Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.3/doc/html/Basic-and-advanced-distribution-interfaces.html @ 95:89f5e221ed7b

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW3
author Chris Cannam <cannam@all-day-breakfast.com>
date Wed, 20 Mar 2013 15:35:50 +0000
parents
children
comparison
equal deleted inserted replaced
94:d278df1123f9 95:89f5e221ed7b
1 <html lang="en">
2 <head>
3 <title>Basic and advanced distribution interfaces - FFTW 3.3.3</title>
4 <meta http-equiv="Content-Type" content="text/html">
5 <meta name="description" content="FFTW 3.3.3">
6 <meta name="generator" content="makeinfo 4.13">
7 <link title="Top" rel="start" href="index.html#Top">
8 <link rel="up" href="MPI-Data-Distribution.html#MPI-Data-Distribution" title="MPI Data Distribution">
9 <link rel="prev" href="MPI-Data-Distribution.html#MPI-Data-Distribution" title="MPI Data Distribution">
10 <link rel="next" href="Load-balancing.html#Load-balancing" title="Load balancing">
11 <link href="http://www.gnu.org/software/texinfo/" rel="generator-home" title="Texinfo Homepage">
12 <!--
13 This manual is for FFTW
14 (version 3.3.3, 25 November 2012).
15
16 Copyright (C) 2003 Matteo Frigo.
17
18 Copyright (C) 2003 Massachusetts Institute of Technology.
19
20 Permission is granted to make and distribute verbatim copies of
21 this manual provided the copyright notice and this permission
22 notice are preserved on all copies.
23
24 Permission is granted to copy and distribute modified versions of
25 this manual under the conditions for verbatim copying, provided
26 that the entire resulting derived work is distributed under the
27 terms of a permission notice identical to this one.
28
29 Permission is granted to copy and distribute translations of this
30 manual into another language, under the above conditions for
31 modified versions, except that this permission notice may be
32 stated in a translation approved by the Free Software Foundation.
33 -->
34 <meta http-equiv="Content-Style-Type" content="text/css">
35 <style type="text/css"><!--
36 pre.display { font-family:inherit }
37 pre.format { font-family:inherit }
38 pre.smalldisplay { font-family:inherit; font-size:smaller }
39 pre.smallformat { font-family:inherit; font-size:smaller }
40 pre.smallexample { font-size:smaller }
41 pre.smalllisp { font-size:smaller }
42 span.sc { font-variant:small-caps }
43 span.roman { font-family:serif; font-weight:normal; }
44 span.sansserif { font-family:sans-serif; font-weight:normal; }
45 --></style>
46 </head>
47 <body>
48 <div class="node">
49 <a name="Basic-and-advanced-distribution-interfaces"></a>
50 <p>
51 Next:&nbsp;<a rel="next" accesskey="n" href="Load-balancing.html#Load-balancing">Load balancing</a>,
52 Previous:&nbsp;<a rel="previous" accesskey="p" href="MPI-Data-Distribution.html#MPI-Data-Distribution">MPI Data Distribution</a>,
53 Up:&nbsp;<a rel="up" accesskey="u" href="MPI-Data-Distribution.html#MPI-Data-Distribution">MPI Data Distribution</a>
54 <hr>
55 </div>
56
57 <h4 class="subsection">6.4.1 Basic and advanced distribution interfaces</h4>
58
59 <p>As with the planner interface, the &lsquo;<samp><span class="samp">fftw_mpi_local_size</span></samp>&rsquo;
60 distribution interface is broken into basic and advanced
61 (&lsquo;<samp><span class="samp">_many</span></samp>&rsquo;) interfaces, where the latter allows you to specify the
62 block size manually and also to request block sizes when computing
63 multiple transforms simultaneously. These functions are documented
64 more exhaustively by the FFTW MPI Reference, but we summarize the
65 basic ideas here using a couple of two-dimensional examples.
66
67 <p>For the 100&nbsp;&times;&nbsp;200 complex-DFT example, above, we would find
68 the distribution by calling the following function in the basic
69 interface:
70
71 <pre class="example"> ptrdiff_t fftw_mpi_local_size_2d(ptrdiff_t n0, ptrdiff_t n1, MPI_Comm comm,
72 ptrdiff_t *local_n0, ptrdiff_t *local_0_start);
73 </pre>
74 <p><a name="index-fftw_005fmpi_005flocal_005fsize_005f2d-370"></a>
75 Given the total size of the data to be transformed (here, <code>n0 =
76 100</code> and <code>n1 = 200</code>) and an MPI communicator (<code>comm</code>), this
77 function provides three numbers.
78
79 <p>First, it describes the shape of the local data: the current process
80 should store a <code>local_n0</code> by <code>n1</code> slice of the overall
81 dataset, in row-major order (<code>n1</code> dimension contiguous), starting
82 at index <code>local_0_start</code>. That is, if the total dataset is
83 viewed as a <code>n0</code> by <code>n1</code> matrix, the current process should
84 store the rows <code>local_0_start</code> to
85 <code>local_0_start+local_n0-1</code>. Obviously, if you are running with
86 only a single MPI process, that process will store the entire array:
87 <code>local_0_start</code> will be zero and <code>local_n0</code> will be
88 <code>n0</code>. See <a href="Row_002dmajor-Format.html#Row_002dmajor-Format">Row-major Format</a>.
89 <a name="index-row_002dmajor-371"></a>
90
91 <p>Second, the return value is the total number of data elements (e.g.,
92 complex numbers for a complex DFT) that should be allocated for the
93 input and output arrays on the current process (ideally with
94 <code>fftw_malloc</code> or an &lsquo;<samp><span class="samp">fftw_alloc</span></samp>&rsquo; function, to ensure optimal
95 alignment). It might seem that this should always be equal to
96 <code>local_n0 * n1</code>, but this is <em>not</em> the case. FFTW's
97 distributed FFT algorithms require data redistributions at
98 intermediate stages of the transform, and in some circumstances this
99 may require slightly larger local storage. This is discussed in more
100 detail below, under <a href="Load-balancing.html#Load-balancing">Load balancing</a>.
101 <a name="index-fftw_005fmalloc-372"></a><a name="index-fftw_005falloc_005fcomplex-373"></a>
102
103 <p><a name="index-advanced-interface-374"></a>The advanced-interface &lsquo;<samp><span class="samp">local_size</span></samp>&rsquo; function for multidimensional
104 transforms returns the same three things (<code>local_n0</code>,
105 <code>local_0_start</code>, and the total number of elements to allocate),
106 but takes more inputs:
107
108 <pre class="example"> ptrdiff_t fftw_mpi_local_size_many(int rnk, const ptrdiff_t *n,
109 ptrdiff_t howmany,
110 ptrdiff_t block0,
111 MPI_Comm comm,
112 ptrdiff_t *local_n0,
113 ptrdiff_t *local_0_start);
114 </pre>
115 <p><a name="index-fftw_005fmpi_005flocal_005fsize_005fmany-375"></a>
116 The two-dimensional case above corresponds to <code>rnk = 2</code> and an
117 array <code>n</code> of length 2 with <code>n[0] = n0</code> and <code>n[1] = n1</code>.
118 This routine is for any <code>rnk &gt; 1</code>; one-dimensional transforms
119 have their own interface because they work slightly differently, as
120 discussed below.
121
122 <p>First, the advanced interface allows you to perform multiple
123 transforms at once, of interleaved data, as specified by the
124 <code>howmany</code> parameter. (<code>hoamany</code> is 1 for a single
125 transform.)
126
127 <p>Second, here you can specify your desired block size in the <code>n0</code>
128 dimension, <code>block0</code>. To use FFTW's default block size, pass
129 <code>FFTW_MPI_DEFAULT_BLOCK</code> (0) for <code>block0</code>. Otherwise, on
130 <code>P</code> processes, FFTW will return <code>local_n0</code> equal to
131 <code>block0</code> on the first <code>P / block0</code> processes (rounded down),
132 return <code>local_n0</code> equal to <code>n0 - block0 * (P / block0)</code> on
133 the next process, and <code>local_n0</code> equal to zero on any remaining
134 processes. In general, we recommend using the default block size
135 (which corresponds to <code>n0 / P</code>, rounded up).
136 <a name="index-FFTW_005fMPI_005fDEFAULT_005fBLOCK-376"></a><a name="index-block-distribution-377"></a>
137
138 <p>For example, suppose you have <code>P = 4</code> processes and <code>n0 =
139 21</code>. The default will be a block size of <code>6</code>, which will give
140 <code>local_n0 = 6</code> on the first three processes and <code>local_n0 =
141 3</code> on the last process. Instead, however, you could specify
142 <code>block0 = 5</code> if you wanted, which would give <code>local_n0 = 5</code>
143 on processes 0 to 2, <code>local_n0 = 6</code> on process 3. (This choice,
144 while it may look superficially more &ldquo;balanced,&rdquo; has the same
145 critical path as FFTW's default but requires more communications.)
146
147 </body></html>
148