reference, declarationdefinition
definition → references, declarations, derived classes, virtual overrides
reference to multiple definitions → definitions
unreferenced
    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
  113
  114
  115
  116
  117
  118
  119
  120
  121
  122
  123
  124
  125
  126
  127
  128
  129
  130
  131
  132
  133
  134
  135
  136
  137
  138
  139
  140
  141
  142
  143
  144
  145
  146
  147
  148
  149
  150
  151
  152
  153
  154
  155
  156
  157
  158
  159
  160
  161
  162
  163
  164
  165
  166
  167
  168
  169
  170
  171
  172
  173
  174
  175
  176
  177
  178
  179
  180
  181
  182
  183
  184
  185
  186
  187
  188
  189
  190
  191
  192
  193
  194
  195
  196
  197
  198
  199
  200
  201
  202
  203
; RUN: llc -march=amdgcn -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=GCN -check-prefix=FUNC %s
; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=VI -check-prefix=GCN -check-prefix=FUNC %s

declare i32 @llvm.r600.read.tidig.x() nounwind readnone

declare i64 @llvm.ctpop.i64(i64) nounwind readnone
declare <2 x i64> @llvm.ctpop.v2i64(<2 x i64>) nounwind readnone
declare <4 x i64> @llvm.ctpop.v4i64(<4 x i64>) nounwind readnone
declare <8 x i64> @llvm.ctpop.v8i64(<8 x i64>) nounwind readnone
declare <16 x i64> @llvm.ctpop.v16i64(<16 x i64>) nounwind readnone

declare i65 @llvm.ctpop.i65(i65) nounwind readnone
declare i128 @llvm.ctpop.i128(i128) nounwind readnone

; FUNC-LABEL: {{^}}s_ctpop_i64:
; SI: s_load_dwordx2 [[SVAL:s\[[0-9]+:[0-9]+\]]], s{{\[[0-9]+:[0-9]+\]}}, 0x13
; VI: s_load_dwordx2 [[SVAL:s\[[0-9]+:[0-9]+\]]], s{{\[[0-9]+:[0-9]+\]}}, 0x4c
; GCN: s_bcnt1_i32_b64 [[SRESULT:s[0-9]+]], [[SVAL]]
; GCN: v_mov_b32_e32 [[VRESULT:v[0-9]+]], [[SRESULT]]
; GCN: buffer_store_dword [[VRESULT]],
; GCN: s_endpgm
define amdgpu_kernel void @s_ctpop_i64(i32 addrspace(1)* noalias %out, [8 x i32], i64 %val) nounwind {
  %ctpop = call i64 @llvm.ctpop.i64(i64 %val) nounwind readnone
  %truncctpop = trunc i64 %ctpop to i32
  store i32 %truncctpop, i32 addrspace(1)* %out, align 4
  ret void
}

; FUNC-LABEL: {{^}}v_ctpop_i64:
; GCN: {{buffer|flat}}_load_dwordx2 v{{\[}}[[LOVAL:[0-9]+]]:[[HIVAL:[0-9]+]]{{\]}},
; GCN: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT:v[0-9]+]], v[[LOVAL]], 0
; SI-NEXT: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
; VI-NEXT: v_bcnt_u32_b32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
define amdgpu_kernel void @v_ctpop_i64(i32 addrspace(1)* noalias %out, i64 addrspace(1)* noalias %in) nounwind {
  %tid = call i32 @llvm.r600.read.tidig.x()
  %in.gep = getelementptr i64, i64 addrspace(1)* %in, i32 %tid
  %val = load i64, i64 addrspace(1)* %in.gep, align 8
  %ctpop = call i64 @llvm.ctpop.i64(i64 %val) nounwind readnone
  %truncctpop = trunc i64 %ctpop to i32
  store i32 %truncctpop, i32 addrspace(1)* %out, align 4
  ret void
}

; FUNC-LABEL: {{^}}v_ctpop_i64_user:
; GCN: {{buffer|flat}}_load_dwordx2 v{{\[}}[[LOVAL:[0-9]+]]:[[HIVAL:[0-9]+]]{{\]}},
; GCN: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT:v[0-9]+]], v[[LOVAL]], 0
; SI-NEXT: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
; VI-NEXT: v_bcnt_u32_b32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
; GCN-DAG: v_or_b32_e32 v[[RESULT_LO:[0-9]+]], s{{[0-9]+}}, [[RESULT]]
; GCN-DAG: v_mov_b32_e32 v[[RESULT_HI:[0-9]+]], s{{[0-9]+}}
; GCN: buffer_store_dwordx2 v{{\[}}[[RESULT_LO]]:[[RESULT_HI]]{{\]}}
; GCN: s_endpgm
define amdgpu_kernel void @v_ctpop_i64_user(i64 addrspace(1)* noalias %out, i64 addrspace(1)* noalias %in, i64 %s.val) nounwind {
  %tid = call i32 @llvm.r600.read.tidig.x()
  %in.gep = getelementptr i64, i64 addrspace(1)* %in, i32 %tid
  %val = load i64, i64 addrspace(1)* %in.gep, align 8
  %ctpop = call i64 @llvm.ctpop.i64(i64 %val) nounwind readnone
  %or = or i64 %ctpop, %s.val
  store i64 %or, i64 addrspace(1)* %out
  ret void
}

; FUNC-LABEL: {{^}}s_ctpop_v2i64:
; GCN: s_bcnt1_i32_b64
; GCN: s_bcnt1_i32_b64
; GCN: s_endpgm
define amdgpu_kernel void @s_ctpop_v2i64(<2 x i32> addrspace(1)* noalias %out, <2 x i64> %val) nounwind {
  %ctpop = call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> %val) nounwind readnone
  %truncctpop = trunc <2 x i64> %ctpop to <2 x i32>
  store <2 x i32> %truncctpop, <2 x i32> addrspace(1)* %out, align 8
  ret void
}

; FUNC-LABEL: {{^}}s_ctpop_v4i64:
; GCN: s_bcnt1_i32_b64
; GCN: s_bcnt1_i32_b64
; GCN: s_bcnt1_i32_b64
; GCN: s_bcnt1_i32_b64
; GCN: s_endpgm
define amdgpu_kernel void @s_ctpop_v4i64(<4 x i32> addrspace(1)* noalias %out, <4 x i64> %val) nounwind {
  %ctpop = call <4 x i64> @llvm.ctpop.v4i64(<4 x i64> %val) nounwind readnone
  %truncctpop = trunc <4 x i64> %ctpop to <4 x i32>
  store <4 x i32> %truncctpop, <4 x i32> addrspace(1)* %out, align 16
  ret void
}

; FUNC-LABEL: {{^}}v_ctpop_v2i64:
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: s_endpgm
define amdgpu_kernel void @v_ctpop_v2i64(<2 x i32> addrspace(1)* noalias %out, <2 x i64> addrspace(1)* noalias %in) nounwind {
  %tid = call i32 @llvm.r600.read.tidig.x()
  %in.gep = getelementptr <2 x i64>, <2 x i64> addrspace(1)* %in, i32 %tid
  %val = load <2 x i64>, <2 x i64> addrspace(1)* %in.gep, align 16
  %ctpop = call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> %val) nounwind readnone
  %truncctpop = trunc <2 x i64> %ctpop to <2 x i32>
  store <2 x i32> %truncctpop, <2 x i32> addrspace(1)* %out, align 8
  ret void
}

; FUNC-LABEL: {{^}}v_ctpop_v4i64:
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: v_bcnt_u32_b32
; GCN: s_endpgm
define amdgpu_kernel void @v_ctpop_v4i64(<4 x i32> addrspace(1)* noalias %out, <4 x i64> addrspace(1)* noalias %in) nounwind {
  %tid = call i32 @llvm.r600.read.tidig.x()
  %in.gep = getelementptr <4 x i64>, <4 x i64> addrspace(1)* %in, i32 %tid
  %val = load <4 x i64>, <4 x i64> addrspace(1)* %in.gep, align 32
  %ctpop = call <4 x i64> @llvm.ctpop.v4i64(<4 x i64> %val) nounwind readnone
  %truncctpop = trunc <4 x i64> %ctpop to <4 x i32>
  store <4 x i32> %truncctpop, <4 x i32> addrspace(1)* %out, align 16
  ret void
}

; FUNC-LABEL: {{^}}ctpop_i64_in_br:
; SI-DAG: s_load_dwordx2 s{{\[}}[[LOVAL:[0-9]+]]:[[HIVAL:[0-9]+]]{{\]}}, s[{{[0-9]+:[0-9]+}}], 0xd
; VI-DAG: s_load_dwordx2 s{{\[}}[[LOVAL:[0-9]+]]:[[HIVAL:[0-9]+]]{{\]}}, s[{{[0-9]+:[0-9]+}}], 0x34
; GCN-DAG: s_bcnt1_i32_b64 [[RESULT:s[0-9]+]], {{s\[}}[[LOVAL]]:[[HIVAL]]{{\]}}
; GCN-DAG: s_mov_b32 [[ZERO:s[0-9]+]], 0
; GCN-DAG: v_mov_b32_e32 v[[VLO:[0-9]+]], [[RESULT]]
; GCN-DAG: v_mov_b32_e32 v[[VHI:[0-9]+]], [[ZERO]]
; GCN: buffer_store_dwordx2 {{v\[}}[[VLO]]:[[VHI]]{{\]}}
; GCN: s_endpgm
define amdgpu_kernel void @ctpop_i64_in_br(i64 addrspace(1)* %out, i64 addrspace(1)* %in, i64 %ctpop_arg, i32 %cond) {
entry:
  %tmp0 = icmp eq i32 %cond, 0
  br i1 %tmp0, label %if, label %else

if:
  %tmp2 = call i64 @llvm.ctpop.i64(i64 %ctpop_arg)
  br label %endif

else:
  %tmp3 = getelementptr i64, i64 addrspace(1)* %in, i32 1
  %tmp4 = load i64, i64 addrspace(1)* %tmp3
  br label %endif

endif:
  %tmp5 = phi i64 [%tmp2, %if], [%tmp4, %else]
  store i64 %tmp5, i64 addrspace(1)* %out
  ret void
}

; FUNC-LABEL: {{^}}s_ctpop_i128:
; GCN: s_bcnt1_i32_b64 [[SRESULT0:s[0-9]+]],
; GCN: s_bcnt1_i32_b64 [[SRESULT1:s[0-9]+]],
; GCN: s_add_i32 s{{[0-9]+}}, [[SRESULT1]], [[SRESULT0]]
; GCN: s_endpgm
define amdgpu_kernel void @s_ctpop_i128(i32 addrspace(1)* noalias %out, i128 %val) nounwind {
  %ctpop = call i128 @llvm.ctpop.i128(i128 %val) nounwind readnone
  %truncctpop = trunc i128 %ctpop to i32
  store i32 %truncctpop, i32 addrspace(1)* %out, align 4
  ret void
}

; FUNC-LABEL: {{^}}s_ctpop_i65:
; GCN: s_bcnt1_i32_b64 [[REG0:s[0-9]+]],
; GCN: s_and_b32
; GCN: s_bcnt1_i32_b64 [[REG1:s[0-9]+]],
; GCN: s_add_i32 {{s[0-9]+}}, [[REG0]], [[REG1]]
; GCN: s_endpgm
define amdgpu_kernel void @s_ctpop_i65(i32 addrspace(1)* noalias %out, i65 %val) nounwind {
  %ctpop = call i65 @llvm.ctpop.i65(i65 %val) nounwind readnone
  %truncctpop = trunc i65 %ctpop to i32
  store i32 %truncctpop, i32 addrspace(1)* %out, align 4
  ret void
}

; FIXME: Should not have extra add

; FUNC-LABEL: {{^}}v_ctpop_i128:
; SI: buffer_load_dwordx4 v{{\[}}[[VAL0:[0-9]+]]:[[VAL3:[0-9]+]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64
; VI: flat_load_dwordx4   v{{\[}}[[VAL0:[0-9]+]]:[[VAL3:[0-9]+]]{{\]}}, v{{\[[0-9]+:[0-9]+\]}}

; GCN-DAG: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT0:v[0-9]+]], v{{[0-9]+}}, 0
; GCN-DAG: v_bcnt_u32_b32{{(_e32)*(_e64)*}} [[MIDRESULT1:v[0-9]+]], v[[VAL3]], [[MIDRESULT0]]

; GCN-DAG: v_bcnt_u32_b32{{(_e64)*}} [[MIDRESULT2:v[0-9]+]], v[[VAL0]], 0
; GCN-DAG: v_bcnt_u32_b32{{(_e32)*(_e64)*}} [[MIDRESULT3:v[0-9]+]], v{{[0-9]+}}, [[MIDRESULT2]]

; GCN: v_add_{{[iu]}}32_e32 [[RESULT:v[0-9]+]], vcc, [[MIDRESULT2]], [[MIDRESULT1]]

; GCN: buffer_store_dword [[RESULT]],
; GCN: s_endpgm
define amdgpu_kernel void @v_ctpop_i128(i32 addrspace(1)* noalias %out, i128 addrspace(1)* noalias %in) nounwind {
  %tid = call i32 @llvm.r600.read.tidig.x()
  %in.gep = getelementptr i128, i128 addrspace(1)* %in, i32 %tid
  %val = load i128, i128 addrspace(1)* %in.gep, align 8
  %ctpop = call i128 @llvm.ctpop.i128(i128 %val) nounwind readnone
  %truncctpop = trunc i128 %ctpop to i32
  store i32 %truncctpop, i32 addrspace(1)* %out, align 4
  ret void
}