Google Git
Sign in
llvm / llvm-project / c49ef1448d1779385e0e0ea88b1d8343f4830a79 / . / llvm / test / CodeGen / AArch64 / fptoui-sat-vector.ll
blob: a903d842ec9b3f30914858b81383d18533f402d6 [file] [log] [blame]
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=aarch64 | FileCheck %s --check-prefixes=CHECK,CHECK-CVT
; RUN: llc < %s -mtriple=aarch64 -mattr=+fullfp16 | FileCheck %s --check-prefixes=CHECK,CHECK-FP16
;
; Float to unsigned 32-bit -- Vector size variation
;
declare <1 x i32> @llvm.fptoui.sat.v1f32.v1i32 (<1 x float>)
declare <2 x i32> @llvm.fptoui.sat.v2f32.v2i32 (<2 x float>)
declare <3 x i32> @llvm.fptoui.sat.v3f32.v3i32 (<3 x float>)
declare <4 x i32> @llvm.fptoui.sat.v4f32.v4i32 (<4 x float>)
declare <5 x i32> @llvm.fptoui.sat.v5f32.v5i32 (<5 x float>)
declare <6 x i32> @llvm.fptoui.sat.v6f32.v6i32 (<6 x float>)
declare <7 x i32> @llvm.fptoui.sat.v7f32.v7i32 (<7 x float>)
declare <8 x i32> @llvm.fptoui.sat.v8f32.v8i32 (<8 x float>)
define <1 x i32> @test_unsigned_v1f32_v1i32(<1 x float> %f) {
; CHECK-LABEL: test_unsigned_v1f32_v1i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.2s, v0.2s
; CHECK-NEXT: ret
%x = call <1 x i32> @llvm.fptoui.sat.v1f32.v1i32(<1 x float> %f)
ret <1 x i32> %x
}
define <2 x i32> @test_unsigned_v2f32_v2i32(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.2s, v0.2s
; CHECK-NEXT: ret
%x = call <2 x i32> @llvm.fptoui.sat.v2f32.v2i32(<2 x float> %f)
ret <2 x i32> %x
}
define <3 x i32> @test_unsigned_v3f32_v3i32(<3 x float> %f) {
; CHECK-LABEL: test_unsigned_v3f32_v3i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: ret
%x = call <3 x i32> @llvm.fptoui.sat.v3f32.v3i32(<3 x float> %f)
ret <3 x i32> %x
}
define <4 x i32> @test_unsigned_v4f32_v4i32(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: ret
%x = call <4 x i32> @llvm.fptoui.sat.v4f32.v4i32(<4 x float> %f)
ret <4 x i32> %x
}
define <5 x i32> @test_unsigned_v5f32_v5i32(<5 x float> %f) {
; CHECK-LABEL: test_unsigned_v5f32_v5i32:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $s0 killed $s0 def $q0
; CHECK-NEXT: // kill: def $s1 killed $s1 def $q1
; CHECK-NEXT: // kill: def $s2 killed $s2 def $q2
; CHECK-NEXT: // kill: def $s3 killed $s3 def $q3
; CHECK-NEXT: // kill: def $s4 killed $s4 def $q4
; CHECK-NEXT: mov v0.s[1], v1.s[0]
; CHECK-NEXT: fcvtzu v4.4s, v4.4s
; CHECK-NEXT: mov v0.s[2], v2.s[0]
; CHECK-NEXT: fmov w4, s4
; CHECK-NEXT: mov v0.s[3], v3.s[0]
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: mov w1, v0.s[1]
; CHECK-NEXT: mov w2, v0.s[2]
; CHECK-NEXT: mov w3, v0.s[3]
; CHECK-NEXT: fmov w0, s0
; CHECK-NEXT: ret
%x = call <5 x i32> @llvm.fptoui.sat.v5f32.v5i32(<5 x float> %f)
ret <5 x i32> %x
}
define <6 x i32> @test_unsigned_v6f32_v6i32(<6 x float> %f) {
; CHECK-LABEL: test_unsigned_v6f32_v6i32:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $s0 killed $s0 def $q0
; CHECK-NEXT: // kill: def $s1 killed $s1 def $q1
; CHECK-NEXT: // kill: def $s2 killed $s2 def $q2
; CHECK-NEXT: // kill: def $s4 killed $s4 def $q4
; CHECK-NEXT: // kill: def $s5 killed $s5 def $q5
; CHECK-NEXT: // kill: def $s3 killed $s3 def $q3
; CHECK-NEXT: mov v0.s[1], v1.s[0]
; CHECK-NEXT: mov v4.s[1], v5.s[0]
; CHECK-NEXT: mov v0.s[2], v2.s[0]
; CHECK-NEXT: fcvtzu v1.4s, v4.4s
; CHECK-NEXT: mov v0.s[3], v3.s[0]
; CHECK-NEXT: mov w5, v1.s[1]
; CHECK-NEXT: fmov w4, s1
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: mov w1, v0.s[1]
; CHECK-NEXT: mov w2, v0.s[2]
; CHECK-NEXT: mov w3, v0.s[3]
; CHECK-NEXT: fmov w0, s0
; CHECK-NEXT: ret
%x = call <6 x i32> @llvm.fptoui.sat.v6f32.v6i32(<6 x float> %f)
ret <6 x i32> %x
}
define <7 x i32> @test_unsigned_v7f32_v7i32(<7 x float> %f) {
; CHECK-LABEL: test_unsigned_v7f32_v7i32:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $s0 killed $s0 def $q0
; CHECK-NEXT: // kill: def $s1 killed $s1 def $q1
; CHECK-NEXT: // kill: def $s4 killed $s4 def $q4
; CHECK-NEXT: // kill: def $s5 killed $s5 def $q5
; CHECK-NEXT: // kill: def $s2 killed $s2 def $q2
; CHECK-NEXT: // kill: def $s6 killed $s6 def $q6
; CHECK-NEXT: // kill: def $s3 killed $s3 def $q3
; CHECK-NEXT: mov v0.s[1], v1.s[0]
; CHECK-NEXT: mov v4.s[1], v5.s[0]
; CHECK-NEXT: mov v0.s[2], v2.s[0]
; CHECK-NEXT: mov v4.s[2], v6.s[0]
; CHECK-NEXT: mov v0.s[3], v3.s[0]
; CHECK-NEXT: fcvtzu v1.4s, v4.4s
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: mov w5, v1.s[1]
; CHECK-NEXT: mov w6, v1.s[2]
; CHECK-NEXT: fmov w4, s1
; CHECK-NEXT: mov w1, v0.s[1]
; CHECK-NEXT: mov w2, v0.s[2]
; CHECK-NEXT: mov w3, v0.s[3]
; CHECK-NEXT: fmov w0, s0
; CHECK-NEXT: ret
%x = call <7 x i32> @llvm.fptoui.sat.v7f32.v7i32(<7 x float> %f)
ret <7 x i32> %x
}
define <8 x i32> @test_unsigned_v8f32_v8i32(<8 x float> %f) {
; CHECK-LABEL: test_unsigned_v8f32_v8i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: fcvtzu v1.4s, v1.4s
; CHECK-NEXT: ret
%x = call <8 x i32> @llvm.fptoui.sat.v8f32.v8i32(<8 x float> %f)
ret <8 x i32> %x
}
;
; Double to unsigned 32-bit -- Vector size variation
;
declare <1 x i32> @llvm.fptoui.sat.v1f64.v1i32 (<1 x double>)
declare <2 x i32> @llvm.fptoui.sat.v2f64.v2i32 (<2 x double>)
declare <3 x i32> @llvm.fptoui.sat.v3f64.v3i32 (<3 x double>)
declare <4 x i32> @llvm.fptoui.sat.v4f64.v4i32 (<4 x double>)
declare <5 x i32> @llvm.fptoui.sat.v5f64.v5i32 (<5 x double>)
declare <6 x i32> @llvm.fptoui.sat.v6f64.v6i32 (<6 x double>)
define <1 x i32> @test_unsigned_v1f64_v1i32(<1 x double> %f) {
; CHECK-LABEL: test_unsigned_v1f64_v1i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu w8, d0
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: ret
%x = call <1 x i32> @llvm.fptoui.sat.v1f64.v1i32(<1 x double> %f)
ret <1 x i32> %x
}
define <2 x i32> @test_unsigned_v2f64_v2i32(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i32:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d1, v0.d[1]
; CHECK-NEXT: fcvtzu w8, d0
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: mov v0.s[1], w8
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
%x = call <2 x i32> @llvm.fptoui.sat.v2f64.v2i32(<2 x double> %f)
ret <2 x i32> %x
}
define <3 x i32> @test_unsigned_v3f64_v3i32(<3 x double> %f) {
; CHECK-LABEL: test_unsigned_v3f64_v3i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu w8, d0
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: mov v0.s[1], w8
; CHECK-NEXT: fcvtzu w8, d2
; CHECK-NEXT: mov v0.s[2], w8
; CHECK-NEXT: fcvtzu w8, d0
; CHECK-NEXT: mov v0.s[3], w8
; CHECK-NEXT: ret
%x = call <3 x i32> @llvm.fptoui.sat.v3f64.v3i32(<3 x double> %f)
ret <3 x i32> %x
}
define <4 x i32> @test_unsigned_v4f64_v4i32(<4 x double> %f) {
; CHECK-LABEL: test_unsigned_v4f64_v4i32:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d2, v0.d[1]
; CHECK-NEXT: fcvtzu w8, d0
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: fcvtzu w8, d2
; CHECK-NEXT: mov v0.s[1], w8
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: mov d1, v1.d[1]
; CHECK-NEXT: mov v0.s[2], w8
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: mov v0.s[3], w8
; CHECK-NEXT: ret
%x = call <4 x i32> @llvm.fptoui.sat.v4f64.v4i32(<4 x double> %f)
ret <4 x i32> %x
}
define <5 x i32> @test_unsigned_v5f64_v5i32(<5 x double> %f) {
; CHECK-LABEL: test_unsigned_v5f64_v5i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu w0, d0
; CHECK-NEXT: fcvtzu w1, d1
; CHECK-NEXT: fcvtzu w2, d2
; CHECK-NEXT: fcvtzu w3, d3
; CHECK-NEXT: fcvtzu w4, d4
; CHECK-NEXT: ret
%x = call <5 x i32> @llvm.fptoui.sat.v5f64.v5i32(<5 x double> %f)
ret <5 x i32> %x
}
define <6 x i32> @test_unsigned_v6f64_v6i32(<6 x double> %f) {
; CHECK-LABEL: test_unsigned_v6f64_v6i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu w0, d0
; CHECK-NEXT: fcvtzu w1, d1
; CHECK-NEXT: fcvtzu w2, d2
; CHECK-NEXT: fcvtzu w3, d3
; CHECK-NEXT: fcvtzu w4, d4
; CHECK-NEXT: fcvtzu w5, d5
; CHECK-NEXT: ret
%x = call <6 x i32> @llvm.fptoui.sat.v6f64.v6i32(<6 x double> %f)
ret <6 x i32> %x
}
;
; FP128 to unsigned 32-bit -- Vector size variation
;
declare <1 x i32> @llvm.fptoui.sat.v1f128.v1i32 (<1 x fp128>)
declare <2 x i32> @llvm.fptoui.sat.v2f128.v2i32 (<2 x fp128>)
declare <3 x i32> @llvm.fptoui.sat.v3f128.v3i32 (<3 x fp128>)
declare <4 x i32> @llvm.fptoui.sat.v4f128.v4i32 (<4 x fp128>)
define <1 x i32> @test_unsigned_v1f128_v1i32(<1 x fp128> %f) {
; CHECK-LABEL: test_unsigned_v1f128_v1i32:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #32
; CHECK-NEXT: stp x30, x19, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 32
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w30, -16
; CHECK-NEXT: adrp x8, .LCPI14_0
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI14_0]
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: adrp x8, .LCPI14_1
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI14_1]
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: csinv w8, w19, wzr, le
; CHECK-NEXT: ldp x30, x19, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: add sp, sp, #32
; CHECK-NEXT: ret
%x = call <1 x i32> @llvm.fptoui.sat.v1f128.v1i32(<1 x fp128> %f)
ret <1 x i32> %x
}
define <2 x i32> @test_unsigned_v2f128_v2i32(<2 x fp128> %f) {
; CHECK-LABEL: test_unsigned_v2f128_v2i32:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #96
; CHECK-NEXT: str x30, [sp, #64] // 8-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 96
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: mov v2.16b, v1.16b
; CHECK-NEXT: adrp x8, .LCPI15_0
; CHECK-NEXT: stp q1, q0, [sp, #32] // 32-byte Folded Spill
; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI15_0]
; CHECK-NEXT: mov v0.16b, v2.16b
; CHECK-NEXT: str q1, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: adrp x8, .LCPI15_1
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI15_1]
; CHECK-NEXT: str q1, [sp] // 16-byte Folded Spill
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: ldr q1, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csinv w20, w19, wzr, le
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: ldr q1, [sp] // 16-byte Folded Reload
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: csinv w8, w19, wzr, le
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: mov v0.s[1], w20
; CHECK-NEXT: ldp x20, x19, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: add sp, sp, #96
; CHECK-NEXT: ret
%x = call <2 x i32> @llvm.fptoui.sat.v2f128.v2i32(<2 x fp128> %f)
ret <2 x i32> %x
}
define <3 x i32> @test_unsigned_v3f128_v3i32(<3 x fp128> %f) {
; CHECK-LABEL: test_unsigned_v3f128_v3i32:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #112
; CHECK-NEXT: str x30, [sp, #80] // 8-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #96] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 112
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: stp q0, q2, [sp, #48] // 32-byte Folded Spill
; CHECK-NEXT: adrp x8, .LCPI16_0
; CHECK-NEXT: mov v2.16b, v1.16b
; CHECK-NEXT: str q1, [sp] // 16-byte Folded Spill
; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI16_0]
; CHECK-NEXT: mov v0.16b, v2.16b
; CHECK-NEXT: str q1, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: adrp x8, .LCPI16_1
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI16_1]
; CHECK-NEXT: str q1, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: ldp q1, q0, [sp, #32] // 32-byte Folded Reload
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: csinv w20, w19, wzr, le
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: ldr q1, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: ldr q1, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: csinv w8, w19, wzr, le
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: mov v0.s[1], w20
; CHECK-NEXT: str q0, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: ldr q0, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: ldr q0, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: ldr q1, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: csinv w8, w19, wzr, le
; CHECK-NEXT: ldr x30, [sp, #80] // 8-byte Folded Reload
; CHECK-NEXT: ldp x20, x19, [sp, #96] // 16-byte Folded Reload
; CHECK-NEXT: mov v0.s[2], w8
; CHECK-NEXT: add sp, sp, #112
; CHECK-NEXT: ret
%x = call <3 x i32> @llvm.fptoui.sat.v3f128.v3i32(<3 x fp128> %f)
ret <3 x i32> %x
}
define <4 x i32> @test_unsigned_v4f128_v4i32(<4 x fp128> %f) {
; CHECK-LABEL: test_unsigned_v4f128_v4i32:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #128
; CHECK-NEXT: str x30, [sp, #96] // 8-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #112] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 128
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: stp q0, q2, [sp, #16] // 32-byte Folded Spill
; CHECK-NEXT: adrp x8, .LCPI17_0
; CHECK-NEXT: mov v2.16b, v1.16b
; CHECK-NEXT: str q1, [sp] // 16-byte Folded Spill
; CHECK-NEXT: str q3, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI17_0]
; CHECK-NEXT: mov v0.16b, v2.16b
; CHECK-NEXT: str q1, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: adrp x8, .LCPI17_1
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: ldr q1, [x8, :lo12:.LCPI17_1]
; CHECK-NEXT: str q1, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: ldr q1, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: csinv w20, w19, wzr, le
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: ldr q1, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: ldr q1, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: csinv w8, w19, wzr, le
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: mov v0.s[1], w20
; CHECK-NEXT: str q0, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: ldp q0, q1, [sp, #32] // 32-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csinv w8, w19, wzr, le
; CHECK-NEXT: mov v0.s[2], w8
; CHECK-NEXT: str q0, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: ldp q1, q0, [sp, #64] // 32-byte Folded Reload
; CHECK-NEXT: bl __getf2
; CHECK-NEXT: ldr q0, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: mov w19, w0
; CHECK-NEXT: bl __fixunstfsi
; CHECK-NEXT: ldr q0, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: cmp w19, #0
; CHECK-NEXT: ldr q1, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: csel w19, wzr, w0, lt
; CHECK-NEXT: bl __gttf2
; CHECK-NEXT: cmp w0, #0
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csinv w8, w19, wzr, le
; CHECK-NEXT: ldr x30, [sp, #96] // 8-byte Folded Reload
; CHECK-NEXT: ldp x20, x19, [sp, #112] // 16-byte Folded Reload
; CHECK-NEXT: mov v0.s[3], w8
; CHECK-NEXT: add sp, sp, #128
; CHECK-NEXT: ret
%x = call <4 x i32> @llvm.fptoui.sat.v4f128.v4i32(<4 x fp128> %f)
ret <4 x i32> %x
}
;
; FP16 to unsigned 32-bit -- Vector size variation
;
declare <1 x i32> @llvm.fptoui.sat.v1f16.v1i32 (<1 x half>)
declare <2 x i32> @llvm.fptoui.sat.v2f16.v2i32 (<2 x half>)
declare <3 x i32> @llvm.fptoui.sat.v3f16.v3i32 (<3 x half>)
declare <4 x i32> @llvm.fptoui.sat.v4f16.v4i32 (<4 x half>)
declare <5 x i32> @llvm.fptoui.sat.v5f16.v5i32 (<5 x half>)
declare <6 x i32> @llvm.fptoui.sat.v6f16.v6i32 (<6 x half>)
declare <7 x i32> @llvm.fptoui.sat.v7f16.v7i32 (<7 x half>)
declare <8 x i32> @llvm.fptoui.sat.v8f16.v8i32 (<8 x half>)
define <1 x i32> @test_unsigned_v1f16_v1i32(<1 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v1f16_v1i32:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvt s0, h0
; CHECK-CVT-NEXT: fcvtzu w8, s0
; CHECK-CVT-NEXT: fmov s0, w8
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v1f16_v1i32:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: fcvtzu w8, h0
; CHECK-FP16-NEXT: fmov s0, w8
; CHECK-FP16-NEXT: ret
%x = call <1 x i32> @llvm.fptoui.sat.v1f16.v1i32(<1 x half> %f)
ret <1 x i32> %x
}
define <2 x i32> @test_unsigned_v2f16_v2i32(<2 x half> %f) {
; CHECK-LABEL: test_unsigned_v2f16_v2i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v0.4s, v0.4h
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
%x = call <2 x i32> @llvm.fptoui.sat.v2f16.v2i32(<2 x half> %f)
ret <2 x i32> %x
}
define <3 x i32> @test_unsigned_v3f16_v3i32(<3 x half> %f) {
; CHECK-LABEL: test_unsigned_v3f16_v3i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v0.4s, v0.4h
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: ret
%x = call <3 x i32> @llvm.fptoui.sat.v3f16.v3i32(<3 x half> %f)
ret <3 x i32> %x
}
define <4 x i32> @test_unsigned_v4f16_v4i32(<4 x half> %f) {
; CHECK-LABEL: test_unsigned_v4f16_v4i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v0.4s, v0.4h
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: ret
%x = call <4 x i32> @llvm.fptoui.sat.v4f16.v4i32(<4 x half> %f)
ret <4 x i32> %x
}
define <5 x i32> @test_unsigned_v5f16_v5i32(<5 x half> %f) {
; CHECK-LABEL: test_unsigned_v5f16_v5i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v1.4s, v0.4h
; CHECK-NEXT: fcvtl2 v0.4s, v0.8h
; CHECK-NEXT: fcvtzu v1.4s, v1.4s
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: mov w1, v1.s[1]
; CHECK-NEXT: mov w2, v1.s[2]
; CHECK-NEXT: mov w3, v1.s[3]
; CHECK-NEXT: fmov w0, s1
; CHECK-NEXT: fmov w4, s0
; CHECK-NEXT: ret
%x = call <5 x i32> @llvm.fptoui.sat.v5f16.v5i32(<5 x half> %f)
ret <5 x i32> %x
}
define <6 x i32> @test_unsigned_v6f16_v6i32(<6 x half> %f) {
; CHECK-LABEL: test_unsigned_v6f16_v6i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v1.4s, v0.4h
; CHECK-NEXT: fcvtl2 v0.4s, v0.8h
; CHECK-NEXT: fcvtzu v1.4s, v1.4s
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: mov w1, v1.s[1]
; CHECK-NEXT: mov w2, v1.s[2]
; CHECK-NEXT: mov w3, v1.s[3]
; CHECK-NEXT: mov w5, v0.s[1]
; CHECK-NEXT: fmov w0, s1
; CHECK-NEXT: fmov w4, s0
; CHECK-NEXT: ret
%x = call <6 x i32> @llvm.fptoui.sat.v6f16.v6i32(<6 x half> %f)
ret <6 x i32> %x
}
define <7 x i32> @test_unsigned_v7f16_v7i32(<7 x half> %f) {
; CHECK-LABEL: test_unsigned_v7f16_v7i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v1.4s, v0.4h
; CHECK-NEXT: fcvtl2 v0.4s, v0.8h
; CHECK-NEXT: fcvtzu v1.4s, v1.4s
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: mov w1, v1.s[1]
; CHECK-NEXT: mov w2, v1.s[2]
; CHECK-NEXT: mov w3, v1.s[3]
; CHECK-NEXT: mov w5, v0.s[1]
; CHECK-NEXT: mov w6, v0.s[2]
; CHECK-NEXT: fmov w0, s1
; CHECK-NEXT: fmov w4, s0
; CHECK-NEXT: ret
%x = call <7 x i32> @llvm.fptoui.sat.v7f16.v7i32(<7 x half> %f)
ret <7 x i32> %x
}
define <8 x i32> @test_unsigned_v8f16_v8i32(<8 x half> %f) {
; CHECK-LABEL: test_unsigned_v8f16_v8i32:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl2 v1.4s, v0.8h
; CHECK-NEXT: fcvtl v0.4s, v0.4h
; CHECK-NEXT: fcvtzu v1.4s, v1.4s
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: ret
%x = call <8 x i32> @llvm.fptoui.sat.v8f16.v8i32(<8 x half> %f)
ret <8 x i32> %x
}
;
; 2-Vector float to unsigned integer -- result size variation
;
declare <2 x i1> @llvm.fptoui.sat.v2f32.v2i1 (<2 x float>)
declare <2 x i8> @llvm.fptoui.sat.v2f32.v2i8 (<2 x float>)
declare <2 x i13> @llvm.fptoui.sat.v2f32.v2i13 (<2 x float>)
declare <2 x i16> @llvm.fptoui.sat.v2f32.v2i16 (<2 x float>)
declare <2 x i19> @llvm.fptoui.sat.v2f32.v2i19 (<2 x float>)
declare <2 x i50> @llvm.fptoui.sat.v2f32.v2i50 (<2 x float>)
declare <2 x i64> @llvm.fptoui.sat.v2f32.v2i64 (<2 x float>)
declare <2 x i100> @llvm.fptoui.sat.v2f32.v2i100(<2 x float>)
declare <2 x i128> @llvm.fptoui.sat.v2f32.v2i128(<2 x float>)
define <2 x i1> @test_unsigned_v2f32_v2i1(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i1:
; CHECK: // %bb.0:
; CHECK-NEXT: movi v1.2s, #1
; CHECK-NEXT: fcvtzu v0.2s, v0.2s
; CHECK-NEXT: umin v0.2s, v0.2s, v1.2s
; CHECK-NEXT: ret
%x = call <2 x i1> @llvm.fptoui.sat.v2f32.v2i1(<2 x float> %f)
ret <2 x i1> %x
}
define <2 x i8> @test_unsigned_v2f32_v2i8(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i8:
; CHECK: // %bb.0:
; CHECK-NEXT: movi d1, #0x0000ff000000ff
; CHECK-NEXT: fcvtzu v0.2s, v0.2s
; CHECK-NEXT: umin v0.2s, v0.2s, v1.2s
; CHECK-NEXT: ret
%x = call <2 x i8> @llvm.fptoui.sat.v2f32.v2i8(<2 x float> %f)
ret <2 x i8> %x
}
define <2 x i13> @test_unsigned_v2f32_v2i13(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i13:
; CHECK: // %bb.0:
; CHECK-NEXT: movi v1.2s, #31, msl #8
; CHECK-NEXT: fcvtzu v0.2s, v0.2s
; CHECK-NEXT: umin v0.2s, v0.2s, v1.2s
; CHECK-NEXT: ret
%x = call <2 x i13> @llvm.fptoui.sat.v2f32.v2i13(<2 x float> %f)
ret <2 x i13> %x
}
define <2 x i16> @test_unsigned_v2f32_v2i16(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i16:
; CHECK: // %bb.0:
; CHECK-NEXT: movi d1, #0x00ffff0000ffff
; CHECK-NEXT: fcvtzu v0.2s, v0.2s
; CHECK-NEXT: umin v0.2s, v0.2s, v1.2s
; CHECK-NEXT: ret
%x = call <2 x i16> @llvm.fptoui.sat.v2f32.v2i16(<2 x float> %f)
ret <2 x i16> %x
}
define <2 x i19> @test_unsigned_v2f32_v2i19(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i19:
; CHECK: // %bb.0:
; CHECK-NEXT: movi v1.2s, #7, msl #16
; CHECK-NEXT: fcvtzu v0.2s, v0.2s
; CHECK-NEXT: umin v0.2s, v0.2s, v1.2s
; CHECK-NEXT: ret
%x = call <2 x i19> @llvm.fptoui.sat.v2f32.v2i19(<2 x float> %f)
ret <2 x i19> %x
}
define <2 x i32> @test_unsigned_v2f32_v2i32_duplicate(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i32_duplicate:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.2s, v0.2s
; CHECK-NEXT: ret
%x = call <2 x i32> @llvm.fptoui.sat.v2f32.v2i32(<2 x float> %f)
ret <2 x i32> %x
}
define <2 x i50> @test_unsigned_v2f32_v2i50(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i50:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: mov s1, v0.s[1]
; CHECK-NEXT: fcvtzu x9, s0
; CHECK-NEXT: mov x10, #1125899906842623
; CHECK-NEXT: fcvtzu x8, s1
; CHECK-NEXT: cmp x8, x10
; CHECK-NEXT: csel x8, x8, x10, lo
; CHECK-NEXT: cmp x9, x10
; CHECK-NEXT: csel x9, x9, x10, lo
; CHECK-NEXT: fmov d0, x9
; CHECK-NEXT: mov v0.d[1], x8
; CHECK-NEXT: ret
%x = call <2 x i50> @llvm.fptoui.sat.v2f32.v2i50(<2 x float> %f)
ret <2 x i50> %x
}
define <2 x i64> @test_unsigned_v2f32_v2i64(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i64:
; CHECK: // %bb.0:
; CHECK-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: mov s1, v0.s[1]
; CHECK-NEXT: fcvtzu x8, s0
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: fcvtzu x8, s1
; CHECK-NEXT: mov v0.d[1], x8
; CHECK-NEXT: ret
%x = call <2 x i64> @llvm.fptoui.sat.v2f32.v2i64(<2 x float> %f)
ret <2 x i64> %x
}
define <2 x i100> @test_unsigned_v2f32_v2i100(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i100:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #64
; CHECK-NEXT: stp d9, d8, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x21, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 64
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w21, -24
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: .cfi_offset b8, -40
; CHECK-NEXT: .cfi_offset b9, -48
; CHECK-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: mov s8, v0.s[1]
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: mov w8, #1904214015
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov x21, #68719476735
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: // kill: def $s0 killed $s0 killed $q0
; CHECK-NEXT: fmov s9, w8
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: csel x19, x21, x9, gt
; CHECK-NEXT: csinv x20, x8, xzr, le
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov x2, x20
; CHECK-NEXT: mov x3, x19
; CHECK-NEXT: ldp x20, x19, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s0, #0.0
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s0, s9
; CHECK-NEXT: ldp d9, d8, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csel x1, x21, x9, gt
; CHECK-NEXT: ldp x30, x21, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: mov v0.d[1], x1
; CHECK-NEXT: fmov x0, d0
; CHECK-NEXT: add sp, sp, #64
; CHECK-NEXT: ret
%x = call <2 x i100> @llvm.fptoui.sat.v2f32.v2i100(<2 x float> %f)
ret <2 x i100> %x
}
define <2 x i128> @test_unsigned_v2f32_v2i128(<2 x float> %f) {
; CHECK-LABEL: test_unsigned_v2f32_v2i128:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #64
; CHECK-NEXT: stp d9, d8, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: str x30, [sp, #32] // 8-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 64
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: .cfi_offset b8, -40
; CHECK-NEXT: .cfi_offset b9, -48
; CHECK-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: mov s8, v0.s[1]
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: mov w8, #2139095039
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: // kill: def $s0 killed $s0 killed $q0
; CHECK-NEXT: fmov s9, w8
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: csinv x19, x9, xzr, le
; CHECK-NEXT: csinv x20, x8, xzr, le
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov x2, x19
; CHECK-NEXT: mov x3, x20
; CHECK-NEXT: ldr x30, [sp, #32] // 8-byte Folded Reload
; CHECK-NEXT: ldp x20, x19, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s0, #0.0
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s0, s9
; CHECK-NEXT: ldp d9, d8, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csinv x1, x9, xzr, le
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: mov v0.d[1], x1
; CHECK-NEXT: fmov x0, d0
; CHECK-NEXT: add sp, sp, #64
; CHECK-NEXT: ret
%x = call <2 x i128> @llvm.fptoui.sat.v2f32.v2i128(<2 x float> %f)
ret <2 x i128> %x
}
;
; 4-Vector float to unsigned integer -- result size variation
;
declare <4 x i1> @llvm.fptoui.sat.v4f32.v4i1 (<4 x float>)
declare <4 x i8> @llvm.fptoui.sat.v4f32.v4i8 (<4 x float>)
declare <4 x i13> @llvm.fptoui.sat.v4f32.v4i13 (<4 x float>)
declare <4 x i16> @llvm.fptoui.sat.v4f32.v4i16 (<4 x float>)
declare <4 x i19> @llvm.fptoui.sat.v4f32.v4i19 (<4 x float>)
declare <4 x i50> @llvm.fptoui.sat.v4f32.v4i50 (<4 x float>)
declare <4 x i64> @llvm.fptoui.sat.v4f32.v4i64 (<4 x float>)
declare <4 x i100> @llvm.fptoui.sat.v4f32.v4i100(<4 x float>)
declare <4 x i128> @llvm.fptoui.sat.v4f32.v4i128(<4 x float>)
define <4 x i1> @test_unsigned_v4f32_v4i1(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i1:
; CHECK: // %bb.0:
; CHECK-NEXT: movi v1.4s, #1
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-NEXT: xtn v0.4h, v0.4s
; CHECK-NEXT: ret
%x = call <4 x i1> @llvm.fptoui.sat.v4f32.v4i1(<4 x float> %f)
ret <4 x i1> %x
}
define <4 x i8> @test_unsigned_v4f32_v4i8(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i8:
; CHECK: // %bb.0:
; CHECK-NEXT: movi v1.2d, #0x0000ff000000ff
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-NEXT: xtn v0.4h, v0.4s
; CHECK-NEXT: ret
%x = call <4 x i8> @llvm.fptoui.sat.v4f32.v4i8(<4 x float> %f)
ret <4 x i8> %x
}
define <4 x i13> @test_unsigned_v4f32_v4i13(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i13:
; CHECK: // %bb.0:
; CHECK-NEXT: movi v1.4s, #31, msl #8
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-NEXT: xtn v0.4h, v0.4s
; CHECK-NEXT: ret
%x = call <4 x i13> @llvm.fptoui.sat.v4f32.v4i13(<4 x float> %f)
ret <4 x i13> %x
}
define <4 x i16> @test_unsigned_v4f32_v4i16(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i16:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: uqxtn v0.4h, v0.4s
; CHECK-NEXT: ret
%x = call <4 x i16> @llvm.fptoui.sat.v4f32.v4i16(<4 x float> %f)
ret <4 x i16> %x
}
define <4 x i19> @test_unsigned_v4f32_v4i19(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i19:
; CHECK: // %bb.0:
; CHECK-NEXT: movi v1.4s, #7, msl #16
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-NEXT: ret
%x = call <4 x i19> @llvm.fptoui.sat.v4f32.v4i19(<4 x float> %f)
ret <4 x i19> %x
}
define <4 x i32> @test_unsigned_v4f32_v4i32_duplicate(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i32_duplicate:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: ret
%x = call <4 x i32> @llvm.fptoui.sat.v4f32.v4i32(<4 x float> %f)
ret <4 x i32> %x
}
define <4 x i50> @test_unsigned_v4f32_v4i50(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i50:
; CHECK: // %bb.0:
; CHECK-NEXT: ext v1.16b, v0.16b, v0.16b, #8
; CHECK-NEXT: mov x8, #1125899906842623
; CHECK-NEXT: mov s3, v0.s[1]
; CHECK-NEXT: fcvtzu x11, s0
; CHECK-NEXT: mov s2, v1.s[1]
; CHECK-NEXT: fcvtzu x9, s1
; CHECK-NEXT: fcvtzu x12, s3
; CHECK-NEXT: cmp x9, x8
; CHECK-NEXT: fcvtzu x10, s2
; CHECK-NEXT: csel x2, x9, x8, lo
; CHECK-NEXT: cmp x10, x8
; CHECK-NEXT: csel x3, x10, x8, lo
; CHECK-NEXT: cmp x11, x8
; CHECK-NEXT: csel x0, x11, x8, lo
; CHECK-NEXT: cmp x12, x8
; CHECK-NEXT: csel x1, x12, x8, lo
; CHECK-NEXT: ret
%x = call <4 x i50> @llvm.fptoui.sat.v4f32.v4i50(<4 x float> %f)
ret <4 x i50> %x
}
define <4 x i64> @test_unsigned_v4f32_v4i64(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i64:
; CHECK: // %bb.0:
; CHECK-NEXT: ext v1.16b, v0.16b, v0.16b, #8
; CHECK-NEXT: mov s3, v0.s[1]
; CHECK-NEXT: fcvtzu x9, s0
; CHECK-NEXT: mov s2, v1.s[1]
; CHECK-NEXT: fcvtzu x8, s1
; CHECK-NEXT: fmov d0, x9
; CHECK-NEXT: fcvtzu x9, s3
; CHECK-NEXT: fmov d1, x8
; CHECK-NEXT: fcvtzu x8, s2
; CHECK-NEXT: mov v0.d[1], x9
; CHECK-NEXT: mov v1.d[1], x8
; CHECK-NEXT: ret
%x = call <4 x i64> @llvm.fptoui.sat.v4f32.v4i64(<4 x float> %f)
ret <4 x i64> %x
}
define <4 x i100> @test_unsigned_v4f32_v4i100(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i100:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #112
; CHECK-NEXT: stp d9, d8, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x25, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x24, x23, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x22, x21, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #96] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 112
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w21, -24
; CHECK-NEXT: .cfi_offset w22, -32
; CHECK-NEXT: .cfi_offset w23, -40
; CHECK-NEXT: .cfi_offset w24, -48
; CHECK-NEXT: .cfi_offset w25, -56
; CHECK-NEXT: .cfi_offset w30, -64
; CHECK-NEXT: .cfi_offset b8, -72
; CHECK-NEXT: .cfi_offset b9, -80
; CHECK-NEXT: mov s8, v0.s[1]
; CHECK-NEXT: str q0, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: mov w8, #1904214015
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov x25, #68719476735
; CHECK-NEXT: fmov s9, w8
; CHECK-NEXT: ext v0.16b, v0.16b, v0.16b, #8
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: // kill: def $s0 killed $s0 killed $q0
; CHECK-NEXT: csel x19, x25, x9, gt
; CHECK-NEXT: csinv x20, x8, xzr, le
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov s8, v0.s[1]
; CHECK-NEXT: fcmp s0, #0.0
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s0, s9
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: csel x21, x25, x9, gt
; CHECK-NEXT: csinv x22, x8, xzr, le
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: // kill: def $s0 killed $s0 killed $q0
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: csel x23, x25, x9, gt
; CHECK-NEXT: csinv x24, x8, xzr, le
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: mov x2, x20
; CHECK-NEXT: mov x3, x19
; CHECK-NEXT: mov x4, x22
; CHECK-NEXT: mov x5, x21
; CHECK-NEXT: mov x6, x24
; CHECK-NEXT: fcmp s0, #0.0
; CHECK-NEXT: mov x7, x23
; CHECK-NEXT: ldp x20, x19, [sp, #96] // 16-byte Folded Reload
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s0, s9
; CHECK-NEXT: ldp x22, x21, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csel x1, x25, x9, gt
; CHECK-NEXT: ldp x24, x23, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: ldp x30, x25, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: mov v0.d[1], x1
; CHECK-NEXT: ldp d9, d8, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: fmov x0, d0
; CHECK-NEXT: add sp, sp, #112
; CHECK-NEXT: ret
%x = call <4 x i100> @llvm.fptoui.sat.v4f32.v4i100(<4 x float> %f)
ret <4 x i100> %x
}
define <4 x i128> @test_unsigned_v4f32_v4i128(<4 x float> %f) {
; CHECK-LABEL: test_unsigned_v4f32_v4i128:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #112
; CHECK-NEXT: stp d9, d8, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: str x30, [sp, #48] // 8-byte Folded Spill
; CHECK-NEXT: stp x24, x23, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x22, x21, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #96] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 112
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w21, -24
; CHECK-NEXT: .cfi_offset w22, -32
; CHECK-NEXT: .cfi_offset w23, -40
; CHECK-NEXT: .cfi_offset w24, -48
; CHECK-NEXT: .cfi_offset w30, -64
; CHECK-NEXT: .cfi_offset b8, -72
; CHECK-NEXT: .cfi_offset b9, -80
; CHECK-NEXT: mov s8, v0.s[1]
; CHECK-NEXT: str q0, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: mov w8, #2139095039
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: fmov s9, w8
; CHECK-NEXT: ext v0.16b, v0.16b, v0.16b, #8
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: // kill: def $s0 killed $s0 killed $q0
; CHECK-NEXT: csinv x19, x9, xzr, le
; CHECK-NEXT: csinv x20, x8, xzr, le
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov s8, v0.s[1]
; CHECK-NEXT: fcmp s0, #0.0
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s0, s9
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: csinv x21, x9, xzr, le
; CHECK-NEXT: csinv x22, x8, xzr, le
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: // kill: def $s0 killed $s0 killed $q0
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: csinv x23, x9, xzr, le
; CHECK-NEXT: csinv x24, x8, xzr, le
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: mov x2, x19
; CHECK-NEXT: mov x3, x20
; CHECK-NEXT: mov x4, x21
; CHECK-NEXT: mov x5, x22
; CHECK-NEXT: mov x6, x23
; CHECK-NEXT: fcmp s0, #0.0
; CHECK-NEXT: mov x7, x24
; CHECK-NEXT: ldp x20, x19, [sp, #96] // 16-byte Folded Reload
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s0, s9
; CHECK-NEXT: ldr x30, [sp, #48] // 8-byte Folded Reload
; CHECK-NEXT: ldp x22, x21, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csinv x1, x9, xzr, le
; CHECK-NEXT: ldp x24, x23, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: ldp d9, d8, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: mov v0.d[1], x1
; CHECK-NEXT: fmov x0, d0
; CHECK-NEXT: add sp, sp, #112
; CHECK-NEXT: ret
%x = call <4 x i128> @llvm.fptoui.sat.v4f32.v4i128(<4 x float> %f)
ret <4 x i128> %x
}
;
; 2-Vector double to unsigned integer -- result size variation
;
declare <2 x i1> @llvm.fptoui.sat.v2f64.v2i1 (<2 x double>)
declare <2 x i8> @llvm.fptoui.sat.v2f64.v2i8 (<2 x double>)
declare <2 x i13> @llvm.fptoui.sat.v2f64.v2i13 (<2 x double>)
declare <2 x i16> @llvm.fptoui.sat.v2f64.v2i16 (<2 x double>)
declare <2 x i19> @llvm.fptoui.sat.v2f64.v2i19 (<2 x double>)
declare <2 x i50> @llvm.fptoui.sat.v2f64.v2i50 (<2 x double>)
declare <2 x i64> @llvm.fptoui.sat.v2f64.v2i64 (<2 x double>)
declare <2 x i100> @llvm.fptoui.sat.v2f64.v2i100(<2 x double>)
declare <2 x i128> @llvm.fptoui.sat.v2f64.v2i128(<2 x double>)
define <2 x i1> @test_unsigned_v2f64_v2i1(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i1:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d1, v0.d[1]
; CHECK-NEXT: fcvtzu w9, d0
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: cmp w8, #1
; CHECK-NEXT: csinc w8, w8, wzr, lo
; CHECK-NEXT: cmp w9, #1
; CHECK-NEXT: csinc w9, w9, wzr, lo
; CHECK-NEXT: fmov s0, w9
; CHECK-NEXT: mov v0.s[1], w8
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
%x = call <2 x i1> @llvm.fptoui.sat.v2f64.v2i1(<2 x double> %f)
ret <2 x i1> %x
}
define <2 x i8> @test_unsigned_v2f64_v2i8(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i8:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d1, v0.d[1]
; CHECK-NEXT: fcvtzu w10, d0
; CHECK-NEXT: mov w8, #255
; CHECK-NEXT: fcvtzu w9, d1
; CHECK-NEXT: cmp w9, #255
; CHECK-NEXT: csel w9, w9, w8, lo
; CHECK-NEXT: cmp w10, #255
; CHECK-NEXT: csel w8, w10, w8, lo
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: mov v0.s[1], w9
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
%x = call <2 x i8> @llvm.fptoui.sat.v2f64.v2i8(<2 x double> %f)
ret <2 x i8> %x
}
define <2 x i13> @test_unsigned_v2f64_v2i13(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i13:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d1, v0.d[1]
; CHECK-NEXT: fcvtzu w9, d0
; CHECK-NEXT: mov w10, #8191
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: cmp w8, w10
; CHECK-NEXT: csel w8, w8, w10, lo
; CHECK-NEXT: cmp w9, w10
; CHECK-NEXT: csel w9, w9, w10, lo
; CHECK-NEXT: fmov s0, w9
; CHECK-NEXT: mov v0.s[1], w8
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
%x = call <2 x i13> @llvm.fptoui.sat.v2f64.v2i13(<2 x double> %f)
ret <2 x i13> %x
}
define <2 x i16> @test_unsigned_v2f64_v2i16(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i16:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d1, v0.d[1]
; CHECK-NEXT: fcvtzu w9, d0
; CHECK-NEXT: mov w10, #65535
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: cmp w8, w10
; CHECK-NEXT: csel w8, w8, w10, lo
; CHECK-NEXT: cmp w9, w10
; CHECK-NEXT: csel w9, w9, w10, lo
; CHECK-NEXT: fmov s0, w9
; CHECK-NEXT: mov v0.s[1], w8
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
%x = call <2 x i16> @llvm.fptoui.sat.v2f64.v2i16(<2 x double> %f)
ret <2 x i16> %x
}
define <2 x i19> @test_unsigned_v2f64_v2i19(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i19:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d1, v0.d[1]
; CHECK-NEXT: fcvtzu w9, d0
; CHECK-NEXT: mov w10, #524287
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: cmp w8, w10
; CHECK-NEXT: csel w8, w8, w10, lo
; CHECK-NEXT: cmp w9, w10
; CHECK-NEXT: csel w9, w9, w10, lo
; CHECK-NEXT: fmov s0, w9
; CHECK-NEXT: mov v0.s[1], w8
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
%x = call <2 x i19> @llvm.fptoui.sat.v2f64.v2i19(<2 x double> %f)
ret <2 x i19> %x
}
define <2 x i32> @test_unsigned_v2f64_v2i32_duplicate(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i32_duplicate:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d1, v0.d[1]
; CHECK-NEXT: fcvtzu w8, d0
; CHECK-NEXT: fmov s0, w8
; CHECK-NEXT: fcvtzu w8, d1
; CHECK-NEXT: mov v0.s[1], w8
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
%x = call <2 x i32> @llvm.fptoui.sat.v2f64.v2i32(<2 x double> %f)
ret <2 x i32> %x
}
define <2 x i50> @test_unsigned_v2f64_v2i50(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i50:
; CHECK: // %bb.0:
; CHECK-NEXT: mov d1, v0.d[1]
; CHECK-NEXT: fcvtzu x9, d0
; CHECK-NEXT: mov x10, #1125899906842623
; CHECK-NEXT: fcvtzu x8, d1
; CHECK-NEXT: cmp x8, x10
; CHECK-NEXT: csel x8, x8, x10, lo
; CHECK-NEXT: cmp x9, x10
; CHECK-NEXT: csel x9, x9, x10, lo
; CHECK-NEXT: fmov d0, x9
; CHECK-NEXT: mov v0.d[1], x8
; CHECK-NEXT: ret
%x = call <2 x i50> @llvm.fptoui.sat.v2f64.v2i50(<2 x double> %f)
ret <2 x i50> %x
}
define <2 x i64> @test_unsigned_v2f64_v2i64(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i64:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtzu v0.2d, v0.2d
; CHECK-NEXT: ret
%x = call <2 x i64> @llvm.fptoui.sat.v2f64.v2i64(<2 x double> %f)
ret <2 x i64> %x
}
define <2 x i100> @test_unsigned_v2f64_v2i100(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i100:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #64
; CHECK-NEXT: stp d9, d8, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x21, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 64
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w21, -24
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: .cfi_offset b8, -40
; CHECK-NEXT: .cfi_offset b9, -48
; CHECK-NEXT: mov d8, v0.d[1]
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: fmov d0, d8
; CHECK-NEXT: bl __fixunsdfti
; CHECK-NEXT: mov x8, #5057542381537067007
; CHECK-NEXT: fcmp d8, #0.0
; CHECK-NEXT: mov x21, #68719476735
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: fmov d9, x8
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp d8, d9
; CHECK-NEXT: csel x19, x21, x9, gt
; CHECK-NEXT: csinv x20, x8, xzr, le
; CHECK-NEXT: bl __fixunsdfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov x2, x20
; CHECK-NEXT: mov x3, x19
; CHECK-NEXT: ldp x20, x19, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp d0, #0.0
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp d0, d9
; CHECK-NEXT: ldp d9, d8, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csel x1, x21, x9, gt
; CHECK-NEXT: ldp x30, x21, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: mov v0.d[1], x1
; CHECK-NEXT: fmov x0, d0
; CHECK-NEXT: add sp, sp, #64
; CHECK-NEXT: ret
%x = call <2 x i100> @llvm.fptoui.sat.v2f64.v2i100(<2 x double> %f)
ret <2 x i100> %x
}
define <2 x i128> @test_unsigned_v2f64_v2i128(<2 x double> %f) {
; CHECK-LABEL: test_unsigned_v2f64_v2i128:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #64
; CHECK-NEXT: stp d9, d8, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: str x30, [sp, #32] // 8-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 64
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: .cfi_offset b8, -40
; CHECK-NEXT: .cfi_offset b9, -48
; CHECK-NEXT: mov d8, v0.d[1]
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: fmov d0, d8
; CHECK-NEXT: bl __fixunsdfti
; CHECK-NEXT: mov x8, #5183643171103440895
; CHECK-NEXT: fcmp d8, #0.0
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: // kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: fmov d9, x8
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp d8, d9
; CHECK-NEXT: csinv x19, x9, xzr, le
; CHECK-NEXT: csinv x20, x8, xzr, le
; CHECK-NEXT: bl __fixunsdfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov x2, x19
; CHECK-NEXT: mov x3, x20
; CHECK-NEXT: ldr x30, [sp, #32] // 8-byte Folded Reload
; CHECK-NEXT: ldp x20, x19, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp d0, #0.0
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp d0, d9
; CHECK-NEXT: ldp d9, d8, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csinv x1, x9, xzr, le
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: mov v0.d[1], x1
; CHECK-NEXT: fmov x0, d0
; CHECK-NEXT: add sp, sp, #64
; CHECK-NEXT: ret
%x = call <2 x i128> @llvm.fptoui.sat.v2f64.v2i128(<2 x double> %f)
ret <2 x i128> %x
}
;
; 4-Vector half to unsigned integer -- result size variation
;
declare <4 x i1> @llvm.fptoui.sat.v4f16.v4i1 (<4 x half>)
declare <4 x i8> @llvm.fptoui.sat.v4f16.v4i8 (<4 x half>)
declare <4 x i13> @llvm.fptoui.sat.v4f16.v4i13 (<4 x half>)
declare <4 x i16> @llvm.fptoui.sat.v4f16.v4i16 (<4 x half>)
declare <4 x i19> @llvm.fptoui.sat.v4f16.v4i19 (<4 x half>)
declare <4 x i50> @llvm.fptoui.sat.v4f16.v4i50 (<4 x half>)
declare <4 x i64> @llvm.fptoui.sat.v4f16.v4i64 (<4 x half>)
declare <4 x i100> @llvm.fptoui.sat.v4f16.v4i100(<4 x half>)
declare <4 x i128> @llvm.fptoui.sat.v4f16.v4i128(<4 x half>)
define <4 x i1> @test_unsigned_v4f16_v4i1(<4 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v4f16_v4i1:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvtl v0.4s, v0.4h
; CHECK-CVT-NEXT: movi v1.4s, #1
; CHECK-CVT-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-CVT-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-CVT-NEXT: xtn v0.4h, v0.4s
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v4f16_v4i1:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: movi v1.4h, #1
; CHECK-FP16-NEXT: fcvtzu v0.4h, v0.4h
; CHECK-FP16-NEXT: umin v0.4h, v0.4h, v1.4h
; CHECK-FP16-NEXT: ret
%x = call <4 x i1> @llvm.fptoui.sat.v4f16.v4i1(<4 x half> %f)
ret <4 x i1> %x
}
define <4 x i8> @test_unsigned_v4f16_v4i8(<4 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v4f16_v4i8:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvtl v0.4s, v0.4h
; CHECK-CVT-NEXT: movi v1.2d, #0x0000ff000000ff
; CHECK-CVT-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-CVT-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-CVT-NEXT: xtn v0.4h, v0.4s
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v4f16_v4i8:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: movi d1, #0xff00ff00ff00ff
; CHECK-FP16-NEXT: fcvtzu v0.4h, v0.4h
; CHECK-FP16-NEXT: umin v0.4h, v0.4h, v1.4h
; CHECK-FP16-NEXT: ret
%x = call <4 x i8> @llvm.fptoui.sat.v4f16.v4i8(<4 x half> %f)
ret <4 x i8> %x
}
define <4 x i13> @test_unsigned_v4f16_v4i13(<4 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v4f16_v4i13:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvtl v0.4s, v0.4h
; CHECK-CVT-NEXT: movi v1.4s, #31, msl #8
; CHECK-CVT-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-CVT-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-CVT-NEXT: xtn v0.4h, v0.4s
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v4f16_v4i13:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: mvni v1.4h, #224, lsl #8
; CHECK-FP16-NEXT: fcvtzu v0.4h, v0.4h
; CHECK-FP16-NEXT: umin v0.4h, v0.4h, v1.4h
; CHECK-FP16-NEXT: ret
%x = call <4 x i13> @llvm.fptoui.sat.v4f16.v4i13(<4 x half> %f)
ret <4 x i13> %x
}
define <4 x i16> @test_unsigned_v4f16_v4i16(<4 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v4f16_v4i16:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvtl v0.4s, v0.4h
; CHECK-CVT-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-CVT-NEXT: uqxtn v0.4h, v0.4s
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v4f16_v4i16:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: fcvtzu v0.4h, v0.4h
; CHECK-FP16-NEXT: ret
%x = call <4 x i16> @llvm.fptoui.sat.v4f16.v4i16(<4 x half> %f)
ret <4 x i16> %x
}
define <4 x i19> @test_unsigned_v4f16_v4i19(<4 x half> %f) {
; CHECK-LABEL: test_unsigned_v4f16_v4i19:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v0.4s, v0.4h
; CHECK-NEXT: movi v1.4s, #7, msl #16
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-NEXT: ret
%x = call <4 x i19> @llvm.fptoui.sat.v4f16.v4i19(<4 x half> %f)
ret <4 x i19> %x
}
define <4 x i32> @test_unsigned_v4f16_v4i32_duplicate(<4 x half> %f) {
; CHECK-LABEL: test_unsigned_v4f16_v4i32_duplicate:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v0.4s, v0.4h
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: ret
%x = call <4 x i32> @llvm.fptoui.sat.v4f16.v4i32(<4 x half> %f)
ret <4 x i32> %x
}
define <4 x i50> @test_unsigned_v4f16_v4i50(<4 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v4f16_v4i50:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-CVT-NEXT: mov h1, v0.h[1]
; CHECK-CVT-NEXT: mov h2, v0.h[2]
; CHECK-CVT-NEXT: mov h3, v0.h[3]
; CHECK-CVT-NEXT: fcvt s0, h0
; CHECK-CVT-NEXT: mov x8, #1125899906842623
; CHECK-CVT-NEXT: fcvt s1, h1
; CHECK-CVT-NEXT: fcvt s2, h2
; CHECK-CVT-NEXT: fcvt s3, h3
; CHECK-CVT-NEXT: fcvtzu x9, s0
; CHECK-CVT-NEXT: fcvtzu x10, s1
; CHECK-CVT-NEXT: fcvtzu x11, s2
; CHECK-CVT-NEXT: cmp x9, x8
; CHECK-CVT-NEXT: fcvtzu x12, s3
; CHECK-CVT-NEXT: csel x0, x9, x8, lo
; CHECK-CVT-NEXT: cmp x10, x8
; CHECK-CVT-NEXT: csel x1, x10, x8, lo
; CHECK-CVT-NEXT: cmp x11, x8
; CHECK-CVT-NEXT: csel x2, x11, x8, lo
; CHECK-CVT-NEXT: cmp x12, x8
; CHECK-CVT-NEXT: csel x3, x12, x8, lo
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v4f16_v4i50:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-FP16-NEXT: mov h1, v0.h[1]
; CHECK-FP16-NEXT: mov h2, v0.h[2]
; CHECK-FP16-NEXT: mov h3, v0.h[3]
; CHECK-FP16-NEXT: fcvtzu x9, h0
; CHECK-FP16-NEXT: mov x8, #1125899906842623
; CHECK-FP16-NEXT: fcvtzu x10, h1
; CHECK-FP16-NEXT: fcvtzu x11, h2
; CHECK-FP16-NEXT: cmp x9, x8
; CHECK-FP16-NEXT: fcvtzu x12, h3
; CHECK-FP16-NEXT: csel x0, x9, x8, lo
; CHECK-FP16-NEXT: cmp x10, x8
; CHECK-FP16-NEXT: csel x1, x10, x8, lo
; CHECK-FP16-NEXT: cmp x11, x8
; CHECK-FP16-NEXT: csel x2, x11, x8, lo
; CHECK-FP16-NEXT: cmp x12, x8
; CHECK-FP16-NEXT: csel x3, x12, x8, lo
; CHECK-FP16-NEXT: ret
%x = call <4 x i50> @llvm.fptoui.sat.v4f16.v4i50(<4 x half> %f)
ret <4 x i50> %x
}
define <4 x i64> @test_unsigned_v4f16_v4i64(<4 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v4f16_v4i64:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-CVT-NEXT: mov h1, v0.h[2]
; CHECK-CVT-NEXT: mov h2, v0.h[1]
; CHECK-CVT-NEXT: fcvt s3, h0
; CHECK-CVT-NEXT: mov h0, v0.h[3]
; CHECK-CVT-NEXT: fcvt s1, h1
; CHECK-CVT-NEXT: fcvt s2, h2
; CHECK-CVT-NEXT: fcvtzu x8, s3
; CHECK-CVT-NEXT: fcvt s3, h0
; CHECK-CVT-NEXT: fcvtzu x9, s1
; CHECK-CVT-NEXT: fmov d0, x8
; CHECK-CVT-NEXT: fcvtzu x8, s2
; CHECK-CVT-NEXT: fmov d1, x9
; CHECK-CVT-NEXT: fcvtzu x9, s3
; CHECK-CVT-NEXT: mov v0.d[1], x8
; CHECK-CVT-NEXT: mov v1.d[1], x9
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v4f16_v4i64:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-FP16-NEXT: mov h1, v0.h[2]
; CHECK-FP16-NEXT: mov h2, v0.h[1]
; CHECK-FP16-NEXT: mov h3, v0.h[3]
; CHECK-FP16-NEXT: fcvtzu x8, h0
; CHECK-FP16-NEXT: fcvtzu x9, h1
; CHECK-FP16-NEXT: fmov d0, x8
; CHECK-FP16-NEXT: fcvtzu x8, h2
; CHECK-FP16-NEXT: fmov d1, x9
; CHECK-FP16-NEXT: fcvtzu x9, h3
; CHECK-FP16-NEXT: mov v0.d[1], x8
; CHECK-FP16-NEXT: mov v1.d[1], x9
; CHECK-FP16-NEXT: ret
%x = call <4 x i64> @llvm.fptoui.sat.v4f16.v4i64(<4 x half> %f)
ret <4 x i64> %x
}
define <4 x i100> @test_unsigned_v4f16_v4i100(<4 x half> %f) {
; CHECK-LABEL: test_unsigned_v4f16_v4i100:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #96
; CHECK-NEXT: stp d9, d8, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x25, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp x24, x23, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x22, x21, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 96
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w21, -24
; CHECK-NEXT: .cfi_offset w22, -32
; CHECK-NEXT: .cfi_offset w23, -40
; CHECK-NEXT: .cfi_offset w24, -48
; CHECK-NEXT: .cfi_offset w25, -56
; CHECK-NEXT: .cfi_offset w30, -64
; CHECK-NEXT: .cfi_offset b8, -72
; CHECK-NEXT: .cfi_offset b9, -80
; CHECK-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: mov h1, v0.h[2]
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: fcvt s8, h1
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov w8, #1904214015
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov x25, #68719476735
; CHECK-NEXT: mov h0, v0.h[1]
; CHECK-NEXT: fmov s9, w8
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csel x19, x25, x9, gt
; CHECK-NEXT: csinv x20, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[3]
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csel x21, x25, x9, gt
; CHECK-NEXT: csinv x22, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csel x23, x25, x9, gt
; CHECK-NEXT: csinv x24, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov x2, x22
; CHECK-NEXT: mov x3, x21
; CHECK-NEXT: mov x4, x20
; CHECK-NEXT: mov x5, x19
; CHECK-NEXT: mov x6, x24
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: mov x7, x23
; CHECK-NEXT: ldp x20, x19, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csel x1, x25, x9, gt
; CHECK-NEXT: ldp x22, x21, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: ldp x24, x23, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: mov v0.d[1], x1
; CHECK-NEXT: ldp x30, x25, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d9, d8, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: fmov x0, d0
; CHECK-NEXT: add sp, sp, #96
; CHECK-NEXT: ret
%x = call <4 x i100> @llvm.fptoui.sat.v4f16.v4i100(<4 x half> %f)
ret <4 x i100> %x
}
define <4 x i128> @test_unsigned_v4f16_v4i128(<4 x half> %f) {
; CHECK-LABEL: test_unsigned_v4f16_v4i128:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #96
; CHECK-NEXT: stp d9, d8, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: str x30, [sp, #32] // 8-byte Folded Spill
; CHECK-NEXT: stp x24, x23, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x22, x21, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 96
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w21, -24
; CHECK-NEXT: .cfi_offset w22, -32
; CHECK-NEXT: .cfi_offset w23, -40
; CHECK-NEXT: .cfi_offset w24, -48
; CHECK-NEXT: .cfi_offset w30, -64
; CHECK-NEXT: .cfi_offset b8, -72
; CHECK-NEXT: .cfi_offset b9, -80
; CHECK-NEXT: // kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: mov h1, v0.h[1]
; CHECK-NEXT: str q0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: fcvt s8, h1
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: mov w8, #2139095039
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[2]
; CHECK-NEXT: fmov s9, w8
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x19, x9, xzr, le
; CHECK-NEXT: csinv x20, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[3]
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x21, x9, xzr, le
; CHECK-NEXT: csinv x22, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x23, x9, xzr, le
; CHECK-NEXT: csinv x24, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov x2, x19
; CHECK-NEXT: mov x3, x20
; CHECK-NEXT: mov x4, x21
; CHECK-NEXT: mov x5, x22
; CHECK-NEXT: mov x6, x23
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: mov x7, x24
; CHECK-NEXT: ldp x20, x19, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csinv x1, x9, xzr, le
; CHECK-NEXT: ldp x22, x21, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: fmov d0, x8
; CHECK-NEXT: ldr x30, [sp, #32] // 8-byte Folded Reload
; CHECK-NEXT: ldp x24, x23, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: mov v0.d[1], x1
; CHECK-NEXT: ldp d9, d8, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: fmov x0, d0
; CHECK-NEXT: add sp, sp, #96
; CHECK-NEXT: ret
%x = call <4 x i128> @llvm.fptoui.sat.v4f16.v4i128(<4 x half> %f)
ret <4 x i128> %x
}
;
; 8-Vector half to unsigned integer -- result size variation
;
declare <8 x i1> @llvm.fptoui.sat.v8f16.v8i1 (<8 x half>)
declare <8 x i8> @llvm.fptoui.sat.v8f16.v8i8 (<8 x half>)
declare <8 x i13> @llvm.fptoui.sat.v8f16.v8i13 (<8 x half>)
declare <8 x i16> @llvm.fptoui.sat.v8f16.v8i16 (<8 x half>)
declare <8 x i19> @llvm.fptoui.sat.v8f16.v8i19 (<8 x half>)
declare <8 x i50> @llvm.fptoui.sat.v8f16.v8i50 (<8 x half>)
declare <8 x i64> @llvm.fptoui.sat.v8f16.v8i64 (<8 x half>)
declare <8 x i100> @llvm.fptoui.sat.v8f16.v8i100(<8 x half>)
declare <8 x i128> @llvm.fptoui.sat.v8f16.v8i128(<8 x half>)
define <8 x i1> @test_unsigned_v8f16_v8i1(<8 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v8f16_v8i1:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvtl2 v1.4s, v0.8h
; CHECK-CVT-NEXT: fcvtl v0.4s, v0.4h
; CHECK-CVT-NEXT: mov s2, v1.s[1]
; CHECK-CVT-NEXT: mov s3, v1.s[2]
; CHECK-CVT-NEXT: mov s4, v1.s[3]
; CHECK-CVT-NEXT: mov s5, v0.s[1]
; CHECK-CVT-NEXT: fcvtzu w9, s1
; CHECK-CVT-NEXT: fcvtzu w10, s0
; CHECK-CVT-NEXT: mov s1, v0.s[2]
; CHECK-CVT-NEXT: mov s0, v0.s[3]
; CHECK-CVT-NEXT: fcvtzu w8, s2
; CHECK-CVT-NEXT: fcvtzu w11, s3
; CHECK-CVT-NEXT: fcvtzu w12, s4
; CHECK-CVT-NEXT: fcvtzu w13, s5
; CHECK-CVT-NEXT: cmp w8, #1
; CHECK-CVT-NEXT: csinc w8, w8, wzr, lo
; CHECK-CVT-NEXT: cmp w9, #1
; CHECK-CVT-NEXT: csinc w9, w9, wzr, lo
; CHECK-CVT-NEXT: cmp w11, #1
; CHECK-CVT-NEXT: csinc w11, w11, wzr, lo
; CHECK-CVT-NEXT: cmp w12, #1
; CHECK-CVT-NEXT: csinc w12, w12, wzr, lo
; CHECK-CVT-NEXT: cmp w13, #1
; CHECK-CVT-NEXT: csinc w13, w13, wzr, lo
; CHECK-CVT-NEXT: cmp w10, #1
; CHECK-CVT-NEXT: csinc w10, w10, wzr, lo
; CHECK-CVT-NEXT: fmov s2, w10
; CHECK-CVT-NEXT: fcvtzu w10, s1
; CHECK-CVT-NEXT: fmov s1, w9
; CHECK-CVT-NEXT: mov v2.s[1], w13
; CHECK-CVT-NEXT: cmp w10, #1
; CHECK-CVT-NEXT: csinc w9, w10, wzr, lo
; CHECK-CVT-NEXT: fcvtzu w10, s0
; CHECK-CVT-NEXT: mov v1.s[1], w8
; CHECK-CVT-NEXT: mov v2.s[2], w9
; CHECK-CVT-NEXT: cmp w10, #1
; CHECK-CVT-NEXT: csinc w8, w10, wzr, lo
; CHECK-CVT-NEXT: mov v1.s[2], w11
; CHECK-CVT-NEXT: mov v2.s[3], w8
; CHECK-CVT-NEXT: mov v1.s[3], w12
; CHECK-CVT-NEXT: xtn v0.4h, v2.4s
; CHECK-CVT-NEXT: xtn2 v0.8h, v1.4s
; CHECK-CVT-NEXT: xtn v0.8b, v0.8h
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v8f16_v8i1:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: movi v1.8h, #1
; CHECK-FP16-NEXT: fcvtzu v0.8h, v0.8h
; CHECK-FP16-NEXT: umin v0.8h, v0.8h, v1.8h
; CHECK-FP16-NEXT: xtn v0.8b, v0.8h
; CHECK-FP16-NEXT: ret
%x = call <8 x i1> @llvm.fptoui.sat.v8f16.v8i1(<8 x half> %f)
ret <8 x i1> %x
}
define <8 x i8> @test_unsigned_v8f16_v8i8(<8 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v8f16_v8i8:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvtl2 v1.4s, v0.8h
; CHECK-CVT-NEXT: fcvtl v0.4s, v0.4h
; CHECK-CVT-NEXT: mov w8, #255
; CHECK-CVT-NEXT: mov s2, v1.s[1]
; CHECK-CVT-NEXT: mov s3, v1.s[2]
; CHECK-CVT-NEXT: mov s4, v1.s[3]
; CHECK-CVT-NEXT: mov s5, v0.s[1]
; CHECK-CVT-NEXT: fcvtzu w10, s1
; CHECK-CVT-NEXT: fcvtzu w11, s0
; CHECK-CVT-NEXT: mov s1, v0.s[2]
; CHECK-CVT-NEXT: mov s0, v0.s[3]
; CHECK-CVT-NEXT: fcvtzu w9, s2
; CHECK-CVT-NEXT: fcvtzu w12, s3
; CHECK-CVT-NEXT: fcvtzu w13, s4
; CHECK-CVT-NEXT: fcvtzu w14, s5
; CHECK-CVT-NEXT: cmp w9, #255
; CHECK-CVT-NEXT: csel w9, w9, w8, lo
; CHECK-CVT-NEXT: cmp w10, #255
; CHECK-CVT-NEXT: csel w10, w10, w8, lo
; CHECK-CVT-NEXT: cmp w12, #255
; CHECK-CVT-NEXT: csel w12, w12, w8, lo
; CHECK-CVT-NEXT: cmp w13, #255
; CHECK-CVT-NEXT: csel w13, w13, w8, lo
; CHECK-CVT-NEXT: cmp w14, #255
; CHECK-CVT-NEXT: csel w14, w14, w8, lo
; CHECK-CVT-NEXT: cmp w11, #255
; CHECK-CVT-NEXT: csel w11, w11, w8, lo
; CHECK-CVT-NEXT: fmov s2, w11
; CHECK-CVT-NEXT: fcvtzu w11, s1
; CHECK-CVT-NEXT: fmov s1, w10
; CHECK-CVT-NEXT: mov v2.s[1], w14
; CHECK-CVT-NEXT: cmp w11, #255
; CHECK-CVT-NEXT: csel w10, w11, w8, lo
; CHECK-CVT-NEXT: fcvtzu w11, s0
; CHECK-CVT-NEXT: mov v1.s[1], w9
; CHECK-CVT-NEXT: mov v2.s[2], w10
; CHECK-CVT-NEXT: cmp w11, #255
; CHECK-CVT-NEXT: csel w8, w11, w8, lo
; CHECK-CVT-NEXT: mov v1.s[2], w12
; CHECK-CVT-NEXT: mov v2.s[3], w8
; CHECK-CVT-NEXT: mov v1.s[3], w13
; CHECK-CVT-NEXT: xtn v0.4h, v2.4s
; CHECK-CVT-NEXT: xtn2 v0.8h, v1.4s
; CHECK-CVT-NEXT: xtn v0.8b, v0.8h
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v8f16_v8i8:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: fcvtzu v0.8h, v0.8h
; CHECK-FP16-NEXT: uqxtn v0.8b, v0.8h
; CHECK-FP16-NEXT: ret
%x = call <8 x i8> @llvm.fptoui.sat.v8f16.v8i8(<8 x half> %f)
ret <8 x i8> %x
}
define <8 x i13> @test_unsigned_v8f16_v8i13(<8 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v8f16_v8i13:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvtl2 v1.4s, v0.8h
; CHECK-CVT-NEXT: fcvtl v0.4s, v0.4h
; CHECK-CVT-NEXT: mov w8, #8191
; CHECK-CVT-NEXT: mov s2, v1.s[1]
; CHECK-CVT-NEXT: mov s3, v1.s[2]
; CHECK-CVT-NEXT: mov s4, v1.s[3]
; CHECK-CVT-NEXT: mov s5, v0.s[1]
; CHECK-CVT-NEXT: fcvtzu w10, s1
; CHECK-CVT-NEXT: fcvtzu w11, s0
; CHECK-CVT-NEXT: mov s1, v0.s[2]
; CHECK-CVT-NEXT: mov s0, v0.s[3]
; CHECK-CVT-NEXT: fcvtzu w9, s2
; CHECK-CVT-NEXT: fcvtzu w12, s3
; CHECK-CVT-NEXT: fcvtzu w13, s4
; CHECK-CVT-NEXT: fcvtzu w14, s5
; CHECK-CVT-NEXT: cmp w9, w8
; CHECK-CVT-NEXT: csel w9, w9, w8, lo
; CHECK-CVT-NEXT: cmp w10, w8
; CHECK-CVT-NEXT: csel w10, w10, w8, lo
; CHECK-CVT-NEXT: cmp w12, w8
; CHECK-CVT-NEXT: csel w12, w12, w8, lo
; CHECK-CVT-NEXT: cmp w13, w8
; CHECK-CVT-NEXT: csel w13, w13, w8, lo
; CHECK-CVT-NEXT: cmp w14, w8
; CHECK-CVT-NEXT: csel w14, w14, w8, lo
; CHECK-CVT-NEXT: cmp w11, w8
; CHECK-CVT-NEXT: csel w11, w11, w8, lo
; CHECK-CVT-NEXT: fmov s2, w11
; CHECK-CVT-NEXT: fcvtzu w11, s1
; CHECK-CVT-NEXT: fmov s1, w10
; CHECK-CVT-NEXT: mov v2.s[1], w14
; CHECK-CVT-NEXT: cmp w11, w8
; CHECK-CVT-NEXT: csel w10, w11, w8, lo
; CHECK-CVT-NEXT: fcvtzu w11, s0
; CHECK-CVT-NEXT: mov v1.s[1], w9
; CHECK-CVT-NEXT: mov v2.s[2], w10
; CHECK-CVT-NEXT: cmp w11, w8
; CHECK-CVT-NEXT: csel w8, w11, w8, lo
; CHECK-CVT-NEXT: mov v1.s[2], w12
; CHECK-CVT-NEXT: mov v2.s[3], w8
; CHECK-CVT-NEXT: mov v1.s[3], w13
; CHECK-CVT-NEXT: xtn v0.4h, v2.4s
; CHECK-CVT-NEXT: xtn2 v0.8h, v1.4s
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v8f16_v8i13:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: mvni v1.8h, #224, lsl #8
; CHECK-FP16-NEXT: fcvtzu v0.8h, v0.8h
; CHECK-FP16-NEXT: umin v0.8h, v0.8h, v1.8h
; CHECK-FP16-NEXT: ret
%x = call <8 x i13> @llvm.fptoui.sat.v8f16.v8i13(<8 x half> %f)
ret <8 x i13> %x
}
define <8 x i16> @test_unsigned_v8f16_v8i16(<8 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v8f16_v8i16:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: fcvtl2 v1.4s, v0.8h
; CHECK-CVT-NEXT: fcvtl v0.4s, v0.4h
; CHECK-CVT-NEXT: mov w8, #65535
; CHECK-CVT-NEXT: mov s2, v1.s[1]
; CHECK-CVT-NEXT: mov s3, v1.s[2]
; CHECK-CVT-NEXT: mov s4, v1.s[3]
; CHECK-CVT-NEXT: mov s5, v0.s[1]
; CHECK-CVT-NEXT: fcvtzu w10, s1
; CHECK-CVT-NEXT: fcvtzu w11, s0
; CHECK-CVT-NEXT: mov s1, v0.s[2]
; CHECK-CVT-NEXT: mov s0, v0.s[3]
; CHECK-CVT-NEXT: fcvtzu w9, s2
; CHECK-CVT-NEXT: fcvtzu w12, s3
; CHECK-CVT-NEXT: fcvtzu w13, s4
; CHECK-CVT-NEXT: fcvtzu w14, s5
; CHECK-CVT-NEXT: cmp w9, w8
; CHECK-CVT-NEXT: csel w9, w9, w8, lo
; CHECK-CVT-NEXT: cmp w10, w8
; CHECK-CVT-NEXT: csel w10, w10, w8, lo
; CHECK-CVT-NEXT: cmp w12, w8
; CHECK-CVT-NEXT: csel w12, w12, w8, lo
; CHECK-CVT-NEXT: cmp w13, w8
; CHECK-CVT-NEXT: csel w13, w13, w8, lo
; CHECK-CVT-NEXT: cmp w14, w8
; CHECK-CVT-NEXT: csel w14, w14, w8, lo
; CHECK-CVT-NEXT: cmp w11, w8
; CHECK-CVT-NEXT: csel w11, w11, w8, lo
; CHECK-CVT-NEXT: fmov s2, w11
; CHECK-CVT-NEXT: fcvtzu w11, s1
; CHECK-CVT-NEXT: fmov s1, w10
; CHECK-CVT-NEXT: mov v2.s[1], w14
; CHECK-CVT-NEXT: cmp w11, w8
; CHECK-CVT-NEXT: csel w10, w11, w8, lo
; CHECK-CVT-NEXT: fcvtzu w11, s0
; CHECK-CVT-NEXT: mov v1.s[1], w9
; CHECK-CVT-NEXT: mov v2.s[2], w10
; CHECK-CVT-NEXT: cmp w11, w8
; CHECK-CVT-NEXT: csel w8, w11, w8, lo
; CHECK-CVT-NEXT: mov v1.s[2], w12
; CHECK-CVT-NEXT: mov v2.s[3], w8
; CHECK-CVT-NEXT: mov v1.s[3], w13
; CHECK-CVT-NEXT: xtn v0.4h, v2.4s
; CHECK-CVT-NEXT: xtn2 v0.8h, v1.4s
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v8f16_v8i16:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: fcvtzu v0.8h, v0.8h
; CHECK-FP16-NEXT: ret
%x = call <8 x i16> @llvm.fptoui.sat.v8f16.v8i16(<8 x half> %f)
ret <8 x i16> %x
}
define <8 x i19> @test_unsigned_v8f16_v8i19(<8 x half> %f) {
; CHECK-LABEL: test_unsigned_v8f16_v8i19:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl v2.4s, v0.4h
; CHECK-NEXT: fcvtl2 v0.4s, v0.8h
; CHECK-NEXT: movi v1.4s, #7, msl #16
; CHECK-NEXT: fcvtzu v2.4s, v2.4s
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: umin v2.4s, v2.4s, v1.4s
; CHECK-NEXT: umin v0.4s, v0.4s, v1.4s
; CHECK-NEXT: mov w1, v2.s[1]
; CHECK-NEXT: mov w2, v2.s[2]
; CHECK-NEXT: mov w5, v0.s[1]
; CHECK-NEXT: mov w3, v2.s[3]
; CHECK-NEXT: mov w6, v0.s[2]
; CHECK-NEXT: mov w7, v0.s[3]
; CHECK-NEXT: fmov w4, s0
; CHECK-NEXT: fmov w0, s2
; CHECK-NEXT: ret
%x = call <8 x i19> @llvm.fptoui.sat.v8f16.v8i19(<8 x half> %f)
ret <8 x i19> %x
}
define <8 x i32> @test_unsigned_v8f16_v8i32_duplicate(<8 x half> %f) {
; CHECK-LABEL: test_unsigned_v8f16_v8i32_duplicate:
; CHECK: // %bb.0:
; CHECK-NEXT: fcvtl2 v1.4s, v0.8h
; CHECK-NEXT: fcvtl v0.4s, v0.4h
; CHECK-NEXT: fcvtzu v1.4s, v1.4s
; CHECK-NEXT: fcvtzu v0.4s, v0.4s
; CHECK-NEXT: ret
%x = call <8 x i32> @llvm.fptoui.sat.v8f16.v8i32(<8 x half> %f)
ret <8 x i32> %x
}
define <8 x i50> @test_unsigned_v8f16_v8i50(<8 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v8f16_v8i50:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: ext v1.16b, v0.16b, v0.16b, #8
; CHECK-CVT-NEXT: mov x8, #1125899906842623
; CHECK-CVT-NEXT: mov h2, v0.h[1]
; CHECK-CVT-NEXT: mov h3, v0.h[2]
; CHECK-CVT-NEXT: mov h5, v0.h[3]
; CHECK-CVT-NEXT: fcvt s0, h0
; CHECK-CVT-NEXT: mov h4, v1.h[1]
; CHECK-CVT-NEXT: mov h6, v1.h[2]
; CHECK-CVT-NEXT: mov h7, v1.h[3]
; CHECK-CVT-NEXT: fcvt s1, h1
; CHECK-CVT-NEXT: fcvt s2, h2
; CHECK-CVT-NEXT: fcvt s3, h3
; CHECK-CVT-NEXT: fcvtzu x9, s0
; CHECK-CVT-NEXT: fcvt s5, h5
; CHECK-CVT-NEXT: fcvt s4, h4
; CHECK-CVT-NEXT: fcvt s6, h6
; CHECK-CVT-NEXT: fcvt s0, h7
; CHECK-CVT-NEXT: fcvtzu x10, s1
; CHECK-CVT-NEXT: fcvtzu x11, s2
; CHECK-CVT-NEXT: fcvtzu x12, s3
; CHECK-CVT-NEXT: fcvtzu x14, s5
; CHECK-CVT-NEXT: fcvtzu x13, s4
; CHECK-CVT-NEXT: fcvtzu x15, s6
; CHECK-CVT-NEXT: cmp x10, x8
; CHECK-CVT-NEXT: fcvtzu x16, s0
; CHECK-CVT-NEXT: csel x4, x10, x8, lo
; CHECK-CVT-NEXT: cmp x13, x8
; CHECK-CVT-NEXT: csel x5, x13, x8, lo
; CHECK-CVT-NEXT: cmp x15, x8
; CHECK-CVT-NEXT: csel x6, x15, x8, lo
; CHECK-CVT-NEXT: cmp x16, x8
; CHECK-CVT-NEXT: csel x7, x16, x8, lo
; CHECK-CVT-NEXT: cmp x9, x8
; CHECK-CVT-NEXT: csel x0, x9, x8, lo
; CHECK-CVT-NEXT: cmp x11, x8
; CHECK-CVT-NEXT: csel x1, x11, x8, lo
; CHECK-CVT-NEXT: cmp x12, x8
; CHECK-CVT-NEXT: csel x2, x12, x8, lo
; CHECK-CVT-NEXT: cmp x14, x8
; CHECK-CVT-NEXT: csel x3, x14, x8, lo
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v8f16_v8i50:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: ext v1.16b, v0.16b, v0.16b, #8
; CHECK-FP16-NEXT: mov x8, #1125899906842623
; CHECK-FP16-NEXT: mov h2, v0.h[1]
; CHECK-FP16-NEXT: mov h3, v0.h[2]
; CHECK-FP16-NEXT: mov h5, v0.h[3]
; CHECK-FP16-NEXT: fcvtzu x9, h0
; CHECK-FP16-NEXT: mov h4, v1.h[1]
; CHECK-FP16-NEXT: mov h6, v1.h[2]
; CHECK-FP16-NEXT: mov h0, v1.h[3]
; CHECK-FP16-NEXT: fcvtzu x10, h1
; CHECK-FP16-NEXT: fcvtzu x11, h2
; CHECK-FP16-NEXT: fcvtzu x12, h3
; CHECK-FP16-NEXT: fcvtzu x14, h5
; CHECK-FP16-NEXT: fcvtzu x13, h4
; CHECK-FP16-NEXT: fcvtzu x15, h6
; CHECK-FP16-NEXT: cmp x10, x8
; CHECK-FP16-NEXT: fcvtzu x16, h0
; CHECK-FP16-NEXT: csel x4, x10, x8, lo
; CHECK-FP16-NEXT: cmp x13, x8
; CHECK-FP16-NEXT: csel x5, x13, x8, lo
; CHECK-FP16-NEXT: cmp x15, x8
; CHECK-FP16-NEXT: csel x6, x15, x8, lo
; CHECK-FP16-NEXT: cmp x16, x8
; CHECK-FP16-NEXT: csel x7, x16, x8, lo
; CHECK-FP16-NEXT: cmp x9, x8
; CHECK-FP16-NEXT: csel x0, x9, x8, lo
; CHECK-FP16-NEXT: cmp x11, x8
; CHECK-FP16-NEXT: csel x1, x11, x8, lo
; CHECK-FP16-NEXT: cmp x12, x8
; CHECK-FP16-NEXT: csel x2, x12, x8, lo
; CHECK-FP16-NEXT: cmp x14, x8
; CHECK-FP16-NEXT: csel x3, x14, x8, lo
; CHECK-FP16-NEXT: ret
%x = call <8 x i50> @llvm.fptoui.sat.v8f16.v8i50(<8 x half> %f)
ret <8 x i50> %x
}
define <8 x i64> @test_unsigned_v8f16_v8i64(<8 x half> %f) {
; CHECK-CVT-LABEL: test_unsigned_v8f16_v8i64:
; CHECK-CVT: // %bb.0:
; CHECK-CVT-NEXT: ext v1.16b, v0.16b, v0.16b, #8
; CHECK-CVT-NEXT: mov h4, v0.h[2]
; CHECK-CVT-NEXT: fcvt s5, h0
; CHECK-CVT-NEXT: fcvt s2, h1
; CHECK-CVT-NEXT: mov h3, v1.h[1]
; CHECK-CVT-NEXT: mov h6, v1.h[2]
; CHECK-CVT-NEXT: fcvt s4, h4
; CHECK-CVT-NEXT: mov h1, v1.h[3]
; CHECK-CVT-NEXT: fcvtzu x9, s5
; CHECK-CVT-NEXT: fcvtzu x8, s2
; CHECK-CVT-NEXT: fcvt s2, h3
; CHECK-CVT-NEXT: mov h3, v0.h[1]
; CHECK-CVT-NEXT: mov h0, v0.h[3]
; CHECK-CVT-NEXT: fcvt s5, h6
; CHECK-CVT-NEXT: fcvt s6, h1
; CHECK-CVT-NEXT: fcvtzu x10, s2
; CHECK-CVT-NEXT: fmov d2, x8
; CHECK-CVT-NEXT: fcvtzu x8, s4
; CHECK-CVT-NEXT: fcvt s3, h3
; CHECK-CVT-NEXT: fcvt s4, h0
; CHECK-CVT-NEXT: fmov d0, x9
; CHECK-CVT-NEXT: mov v2.d[1], x10
; CHECK-CVT-NEXT: fcvtzu x10, s5
; CHECK-CVT-NEXT: fmov d1, x8
; CHECK-CVT-NEXT: fcvtzu x9, s3
; CHECK-CVT-NEXT: fcvtzu x8, s4
; CHECK-CVT-NEXT: fmov d3, x10
; CHECK-CVT-NEXT: fcvtzu x10, s6
; CHECK-CVT-NEXT: mov v0.d[1], x9
; CHECK-CVT-NEXT: mov v1.d[1], x8
; CHECK-CVT-NEXT: mov v3.d[1], x10
; CHECK-CVT-NEXT: ret
;
; CHECK-FP16-LABEL: test_unsigned_v8f16_v8i64:
; CHECK-FP16: // %bb.0:
; CHECK-FP16-NEXT: ext v1.16b, v0.16b, v0.16b, #8
; CHECK-FP16-NEXT: mov h3, v0.h[2]
; CHECK-FP16-NEXT: mov h5, v0.h[3]
; CHECK-FP16-NEXT: fcvtzu x9, h0
; CHECK-FP16-NEXT: mov h2, v1.h[1]
; CHECK-FP16-NEXT: fcvtzu x8, h1
; CHECK-FP16-NEXT: mov h4, v1.h[2]
; CHECK-FP16-NEXT: mov h6, v1.h[3]
; CHECK-FP16-NEXT: fcvtzu x10, h2
; CHECK-FP16-NEXT: fmov d2, x8
; CHECK-FP16-NEXT: fcvtzu x8, h3
; CHECK-FP16-NEXT: mov h3, v0.h[1]
; CHECK-FP16-NEXT: fmov d0, x9
; CHECK-FP16-NEXT: mov v2.d[1], x10
; CHECK-FP16-NEXT: fcvtzu x10, h4
; CHECK-FP16-NEXT: fmov d1, x8
; CHECK-FP16-NEXT: fcvtzu x9, h3
; CHECK-FP16-NEXT: fcvtzu x8, h5
; CHECK-FP16-NEXT: fmov d3, x10
; CHECK-FP16-NEXT: fcvtzu x10, h6
; CHECK-FP16-NEXT: mov v0.d[1], x9
; CHECK-FP16-NEXT: mov v1.d[1], x8
; CHECK-FP16-NEXT: mov v3.d[1], x10
; CHECK-FP16-NEXT: ret
%x = call <8 x i64> @llvm.fptoui.sat.v8f16.v8i64(<8 x half> %f)
ret <8 x i64> %x
}
define <8 x i100> @test_unsigned_v8f16_v8i100(<8 x half> %f) {
; CHECK-LABEL: test_unsigned_v8f16_v8i100:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #176
; CHECK-NEXT: stp d9, d8, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x29, x30, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: stp x28, x27, [sp, #96] // 16-byte Folded Spill
; CHECK-NEXT: stp x26, x25, [sp, #112] // 16-byte Folded Spill
; CHECK-NEXT: stp x24, x23, [sp, #128] // 16-byte Folded Spill
; CHECK-NEXT: stp x22, x21, [sp, #144] // 16-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #160] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 176
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w21, -24
; CHECK-NEXT: .cfi_offset w22, -32
; CHECK-NEXT: .cfi_offset w23, -40
; CHECK-NEXT: .cfi_offset w24, -48
; CHECK-NEXT: .cfi_offset w25, -56
; CHECK-NEXT: .cfi_offset w26, -64
; CHECK-NEXT: .cfi_offset w27, -72
; CHECK-NEXT: .cfi_offset w28, -80
; CHECK-NEXT: .cfi_offset w30, -88
; CHECK-NEXT: .cfi_offset w29, -96
; CHECK-NEXT: .cfi_offset b8, -104
; CHECK-NEXT: .cfi_offset b9, -112
; CHECK-NEXT: str q0, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: mov x19, x8
; CHECK-NEXT: ext v0.16b, v0.16b, v0.16b, #8
; CHECK-NEXT: str q0, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: mov h0, v0.h[1]
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: mov w8, #1904214015
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov x21, #68719476735
; CHECK-NEXT: mov h0, v0.h[3]
; CHECK-NEXT: fmov s9, w8
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x9, x9, xzr, le
; CHECK-NEXT: csel x20, x21, x8, gt
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: str x9, [sp, #24] // 8-byte Folded Spill
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x9, x9, xzr, le
; CHECK-NEXT: csel x23, x21, x8, gt
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: str x9, [sp, #16] // 8-byte Folded Spill
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[2]
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csel x24, x21, x9, gt
; CHECK-NEXT: str x8, [sp, #32] // 8-byte Folded Spill
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[1]
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csel x26, x21, x9, gt
; CHECK-NEXT: str x8, [sp, #8] // 8-byte Folded Spill
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[3]
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x29, x9, xzr, le
; CHECK-NEXT: csel x28, x21, x8, gt
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x27, x9, xzr, le
; CHECK-NEXT: csel x22, x21, x8, gt
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[2]
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csel x25, x21, x9, gt
; CHECK-NEXT: str x8, [sp] // 8-byte Folded Spill
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr x11, [sp, #8] // 8-byte Folded Reload
; CHECK-NEXT: fmov d0, x27
; CHECK-NEXT: fmov d1, x29
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: lsr x10, x22, #28
; CHECK-NEXT: stur x11, [x19, #75]
; CHECK-NEXT: lsr x11, x28, #28
; CHECK-NEXT: mov v0.d[1], x22
; CHECK-NEXT: ldr x12, [sp, #32] // 8-byte Folded Reload
; CHECK-NEXT: mov v1.d[1], x28
; CHECK-NEXT: csel x8, xzr, x0, lt
; CHECK-NEXT: csel x9, xzr, x1, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: stur x12, [x19, #50]
; CHECK-NEXT: fmov x12, d0
; CHECK-NEXT: fmov x13, d1
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: ldp d0, d1, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: csel x9, x21, x9, gt
; CHECK-NEXT: strb w10, [x19, #49]
; CHECK-NEXT: extr x10, x22, x12, #28
; CHECK-NEXT: bfi x9, x12, #36, #28
; CHECK-NEXT: stur x8, [x19, #25]
; CHECK-NEXT: extr x8, x28, x13, #28
; CHECK-NEXT: mov v0.d[1], x23
; CHECK-NEXT: strb w11, [x19, #24]
; CHECK-NEXT: mov v1.d[1], x20
; CHECK-NEXT: stur x10, [x19, #41]
; CHECK-NEXT: stur x9, [x19, #33]
; CHECK-NEXT: bfi x25, x13, #36, #28
; CHECK-NEXT: str x8, [x19, #16]
; CHECK-NEXT: lsr x9, x23, #28
; CHECK-NEXT: fmov x8, d0
; CHECK-NEXT: ldr x12, [sp] // 8-byte Folded Reload
; CHECK-NEXT: fmov x11, d1
; CHECK-NEXT: lsr x10, x20, #28
; CHECK-NEXT: strb w9, [x19, #99]
; CHECK-NEXT: stp x12, x25, [x19]
; CHECK-NEXT: extr x12, x23, x8, #28
; CHECK-NEXT: bfi x26, x8, #36, #28
; CHECK-NEXT: extr x8, x20, x11, #28
; CHECK-NEXT: bfi x24, x11, #36, #28
; CHECK-NEXT: strb w10, [x19, #74]
; CHECK-NEXT: stur x12, [x19, #91]
; CHECK-NEXT: stur x26, [x19, #83]
; CHECK-NEXT: stur x8, [x19, #66]
; CHECK-NEXT: stur x24, [x19, #58]
; CHECK-NEXT: ldp x20, x19, [sp, #160] // 16-byte Folded Reload
; CHECK-NEXT: ldp x22, x21, [sp, #144] // 16-byte Folded Reload
; CHECK-NEXT: ldp x24, x23, [sp, #128] // 16-byte Folded Reload
; CHECK-NEXT: ldp x26, x25, [sp, #112] // 16-byte Folded Reload
; CHECK-NEXT: ldp x28, x27, [sp, #96] // 16-byte Folded Reload
; CHECK-NEXT: ldp x29, x30, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: ldp d9, d8, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: add sp, sp, #176
; CHECK-NEXT: ret
%x = call <8 x i100> @llvm.fptoui.sat.v8f16.v8i100(<8 x half> %f)
ret <8 x i100> %x
}
define <8 x i128> @test_unsigned_v8f16_v8i128(<8 x half> %f) {
; CHECK-LABEL: test_unsigned_v8f16_v8i128:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #176
; CHECK-NEXT: stp d9, d8, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x29, x30, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: stp x28, x27, [sp, #96] // 16-byte Folded Spill
; CHECK-NEXT: stp x26, x25, [sp, #112] // 16-byte Folded Spill
; CHECK-NEXT: stp x24, x23, [sp, #128] // 16-byte Folded Spill
; CHECK-NEXT: stp x22, x21, [sp, #144] // 16-byte Folded Spill
; CHECK-NEXT: stp x20, x19, [sp, #160] // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 176
; CHECK-NEXT: .cfi_offset w19, -8
; CHECK-NEXT: .cfi_offset w20, -16
; CHECK-NEXT: .cfi_offset w21, -24
; CHECK-NEXT: .cfi_offset w22, -32
; CHECK-NEXT: .cfi_offset w23, -40
; CHECK-NEXT: .cfi_offset w24, -48
; CHECK-NEXT: .cfi_offset w25, -56
; CHECK-NEXT: .cfi_offset w26, -64
; CHECK-NEXT: .cfi_offset w27, -72
; CHECK-NEXT: .cfi_offset w28, -80
; CHECK-NEXT: .cfi_offset w30, -88
; CHECK-NEXT: .cfi_offset w29, -96
; CHECK-NEXT: .cfi_offset b8, -104
; CHECK-NEXT: .cfi_offset b9, -112
; CHECK-NEXT: str q0, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: mov x19, x8
; CHECK-NEXT: ext v0.16b, v0.16b, v0.16b, #8
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: str q0, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: mov w8, #2139095039
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[1]
; CHECK-NEXT: fmov s9, w8
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x9, x9, xzr, le
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: stp x8, x9, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[2]
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x9, x9, xzr, le
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: stp x8, x9, [sp] // 16-byte Folded Spill
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[3]
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x24, x9, xzr, le
; CHECK-NEXT: csinv x25, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x26, x9, xzr, le
; CHECK-NEXT: csinv x27, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[1]
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x28, x9, xzr, le
; CHECK-NEXT: csinv x29, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[2]
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x20, x9, xzr, le
; CHECK-NEXT: csinv x21, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: ldr q0, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: mov h0, v0.h[3]
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: fcvt s8, h0
; CHECK-NEXT: csinv x22, x9, xzr, le
; CHECK-NEXT: csinv x23, x8, xzr, le
; CHECK-NEXT: fmov s0, s8
; CHECK-NEXT: bl __fixunssfti
; CHECK-NEXT: fcmp s8, #0.0
; CHECK-NEXT: stp x22, x23, [x19, #32]
; CHECK-NEXT: stp x20, x21, [x19, #16]
; CHECK-NEXT: stp x28, x29, [x19]
; CHECK-NEXT: csel x8, xzr, x1, lt
; CHECK-NEXT: csel x9, xzr, x0, lt
; CHECK-NEXT: fcmp s8, s9
; CHECK-NEXT: stp x26, x27, [x19, #112]
; CHECK-NEXT: stp x24, x25, [x19, #96]
; CHECK-NEXT: csinv x8, x8, xzr, le
; CHECK-NEXT: csinv x9, x9, xzr, le
; CHECK-NEXT: stp x9, x8, [x19, #48]
; CHECK-NEXT: ldr x8, [sp] // 8-byte Folded Reload
; CHECK-NEXT: str x8, [x19, #88]
; CHECK-NEXT: ldr x8, [sp, #8] // 8-byte Folded Reload
; CHECK-NEXT: str x8, [x19, #80]
; CHECK-NEXT: ldr x8, [sp, #16] // 8-byte Folded Reload
; CHECK-NEXT: str x8, [x19, #72]
; CHECK-NEXT: ldr x8, [sp, #24] // 8-byte Folded Reload
; CHECK-NEXT: str x8, [x19, #64]
; CHECK-NEXT: ldp x20, x19, [sp, #160] // 16-byte Folded Reload
; CHECK-NEXT: ldp x22, x21, [sp, #144] // 16-byte Folded Reload
; CHECK-NEXT: ldp x24, x23, [sp, #128] // 16-byte Folded Reload
; CHECK-NEXT: ldp x26, x25, [sp, #112] // 16-byte Folded Reload
; CHECK-NEXT: ldp x28, x27, [sp, #96] // 16-byte Folded Reload
; CHECK-NEXT: ldp x29, x30, [sp, #80] // 16-byte Folded Reload
; CHECK-NEXT: ldp d9, d8, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: add sp, sp, #176
; CHECK-NEXT: ret
%x = call <8 x i128> @llvm.fptoui.sat.v8f16.v8i128(<8 x half> %f)
ret <8 x i128> %x
}