bitcast <4 x i32> %masked to <4 x float>
store <4 x float> %fvec, ptr %p
%elt = extractelement <4 x float> %fvec, i32 0
%res = call i1 @llvm.is.fpclass.f32(float %elt, i32 960) ; 0x3C0 = any positive
ret i1 %res
}
; elts 0,2 sign set, idx=1 unknown sign, ispos -> cannot fold
define i1 @extract_bitcast_unknown_sign_ispos(<4 x i32> %bits, ptr %p) nounwind {
; CHECK-LABEL: extract_bitcast_unknown_sign_ispos:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vid.v v9
; CHECK-NEXT: vsll.vi v9, v9, 31
; CHECK-NEXT: lui a1, 524288
; CHECK-NEXT: vrsub.vx v9, v9, a1
; CHECK-NEXT: vor.vv v8, v8, v9
; CHECK-NEXT: vslidedown.vi v9, v8, 1
; CHECK-NEXT: vfmv.f.s fa5, v9
; CHECK-NEXT: fclass.s a1, fa5
; CHECK-NEXT: andi a1, a1, 240
; CHECK-NEXT: snez a1, a1
; CHECK-NEXT: vse32.v v8, (a0)
; CHECK-NEXT: mv a0, a1
; CHECK-NEXT: ret
%masked = or <4 x i32> %bits, <i32 u0x80000000, i32 0, i32 u0x80000000, i32 0>
%fvec = bitcast <4 x i32> %masked to <4 x float>
store <4 x float> %fvec, ptr %p
%elt = extractelement <4 x float> %fvec, i32 1
%res = call i1 @llvm.is.fpclass.f32(float %elt, i32 960) ; 0x3C0 = any positive
ret i1 %res
}
; elts 0,2 exp bit cleared (not NaN), idx=0, isnan -> false
define i1 @extract_bitcast_not_nan(<4 x i32> %bits, ptr %p) nounwind {
; CHECK-LABEL: extract_bitcast_not_nan:
; CHECK: # %bb.0:
; CHECK-NEXT: lui a1, 786432
; CHECK-NEXT: addi a1, a1, -1
; CHECK-NEXT: vsetivli zero, 2, e64, m1, ta, ma
; CHECK-NEXT: vmv.v.x v9, a1
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vand.vv v8, v8, v9
; CHECK-NEXT: vse32.v v8, (a0)
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%masked = and <4 x i32> %bits, <i32 u0xBFFFFFFF, i32 -1, i32 u0xBFFFFFFF, i32 -1>
%fvec = bitcast <4 x i32> %masked to <4 x float>
store <4 x float> %fvec, ptr %p
%elt = extractelement <4 x float> %fvec, i32 0
%res = call i1 @llvm.is.fpclass.f32(float %elt, i32 3) ; 0x3 = nan
ret i1 %res
}
; elts 0,2 exp bit cleared, idx=1 unchanged, isnan -> cannot fold
define i1 @extract_bitcast_maybe_nan(<4 x i32> %bits, ptr %p) nounwind {
; CHECK-LABEL: extract_bitcast_maybe_nan:
; CHECK: # %bb.0:
; CHECK-NEXT: lui a1, 786432
; CHECK-NEXT: addi a1, a1, -1
; CHECK-NEXT: vsetivli zero, 2, e64, m1, ta, ma
; CHECK-NEXT: vmv.v.x v9, a1
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vand.vv v8, v8, v9
; CHECK-NEXT: vslidedown.vi v9, v8, 1
; CHECK-NEXT: vfmv.f.s fa5, v9
; CHECK-NEXT: fclass.s a1, fa5
; CHECK-NEXT: andi a1, a1, 768
; CHECK-NEXT: snez a1, a1
; CHECK-NEXT: vse32.v v8, (a0)
; CHECK-NEXT: mv a0, a1
; CHECK-NEXT: ret
%masked = and <4 x i32> %bits, <i32 u0xBFFFFFFF, i32 -1, i32 u0xBFFFFFFF, i32 -1>
%fvec = bitcast <4 x i32> %masked to <4 x float>
store <4 x float> %fvec, ptr %p
%elt = extractelement <4 x float> %fvec, i32 1
%res = call i1 @llvm.is.fpclass.f32(float %elt, i32 3) ; 0x3 = nan
ret i1 %res
}
; mixed sign, unknown idx, ispos -> cannot fold
define i1 @extract_bitcast_unknown_idx_not_pos(<4 x i32> %bits, ptr %p, i32 %idx) nounwind {
; CHECK-LABEL: extract_bitcast_unknown_idx_not_pos:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vid.v v9
; CHECK-NEXT: lui a2, 524288
; CHECK-NEXT: slli a1, a1, 32
; CHECK-NEXT: vsll.vi v9, v9, 31
; CHECK-NEXT: vrsub.vx v9, v9, a2
; CHECK-NEXT: vor.vv v8, v8, v9
; CHECK-NEXT: srli a1, a1, 32
; CHECK-NEXT: vslidedown.vx v9, v8, a1
; CHECK-NEXT: vfmv.f.s fa5, v9
; CHECK-NEXT: fclass.s a1, fa5
; CHECK-NEXT: andi a1, a1, 240
; CHECK-NEXT: snez a1, a1
; CHECK-NEXT: vse32.v v8, (a0)
; CHECK-NEXT: mv a0, a1
; CHECK-NEXT: ret
%masked = or <4 x i32> %bits, <i32 u0x80000000, i32 0, i32 u0x80000000, i32 0>
%fvec = bitcast <4 x i32> %masked to <4 x float>
store <4 x float> %fvec, ptr %p
%elt = extractelement <4 x float> %fvec, i32 %idx
%res = call i1 @llvm.is.fpclass.f32(float %elt, i32 960) ; 0x3C0 = any positive
ret i1 %res
}
; idx=6 (OOB) -> poison, isnan -> false
define i1 @extract_bitcast_oob_idx(<4 x i32> %bits, ptr %p) nounwind {
; CHECK-LABEL: extract_bitcast_oob_idx:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vid.v v9
; CHECK-NEXT: vsll.vi v9, v9, 31
; CHECK-NEXT: lui a1, 524288
; CHECK-NEXT: vrsub.vx v9, v9, a1
; CHECK-NEXT: vor.vv v8, v8, v9
; CHECK-NEXT: vse32.v v8, (a0)
; CHECK-NEXT: ret
%masked = or <4 x i32> %bits, <i32 u0x80000000, i32 0, i32 u0x80000000, i32 0>
%fvec = bitcast <4 x i32> %masked to <4 x float>
store <4 x float> %fvec, ptr %p
%elt = extractelement <4 x float> %fvec, i32 6
%res = call i1 @llvm.is.fpclass.f32(float %elt, i32 3) ; 0x3 = nan
ret i1 %res
}
define i1 @extract_fabs_unknown_idx_isneg(<4 x float> %a0, i32 %idx) nounwind {
; CHECK-LABEL: extract_fabs_unknown_idx_isneg:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%abs = call <4 x float> @llvm.fabs.v4f32(<4 x float> %a0)
%elt = extractelement <4 x float> %abs, i32 %idx
%res = call i1 @llvm.is.fpclass.f32(float %elt, i32 60) ; 0x3C = negative
ret i1 %res
}
define i1 @extract_scalable_fabs_isneg(<vscale x 4 x float> %a0) nounwind {
; CHECK-LABEL: extract_scalable_fabs_isneg:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%abs = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> %a0)
%elt = extractelement <vscale x 4 x float> %abs, i32 0
%res = call i1 @llvm.is.fpclass.f32(float %elt, i32 60) ; 0x3C = negative
ret i1 %res
}
define i1 @fabs_is_nonneg_or_nan(float %x) nounwind {
; CHECK-LABEL: fabs_is_nonneg_or_nan:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 1
; CHECK-NEXT: ret
%a = call float @llvm.fabs.f32(float %x)
%res = call i1 @llvm.is.fpclass.f32(float %a, i32 963) ; 0x3C3 = any positive | nan
ret i1 %res
}
define <vscale x 4 x i1> @splat_constant_is_pos_normal() {
; CHECK-LABEL: splat_constant_is_pos_normal:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e8, mf2, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%res = call <vscale x 4 x i1> @llvm.is.fpclass.nxv4f32(<vscale x 4 x float> splat (float 1.0), i32 128) ; 128 = pos_normal
ret <vscale x 4 x i1> %res
}
define <vscale x 4 x i1> @splat_constant_isnan_false() {
; CHECK-LABEL: splat_constant_isnan_false:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e8, mf2, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%res = call <vscale x 4 x i1> @llvm.is.fpclass.nxv4f32(<vscale x 4 x float> splat (float 1.0), i32 3) ; 3 = nan
ret <vscale x 4 x i1> %res
}
define <vscale x 2 x i1> @splat_constant_f64_isinf_false() {
; CHECK-LABEL: splat_constant_f64_isinf_false:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e8, mf4, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%res = call <vscale x 2 x i1> @llvm.is.fpclass.nxv2f64(<vscale x 2 x double> splat (double 1.0), i32 516) ; 516 = inf
ret <vscale x 2 x i1> %res
}
define <4 x i1> @fneg_vec_not_negative(<4 x i32> %vecbits) nounwind {
; CHECK-LABEL: fneg_vec_not_negative:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%masked = or <4 x i32> %vecbits, splat(i32 u0x80000000) ; set sign bits
%fvec = bitcast <4 x i32> %masked to <4 x float>
%neg = fneg <4 x float> %fvec
%res = call <4 x i1> @llvm.is.fpclass.v4f32(<4 x float> %neg, i32 60) ; any negative
ret <4 x i1> %res
}
define <4 x i1> @fneg_vec_not_positive(<4 x i32> %vecbits) nounwind {
; CHECK-LABEL: fneg_vec_not_positive:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%masked = and <4 x i32> %vecbits, splat(i32 u0x7fffffff) ; clear sign bit
%fvec = bitcast <4 x i32> %masked to <4 x float>
%neg = fneg <4 x float> %fvec
%res = call <4 x i1> @llvm.is.fpclass.v4f32(<4 x float> %neg, i32 960) ; any positive
ret <4 x i1> %res
}
define <4 x i1> @fneg_vec_unknown_sign(<4 x i32> %vecbits) nounwind {
; CHECK-LABEL: fneg_vec_unknown_sign:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vfneg.v v8, v8
; CHECK-NEXT: vfclass.v v8, v8
; CHECK-NEXT: vand.vi v8, v8, 15
; CHECK-NEXT: vmsne.vi v0, v8, 0
; CHECK-NEXT: ret
%fvec = bitcast <4 x i32> %vecbits to <4 x float> ; no modification
%neg = fneg <4 x float> %fvec
%res = call <4 x i1> @llvm.is.fpclass.v4f32(<4 x float> %neg, i32 60) ; any negative
ret <4 x i1> %res
}
define i1 @copysign_nan_mag_is_finite_or_inf(float %y) nounwind {
; CHECK-LABEL: copysign_nan_mag_is_finite_or_inf:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%r = call float @llvm.copysign.f32(float 0x7FF8000000000000, float %y)
%res = call i1 @llvm.is.fpclass.f32(float %r, i32 1020) ; 0x3FC = finite | inf
ret i1 %res
}
define i1 @copysign_zero_mag_isnan(float %y) nounwind {
; CHECK-LABEL: copysign_zero_mag_isnan:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%r = call float @llvm.copysign.f32(float 0.0, float %y)
%res = call i1 @llvm.is.fpclass.f32(float %r, i32 3) ; 0x3 = nan
ret i1 %res
}
define i1 @copysign_inf_mag_isfinite_or_nan(float %y) nounwind {
; CHECK-LABEL: copysign_inf_mag_isfinite_or_nan:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%r = call float @llvm.copysign.f32(float 0x7FF0000000000000, float %y)
%res = call i1 @llvm.is.fpclass.f32(float %r, i32 507) ; 0x1FB = finite (0x1F8) | nan (0x3)
ret i1 %res
}
define i1 @copysign_normal_mag_not_normal(float %y) nounwind {
; CHECK-LABEL: copysign_normal_mag_not_normal:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 0
; CHECK-NEXT: ret
%r = call float @llvm.copysign.f32(float 1.0, float %y)
%res = call i1 @llvm.is.fpclass.f32(float %r, i32 759) ; 0x2F7 = all classes except +/-normal (0x108)
ret i1 %res
}
define i1 @copysign_normal_mag_is_normal(float %y) nounwind {
; CHECK-LABEL: copysign_normal_mag_is_normal:
; CHECK: # %bb.0:
; CHECK-NEXT: li a0, 1
; CHECK-NEXT: ret
%r = call float @llvm.copysign.f32(float 1.0, float %y)
%res = call i1 @llvm.is.fpclass.f32(float %r, i32 264) ; 0x108 = +/-normal
ret i1 %res
}
define i1 @copysign_unknown_sign_no_fold(float %x, float %y) nounwind {
; CHECK-LABEL: copysign_unknown_sign_no_fold:
; CHECK: # %bb.0:
; CHECK-NEXT: fsgnj.s fa5, fa0, fa1
; CHECK-NEXT: fclass.s a0, fa5
; CHECK-NEXT: andi a0, a0, 15
; CHECK-NEXT: snez a0, a0
; CHECK-NEXT: ret
%r = call float @llvm.copysign.f32(float %x, float %y)
%res = call i1 @llvm.is.fpclass.f32(float %r, i32 60) ; 0x3C = any negative
ret i1 %res
}
define <vscale x 4 x i1> @extract_subvec_scalable_isneg_false(<vscale x 8 x float> %a0) {
; CHECK-LABEL: extract_subvec_scalable_isneg_false:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e8, mf2, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%abs = call <vscale x 8 x float> @llvm.fabs.nxv8f32(<vscale x 8 x float> %a0)
%sub = call <vscale x 4 x float> @llvm.vector.extract.nxv4f32.nxv8f32(<vscale x 8 x float> %abs, i64 0)
%res = call <vscale x 4 x i1> @llvm.is.fpclass.nxv4f32(<vscale x 4 x float> %sub, i32 60) ; 60 = neginf,negnorm,negsub,negzero
ret <vscale x 4 x i1> %res
}
define <2 x i1> @extract_subvec_fixed_isneg_false(<4 x float> %a0) {
; CHECK-LABEL: extract_subvec_fixed_isneg_false:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%abs = call <4 x float> @llvm.fabs.v4f32(<4 x float> %a0)
%sub = call <2 x float> @llvm.vector.extract.v2f32.v4f32(<4 x float> %abs, i64 0)
%res = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> %sub, i32 60) ; 60 = neginf,negnorm,negsub,negzero
ret <2 x i1> %res
}
define <vscale x 4 x i1> @insert_subvec_scalable_both_isneg_false(<vscale x 4 x float> %base, <vscale x 2 x float> %sub) {
; CHECK-LABEL: insert_subvec_scalable_both_isneg_false:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetvli a0, zero, e8, mf2, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%absbase = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> %base)
%abssub = call <vscale x 2 x float> @llvm.fabs.nxv2f32(<vscale x 2 x float> %sub)
%ins = call <vscale x 4 x float> @llvm.vector.insert.nxv4f32.nxv2f32(<vscale x 4 x float> %absbase, <vscale x 2 x float> %abssub, i64 0)
%res = call <vscale x 4 x i1> @llvm.is.fpclass.nxv4f32(<vscale x 4 x float> %ins, i32 60) ; 60 = neginf,negnorm,negsub,negzero
ret <vscale x 4 x i1> %res
}
define <4 x i1> @insert_subvec_fixed_both_isneg_false(<4 x float> %base, <2 x float> %sub) {
; CHECK-LABEL: insert_subvec_fixed_both_isneg_false:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%absbase = call <4 x float> @llvm.fabs.v4f32(<4 x float> %base)
%abssub = call <2 x float> @llvm.fabs.v2f32(<2 x float> %sub)
%ins = call <4 x float> @llvm.vector.insert.v4f32.v2f32(<4 x float> %absbase, <2 x float> %abssub, i64 0)
%res = call <4 x i1> @llvm.is.fpclass.v4f32(<4 x float> %ins, i32 60) ; 60 = neginf,negnorm,negsub,negzero
ret <4 x i1> %res
}
define <2 x i1> @extract_subvec_fixed_mixed_to_pos_isneg_false(<4 x float> %a) {
; CHECK-LABEL: extract_subvec_fixed_mixed_to_pos_isneg_false:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 2, e8, mf8, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%abs = call <4 x float> @llvm.fabs.v4f32(<4 x float> %a)
%neg = fneg <4 x float> %abs
%mixed = shufflevector <4 x float> %abs, <4 x float> %neg, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
%sub = call <2 x float> @llvm.vector.extract.v2f32.v4f32(<4 x float> %mixed, i64 0)
%res = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> %sub, i32 60) ; 60 = neginf,negnorm,negsub,negzero
ret <2 x i1> %res
}
define <2 x i1> @extract_subvec_fixed_straddle_isneg_unknown(<4 x float> %a) {
; CHECK-LABEL: extract_subvec_fixed_straddle_isneg_unknown:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vfabs.v v8, v8
; CHECK-NEXT: vfneg.v v9, v8
; CHECK-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT: vslideup.vi v8, v9, 1
; CHECK-NEXT: vfclass.v v8, v8
; CHECK-NEXT: vand.vi v8, v8, 15
; CHECK-NEXT: vmsne.vi v0, v8, 0
; CHECK-NEXT: ret
%abs = call <4 x float> @llvm.fabs.v4f32(<4 x float> %a)
%neg = fneg <4 x float> %abs
%mixed = shufflevector <4 x float> %abs, <4 x float> %neg, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
%sub = call <2 x float> @llvm.vector.extract.v2f32.v4f32(<4 x float> %mixed, i64 0)
%res = call <2 x i1> @llvm.is.fpclass.v2f32(<2 x float> %sub, i32 60) ; 60 = neginf,negnorm,negsub,negzero
ret <2 x i1> %res
}
define <4 x i1> @insert_subvec_fixed_pos_base_neg_sub_isneg_unknown(<4 x float> %base, <2 x float> %sub) {
; CHECK-LABEL: insert_subvec_fixed_pos_base_neg_sub_isneg_unknown:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vfabs.v v8, v8
; CHECK-NEXT: vsetivli zero, 2, e32, mf2, ta, ma
; CHECK-NEXT: vfneg.v v9, v9
; CHECK-NEXT: vsetivli zero, 4, e32, m1, ta, ma
; CHECK-NEXT: vslideup.vi v8, v9, 2
; CHECK-NEXT: vfclass.v v8, v8
; CHECK-NEXT: vand.vi v8, v8, 15
; CHECK-NEXT: vmsne.vi v0, v8, 0
; CHECK-NEXT: ret
%abs_base = call <4 x float> @llvm.fabs.v4f32(<4 x float> %base)
%neg_sub = fneg <2 x float> %sub
%ins = call <4 x float> @llvm.vector.insert.v4f32.v2f32(<4 x float> %abs_base, <2 x float> %neg_sub, i64 2)
%res = call <4 x i1> @llvm.is.fpclass.v4f32(<4 x float> %ins, i32 60) ; 60 = neginf,negnorm,negsub,negzero
ret <4 x i1> %res
}
define <4 x i1> @insert_subvec_fixed_pos_base_pos_sub_isneg_false(<4 x float> %base, <2 x float> %sub) {
; CHECK-LABEL: insert_subvec_fixed_pos_base_pos_sub_isneg_false:
; CHECK: # %bb.0:
; CHECK-NEXT: vsetivli zero, 4, e8, mf4, ta, ma
; CHECK-NEXT: vmclr.m v0
; CHECK-NEXT: ret
%abs_base = call <4 x float> @llvm.fabs.v4f32(<4 x float> %base)
%abs_sub = call <2 x float> @llvm.fabs.v2f32(<2 x float> %sub)
%ins = call <4 x float> @llvm.vector.insert.v4f32.v2f32(<4 x float> %abs_base, <2 x float> %abs_sub, i64 2)
%res = call <4 x i1> @llvm.is.fpclass.v4f32(<4 x float> %ins, i32 60) ; 60 = neginf,negnorm,negsub,negzero
ret <4 x i1> %res
}
Reference in new issue View Git Blame Copy Permalink