Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.3/doc/html/MPI-Plan-Creation.html @ 10:37bf6b4a2645

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW3
author Chris Cannam
date Wed, 20 Mar 2013 15:35:50 +0000
parents
children
rev   line source
Chris@10 1 <html lang="en">
Chris@10 2 <head>
Chris@10 3 <title>MPI Plan Creation - 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="FFTW-MPI-Reference.html#FFTW-MPI-Reference" title="FFTW MPI Reference">
Chris@10 9 <link rel="prev" href="MPI-Data-Distribution-Functions.html#MPI-Data-Distribution-Functions" title="MPI Data Distribution Functions">
Chris@10 10 <link rel="next" href="MPI-Wisdom-Communication.html#MPI-Wisdom-Communication" title="MPI Wisdom Communication">
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="MPI-Plan-Creation"></a>
Chris@10 50 <p>
Chris@10 51 Next:&nbsp;<a rel="next" accesskey="n" href="MPI-Wisdom-Communication.html#MPI-Wisdom-Communication">MPI Wisdom Communication</a>,
Chris@10 52 Previous:&nbsp;<a rel="previous" accesskey="p" href="MPI-Data-Distribution-Functions.html#MPI-Data-Distribution-Functions">MPI Data Distribution Functions</a>,
Chris@10 53 Up:&nbsp;<a rel="up" accesskey="u" href="FFTW-MPI-Reference.html#FFTW-MPI-Reference">FFTW MPI Reference</a>
Chris@10 54 <hr>
Chris@10 55 </div>
Chris@10 56
Chris@10 57 <h4 class="subsection">6.12.5 MPI Plan Creation</h4>
Chris@10 58
Chris@10 59 <h5 class="subsubheading">Complex-data MPI DFTs</h5>
Chris@10 60
Chris@10 61 <p>Plans for complex-data DFTs (see <a href="2d-MPI-example.html#g_t2d-MPI-example">2d MPI example</a>) are created by:
Chris@10 62
Chris@10 63 <p><a name="index-fftw_005fmpi_005fplan_005fdft_005f1d-461"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005f2d-462"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005f3d-463"></a><a name="index-fftw_005fmpi_005fplan_005fdft-464"></a><a name="index-fftw_005fmpi_005fplan_005fmany_005fdft-465"></a>
Chris@10 64 <pre class="example"> fftw_plan fftw_mpi_plan_dft_1d(ptrdiff_t n0, fftw_complex *in, fftw_complex *out,
Chris@10 65 MPI_Comm comm, int sign, unsigned flags);
Chris@10 66 fftw_plan fftw_mpi_plan_dft_2d(ptrdiff_t n0, ptrdiff_t n1,
Chris@10 67 fftw_complex *in, fftw_complex *out,
Chris@10 68 MPI_Comm comm, int sign, unsigned flags);
Chris@10 69 fftw_plan fftw_mpi_plan_dft_3d(ptrdiff_t n0, ptrdiff_t n1, ptrdiff_t n2,
Chris@10 70 fftw_complex *in, fftw_complex *out,
Chris@10 71 MPI_Comm comm, int sign, unsigned flags);
Chris@10 72 fftw_plan fftw_mpi_plan_dft(int rnk, const ptrdiff_t *n,
Chris@10 73 fftw_complex *in, fftw_complex *out,
Chris@10 74 MPI_Comm comm, int sign, unsigned flags);
Chris@10 75 fftw_plan fftw_mpi_plan_many_dft(int rnk, const ptrdiff_t *n,
Chris@10 76 ptrdiff_t howmany, ptrdiff_t block, ptrdiff_t tblock,
Chris@10 77 fftw_complex *in, fftw_complex *out,
Chris@10 78 MPI_Comm comm, int sign, unsigned flags);
Chris@10 79 </pre>
Chris@10 80 <p><a name="index-MPI-communicator-466"></a><a name="index-collective-function-467"></a>These are similar to their serial counterparts (see <a href="Complex-DFTs.html#Complex-DFTs">Complex DFTs</a>)
Chris@10 81 in specifying the dimensions, sign, and flags of the transform. The
Chris@10 82 <code>comm</code> argument gives an MPI communicator that specifies the set
Chris@10 83 of processes to participate in the transform; plan creation is a
Chris@10 84 collective function that must be called for all processes in the
Chris@10 85 communicator. The <code>in</code> and <code>out</code> pointers refer only to a
Chris@10 86 portion of the overall transform data (see <a href="MPI-Data-Distribution.html#MPI-Data-Distribution">MPI Data Distribution</a>)
Chris@10 87 as specified by the &lsquo;<samp><span class="samp">local_size</span></samp>&rsquo; functions in the previous
Chris@10 88 section. Unless <code>flags</code> contains <code>FFTW_ESTIMATE</code>, these
Chris@10 89 arrays are overwritten during plan creation as for the serial
Chris@10 90 interface. For multi-dimensional transforms, any dimensions <code>&gt;
Chris@10 91 1</code> are supported; for one-dimensional transforms, only composite
Chris@10 92 (non-prime) <code>n0</code> are currently supported (unlike the serial
Chris@10 93 FFTW). Requesting an unsupported transform size will yield a
Chris@10 94 <code>NULL</code> plan. (As in the serial interface, highly composite sizes
Chris@10 95 generally yield the best performance.)
Chris@10 96
Chris@10 97 <p><a name="index-advanced-interface-468"></a><a name="index-FFTW_005fMPI_005fDEFAULT_005fBLOCK-469"></a><a name="index-stride-470"></a>The advanced-interface <code>fftw_mpi_plan_many_dft</code> additionally
Chris@10 98 allows you to specify the block sizes for the first dimension
Chris@10 99 (<code>block</code>) of the n<sub>0</sub>&nbsp;&times;&nbsp;n<sub>1</sub>&nbsp;&times;&nbsp;n<sub>2</sub>&nbsp;&times;&nbsp;&hellip;&nbsp;&times;&nbsp;n<sub>d-1</sub> input data and the first dimension
Chris@10 100 (<code>tblock</code>) of the n<sub>1</sub>&nbsp;&times;&nbsp;n<sub>0</sub>&nbsp;&times;&nbsp;n<sub>2</sub>&nbsp;&times;&hellip;&times;&nbsp;n<sub>d-1</sub> transposed data (at intermediate
Chris@10 101 steps of the transform, and for the output if
Chris@10 102 <code>FFTW_TRANSPOSED_OUT</code> is specified in <code>flags</code>). These must
Chris@10 103 be the same block sizes as were passed to the corresponding
Chris@10 104 &lsquo;<samp><span class="samp">local_size</span></samp>&rsquo; function; you can pass <code>FFTW_MPI_DEFAULT_BLOCK</code>
Chris@10 105 to use FFTW's default block size as in the basic interface. Also, the
Chris@10 106 <code>howmany</code> parameter specifies that the transform is of contiguous
Chris@10 107 <code>howmany</code>-tuples rather than individual complex numbers; this
Chris@10 108 corresponds to the same parameter in the serial advanced interface
Chris@10 109 (see <a href="Advanced-Complex-DFTs.html#Advanced-Complex-DFTs">Advanced Complex DFTs</a>) with <code>stride = howmany</code> and
Chris@10 110 <code>dist = 1</code>.
Chris@10 111
Chris@10 112 <h5 class="subsubheading">MPI flags</h5>
Chris@10 113
Chris@10 114 <p>The <code>flags</code> can be any of those for the serial FFTW
Chris@10 115 (see <a href="Planner-Flags.html#Planner-Flags">Planner Flags</a>), and in addition may include one or more of
Chris@10 116 the following MPI-specific flags, which improve performance at the
Chris@10 117 cost of changing the output or input data formats.
Chris@10 118
Chris@10 119 <ul>
Chris@10 120 <li><a name="index-FFTW_005fMPI_005fSCRAMBLED_005fOUT-471"></a><a name="index-FFTW_005fMPI_005fSCRAMBLED_005fIN-472"></a><code>FFTW_MPI_SCRAMBLED_OUT</code>, <code>FFTW_MPI_SCRAMBLED_IN</code>: valid for
Chris@10 121 1d transforms only, these flags indicate that the output/input of the
Chris@10 122 transform are in an undocumented &ldquo;scrambled&rdquo; order. A forward
Chris@10 123 <code>FFTW_MPI_SCRAMBLED_OUT</code> transform can be inverted by a backward
Chris@10 124 <code>FFTW_MPI_SCRAMBLED_IN</code> (times the usual 1/<i>N</i> normalization).
Chris@10 125 See <a href="One_002ddimensional-distributions.html#One_002ddimensional-distributions">One-dimensional distributions</a>.
Chris@10 126
Chris@10 127 <li><a name="index-FFTW_005fMPI_005fTRANSPOSED_005fOUT-473"></a><a name="index-FFTW_005fMPI_005fTRANSPOSED_005fIN-474"></a><code>FFTW_MPI_TRANSPOSED_OUT</code>, <code>FFTW_MPI_TRANSPOSED_IN</code>: valid
Chris@10 128 for multidimensional (<code>rnk &gt; 1</code>) transforms only, these flags
Chris@10 129 specify that the output or input of an n<sub>0</sub>&nbsp;&times;&nbsp;n<sub>1</sub>&nbsp;&times;&nbsp;n<sub>2</sub>&nbsp;&times;&nbsp;&hellip;&nbsp;&times;&nbsp;n<sub>d-1</sub> transform is
Chris@10 130 transposed to n<sub>1</sub>&nbsp;&times;&nbsp;n<sub>0</sub>&nbsp;&times;&nbsp;n<sub>2</sub>&nbsp;&times;&hellip;&times;&nbsp;n<sub>d-1</sub>. See <a href="Transposed-distributions.html#Transposed-distributions">Transposed distributions</a>.
Chris@10 131
Chris@10 132 </ul>
Chris@10 133
Chris@10 134 <h5 class="subsubheading">Real-data MPI DFTs</h5>
Chris@10 135
Chris@10 136 <p><a name="index-r2c-475"></a>Plans for real-input/output (r2c/c2r) DFTs (see <a href="Multi_002ddimensional-MPI-DFTs-of-Real-Data.html#Multi_002ddimensional-MPI-DFTs-of-Real-Data">Multi-dimensional MPI DFTs of Real Data</a>) are created by:
Chris@10 137
Chris@10 138 <p><a name="index-fftw_005fmpi_005fplan_005fdft_005fr2c_005f2d-476"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005fr2c_005f2d-477"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005fr2c_005f3d-478"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005fr2c-479"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005fc2r_005f2d-480"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005fc2r_005f2d-481"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005fc2r_005f3d-482"></a><a name="index-fftw_005fmpi_005fplan_005fdft_005fc2r-483"></a>
Chris@10 139 <pre class="example"> fftw_plan fftw_mpi_plan_dft_r2c_2d(ptrdiff_t n0, ptrdiff_t n1,
Chris@10 140 double *in, fftw_complex *out,
Chris@10 141 MPI_Comm comm, unsigned flags);
Chris@10 142 fftw_plan fftw_mpi_plan_dft_r2c_2d(ptrdiff_t n0, ptrdiff_t n1,
Chris@10 143 double *in, fftw_complex *out,
Chris@10 144 MPI_Comm comm, unsigned flags);
Chris@10 145 fftw_plan fftw_mpi_plan_dft_r2c_3d(ptrdiff_t n0, ptrdiff_t n1, ptrdiff_t n2,
Chris@10 146 double *in, fftw_complex *out,
Chris@10 147 MPI_Comm comm, unsigned flags);
Chris@10 148 fftw_plan fftw_mpi_plan_dft_r2c(int rnk, const ptrdiff_t *n,
Chris@10 149 double *in, fftw_complex *out,
Chris@10 150 MPI_Comm comm, unsigned flags);
Chris@10 151 fftw_plan fftw_mpi_plan_dft_c2r_2d(ptrdiff_t n0, ptrdiff_t n1,
Chris@10 152 fftw_complex *in, double *out,
Chris@10 153 MPI_Comm comm, unsigned flags);
Chris@10 154 fftw_plan fftw_mpi_plan_dft_c2r_2d(ptrdiff_t n0, ptrdiff_t n1,
Chris@10 155 fftw_complex *in, double *out,
Chris@10 156 MPI_Comm comm, unsigned flags);
Chris@10 157 fftw_plan fftw_mpi_plan_dft_c2r_3d(ptrdiff_t n0, ptrdiff_t n1, ptrdiff_t n2,
Chris@10 158 fftw_complex *in, double *out,
Chris@10 159 MPI_Comm comm, unsigned flags);
Chris@10 160 fftw_plan fftw_mpi_plan_dft_c2r(int rnk, const ptrdiff_t *n,
Chris@10 161 fftw_complex *in, double *out,
Chris@10 162 MPI_Comm comm, unsigned flags);
Chris@10 163 </pre>
Chris@10 164 <p>Similar to the serial interface (see <a href="Real_002ddata-DFTs.html#Real_002ddata-DFTs">Real-data DFTs</a>), these
Chris@10 165 transform logically n<sub>0</sub>&nbsp;&times;&nbsp;n<sub>1</sub>&nbsp;&times;&nbsp;n<sub>2</sub>&nbsp;&times;&nbsp;&hellip;&nbsp;&times;&nbsp;n<sub>d-1</sub> real data to/from n<sub>0</sub>&nbsp;&times;&nbsp;n<sub>1</sub>&nbsp;&times;&nbsp;n<sub>2</sub>&nbsp;&times;&nbsp;&hellip;&nbsp;&times;&nbsp;(n<sub>d-1</sub>/2 + 1) complex
Chris@10 166 data, representing the non-redundant half of the conjugate-symmetry
Chris@10 167 output of a real-input DFT (see <a href="Multi_002ddimensional-Transforms.html#Multi_002ddimensional-Transforms">Multi-dimensional Transforms</a>).
Chris@10 168 However, the real array must be stored within a padded n<sub>0</sub>&nbsp;&times;&nbsp;n<sub>1</sub>&nbsp;&times;&nbsp;n<sub>2</sub>&nbsp;&times;&nbsp;&hellip;&nbsp;&times;&nbsp;[2&nbsp;(n<sub>d-1</sub>/2 + 1)]
Chris@10 169
Chris@10 170 <p>array (much like the in-place serial r2c transforms, but here for
Chris@10 171 out-of-place transforms as well). Currently, only multi-dimensional
Chris@10 172 (<code>rnk &gt; 1</code>) r2c/c2r transforms are supported (requesting a plan
Chris@10 173 for <code>rnk = 1</code> will yield <code>NULL</code>). As explained above
Chris@10 174 (see <a href="Multi_002ddimensional-MPI-DFTs-of-Real-Data.html#Multi_002ddimensional-MPI-DFTs-of-Real-Data">Multi-dimensional MPI DFTs of Real Data</a>), the data
Chris@10 175 distribution of both the real and complex arrays is given by the
Chris@10 176 &lsquo;<samp><span class="samp">local_size</span></samp>&rsquo; function called for the dimensions of the
Chris@10 177 <em>complex</em> array. Similar to the other planning functions, the
Chris@10 178 input and output arrays are overwritten when the plan is created
Chris@10 179 except in <code>FFTW_ESTIMATE</code> mode.
Chris@10 180
Chris@10 181 <p>As for the complex DFTs above, there is an advance interface that
Chris@10 182 allows you to manually specify block sizes and to transform contiguous
Chris@10 183 <code>howmany</code>-tuples of real/complex numbers:
Chris@10 184
Chris@10 185 <p><a name="index-fftw_005fmpi_005fplan_005fmany_005fdft_005fr2c-484"></a><a name="index-fftw_005fmpi_005fplan_005fmany_005fdft_005fc2r-485"></a>
Chris@10 186 <pre class="example"> fftw_plan fftw_mpi_plan_many_dft_r2c
Chris@10 187 (int rnk, const ptrdiff_t *n, ptrdiff_t howmany,
Chris@10 188 ptrdiff_t iblock, ptrdiff_t oblock,
Chris@10 189 double *in, fftw_complex *out,
Chris@10 190 MPI_Comm comm, unsigned flags);
Chris@10 191 fftw_plan fftw_mpi_plan_many_dft_c2r
Chris@10 192 (int rnk, const ptrdiff_t *n, ptrdiff_t howmany,
Chris@10 193 ptrdiff_t iblock, ptrdiff_t oblock,
Chris@10 194 fftw_complex *in, double *out,
Chris@10 195 MPI_Comm comm, unsigned flags);
Chris@10 196 </pre>
Chris@10 197 <h5 class="subsubheading">MPI r2r transforms</h5>
Chris@10 198
Chris@10 199 <p><a name="index-r2r-486"></a>There are corresponding plan-creation routines for r2r
Chris@10 200 transforms (see <a href="More-DFTs-of-Real-Data.html#More-DFTs-of-Real-Data">More DFTs of Real Data</a>), currently supporting
Chris@10 201 multidimensional (<code>rnk &gt; 1</code>) transforms only (<code>rnk = 1</code> will
Chris@10 202 yield a <code>NULL</code> plan):
Chris@10 203
Chris@10 204 <pre class="example"> fftw_plan fftw_mpi_plan_r2r_2d(ptrdiff_t n0, ptrdiff_t n1,
Chris@10 205 double *in, double *out,
Chris@10 206 MPI_Comm comm,
Chris@10 207 fftw_r2r_kind kind0, fftw_r2r_kind kind1,
Chris@10 208 unsigned flags);
Chris@10 209 fftw_plan fftw_mpi_plan_r2r_3d(ptrdiff_t n0, ptrdiff_t n1, ptrdiff_t n2,
Chris@10 210 double *in, double *out,
Chris@10 211 MPI_Comm comm,
Chris@10 212 fftw_r2r_kind kind0, fftw_r2r_kind kind1, fftw_r2r_kind kind2,
Chris@10 213 unsigned flags);
Chris@10 214 fftw_plan fftw_mpi_plan_r2r(int rnk, const ptrdiff_t *n,
Chris@10 215 double *in, double *out,
Chris@10 216 MPI_Comm comm, const fftw_r2r_kind *kind,
Chris@10 217 unsigned flags);
Chris@10 218 fftw_plan fftw_mpi_plan_many_r2r(int rnk, const ptrdiff_t *n,
Chris@10 219 ptrdiff_t iblock, ptrdiff_t oblock,
Chris@10 220 double *in, double *out,
Chris@10 221 MPI_Comm comm, const fftw_r2r_kind *kind,
Chris@10 222 unsigned flags);
Chris@10 223 </pre>
Chris@10 224 <p>The parameters are much the same as for the complex DFTs above, except
Chris@10 225 that the arrays are of real numbers (and hence the outputs of the
Chris@10 226 &lsquo;<samp><span class="samp">local_size</span></samp>&rsquo; data-distribution functions should be interpreted as
Chris@10 227 counts of real rather than complex numbers). Also, the <code>kind</code>
Chris@10 228 parameters specify the r2r kinds along each dimension as for the
Chris@10 229 serial interface (see <a href="Real_002dto_002dReal-Transform-Kinds.html#Real_002dto_002dReal-Transform-Kinds">Real-to-Real Transform Kinds</a>). See <a href="Other-Multi_002ddimensional-Real_002ddata-MPI-Transforms.html#Other-Multi_002ddimensional-Real_002ddata-MPI-Transforms">Other Multi-dimensional Real-data MPI Transforms</a>.
Chris@10 230
Chris@10 231 <h5 class="subsubheading">MPI transposition</h5>
Chris@10 232
Chris@10 233 <p><a name="index-transpose-487"></a>
Chris@10 234 FFTW also provides routines to plan a transpose of a distributed
Chris@10 235 <code>n0</code> by <code>n1</code> array of real numbers, or an array of
Chris@10 236 <code>howmany</code>-tuples of real numbers with specified block sizes
Chris@10 237 (see <a href="FFTW-MPI-Transposes.html#FFTW-MPI-Transposes">FFTW MPI Transposes</a>):
Chris@10 238
Chris@10 239 <p><a name="index-fftw_005fmpi_005fplan_005ftranspose-488"></a><a name="index-fftw_005fmpi_005fplan_005fmany_005ftranspose-489"></a>
Chris@10 240 <pre class="example"> fftw_plan fftw_mpi_plan_transpose(ptrdiff_t n0, ptrdiff_t n1,
Chris@10 241 double *in, double *out,
Chris@10 242 MPI_Comm comm, unsigned flags);
Chris@10 243 fftw_plan fftw_mpi_plan_many_transpose
Chris@10 244 (ptrdiff_t n0, ptrdiff_t n1, ptrdiff_t howmany,
Chris@10 245 ptrdiff_t block0, ptrdiff_t block1,
Chris@10 246 double *in, double *out, MPI_Comm comm, unsigned flags);
Chris@10 247 </pre>
Chris@10 248 <p><a name="index-new_002darray-execution-490"></a><a name="index-fftw_005fmpi_005fexecute_005fr2r-491"></a>These plans are used with the <code>fftw_mpi_execute_r2r</code> new-array
Chris@10 249 execute function (see <a href="Using-MPI-Plans.html#Using-MPI-Plans">Using MPI Plans</a>), since they count as (rank
Chris@10 250 zero) r2r plans from FFTW's perspective.
Chris@10 251
Chris@10 252 </body></html>
Chris@10 253