Codebase list cppad / lintian-fixes/main omh / multi_thread.omh
lintian-fixes/main

Tree @lintian-fixes/main (Download .tar.gz)

multi_thread.omh @lintian-fixes/main — raw · history · blame 

 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
/* --------------------------------------------------------------------------
CppAD: C++ Algorithmic Differentiation: Copyright (C) 2003-17 Bradley M. Bell

  CppAD is distributed under the terms of the
               Eclipse Public License Version 2.0.

  This Source Code may also be made available under the following
  Secondary License when the conditions for such availability set forth
  in the Eclipse Public License, Version 2.0 are satisfied:
        GNU General Public License, Version 2.0 or later.
-------------------------------------------------------------------------- */
$begin multi_thread$$
$spell
    cppad_max_num_threads
    alloc
    CppAD
    Rosen
    Runge
$$

$section Using CppAD in a Multi-Threading Environment$$

$head Purpose$$
Extra steps and care must be taken to use CppAD
in $cref/parallel/ta_in_parallel/$$ execution mode.
This section collects this information in one place.

$head CPPAD_MAX_NUM_THREADS$$
The value $code CPPAD_MAX_NUM_THREADS$$ is an absolute maximum for
the number of threads that CppAD should support.
If this preprocessor symbol is defined
before including any CppAD header files,
it must be an integer greater than or equal to one.
Otherwise, $cref/cppad_max_num_threads/cmake/cppad_max_num_threads/$$
is used to define this preprocessor symbol.
Note that the minimum allowable value for
$icode cppad_max_num_threads$$ is 4; i.e., you can only
get smaller values for $code CPPAD_MAX_NUM_THREADS$$ by
defining it before including the CppAD header files.

$head parallel_setup$$
Using any of the following routines in a multi-threading environment
requires that $cref/thread_alloc::parallel_setup/ta_parallel_setup/$$
has been completed:
$cref/CppAD::vector/CppAD_vector/Memory and Parallel Mode/$$,
$cref/CheckSimpleVector/CheckSimpleVector/Parallel Mode/$$,
$cref/CheckNumericType/CheckNumericType/Parallel Mode/$$,
$cref parallel_ad$$.

$head hold_memory$$
Memory allocation should be much faster after calling
$code hold_memory$$ with
$cref/value/ta_hold_memory/value/$$ equal to true.
This may even be true if there is only one thread.

$head Parallel AD$$
One must first call
$cref/thread_alloc::parallel_setup/ta_parallel_setup/$$
and then call $cref parallel_ad$$ before using $code AD$$ types
in $cref/parallel/ta_in_parallel/$$ execution mode.

$head Initialization$$
The following routines must be called at least once before being used
in parallel mode:
$cref/CheckSimpleVector/CheckSimpleVector/Parallel Mode/$$,
$cref/CheckNumericType/CheckNumericType/Parallel Mode/$$,
$cref/discrete functions/Discrete/Parallel Mode/$$,
$cref/Rosen34/Rosen34/Parallel Mode/$$,
$cref/Runge45/Runge45/Parallel Mode/$$.

$head Same Thread$$
Some operations must be preformed by the same thread:
$cref/ADFun/FunConstruct/Parallel Mode/$$,
$cref Independent$$, $cref Dependent$$.


$head Parallel Prohibited$$
The following routine cannot be called in parallel mode:
$cref/ErrorHandler constructor/ErrorHandler/Constructor/Parallel Mode/$$.

$childtable%
    include/cppad/core/parallel_ad.hpp%
    example/multi_thread/thread_test.cpp
%$$

$end