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
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
;RUN: llc < %s -mtriple=amdgcn-amd-mesa3d -mcpu=fiji -verify-machineinstrs | FileCheck -check-prefix=VI %s
;RUN: llc < %s -mtriple=amdgcn-amd-mesa3d -mcpu=gfx900 -verify-machineinstrs | FileCheck -check-prefix=GFX9 %s
; ===================================================================================
; V_AND_OR_B32
; ===================================================================================
define amdgpu_ps float @and_or(i32 %a, i32 %b, i32 %c) {
; VI-LABEL: and_or:
; VI: ; %bb.0:
; VI-NEXT: v_and_b32_e32 v0, v0, v1
; VI-NEXT: v_or_b32_e32 v0, v0, v2
; VI-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: and_or:
; GFX9: ; %bb.0:
; GFX9-NEXT: v_and_or_b32 v0, v0, v1, v2
; GFX9-NEXT: ; return to shader part epilog
%x = and i32 %a, %b
%result = or i32 %x, %c
%bc = bitcast i32 %result to float
ret float %bc
}
; ThreeOp instruction variant not used due to Constant Bus Limitations
define amdgpu_ps float @and_or_vgpr_b(i32 inreg %a, i32 %b, i32 inreg %c) {
; VI-LABEL: and_or_vgpr_b:
; VI: ; %bb.0:
; VI-NEXT: v_and_b32_e32 v0, s2, v0
; VI-NEXT: v_or_b32_e32 v0, s3, v0
; VI-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: and_or_vgpr_b:
; GFX9: ; %bb.0:
; GFX9-NEXT: v_and_b32_e32 v0, s2, v0
; GFX9-NEXT: v_or_b32_e32 v0, s3, v0
; GFX9-NEXT: ; return to shader part epilog
%x = and i32 %a, %b
%result = or i32 %x, %c
%bc = bitcast i32 %result to float
ret float %bc
}
define amdgpu_ps float @and_or_vgpr_ab(i32 %a, i32 %b, i32 inreg %c) {
; VI-LABEL: and_or_vgpr_ab:
; VI: ; %bb.0:
; VI-NEXT: v_and_b32_e32 v0, v0, v1
; VI-NEXT: v_or_b32_e32 v0, s2, v0
; VI-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: and_or_vgpr_ab:
; GFX9: ; %bb.0:
; GFX9-NEXT: v_and_or_b32 v0, v0, v1, s2
; GFX9-NEXT: ; return to shader part epilog
%x = and i32 %a, %b
%result = or i32 %x, %c
%bc = bitcast i32 %result to float
ret float %bc
}
define amdgpu_ps float @and_or_vgpr_const(i32 %a, i32 %b) {
; VI-LABEL: and_or_vgpr_const:
; VI: ; %bb.0:
; VI-NEXT: v_and_b32_e32 v0, 4, v0
; VI-NEXT: v_or_b32_e32 v0, v0, v1
; VI-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: and_or_vgpr_const:
; GFX9: ; %bb.0:
; GFX9-NEXT: v_and_or_b32 v0, v0, 4, v1
; GFX9-NEXT: ; return to shader part epilog
%x = and i32 4, %a
%result = or i32 %x, %b
%bc = bitcast i32 %result to float
ret float %bc
}
define amdgpu_ps float @and_or_vgpr_const_inline_const(i32 %a) {
; VI-LABEL: and_or_vgpr_const_inline_const:
; VI: ; %bb.0:
; VI-NEXT: v_and_b32_e32 v0, 20, v0
; VI-NEXT: v_or_b32_e32 v0, 0x808, v0
; VI-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: and_or_vgpr_const_inline_const:
; GFX9: ; %bb.0:
; GFX9-NEXT: v_mov_b32_e32 v1, 0x808
; GFX9-NEXT: v_and_or_b32 v0, v0, 20, v1
; GFX9-NEXT: ; return to shader part epilog
%x = and i32 20, %a
%result = or i32 %x, 2056
%bc = bitcast i32 %result to float
ret float %bc
}
define amdgpu_ps float @and_or_vgpr_inline_const_x2(i32 %a) {
; VI-LABEL: and_or_vgpr_inline_const_x2:
; VI: ; %bb.0:
; VI-NEXT: v_and_b32_e32 v0, 4, v0
; VI-NEXT: v_or_b32_e32 v0, 1, v0
; VI-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: and_or_vgpr_inline_const_x2:
; GFX9: ; %bb.0:
; GFX9-NEXT: v_and_or_b32 v0, v0, 4, 1
; GFX9-NEXT: ; return to shader part epilog
%x = and i32 4, %a
%result = or i32 %x, 1
%bc = bitcast i32 %result to float
ret float %bc
}
|
