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
|
; RUN: llc -amdgpu-scalarize-global-loads=false -stress-early-ifcvt -amdgpu-early-ifcvt=1 -march=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,SI %s
; RUN: llc -amdgpu-scalarize-global-loads=false -stress-early-ifcvt -amdgpu-early-ifcvt=1 -march=amdgcn -mcpu=gfx700 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN,GCNX3 %s
; FIXME: Most of these cases that don't trigger because of broken cost
; heuristics. Should not need -stress-early-ifcvt
; GCN-LABEL: {{^}}test_vccnz_ifcvt_triangle64:
; GCN: buffer_load_dwordx2 v[[[VAL_LO:[0-9]+]]:[[VAL_HI:[0-9]+]]]
; GCN: v_cmp_neq_f64_e32 vcc, 1.0, v[[[VAL_LO]]:[[VAL_HI]]]
; GCN: v_add_f64 v[[[ADD_LO:[0-9]+]]:[[ADD_HI:[0-9]+]]], v[[[VAL_LO]]:[[VAL_HI]]], v[[[VAL_LO]]:[[VAL_HI]]]
; GCN-DAG: v_cndmask_b32_e32 v[[RESULT_LO:[0-9]+]], v[[ADD_LO]], v[[VAL_LO]], vcc
; GCN-DAG: v_cndmask_b32_e32 v[[RESULT_HI:[0-9]+]], v[[ADD_HI]], v[[VAL_HI]], vcc
; GCN: buffer_store_dwordx2 v[[[RESULT_LO]]:[[RESULT_HI]]]
define amdgpu_kernel void @test_vccnz_ifcvt_triangle64(ptr addrspace(1) %out, ptr addrspace(1) %in) #0 {
entry:
%v = load double, ptr addrspace(1) %in
%cc = fcmp oeq double %v, 1.000000e+00
br i1 %cc, label %if, label %endif
if:
%u = fadd double %v, %v
br label %endif
endif:
%r = phi double [ %v, %entry ], [ %u, %if ]
store double %r, ptr addrspace(1) %out
ret void
}
; vcc branch with SGPR inputs
; GCN-LABEL: {{^}}test_vccnz_sgpr_ifcvt_triangle64:
; GCN: v_cmp_neq_f64
; GCN: v_add_f64
; GCN: v_cndmask_b32_e32
; GCN: v_cndmask_b32_e32
define amdgpu_kernel void @test_vccnz_sgpr_ifcvt_triangle64(ptr addrspace(1) %out, ptr addrspace(4) %in) #0 {
entry:
%v = load double, ptr addrspace(4) %in
%cc = fcmp oeq double %v, 1.000000e+00
br i1 %cc, label %if, label %endif
if:
%u = fadd double %v, %v
br label %endif
endif:
%r = phi double [ %v, %entry ], [ %u, %if ]
store double %r, ptr addrspace(1) %out
ret void
}
; GCN-LABEL: {{^}}test_vccnz_ifcvt_triangle96:
; GCN: v_cmp_neq_f32_e64 [[CMP:s\[[0-9]+:[0-9]+\]]], s{{[0-9]+}}, 1.0
; GCN: v_add_i32_e32
; GCN: v_add_i32_e32
; GCN: v_add_i32_e32
; GCN: s_mov_b64 vcc, [[CMP]]
; GCN: v_cndmask_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}, vcc
; GCN: v_cndmask_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}, vcc
; GCN: v_cndmask_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}, vcc
; SI-DAG: buffer_store_dwordx2
; SI-DAG: buffer_store_dword v
; GCNX3: buffer_store_dwordx3
define amdgpu_kernel void @test_vccnz_ifcvt_triangle96(ptr addrspace(1) %out, ptr addrspace(1) %in, float %cnd) #0 {
entry:
%v = load <3 x i32>, ptr addrspace(1) %in
%cc = fcmp oeq float %cnd, 1.000000e+00
br i1 %cc, label %if, label %endif
if:
%u = add <3 x i32> %v, %v
br label %endif
endif:
%r = phi <3 x i32> [ %v, %entry ], [ %u, %if ]
store <3 x i32> %r, ptr addrspace(1) %out
ret void
}
; GCN-LABEL: {{^}}test_vccnz_ifcvt_triangle128:
; GCN: v_cmp_neq_f32_e64 [[CMP:s\[[0-9]+:[0-9]+\]]], s{{[0-9]+}}, 1.0
; GCN: v_add_i32_e32
; GCN: v_add_i32_e32
; GCN: v_add_i32_e32
; GCN: v_add_i32_e32
; GCN: s_mov_b64 vcc, [[CMP]]
; GCN: v_cndmask_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}, vcc
; GCN: v_cndmask_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}, vcc
; GCN: v_cndmask_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}, vcc
; GCN: v_cndmask_b32_e32 v{{[0-9]+}}, v{{[0-9]+}}, v{{[0-9]+}}, vcc
; GCN: buffer_store_dwordx4
define amdgpu_kernel void @test_vccnz_ifcvt_triangle128(ptr addrspace(1) %out, ptr addrspace(1) %in, float %cnd) #0 {
entry:
%v = load <4 x i32>, ptr addrspace(1) %in
%cc = fcmp oeq float %cnd, 1.000000e+00
br i1 %cc, label %if, label %endif
if:
%u = add <4 x i32> %v, %v
br label %endif
endif:
%r = phi <4 x i32> [ %v, %entry ], [ %u, %if ]
store <4 x i32> %r, ptr addrspace(1) %out
ret void
}
|
