xref: /llvm-project/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fptoui-vp-mask.ll (revision 1cb599835ccf7ee8b2d1d5a7f3107e19a26fc6f5)
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc -mtriple=riscv32 -mattr=+m,+v,+zvfh < %s | FileCheck %s --check-prefixes=CHECK,ZVFH
3; RUN: llc -mtriple=riscv64 -mattr=+m,+v,+zvfh < %s | FileCheck %s --check-prefixes=CHECK,ZVFH
4; RUN: llc -mtriple=riscv32 -mattr=+m,+v,+zvfhmin < %s | FileCheck %s --check-prefixes=CHECK,ZVFHMIN
5; RUN: llc -mtriple=riscv64 -mattr=+m,+v,+zvfhmin < %s | FileCheck %s --check-prefixes=CHECK,ZVFHMIN
6
7declare <4 x i1> @llvm.vp.fptoui.v4i1.v4f16(<4 x half>, <4 x i1>, i32)
8
9define <4 x i1> @vfptoui_v4i1_v4f16(<4 x half> %va, <4 x i1> %m, i32 zeroext %evl) {
10; ZVFH-LABEL: vfptoui_v4i1_v4f16:
11; ZVFH:       # %bb.0:
12; ZVFH-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
13; ZVFH-NEXT:    vfcvt.rtz.xu.f.v v8, v8, v0.t
14; ZVFH-NEXT:    vmsne.vi v0, v8, 0, v0.t
15; ZVFH-NEXT:    ret
16;
17; ZVFHMIN-LABEL: vfptoui_v4i1_v4f16:
18; ZVFHMIN:       # %bb.0:
19; ZVFHMIN-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
20; ZVFHMIN-NEXT:    vfwcvt.f.f.v v9, v8
21; ZVFHMIN-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
22; ZVFHMIN-NEXT:    vfcvt.rtz.xu.f.v v8, v9, v0.t
23; ZVFHMIN-NEXT:    vmsne.vi v0, v8, 0, v0.t
24; ZVFHMIN-NEXT:    ret
25  %v = call <4 x i1> @llvm.vp.fptoui.v4i1.v4f16(<4 x half> %va, <4 x i1> %m, i32 %evl)
26  ret <4 x i1> %v
27}
28
29define <4 x i1> @vfptoui_v4i1_v4f16_unmasked(<4 x half> %va, i32 zeroext %evl) {
30; ZVFH-LABEL: vfptoui_v4i1_v4f16_unmasked:
31; ZVFH:       # %bb.0:
32; ZVFH-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
33; ZVFH-NEXT:    vfcvt.rtz.xu.f.v v8, v8
34; ZVFH-NEXT:    vmsne.vi v0, v8, 0
35; ZVFH-NEXT:    ret
36;
37; ZVFHMIN-LABEL: vfptoui_v4i1_v4f16_unmasked:
38; ZVFHMIN:       # %bb.0:
39; ZVFHMIN-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
40; ZVFHMIN-NEXT:    vfwcvt.f.f.v v9, v8
41; ZVFHMIN-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
42; ZVFHMIN-NEXT:    vfcvt.rtz.xu.f.v v8, v9
43; ZVFHMIN-NEXT:    vmsne.vi v0, v8, 0
44; ZVFHMIN-NEXT:    ret
45  %v = call <4 x i1> @llvm.vp.fptoui.v4i1.v4f16(<4 x half> %va, <4 x i1> splat (i1 true), i32 %evl)
46  ret <4 x i1> %v
47}
48
49declare <4 x i1> @llvm.vp.fptoui.v4i1.v4f32(<4 x float>, <4 x i1>, i32)
50
51define <4 x i1> @vfptoui_v4i1_v4f32(<4 x float> %va, <4 x i1> %m, i32 zeroext %evl) {
52; CHECK-LABEL: vfptoui_v4i1_v4f32:
53; CHECK:       # %bb.0:
54; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
55; CHECK-NEXT:    vfcvt.rtz.xu.f.v v8, v8, v0.t
56; CHECK-NEXT:    vmsne.vi v0, v8, 0, v0.t
57; CHECK-NEXT:    ret
58  %v = call <4 x i1> @llvm.vp.fptoui.v4i1.v4f32(<4 x float> %va, <4 x i1> %m, i32 %evl)
59  ret <4 x i1> %v
60}
61
62define <4 x i1> @vfptoui_v4i1_v4f32_unmasked(<4 x float> %va, i32 zeroext %evl) {
63; CHECK-LABEL: vfptoui_v4i1_v4f32_unmasked:
64; CHECK:       # %bb.0:
65; CHECK-NEXT:    vsetvli zero, a0, e32, m1, ta, ma
66; CHECK-NEXT:    vfcvt.rtz.xu.f.v v8, v8
67; CHECK-NEXT:    vmsne.vi v0, v8, 0
68; CHECK-NEXT:    ret
69  %v = call <4 x i1> @llvm.vp.fptoui.v4i1.v4f32(<4 x float> %va, <4 x i1> splat (i1 true), i32 %evl)
70  ret <4 x i1> %v
71}
72
73declare <4 x i1> @llvm.vp.fptoui.v4i1.v4f64(<4 x double>, <4 x i1>, i32)
74
75define <4 x i1> @vfptoui_v4i1_v4f64(<4 x double> %va, <4 x i1> %m, i32 zeroext %evl) {
76; CHECK-LABEL: vfptoui_v4i1_v4f64:
77; CHECK:       # %bb.0:
78; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, ma
79; CHECK-NEXT:    vfcvt.rtz.xu.f.v v10, v8, v0.t
80; CHECK-NEXT:    vmsne.vi v8, v10, 0, v0.t
81; CHECK-NEXT:    vmv1r.v v0, v8
82; CHECK-NEXT:    ret
83  %v = call <4 x i1> @llvm.vp.fptoui.v4i1.v4f64(<4 x double> %va, <4 x i1> %m, i32 %evl)
84  ret <4 x i1> %v
85}
86
87define <4 x i1> @vfptoui_v4i1_v4f64_unmasked(<4 x double> %va, i32 zeroext %evl) {
88; CHECK-LABEL: vfptoui_v4i1_v4f64_unmasked:
89; CHECK:       # %bb.0:
90; CHECK-NEXT:    vsetvli zero, a0, e64, m2, ta, ma
91; CHECK-NEXT:    vfcvt.rtz.xu.f.v v8, v8
92; CHECK-NEXT:    vmsne.vi v0, v8, 0
93; CHECK-NEXT:    ret
94  %v = call <4 x i1> @llvm.vp.fptoui.v4i1.v4f64(<4 x double> %va, <4 x i1> splat (i1 true), i32 %evl)
95  ret <4 x i1> %v
96}
97