cannam@95
|
1 .\"
|
cannam@95
|
2 .\" Copyright (c) 2003, 2007-11 Matteo Frigo
|
cannam@95
|
3 .\" Copyright (c) 2003, 2007-11 Massachusetts Institute of Technology
|
cannam@95
|
4 .\"
|
cannam@95
|
5 .\" This program is free software; you can redistribute it and/or modify
|
cannam@95
|
6 .\" it under the terms of the GNU General Public License as published by
|
cannam@95
|
7 .\" the Free Software Foundation; either version 2 of the License, or
|
cannam@95
|
8 .\" (at your option) any later version.
|
cannam@95
|
9 .\"
|
cannam@95
|
10 .\" This program is distributed in the hope that it will be useful,
|
cannam@95
|
11 .\" but WITHOUT ANY WARRANTY; without even the implied warranty of
|
cannam@95
|
12 .\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
cannam@95
|
13 .\" GNU General Public License for more details.
|
cannam@95
|
14 .\"
|
cannam@95
|
15 .\" You should have received a copy of the GNU General Public License
|
cannam@95
|
16 .\" along with this program; if not, write to the Free Software
|
cannam@95
|
17 .\" Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
cannam@95
|
18 .\"
|
cannam@95
|
19 .TH FFTW-WISDOM 1 "February, 2003" "fftw" "fftw"
|
cannam@95
|
20 .SH NAME
|
cannam@95
|
21 fftwf-wisdom \- create wisdom (pre-optimized FFTs)
|
cannam@95
|
22 .SH SYNOPSIS
|
cannam@95
|
23 .B fftwf-wisdom
|
cannam@95
|
24 [\fIOPTION\fR]... [\fISIZE\fR]...
|
cannam@95
|
25 .SH DESCRIPTION
|
cannam@95
|
26 .PP
|
cannam@95
|
27 ." Add any additional description here
|
cannam@95
|
28 .I fftwf-wisdom
|
cannam@95
|
29 is a utility to generate FFTW
|
cannam@95
|
30 .B wisdom
|
cannam@95
|
31 files, which contain saved information about how to optimally compute
|
cannam@95
|
32 (Fourier) transforms of various sizes. FFTW is a free library to
|
cannam@95
|
33 compute discrete Fourier transforms in one or more dimensions, for
|
cannam@95
|
34 arbitrary sizes, and of both real and complex data, among other
|
cannam@95
|
35 related operations. More information on FFTW can be found at the FFTW
|
cannam@95
|
36 home page:
|
cannam@95
|
37 .I http://www.fftw.org
|
cannam@95
|
38
|
cannam@95
|
39 Programs using FFTW can be written to load wisdom from an arbitrary file,
|
cannam@95
|
40 string, or other source. Moreover, it is likely that many FFTW-using
|
cannam@95
|
41 programs will load the \fBsystem wisdom\fR file, which is stored in
|
cannam@95
|
42 .I /etc/fftw/wisdomf
|
cannam@95
|
43 by default.
|
cannam@95
|
44 .I fftwf-wisdom
|
cannam@95
|
45 can be used to create or add to such wisdom files. In its most
|
cannam@95
|
46 typical usage, the wisdom file can be created to pre-plan a canonical
|
cannam@95
|
47 set of sizes (see below) via:
|
cannam@95
|
48
|
cannam@95
|
49 .ce
|
cannam@95
|
50 fftwf-wisdom -v -c -o wisdomf
|
cannam@95
|
51
|
cannam@95
|
52 (this will take many hours, which can be limited by the
|
cannam@95
|
53 .B -t
|
cannam@95
|
54 option) and the output
|
cannam@95
|
55 .I wisdomf
|
cannam@95
|
56 file can then be copied (as root) to
|
cannam@95
|
57 .I /etc/fftw/
|
cannam@95
|
58 or whatever.
|
cannam@95
|
59
|
cannam@95
|
60 The
|
cannam@95
|
61 .I fftwf-wisdom
|
cannam@95
|
62 program normally writes the wisdom directly to standard output, but this
|
cannam@95
|
63 can be changed via the
|
cannam@95
|
64 .B -o
|
cannam@95
|
65 option, as in the example above.
|
cannam@95
|
66
|
cannam@95
|
67 If the system wisdom file
|
cannam@95
|
68 .I /etc/fftw/wisdomf
|
cannam@95
|
69 already exists, then
|
cannam@95
|
70 .I fftwf-wisdom
|
cannam@95
|
71 reads this existing wisdom (unless the
|
cannam@95
|
72 .B -n
|
cannam@95
|
73 option is specified) and outputs both the old wisdom and any
|
cannam@95
|
74 newly created wisdom. In this way, it can be used to add new transform
|
cannam@95
|
75 sizes to the existing system wisdom (or other wisdom file, with the
|
cannam@95
|
76 .B -w
|
cannam@95
|
77 option).
|
cannam@95
|
78 .SH SPECIFYING SIZES
|
cannam@95
|
79 Although a canonical set of sizes to optimize is specified by the
|
cannam@95
|
80 .B -c
|
cannam@95
|
81 option, the user can also specify zero or more non-canonical transform
|
cannam@95
|
82 sizes and types to optimize, via the
|
cannam@95
|
83 .I SIZE
|
cannam@95
|
84 arguments following the option flags. Alternatively, the sizes to
|
cannam@95
|
85 optimize can be read from standard input (whitespace-separated), if a
|
cannam@95
|
86 .I SIZE
|
cannam@95
|
87 argument of "-" is supplied.
|
cannam@95
|
88
|
cannam@95
|
89 Sizes are specified by the syntax:
|
cannam@95
|
90
|
cannam@95
|
91 .ce
|
cannam@95
|
92 <\fItype\fR><\fIinplace\fR><\fIdirection\fR><\fIgeometry\fR>
|
cannam@95
|
93
|
cannam@95
|
94 <\fItype\fR> is either \'c\' (complex), \'r\' (real, r2c/c2r), or
|
cannam@95
|
95 \'k\' (r2r, per-dimension kinds, specified in the geometry, below).
|
cannam@95
|
96
|
cannam@95
|
97 <\fIinplace\fR> is either \'i\' (in place) or \'o\' (out of place).
|
cannam@95
|
98
|
cannam@95
|
99 <\fIdirection\fR> is either \'f\' (forward) or \'b\' (backward). The
|
cannam@95
|
100 <\fIdirection\fR> should be omitted for \'k\' transforms, where it is
|
cannam@95
|
101 specified via the geometry instead.
|
cannam@95
|
102
|
cannam@95
|
103 <\fIgeometry\fR> is the size and dimensionality of the transform,
|
cannam@95
|
104 where different dimensions are separated by \'x\' (e.g. \'16x32\' for
|
cannam@95
|
105 a two-dimensional 16 by 32 transform). In the case of \'k\'
|
cannam@95
|
106 transforms, the size of each dimension is followed by a "type" string,
|
cannam@95
|
107 which can be one of f/b/h/e00/e01/e10/e11/o00/o01/o10/o11 for
|
cannam@95
|
108 R2HC/HC2R/DHT/REDFT00/.../RODFT11, respectively, as defined in the
|
cannam@95
|
109 FFTW manual.
|
cannam@95
|
110
|
cannam@95
|
111 For example, \'cif12x13x14\' is a three-dimensional 12 by 13 x 14
|
cannam@95
|
112 complex DFT operating in-place. \'rob65536\' is a one-dimensional
|
cannam@95
|
113 size-65536 out-of-place complex-to-real (backwards) transform
|
cannam@95
|
114 operating on Hermitian-symmetry input. \'ki10hx20e01\' is a
|
cannam@95
|
115 two-dimensional 10 by 20 r2r transform where the first dimension is a
|
cannam@95
|
116 DHT and the second dimension is an REDFT01 (DCT-III).
|
cannam@95
|
117
|
cannam@95
|
118 .SH OPTIONS
|
cannam@95
|
119 .TP
|
cannam@95
|
120 \fB\-h\fR, \fB\--help\fR
|
cannam@95
|
121 Display help on the command-line options and usage.
|
cannam@95
|
122 .TP
|
cannam@95
|
123 \fB\-V\fR, \fB\--version\fR
|
cannam@95
|
124 Print the version number and copyright information.
|
cannam@95
|
125 .TP
|
cannam@95
|
126 \fB\-v\fR, \fB\--verbose\fR
|
cannam@95
|
127 Verbose output. (You can specify this multiple times, or supply a numeric
|
cannam@95
|
128 argument greater than 1, to increase the verbosity level.) Note that the
|
cannam@95
|
129 verbose output will be mixed with the wisdom output (making it impossible
|
cannam@95
|
130 to import), unless you write the wisdom to a file via the
|
cannam@95
|
131 .B -o
|
cannam@95
|
132 option.
|
cannam@95
|
133 .TP
|
cannam@95
|
134 \fB\-c\fR, \fB\--canonical\fR
|
cannam@95
|
135 Optimize/pre-plan a canonical set of sizes: all powers of two and ten
|
cannam@95
|
136 up to 2^20 (1048576), including both real and complex, forward and
|
cannam@95
|
137 backwards, in-place and out-of-place transforms. Also includes two-
|
cannam@95
|
138 and three-dimensional transforms of equal-size dimensions
|
cannam@95
|
139 (e.g. 16x16x16).
|
cannam@95
|
140 .TP
|
cannam@95
|
141 \fB\-t\fR \fIhours\fR, \fB\--time-limit\fR=\fIhours\fR
|
cannam@95
|
142 Stop after a time of
|
cannam@95
|
143 .I hours
|
cannam@95
|
144 (hours) has elapsed, outputting accumulated wisdom. (The problems are planned
|
cannam@95
|
145 in increasing order of size.) Defaults to 0, indicating no time limit.
|
cannam@95
|
146 .TP
|
cannam@95
|
147 \fB\-o\fR \fIfile\fR, \fB\--output-file\fR=\fIfile\fR
|
cannam@95
|
148 Send wisdom output to
|
cannam@95
|
149 .I file
|
cannam@95
|
150 rather than to standard output (the default).
|
cannam@95
|
151 .TP
|
cannam@95
|
152 \fB\-m\fR, \fB\--measure\fR; \fB\-e\fR, \fB\--estimate\fR; \fB\-x\fR, \fB\--exhaustive\fR
|
cannam@95
|
153 Normally,
|
cannam@95
|
154 .I fftwf-wisdom
|
cannam@95
|
155 creates plans in FFTW_PATIENT mode, but with these options you can instead
|
cannam@95
|
156 use FFTW_MEASURE, FFTW_ESTIMATE, or FFTW_EXHAUSTIVE modes, respectively,
|
cannam@95
|
157 as described in more detail by the FFTW manual.
|
cannam@95
|
158
|
cannam@95
|
159 Note that wisdom is tagged with the planning patience level, and a
|
cannam@95
|
160 single file can mix different levels of wisdom (e.g. you can mostly
|
cannam@95
|
161 use the patient default, but plan a few sizes that you especially care
|
cannam@95
|
162 about in
|
cannam@95
|
163 .B --exhaustive
|
cannam@95
|
164 mode).
|
cannam@95
|
165 .TP
|
cannam@95
|
166 \fB\-n\fR, \fB\--no-system-wisdom\fR
|
cannam@95
|
167 Do not import the system wisdom from
|
cannam@95
|
168 .I /etc/fftw/wisdomf
|
cannam@95
|
169 (which is normally read by default).
|
cannam@95
|
170 .TP
|
cannam@95
|
171 \fB\-w\fR \fIfile\fR, \fB\--wisdom-file\fR=\fIfile\fR
|
cannam@95
|
172 Import wisdom from
|
cannam@95
|
173 .I file
|
cannam@95
|
174 (in addition to the system wisdom, unless
|
cannam@95
|
175 .B -n
|
cannam@95
|
176 is specified). Multiple wisdom files can be read via multiple
|
cannam@95
|
177 .B -w
|
cannam@95
|
178 options. If
|
cannam@95
|
179 .I file
|
cannam@95
|
180 is "-", then read wisdom from standard input.
|
cannam@95
|
181 .SH BUGS
|
cannam@95
|
182 Send bug reports to fftw@fftw.org.
|
cannam@95
|
183 .SH AUTHORS
|
cannam@95
|
184 Written by Steven G. Johnson and Matteo Frigo.
|
cannam@95
|
185
|
cannam@95
|
186 Copyright (c) 2003, 2007-11 Matteo Frigo
|
cannam@95
|
187 .br
|
cannam@95
|
188 Copyright (c) 2003, 2007-11 Massachusetts Institute of Technology
|
cannam@95
|
189 .SH "SEE ALSO"
|
cannam@95
|
190 fftw-wisdom-to-conf(1)
|