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
|
SuiteSparse:GraphBLAS, Timothy A. Davis, (c) 2017-2025, All Rights Reserved.
SPDX-License-Identifier: Apache-2.0
This folder, GraphBLAS/Source, contains all the primary source files for
GraphBLAS, and internal include files that are meant for internal inclusion in
GraphBLAS itself. They should not be included in end-user applications.
The Source/*/template/* files are not compiled separately, but are #include'd
into files in this folder instead. The are templatized source files. The
Source/*/include/* files are header files for template methods. Both template
and include files are merged into the JITpackage and then exported to
~/.SuiteSparse/GrB.x.y.z/src.
Source/*/factory/* files are not used in the JIT.
Files and folders in Source/
add GrB_eWiseAdd and GxB_eWiseUnion (see also ewise folder)
aliased methods for determining if any components of a matrix
are aliased (shared) with another matrix. Such a matrix
is called shallow (an array such as A->i is shallow
if A->i_shallow is true, and is owned by another matrix).
Such matrices are never returned to the user; they are
only internal.
apply GrB_apply
assign GrB_assign and GxB_assign
binaryop GrB_BinaryOp object
builder GrB_build
builtin builtin types, operators, monoids, and semirings
callback JIT kernels are not linked against -lgraphblas and thus
cannot call back into GraphBLAS directly. Instead, a
struct containing function pointers to various utility
functions is passed to each JIT kernel.
cast typecasting
clear GrB_clear
codegen MATLAB scripts for creating ../FactoryKernels,
Source/mxm/GB_AxB__any_pair_iso.c, and
Source/mxm/GB_AxB__include1.h.
compatible testing if operators and types are compatible (if they can
be typecasted to another type.
concat GxB_concat
context GxB_Context object
convert convert between matrix formats (sparse, hypersparse, bitmap,
and full), and conform a matrix to its desired format.
cpu wrapper for Google's cpu_features package
cumsum cumulative sum
descriptor GrB_Descriptor object
diag GrB_diag and GxB_diag
dup GrB_dup
element GrB_setElement, GrB_extractElement, GrB_removeElement
emult GrB_eWiseMult (see also ewise folder)
ewise element-wise methods for ewise add, emult, and union
extract GrB_extract
extractTuples GrB_extractTuples
gateway definitions for calling methods in the CUDA folder
GB.h primary internal include file
GB_control.h controls which FactoryKernels are compiled
generic definitions for generic methods
get_set GrB_get, GrB_set
global global variables and the routines for accessing them
helper methods used in MATLAB/Octave @GrB interface only
hyper methods for hypersparse matrices
ij determining properities of I and J index lists for
GrB_extract, GrB_assign, and GxB_subassign
import_export GrB_import, GrB_export, GxB_import, and GxB_export
include general-purpose header files that do not fit into any
particular subfolder of GraphBLAS/Source, such as compiler
settings, and GB_include.h which is a primary internal
include file.
indexbinaryop GxB_IndexBinaryOp object
indexunaryop GrB_IndexUnaryOp object
init GrB_init, GxB_init, GrB_finalize, GrB_error, GrB_getVersion
iso iso-valued matrices
iterator GxB_Iterator object and its methods
jitifyer the GraphBLAS Just-In-Time compiler for CPU and CUDA JIT
kernels
jit_kernels templates for all CPU JIT kernels. These are not compiled
when the libgraphblas.so library is built. Instead, they
are compiled at run time when requested by the JIT.
jit_wrappers interfaces that access the JIT to call/load/compile each
CPU JIT kernel
kronecker GrB_kronecker
lz4_wrapper wrapper for the lz4 compression package
mask mask/accum methods, for computing C<M>+=T
math basic mathematical functions
matrix GrB_Matrix object
memory memory allocators (wrappers for malloc, free, etc),
memset, memcpy
monoid GrB_Monoid object
mxm GrB_mxm, GrB_vxm, and GrB_mxv
nvals GrB_nvals
ok debugging assertions
omp OpenMP interface and atomics
op methods for operators (GrB_UnaryOp, GrB_BinaryOp, etc)
pack_unpack GxB_pack, GxB_unpack
pending pending tuples for updates to matrices, vectors, and scalars
positional methods for positional operators
print GxB_print, GxB_fprint
README.txt this file
reduce GrB_reduce, to scalar and vector
reshape GrB_reshape
resize GrB_resize
scalar GrB_Scalar object
select GrB_select
semiring GrB_Semiring object
serialize GrB_serialize, GrB_deserialze
slice methods for slicing matrices to create parallel tasks
sort GxB_sort, and internal sorting methods
split GxB_split
transplant transplant contents of one matrix into another
transpose GrB_transpose
type GrB_Type object
unaryop GrB_UnaryOp object
vector GrB_Vector object
wait GrB_wait
werk the Werk space is a small amount of space on the stack
use for small things (scalars, O(# parallel OpenMP tasks),
and such. It is spell differently for easier 'grep'.
zstd_wrapper wrapper for the zstd compression package
|
