fftw_5.html
来自「FFTW, a collection of fast C routines to」· HTML 代码 · 共 344 行
HTML
344 行
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"><HTML><HEAD><!-- This HTML file has been created by texi2html 1.52 from fftw.texi on 24 March 2003 --><TITLE>FFTW - Calling FFTW from Fortran</TITLE></HEAD><BODY TEXT="#000000" BGCOLOR="#FFFFFF">Go to the <A HREF="fftw_1.html">first</A>, <A HREF="fftw_4.html">previous</A>, <A HREF="fftw_6.html">next</A>, <A HREF="fftw_10.html">last</A> section, <A HREF="fftw_toc.html">table of contents</A>.<P><HR><P><H1><A NAME="SEC62">Calling FFTW from Fortran</A></H1><P><A NAME="IDX290"></A>The standard FFTW libraries include special wrapper functions that allowFortran programs to call FFTW subroutines. This chapter describes howthose functions may be employed to use FFTW from Fortran. We assumehere that the reader is already familiar with the usage of FFTW in C, asdescribed elsewhere in this manual.<P>In general, it is not possible to call C functions directly fromFortran, due to Fortran's inability to pass arguments by value and alsobecause Fortran compilers typically expect identifiers to be mangled<A NAME="IDX291"></A>somehow for linking. However, if C functions are written in a specialway, they <EM>are</EM> callable from Fortran, and we have employed thistechnique to create Fortran-callable "wrapper" functions around themain FFTW routines. These wrapper functions are included in the FFTWlibraries by default, unless a Fortran compiler isn't found on yoursystem or <CODE>--disable-fortran</CODE> is included in the <CODE>configure</CODE>flags.<P>As a result, calling FFTW from Fortran requires little more thanappending <SAMP>`<CODE>_f77</CODE>'</SAMP> to the function names and then linkingnormally with the FFTW libraries. There are a few wrinkles, however, aswe shall discuss below.<H2><A NAME="SEC63">Wrapper Routines</A></H2><P>All of the uniprocessor and multi-threaded transform routines haveFortran-callable wrappers, except for the wisdom import/export functions(since it is not possible to exchange string and file arguments portablywith Fortran) and the specific planner routines (see Section <A HREF="fftw_3.html#SEC20">Discussion on Specific Plans</A>). The name of the wrapper routine is the same as thatof the corresponding C routine, but with<CODE>fftw/fftwnd/rfftw/rfftwnd</CODE> replaced by<CODE>fftw_f77/fftwnd_f77/rfftw_f77/rfftwnd_f77</CODE>. For example, inFortran, instead of calling <CODE>fftw_one</CODE> you would call<CODE>fftw_f77_one</CODE>.<A NAME="DOCF8" HREF="fftw_foot.html#FOOT8">(8)</A><A NAME="IDX292"></A>For the most part, all of the arguments to the functions are the same,with the following exceptions:<UL><LI><CODE>plan</CODE> variables (what would be of type <CODE>fftw_plan</CODE>,<CODE>rfftwnd_plan</CODE>, etcetera, in C), must be declared as a type that isthe same size as a pointer (address) on your machine. (Fortran has nogeneric pointer type.) The Fortran <CODE>integer</CODE> type is usually thesame size as a pointer, but you need to be wary (especially on 64-bitmachines). (You could also use <CODE>integer*4</CODE> on a 32-bit machine and<CODE>integer*8</CODE> on a 64-bit machine.) Ugh. (<CODE>g77</CODE> has a specialtype, <CODE>integer(kind=7)</CODE>, that is defined to be the same size as apointer.)<LI>Any function that returns a value (e.g. <CODE>fftw_create_plan</CODE>) isconverted into a subroutine. The return value is converted into anadditional (first) parameter of the wrapper subroutine. (The reason forthis is that some Fortran implementations seem to have trouble with Cfunction return values.)<LI><A NAME="IDX293"></A>When performing one-dimensional <CODE>FFTW_IN_PLACE</CODE> transforms, youdon't have the option of passing <CODE>NULL</CODE> for the <CODE>out</CODE> argument(since there is no way to pass <CODE>NULL</CODE> from Fortran). Therefore,when performing such transforms, you <EM>must</EM> allocate and pass acontiguous scratch array of the same size as the transform. Note thatfor in-place multi-dimensional (<CODE>(r)fftwnd</CODE>) transforms, the<CODE>out</CODE> argument is ignored, so you can pass anything for thatparameter.<LI><A NAME="IDX294"></A>The wrapper routines expect multi-dimensional arrays to be incolumn-major order, which is the ordinary format of Fortran arrays.They do this transparently and costlessly simply by reversing the orderof the dimensions passed to FFTW, but this has one important consequencefor multi-dimensional real-complex transforms, discussed below.</UL><P><A NAME="IDX295"></A>In general, you should take care to use Fortran data types thatcorrespond to (i.e. are the same size as) the C types used by FFTW. Ifyour C and Fortran compilers are made by the same vendor, thecorrespondence is usually straightforward (i.e. <CODE>integer</CODE>corresponds to <CODE>int</CODE>, <CODE>real</CODE> corresponds to <CODE>float</CODE>,etcetera). Such simple correspondences are assumed in the examplesbelow. The examples also assume that FFTW was compiled indouble precision (the default).<H2><A NAME="SEC64">FFTW Constants in Fortran</A></H2><P>When creating plans in FFTW, a number of constants are used to specifyoptions, such as <CODE>FFTW_FORWARD</CODE> or <CODE>FFTW_USE_WISDOM</CODE>. Thesame constants must be used with the wrapper routines, but of course theC header files where the constants are defined can't be incorporateddirectly into Fortran code.<P>Instead, we have placed Fortran equivalents of the FFTW constantdefinitions in the file <CODE>fortran/fftw_f77.i</CODE> of the FFTW package.If your Fortran compiler supports a preprocessor, you can use that toincorporate this file into your code whenever you need to call FFTW.Otherwise, you will have to paste the constant definitions in directly.They are:<PRE> integer FFTW_FORWARD,FFTW_BACKWARD parameter (FFTW_FORWARD=-1,FFTW_BACKWARD=1) integer FFTW_REAL_TO_COMPLEX,FFTW_COMPLEX_TO_REAL parameter (FFTW_REAL_TO_COMPLEX=-1,FFTW_COMPLEX_TO_REAL=1) integer FFTW_ESTIMATE,FFTW_MEASURE parameter (FFTW_ESTIMATE=0,FFTW_MEASURE=1) integer FFTW_OUT_OF_PLACE,FFTW_IN_PLACE,FFTW_USE_WISDOM parameter (FFTW_OUT_OF_PLACE=0) parameter (FFTW_IN_PLACE=8,FFTW_USE_WISDOM=16) integer FFTW_THREADSAFE parameter (FFTW_THREADSAFE=128)</PRE><P><A NAME="IDX296"></A>In C, you combine different flags (like <CODE>FFTW_USE_WISDOM</CODE> and<CODE>FFTW_MEASURE</CODE>) using the <SAMP>`<CODE>|</CODE>'</SAMP> operator; in Fortran youshould just use <SAMP>`<CODE>+</CODE>'</SAMP>.<H2><A NAME="SEC65">Fortran Examples</A></H2><P>In C you might have something like the following to transform aone-dimensional complex array:<PRE> fftw_complex in[N], *out[N]; fftw_plan plan; plan = fftw_create_plan(N,FFTW_FORWARD,FFTW_ESTIMATE); fftw_one(plan,in,out); fftw_destroy_plan(plan);</PRE><P>In Fortran, you use the following to accomplish the same thing:<PRE> double complex in, out dimension in(N), out(N) integer plan call fftw_f77_create_plan(plan,N,FFTW_FORWARD,FFTW_ESTIMATE) call fftw_f77_one(plan,in,out) call fftw_f77_destroy_plan(plan)</PRE><P><A NAME="IDX297"></A><A NAME="IDX298"></A><A NAME="IDX299"></A><P>Notice how all routines are called as Fortran subroutines, and the planis returned via the first argument to <CODE>fftw_f77_create_plan</CODE>.<EM>Important:</EM> these examples assume that <CODE>integer</CODE> is the samesize as a pointer, and may need modification on a 64-bit machine.See Section <A HREF="fftw_5.html#SEC63">Wrapper Routines</A>, above. To do the same thing, but using 8threads in parallel (see Section <A HREF="fftw_4.html#SEC48">Multi-threaded FFTW</A>), you would simplyreplace the call to <CODE>fftw_f77_one</CODE> with:<PRE> call fftw_f77_threads_one(8,plan,in,out)</PRE><P><A NAME="IDX300"></A><P>To transform a three-dimensional array in-place with C, you might do:<PRE> fftw_complex arr[L][M][N]; fftwnd_plan plan; int n[3] = {L,M,N}; plan = fftwnd_create_plan(3,n,FFTW_FORWARD, FFTW_ESTIMATE | FFTW_IN_PLACE); fftwnd_one(plan, arr, 0); fftwnd_destroy_plan(plan);</PRE><P>In Fortran, you would use this instead:<PRE> double complex arr dimension arr(L,M,N) integer n dimension n(3) integer plan n(1) = L n(2) = M n(3) = N call fftwnd_f77_create_plan(plan,3,n,FFTW_FORWARD, + FFTW_ESTIMATE + FFTW_IN_PLACE) call fftwnd_f77_one(plan, arr, 0) call fftwnd_f77_destroy_plan(plan)</PRE><P><A NAME="IDX301"></A><A NAME="IDX302"></A><A NAME="IDX303"></A><P>Instead of calling <CODE>fftwnd_f77_create_plan(plan,3,n,...)</CODE>, we couldalso have called <CODE>fftw3d_f77_create_plan(plan,L,M,N,...)</CODE>.<A NAME="IDX304"></A><P>Note that we pass the array dimensions in the "natural" order; alsonote that the last argument to <CODE>fftwnd_f77</CODE> is ignored since thetransform is <CODE>FFTW_IN_PLACE</CODE>.<P>To transform a one-dimensional real array in Fortran, you might do:<PRE> double precision in, out dimension in(N), out(N) integer plan call rfftw_f77_create_plan(plan,N,FFTW_REAL_TO_COMPLEX, + FFTW_ESTIMATE) call rfftw_f77_one(plan,in,out) call rfftw_f77_destroy_plan(plan)</PRE><P><A NAME="IDX305"></A><A NAME="IDX306"></A><A NAME="IDX307"></A><P>To transform a two-dimensional real array, out of place, you might usethe following:<PRE> double precision in double complex out dimension in(M,N), out(M/2 + 1, N) integer plan call rfftw2d_f77_create_plan(plan,M,N,FFTW_REAL_TO_COMPLEX, + FFTW_ESTIMATE) call rfftwnd_f77_one_real_to_complex(plan, in, out) call rfftwnd_f77_destroy_plan(plan)</PRE><P><A NAME="IDX308"></A><A NAME="IDX309"></A><A NAME="IDX310"></A><P><B>Important:</B> Notice that it is the <EM>first</EM> dimension of thecomplex output array that is cut in half in Fortran, rather than thelast dimension as in C. This is a consequence of the wrapper routinesreversing the order of the array dimensions passed to FFTW so that theFortran program can use its ordinary column-major order.<A NAME="IDX311"></A><A NAME="IDX312"></A><P><HR><P>Go to the <A HREF="fftw_1.html">first</A>, <A HREF="fftw_4.html">previous</A>, <A HREF="fftw_6.html">next</A>, <A HREF="fftw_10.html">last</A> section, <A HREF="fftw_toc.html">table of contents</A>.</BODY></HTML>⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?