1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=riscv32 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=RV32
; RUN: llc -mtriple=riscv64 -verify-machineinstrs < %s | FileCheck %s --check-prefixes=RV64
; Should expand to scalar instructions and do not crash
declare i32 @llvm.vp.reduce.add.v4i32(i32, <4 x i32>, <4 x i1>, i32)
define i32 @vpreduce_add_v4i32(i32 %s, <4 x i32> %v, <4 x i1> %m, i32 %evl) {
; RV32-LABEL: vpreduce_add_v4i32:
; RV32: # %bb.0:
; RV32-NEXT: lw a4, 0(a1)
; RV32-NEXT: lw a5, 4(a1)
; RV32-NEXT: lw a6, 8(a1)
; RV32-NEXT: lw a1, 12(a1)
; RV32-NEXT: lw a7, 0(a2)
; RV32-NEXT: lw t0, 4(a2)
; RV32-NEXT: lw t1, 8(a2)
; RV32-NEXT: lw a2, 12(a2)
; RV32-NEXT: snez t2, a3
; RV32-NEXT: sltiu t3, a3, 3
; RV32-NEXT: sltiu t4, a3, 4
; RV32-NEXT: sltiu a3, a3, 2
; RV32-NEXT: xori t3, t3, 1
; RV32-NEXT: xori t4, t4, 1
; RV32-NEXT: xori a3, a3, 1
; RV32-NEXT: and a3, a3, t0
; RV32-NEXT: and a2, t4, a2
; RV32-NEXT: and t0, t3, t1
; RV32-NEXT: and a7, t2, a7
; RV32-NEXT: neg a7, a7
; RV32-NEXT: neg t0, t0
; RV32-NEXT: neg a2, a2
; RV32-NEXT: neg a3, a3
; RV32-NEXT: and a4, a7, a4
; RV32-NEXT: and a6, t0, a6
; RV32-NEXT: and a1, a2, a1
; RV32-NEXT: and a3, a3, a5
; RV32-NEXT: add a1, a3, a1
; RV32-NEXT: add a4, a4, a6
; RV32-NEXT: add a1, a4, a1
; RV32-NEXT: add a0, a1, a0
; RV32-NEXT: ret
;
; RV64-LABEL: vpreduce_add_v4i32:
; RV64: # %bb.0:
; RV64-NEXT: lw a4, 0(a1)
; RV64-NEXT: lw a5, 8(a1)
; RV64-NEXT: lw a6, 16(a1)
; RV64-NEXT: lw a1, 24(a1)
; RV64-NEXT: ld a7, 0(a2)
; RV64-NEXT: ld t0, 8(a2)
; RV64-NEXT: ld t1, 16(a2)
; RV64-NEXT: ld a2, 24(a2)
; RV64-NEXT: sext.w a3, a3
; RV64-NEXT: snez t2, a3
; RV64-NEXT: sltiu t3, a3, 3
; RV64-NEXT: sltiu t4, a3, 4
; RV64-NEXT: sltiu a3, a3, 2
; RV64-NEXT: xori t3, t3, 1
; RV64-NEXT: xori t4, t4, 1
; RV64-NEXT: xori a3, a3, 1
; RV64-NEXT: and a3, a3, t0
; RV64-NEXT: and a2, t4, a2
; RV64-NEXT: and t0, t3, t1
; RV64-NEXT: and a7, t2, a7
; RV64-NEXT: negw a7, a7
; RV64-NEXT: negw t0, t0
; RV64-NEXT: negw a2, a2
; RV64-NEXT: negw a3, a3
; RV64-NEXT: and a4, a7, a4
; RV64-NEXT: and a6, t0, a6
; RV64-NEXT: and a1, a2, a1
; RV64-NEXT: and a3, a3, a5
; RV64-NEXT: add a1, a3, a1
; RV64-NEXT: add a4, a4, a6
; RV64-NEXT: add a1, a4, a1
; RV64-NEXT: addw a0, a1, a0
; RV64-NEXT: ret
%r = call i32 @llvm.vp.reduce.add.v4i32(i32 %s, <4 x i32> %v, <4 x i1> %m, i32 %evl)
ret i32 %r
}
|
