xref: /llvm-project/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-vpmerge-bf16.ll (revision ff0f2011e475141454028bce9cf7c6ff37a49620) 
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc -mtriple=riscv32 -mattr=+d,+v,+m,+zfbfmin,+zvfbfmin -target-abi=ilp32d \
3; RUN:   -verify-machineinstrs < %s | FileCheck %s
4; RUN: llc -mtriple=riscv64 -mattr=+d,+v,+m,+zfbfmin,+zvfbfmin -target-abi=lp64d \
5; RUN:   -verify-machineinstrs < %s | FileCheck %s
6; RUN: llc -mtriple=riscv32 -mattr=+d,+v,+zvfh,+m,+zfbfmin,+zvfbfmin -target-abi=ilp32d \
7; RUN:   -verify-machineinstrs < %s | FileCheck %s
8; RUN: llc -mtriple=riscv64 -mattr=+d,+v,+m,+zvfh,+zfbfmin,+zvfbfmin -target-abi=lp64d \
9; RUN:   -verify-machineinstrs < %s | FileCheck %s
10
11declare <2 x bfloat> @llvm.vp.merge.v2bf16(<2 x i1>, <2 x bfloat>, <2 x bfloat>, i32)
12
13define <2 x bfloat> @vpmerge_vv_v2bf16(<2 x bfloat> %va, <2 x bfloat> %vb, <2 x i1> %m, i32 zeroext %evl) {
14; CHECK-LABEL: vpmerge_vv_v2bf16:
15; CHECK:       # %bb.0:
16; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, tu, ma
17; CHECK-NEXT:    vmerge.vvm v9, v9, v8, v0
18; CHECK-NEXT:    vmv1r.v v8, v9
19; CHECK-NEXT:    ret
20  %v = call <2 x bfloat> @llvm.vp.merge.v2bf16(<2 x i1> %m, <2 x bfloat> %va, <2 x bfloat> %vb, i32 %evl)
21  ret <2 x bfloat> %v
22}
23
24define <2 x bfloat> @vpmerge_vf_v2bf16(bfloat %a, <2 x bfloat> %vb, <2 x i1> %m, i32 zeroext %evl) {
25; CHECK-LABEL: vpmerge_vf_v2bf16:
26; CHECK:       # %bb.0:
27; CHECK-NEXT:    fmv.x.h a1, fa0
28; CHECK-NEXT:    vsetivli zero, 2, e16, mf4, ta, ma
29; CHECK-NEXT:    vmv.v.x v9, a1
30; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, tu, ma
31; CHECK-NEXT:    vmerge.vvm v8, v8, v9, v0
32; CHECK-NEXT:    ret
33  %elt.head = insertelement <2 x bfloat> poison, bfloat %a, i32 0
34  %va = shufflevector <2 x bfloat> %elt.head, <2 x bfloat> poison, <2 x i32> zeroinitializer
35  %v = call <2 x bfloat> @llvm.vp.merge.v2bf16(<2 x i1> %m, <2 x bfloat> %va, <2 x bfloat> %vb, i32 %evl)
36  ret <2 x bfloat> %v
37}
38
39declare <4 x bfloat> @llvm.vp.merge.v4bf16(<4 x i1>, <4 x bfloat>, <4 x bfloat>, i32)
40
41define <4 x bfloat> @vpmerge_vv_v4bf16(<4 x bfloat> %va, <4 x bfloat> %vb, <4 x i1> %m, i32 zeroext %evl) {
42; CHECK-LABEL: vpmerge_vv_v4bf16:
43; CHECK:       # %bb.0:
44; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, tu, ma
45; CHECK-NEXT:    vmerge.vvm v9, v9, v8, v0
46; CHECK-NEXT:    vmv1r.v v8, v9
47; CHECK-NEXT:    ret
48  %v = call <4 x bfloat> @llvm.vp.merge.v4bf16(<4 x i1> %m, <4 x bfloat> %va, <4 x bfloat> %vb, i32 %evl)
49  ret <4 x bfloat> %v
50}
51
52define <4 x bfloat> @vpmerge_vf_v4bf16(bfloat %a, <4 x bfloat> %vb, <4 x i1> %m, i32 zeroext %evl) {
53; CHECK-LABEL: vpmerge_vf_v4bf16:
54; CHECK:       # %bb.0:
55; CHECK-NEXT:    fmv.x.h a1, fa0
56; CHECK-NEXT:    vsetivli zero, 4, e16, mf2, ta, ma
57; CHECK-NEXT:    vmv.v.x v9, a1
58; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, tu, ma
59; CHECK-NEXT:    vmerge.vvm v8, v8, v9, v0
60; CHECK-NEXT:    ret
61  %elt.head = insertelement <4 x bfloat> poison, bfloat %a, i32 0
62  %va = shufflevector <4 x bfloat> %elt.head, <4 x bfloat> poison, <4 x i32> zeroinitializer
63  %v = call <4 x bfloat> @llvm.vp.merge.v4bf16(<4 x i1> %m, <4 x bfloat> %va, <4 x bfloat> %vb, i32 %evl)
64  ret <4 x bfloat> %v
65}
66
67declare <8 x bfloat> @llvm.vp.merge.v8bf16(<8 x i1>, <8 x bfloat>, <8 x bfloat>, i32)
68
69define <8 x bfloat> @vpmerge_vv_v8bf16(<8 x bfloat> %va, <8 x bfloat> %vb, <8 x i1> %m, i32 zeroext %evl) {
70; CHECK-LABEL: vpmerge_vv_v8bf16:
71; CHECK:       # %bb.0:
72; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
73; CHECK-NEXT:    vmerge.vvm v9, v9, v8, v0
74; CHECK-NEXT:    vmv1r.v v8, v9
75; CHECK-NEXT:    ret
76  %v = call <8 x bfloat> @llvm.vp.merge.v8bf16(<8 x i1> %m, <8 x bfloat> %va, <8 x bfloat> %vb, i32 %evl)
77  ret <8 x bfloat> %v
78}
79
80define <8 x bfloat> @vpmerge_vf_v8bf16(bfloat %a, <8 x bfloat> %vb, <8 x i1> %m, i32 zeroext %evl) {
81; CHECK-LABEL: vpmerge_vf_v8bf16:
82; CHECK:       # %bb.0:
83; CHECK-NEXT:    fmv.x.h a1, fa0
84; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
85; CHECK-NEXT:    vmv.v.x v9, a1
86; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
87; CHECK-NEXT:    vmerge.vvm v8, v8, v9, v0
88; CHECK-NEXT:    ret
89  %elt.head = insertelement <8 x bfloat> poison, bfloat %a, i32 0
90  %va = shufflevector <8 x bfloat> %elt.head, <8 x bfloat> poison, <8 x i32> zeroinitializer
91  %v = call <8 x bfloat> @llvm.vp.merge.v8bf16(<8 x i1> %m, <8 x bfloat> %va, <8 x bfloat> %vb, i32 %evl)
92  ret <8 x bfloat> %v
93}
94
95declare <16 x bfloat> @llvm.vp.merge.v16bf16(<16 x i1>, <16 x bfloat>, <16 x bfloat>, i32)
96
97define <16 x bfloat> @vpmerge_vv_v16bf16(<16 x bfloat> %va, <16 x bfloat> %vb, <16 x i1> %m, i32 zeroext %evl) {
98; CHECK-LABEL: vpmerge_vv_v16bf16:
99; CHECK:       # %bb.0:
100; CHECK-NEXT:    vsetvli zero, a0, e16, m2, tu, ma
101; CHECK-NEXT:    vmerge.vvm v10, v10, v8, v0
102; CHECK-NEXT:    vmv2r.v v8, v10
103; CHECK-NEXT:    ret
104  %v = call <16 x bfloat> @llvm.vp.merge.v16bf16(<16 x i1> %m, <16 x bfloat> %va, <16 x bfloat> %vb, i32 %evl)
105  ret <16 x bfloat> %v
106}
107
108define <16 x bfloat> @vpmerge_vf_v16bf16(bfloat %a, <16 x bfloat> %vb, <16 x i1> %m, i32 zeroext %evl) {
109; CHECK-LABEL: vpmerge_vf_v16bf16:
110; CHECK:       # %bb.0:
111; CHECK-NEXT:    fmv.x.h a1, fa0
112; CHECK-NEXT:    vsetivli zero, 16, e16, m2, ta, ma
113; CHECK-NEXT:    vmv.v.x v10, a1
114; CHECK-NEXT:    vsetvli zero, a0, e16, m2, tu, ma
115; CHECK-NEXT:    vmerge.vvm v8, v8, v10, v0
116; CHECK-NEXT:    ret
117  %elt.head = insertelement <16 x bfloat> poison, bfloat %a, i32 0
118  %va = shufflevector <16 x bfloat> %elt.head, <16 x bfloat> poison, <16 x i32> zeroinitializer
119  %v = call <16 x bfloat> @llvm.vp.merge.v16bf16(<16 x i1> %m, <16 x bfloat> %va, <16 x bfloat> %vb, i32 %evl)
120  ret <16 x bfloat> %v
121}
122