|
Chris@10
|
1 <html lang="en">
|
|
Chris@10
|
2 <head>
|
|
Chris@10
|
3 <title>Transposed distributions - FFTW 3.3.3</title>
|
|
Chris@10
|
4 <meta http-equiv="Content-Type" content="text/html">
|
|
Chris@10
|
5 <meta name="description" content="FFTW 3.3.3">
|
|
Chris@10
|
6 <meta name="generator" content="makeinfo 4.13">
|
|
Chris@10
|
7 <link title="Top" rel="start" href="index.html#Top">
|
|
Chris@10
|
8 <link rel="up" href="MPI-Data-Distribution.html#MPI-Data-Distribution" title="MPI Data Distribution">
|
|
Chris@10
|
9 <link rel="prev" href="Load-balancing.html#Load-balancing" title="Load balancing">
|
|
Chris@10
|
10 <link rel="next" href="One_002ddimensional-distributions.html#One_002ddimensional-distributions" title="One-dimensional distributions">
|
|
Chris@10
|
11 <link href="http://www.gnu.org/software/texinfo/" rel="generator-home" title="Texinfo Homepage">
|
|
Chris@10
|
12 <!--
|
|
Chris@10
|
13 This manual is for FFTW
|
|
Chris@10
|
14 (version 3.3.3, 25 November 2012).
|
|
Chris@10
|
15
|
|
Chris@10
|
16 Copyright (C) 2003 Matteo Frigo.
|
|
Chris@10
|
17
|
|
Chris@10
|
18 Copyright (C) 2003 Massachusetts Institute of Technology.
|
|
Chris@10
|
19
|
|
Chris@10
|
20 Permission is granted to make and distribute verbatim copies of
|
|
Chris@10
|
21 this manual provided the copyright notice and this permission
|
|
Chris@10
|
22 notice are preserved on all copies.
|
|
Chris@10
|
23
|
|
Chris@10
|
24 Permission is granted to copy and distribute modified versions of
|
|
Chris@10
|
25 this manual under the conditions for verbatim copying, provided
|
|
Chris@10
|
26 that the entire resulting derived work is distributed under the
|
|
Chris@10
|
27 terms of a permission notice identical to this one.
|
|
Chris@10
|
28
|
|
Chris@10
|
29 Permission is granted to copy and distribute translations of this
|
|
Chris@10
|
30 manual into another language, under the above conditions for
|
|
Chris@10
|
31 modified versions, except that this permission notice may be
|
|
Chris@10
|
32 stated in a translation approved by the Free Software Foundation.
|
|
Chris@10
|
33 -->
|
|
Chris@10
|
34 <meta http-equiv="Content-Style-Type" content="text/css">
|
|
Chris@10
|
35 <style type="text/css"><!--
|
|
Chris@10
|
36 pre.display { font-family:inherit }
|
|
Chris@10
|
37 pre.format { font-family:inherit }
|
|
Chris@10
|
38 pre.smalldisplay { font-family:inherit; font-size:smaller }
|
|
Chris@10
|
39 pre.smallformat { font-family:inherit; font-size:smaller }
|
|
Chris@10
|
40 pre.smallexample { font-size:smaller }
|
|
Chris@10
|
41 pre.smalllisp { font-size:smaller }
|
|
Chris@10
|
42 span.sc { font-variant:small-caps }
|
|
Chris@10
|
43 span.roman { font-family:serif; font-weight:normal; }
|
|
Chris@10
|
44 span.sansserif { font-family:sans-serif; font-weight:normal; }
|
|
Chris@10
|
45 --></style>
|
|
Chris@10
|
46 </head>
|
|
Chris@10
|
47 <body>
|
|
Chris@10
|
48 <div class="node">
|
|
Chris@10
|
49 <a name="Transposed-distributions"></a>
|
|
Chris@10
|
50 <p>
|
|
Chris@10
|
51 Next: <a rel="next" accesskey="n" href="One_002ddimensional-distributions.html#One_002ddimensional-distributions">One-dimensional distributions</a>,
|
|
Chris@10
|
52 Previous: <a rel="previous" accesskey="p" href="Load-balancing.html#Load-balancing">Load balancing</a>,
|
|
Chris@10
|
53 Up: <a rel="up" accesskey="u" href="MPI-Data-Distribution.html#MPI-Data-Distribution">MPI Data Distribution</a>
|
|
Chris@10
|
54 <hr>
|
|
Chris@10
|
55 </div>
|
|
Chris@10
|
56
|
|
Chris@10
|
57 <h4 class="subsection">6.4.3 Transposed distributions</h4>
|
|
Chris@10
|
58
|
|
Chris@10
|
59 <p>Internally, FFTW's MPI transform algorithms work by first computing
|
|
Chris@10
|
60 transforms of the data local to each process, then by globally
|
|
Chris@10
|
61 <em>transposing</em> the data in some fashion to redistribute the data
|
|
Chris@10
|
62 among the processes, transforming the new data local to each process,
|
|
Chris@10
|
63 and transposing back. For example, a two-dimensional <code>n0</code> by
|
|
Chris@10
|
64 <code>n1</code> array, distributed across the <code>n0</code> dimension, is
|
|
Chris@10
|
65 transformd by: (i) transforming the <code>n1</code> dimension, which are
|
|
Chris@10
|
66 local to each process; (ii) transposing to an <code>n1</code> by <code>n0</code>
|
|
Chris@10
|
67 array, distributed across the <code>n1</code> dimension; (iii) transforming
|
|
Chris@10
|
68 the <code>n0</code> dimension, which is now local to each process; (iv)
|
|
Chris@10
|
69 transposing back.
|
|
Chris@10
|
70 <a name="index-transpose-379"></a>
|
|
Chris@10
|
71
|
|
Chris@10
|
72 <p>However, in many applications it is acceptable to compute a
|
|
Chris@10
|
73 multidimensional DFT whose results are produced in transposed order
|
|
Chris@10
|
74 (e.g., <code>n1</code> by <code>n0</code> in two dimensions). This provides a
|
|
Chris@10
|
75 significant performance advantage, because it means that the final
|
|
Chris@10
|
76 transposition step can be omitted. FFTW supports this optimization,
|
|
Chris@10
|
77 which you specify by passing the flag <code>FFTW_MPI_TRANSPOSED_OUT</code>
|
|
Chris@10
|
78 to the planner routines. To compute the inverse transform of
|
|
Chris@10
|
79 transposed output, you specify <code>FFTW_MPI_TRANSPOSED_IN</code> to tell
|
|
Chris@10
|
80 it that the input is transposed. In this section, we explain how to
|
|
Chris@10
|
81 interpret the output format of such a transform.
|
|
Chris@10
|
82 <a name="index-FFTW_005fMPI_005fTRANSPOSED_005fOUT-380"></a><a name="index-FFTW_005fMPI_005fTRANSPOSED_005fIN-381"></a>
|
|
Chris@10
|
83
|
|
Chris@10
|
84 <p>Suppose you have are transforming multi-dimensional data with (at
|
|
Chris@10
|
85 least two) dimensions n<sub>0</sub> × n<sub>1</sub> × n<sub>2</sub> × … × n<sub>d-1</sub>. As always, it is distributed along
|
|
Chris@10
|
86 the first dimension n<sub>0</sub>. Now, if we compute its DFT with the
|
|
Chris@10
|
87 <code>FFTW_MPI_TRANSPOSED_OUT</code> flag, the resulting output data are stored
|
|
Chris@10
|
88 with the first <em>two</em> dimensions transposed: n<sub>1</sub> × n<sub>0</sub> × n<sub>2</sub> ×…× n<sub>d-1</sub>,
|
|
Chris@10
|
89 distributed along the n<sub>1</sub> dimension. Conversely, if we take the
|
|
Chris@10
|
90 n<sub>1</sub> × n<sub>0</sub> × n<sub>2</sub> ×…× n<sub>d-1</sub> data and transform it with the
|
|
Chris@10
|
91 <code>FFTW_MPI_TRANSPOSED_IN</code> flag, then the format goes back to the
|
|
Chris@10
|
92 original n<sub>0</sub> × n<sub>1</sub> × n<sub>2</sub> × … × n<sub>d-1</sub> array.
|
|
Chris@10
|
93
|
|
Chris@10
|
94 <p>There are two ways to find the portion of the transposed array that
|
|
Chris@10
|
95 resides on the current process. First, you can simply call the
|
|
Chris@10
|
96 appropriate ‘<samp><span class="samp">local_size</span></samp>’ function, passing n<sub>1</sub> × n<sub>0</sub> × n<sub>2</sub> ×…× n<sub>d-1</sub> (the
|
|
Chris@10
|
97 transposed dimensions). This would mean calling the ‘<samp><span class="samp">local_size</span></samp>’
|
|
Chris@10
|
98 function twice, once for the transposed and once for the
|
|
Chris@10
|
99 non-transposed dimensions. Alternatively, you can call one of the
|
|
Chris@10
|
100 ‘<samp><span class="samp">local_size_transposed</span></samp>’ functions, which returns both the
|
|
Chris@10
|
101 non-transposed and transposed data distribution from a single call.
|
|
Chris@10
|
102 For example, for a 3d transform with transposed output (or input), you
|
|
Chris@10
|
103 might call:
|
|
Chris@10
|
104
|
|
Chris@10
|
105 <pre class="example"> ptrdiff_t fftw_mpi_local_size_3d_transposed(
|
|
Chris@10
|
106 ptrdiff_t n0, ptrdiff_t n1, ptrdiff_t n2, MPI_Comm comm,
|
|
Chris@10
|
107 ptrdiff_t *local_n0, ptrdiff_t *local_0_start,
|
|
Chris@10
|
108 ptrdiff_t *local_n1, ptrdiff_t *local_1_start);
|
|
Chris@10
|
109 </pre>
|
|
Chris@10
|
110 <p><a name="index-fftw_005fmpi_005flocal_005fsize_005f3d_005ftransposed-382"></a>
|
|
Chris@10
|
111 Here, <code>local_n0</code> and <code>local_0_start</code> give the size and
|
|
Chris@10
|
112 starting index of the <code>n0</code> dimension for the
|
|
Chris@10
|
113 <em>non</em>-transposed data, as in the previous sections. For
|
|
Chris@10
|
114 <em>transposed</em> data (e.g. the output for
|
|
Chris@10
|
115 <code>FFTW_MPI_TRANSPOSED_OUT</code>), <code>local_n1</code> and
|
|
Chris@10
|
116 <code>local_1_start</code> give the size and starting index of the <code>n1</code>
|
|
Chris@10
|
117 dimension, which is the first dimension of the transposed data
|
|
Chris@10
|
118 (<code>n1</code> by <code>n0</code> by <code>n2</code>).
|
|
Chris@10
|
119
|
|
Chris@10
|
120 <p>(Note that <code>FFTW_MPI_TRANSPOSED_IN</code> is completely equivalent to
|
|
Chris@10
|
121 performing <code>FFTW_MPI_TRANSPOSED_OUT</code> and passing the first two
|
|
Chris@10
|
122 dimensions to the planner in reverse order, or vice versa. If you
|
|
Chris@10
|
123 pass <em>both</em> the <code>FFTW_MPI_TRANSPOSED_IN</code> and
|
|
Chris@10
|
124 <code>FFTW_MPI_TRANSPOSED_OUT</code> flags, it is equivalent to swapping the
|
|
Chris@10
|
125 first two dimensions passed to the planner and passing <em>neither</em>
|
|
Chris@10
|
126 flag.)
|
|
Chris@10
|
127
|
|
Chris@10
|
128 </body></html>
|
|
Chris@10
|
129
|
