1GFORTRAN(1) GNU GFORTRAN(1)
2
6 gfortran - GNU Fortran 95 compiler
7
9 gfortran [-c⎪-S⎪-E]
10 [-g] [-pg] [-Olevel]
11 [-Wwarn...] [-pedantic]
12 [-Idir...] [-Ldir...]
13 [-Dmacro[=defn]...] [-Umacro]
14 [-foption...] [-mmachine-option...]
15 [-o outfile] infile...
16 Only the most useful options are listed here; see below for the remain‐
18 der.
19
21 The gfortran command supports all the options supported by the gcc com‐
22 mand. Only options specific to gfortran are documented here.
23 All gcc and gfortran options are accepted both by gfortran and by gcc
25 (as well as any other drivers built at the same time, such as g++),
26 since adding gfortran to the gcc distribution enables acceptance of
27 gfortran options by all of the relevant drivers.
28 In some cases, options have positive and negative forms; the negative
30 form of -ffoo would be -fno-foo. This manual documents only one of
31 these two forms, whichever one is not the default.
32
34 Here is a summary of all the options specific to GNU Fortran, grouped
35 by type. Explanations are in the following sections.
36 Fortran Language Options
38 -ffree-form -fno-fixed-form -fdollar-ok -fimplicit-none
39 -fmax-identifier-length -std=std -fd-lines-as-code
40 -fd-lines-as-comments -ffixed-line-length-n
41 -ffixed-line-length-none -ffree-line-length-n
42 -ffree-line-length-none -fdefault-double-8 -fdefault-integer-8
43 -fdefault-real-8 -fcray-pointer -fopenmp -frange-check
44 Warning Options
46 -fsyntax-only -pedantic -pedantic-errors -w -Wall -Waliasing
47 -Wampersand -Wconversion -Wimplicit-interface -Wnonstd-intrinsics
48 -Wsurprising -Wunderflow -Wunused-labels -Wline-truncation -W
49 Debugging Options
51 -fdump-parse-tree -ffpe-trap=list
52 Directory Options
54 -Idir -Mdir
55 Runtime Options
57 -fconvert=conversion -frecord-marker=length
58 Code Generation Options
60 -fno-automatic -ff2c -fno-underscoring -fsecond-underscore
61 -fbounds-check -fmax-stack-var-size=n -fpackderived
62 -frepack-arrays -fshort-enums
63 Options Controlling Fortran Dialect
65 The following options control the dialect of Fortran that the compiler
67 accepts:
68 -ffree-form
70 -ffixed-form
71 Specify the layout used by the source file. The free form layout
72 was introduced in Fortran 90. Fixed form was traditionally used in
73 older Fortran programs.
74 -fd-lines-as-code
76 -fd-lines-as-comments
77 Enables special treating for lines with d or D in fixed form
78 sources. If the -fd-lines-as-code option is given they are treated
79 as if the first column contained a blank. If the -fd-lines-as-com‐
80 ments option is given, they are treated as comment lines.
81 -fdefault-double-8
83 Set the "DOUBLE PRECISION" type to an 8 byte wide.
84 -fdefault-integer-8
86 Set the default integer and logical types to an 8 byte wide type.
87 Do nothing if this is already the default.
88 -fdefault-real-8
90 Set the default real type to an 8 byte wide type. Do nothing if
91 this is already the default.
92 -fdollar-ok
94 Allow $ as a valid character in a symbol name.
95 -fno-backslash
97 Compile switch to change the interpretation of a backslash from
98 "C"-style escape characters to a single backslash character.
99 -ffixed-line-length-n
101 Set column after which characters are ignored in typical fixed-form
102 lines in the source file, and through which spaces are assumed (as
103 if padded to that length) after the ends of short fixed-form lines.
104 Popular values for n include 72 (the standard and the default), 80
106 (card image), and 132 (corresponds to "extended-source" options in
107 some popular compilers). n may be none, meaning that the entire
108 line is meaningful and that continued character constants never
109 have implicit spaces appended to them to fill out the line.
110 -ffixed-line-length-0 means the same thing as
111 -ffixed-line-length-none.
112 -ffree-line-length-n
114 Set column after which characters are ignored in typical free-form
115 lines in the source file. For free-form, the default value is 132.
116 n may be none, meaning that the entire line is meaningful.
117 -ffree-line-length-0 means the same thing as
118 -ffree-line-length-none.
119 -fmax-identifier-length=n
121 Specify the maximum allowed identifier length. Typical values are
122 31 (Fortran 95) and 63 (Fortran 200x).
123 -fimplicit-none
125 Specify that no implicit typing is allowed, unless overridden by
126 explicit IMPLICIT statements. This is the equivalent of adding
127 implicit none to the start of every procedure.
128 -fcray-pointer
130 Enables the Cray pointer extension, which provides a C-like
131 pointer.
132 -fopenmp
134 Enables handling of OpenMP "!$omp" directives in free form and
135 "c$omp", *$omp and "!$omp" directives in fixed form, enables "!$"
136 conditional compilation sentinels in free form and "c$", "*$" and
137 "!$" sentinels in fixed form and when linking arranges for the
138 OpenMP runtime library to be linked in.
139 -frange-check
141 Enable range checking on results of simplification of constant
142 expressions during compilation. For example, by default, gfortran
143 will give an overflow error at compile time when simplifying "a =
144 EXP(1000)". With -fno-range-check, no error will be given and the
145 variable "a" will be assigned the value "+Infinity".
146 -std=std
148 Conform to the specified standard. Allowed values for std are gnu,
149 f95, f2003 and legacy.
150 Options to Request or Suppress Warnings
152 Warnings are diagnostic messages that report constructions which are
154 not inherently erroneous but which are risky or suggest there might
155 have been an error.
156 You can request many specific warnings with options beginning -W, for
158 example -Wimplicit to request warnings on implicit declarations. Each
159 of these specific warning options also has a negative form beginning
160 -Wno- to turn off warnings; for example, -Wno-implicit. This manual
161 lists only one of the two forms, whichever is not the default.
162 These options control the amount and kinds of warnings produced by GNU
164 Fortran:
165 -fsyntax-only
167 Check the code for syntax errors, but don't do anything beyond
168 that.
169 -pedantic
171 Issue warnings for uses of extensions to FORTRAN 95. -pedantic
172 also applies to C-language constructs where they occur in GNU For‐
173 tran source files, such as use of \e in a character constant within
174 a directive like #include.
175 Valid FORTRAN 95 programs should compile properly with or without
177 this option. However, without this option, certain GNU extensions
178 and traditional Fortran features are supported as well. With this
179 option, many of them are rejected.
180 Some users try to use -pedantic to check programs for conformance.
182 They soon find that it does not do quite what they want---it finds
183 some nonstandard practices, but not all. However, improvements to
184 gfortran in this area are welcome.
185 This should be used in conjunction with -std=std.
187 -pedantic-errors
189 Like -pedantic, except that errors are produced rather than warn‐
190 ings.
191 -w Inhibit all warning messages.
193 -Wall
195 Enables commonly used warning options pertaining to usage that we
196 recommend avoiding and that we believe are easy to avoid. This cur‐
197 rently includes -Wunused-labels, -Waliasing, -Wampersand, -Wsur‐
198 prising, -Wnonstd-intrinsic, and -Wline-truncation.
199 -Waliasing
201 Warn about possible aliasing of dummy arguments. Specifically, it
202 warns if the same actual argument is associated with a dummy argu‐
203 ment with "intent(in)" and a dummy argument with "intent(out)" in a
204 call with an explicit interface.
205 The following example will trigger the warning.
207 interface
209 subroutine bar(a,b)
210 integer, intent(in) :: a
211 integer, intent(out) :: b
212 end subroutine
213 end interface
214 integer :: a
215 call bar(a,a)
217 -Wampersand
219 Warn about missing ampersand in continued character literals. The
220 warning is given with -Wampersand, -pedantic, and -std=f95. Note:
221 With no ampersand given in a continued character literal, gfortran
222 assumes continuation at the first non-comment, non-whitespace char‐
223 acter.
224 -Wconversion
226 Warn about implicit conversions between different types.
227 -Wimplicit-interface
229 Warn about when procedure are called without an explicit interface.
230 Note this only checks that an explicit interface is present. It
231 does not check that the declared interfaces are consistent across
232 program units.
233 -Wnonstd-intrinsic
235 Warn if the user tries to use an intrinsic that does not belong to
236 the standard the user has chosen via the -std option.
237 -Wsurprising
239 Produce a warning when "suspicious" code constructs are encoun‐
240 tered. While technically legal these usually indicate that an
241 error has been made.
242 This currently produces a warning under the following circum‐
244 stances:
245 * An INTEGER SELECT construct has a CASE that can never be
247 matched as its lower value is greater than its upper value.
248 * A LOGICAL SELECT construct has three CASE statements.
250 -Wunderflow
252 Produce a warning when numerical constant expressions are encoun‐
253 tered, which yield an UNDERFLOW during compilation.
254 -Wunused-labels
256 Warn whenever a label is defined but never referenced.
257 -Werror
259 Turns all warnings into errors.
260 -W Turns on "extra warnings" and, if optimization is specified via -O,
262 the -Wuninitialized option. (This might change in future versions
263 of gfortran
264 Some of these have no effect when compiling programs written in For‐
266 tran.
267 Options for Debugging Your Program or GNU Fortran
269 GNU Fortran has various special options that are used for debugging
271 either your program or gfortran
272 -fdump-parse-tree
274 Output the internal parse tree before starting code generation.
275 Only really useful for debugging gfortran itself.
276 -ffpe-trap=list
278 Specify a list of IEEE exceptions when a Floating Point Exception
279 (FPE) should be raised. On most systems, this will result in a
280 SIGFPE signal being sent and the program being interrupted, produc‐
281 ing a core file useful for debugging. list is a (possibly empty)
282 comma-separated list of the following IEEE exceptions: invalid
283 (invalid floating point operation, such as "sqrt(-1.0)"), zero
284 (division by zero), overflow (overflow in a floating point opera‐
285 tion), underflow (underflow in a floating point operation), preci‐
286 sion (loss of precision during operation) and denormal (operation
287 produced a denormal denormal value).
288 Options for Directory Search
290 These options affect how gfortran searches for files specified by the
292 "INCLUDE" directive and where it searches for previously compiled mod‐
293 ules.
294 It also affects the search paths used by cpp when used to preprocess
296 Fortran source.
297 -Idir
299 These affect interpretation of the "INCLUDE" directive (as well as
300 of the "#include" directive of the cpp preprocessor).
301 Also note that the general behavior of -I and "INCLUDE" is pretty
303 much the same as of -I with "#include" in the cpp preprocessor,
304 with regard to looking for header.gcc files and other such things.
305 This path is also used to search for .mod files when previously
307 compiled modules are required by a "USE" statement.
308 -Mdir
310 -Jdir
311 This option specifies where to put .mod files for compiled modules.
312 It is also added to the list of directories to searched by an "USE"
313 statement.
314 The default is the current directory.
316 -J is an alias for -M to avoid conflicts with existing GCC options.
318 Influencing runtime behavior
320 These options affect the runtime behavior of gfortran.
322 -fconvert=conversion
324 Specify the representation of data for unformatted files. Valid
325 values for conversion are: native, the default; swap, swap between
326 big- and little-endian; big-endian, use big-endian representation
327 for unformatted files; little-endian, use little-endian representa‐
328 tion for unformatted files.
329 This option has an effect only when used in the main program. The
331 "CONVERT" specifier and the GFORTRAN_CONVERT_UNIT environment vari‐
332 able override the default specified by -fconvert.
333 -frecord-marker=length
335 Specify the length of record markers for unformatted files. Valid
336 values for length are 4 and 8. Default is whatever "off_t" is
337 specified to be on that particular system. Note that specifying
338 length as 4 limits the record length of unformatted files to 2 GB.
339 This option does not extend the maximum possible record length on
340 systems where "off_t" is a four_byte quantity.
341 Options for Code Generation Conventions
343 These machine-independent options control the interface conventions
345 used in code generation.
346 Most of them have both positive and negative forms; the negative form
348 of -ffoo would be -fno-foo. In the table below, only one of the forms
349 is listed---the one which is not the default. You can figure out the
350 other form by either removing no- or adding it.
351 -fno-automatic
353 Treat each program unit as if the "SAVE" statement was specified
354 for every local variable and array referenced in it. Does not
355 affect common blocks. (Some Fortran compilers provide this option
356 under the name -static.)
357 -ff2c
359 Generate code designed to be compatible with code generated by g77
360 and f2c.
361 The calling conventions used by g77 (originally implemented in f2c)
363 require functions that return type default "REAL" to actually
364 return the C type "double", and functions that return type "COM‐
365 PLEX" to return the values via an extra argument in the calling
366 sequence that points to where to store the return value. Under the
367 default GNU calling conventions, such functions simply return their
368 results as they would in GNU C -- default "REAL" functions return
369 the C type "float", and "COMPLEX" functions return the GNU C type
370 "complex". Additionally, this option implies the -fsecond-under‐
371 score option, unless -fno-second-underscore is explicitly
372 requested.
373 This does not affect the generation of code that interfaces with
375 the libgfortran library.
376 Caution: It is not a good idea to mix Fortran code compiled with
378 "-ff2c" with code compiled with the default "-fno-f2c" calling con‐
379 ventions as, calling "COMPLEX" or default "REAL" functions between
380 program parts which were compiled with different calling conven‐
381 tions will break at execution time.
382 Caution: This will break code which passes intrinsic functions of
384 type default "REAL" or "COMPLEX" as actual arguments, as the
385 library implementations use the -fno-f2c calling conventions.
386 -fno-underscoring
388 Do not transform names of entities specified in the Fortran source
389 file by appending underscores to them.
390 With -funderscoring in effect, gfortran appends one underscore to
392 external names with no underscores. This is done to ensure compat‐
393 ibility with code produced by many UNIX Fortran compilers.
394 Caution: The default behavior of gfortran is incompatible with f2c
396 and g77, please use the -ff2c option if you want object files com‐
397 piled with gfortran to be compatible with object code created with
398 these tools.
399 Use of -fno-underscoring is not recommended unless you are experi‐
401 menting with issues such as integration of (GNU) Fortran into
402 existing system environments (vis-a-vis existing libraries, tools,
403 and so on).
404 For example, with -funderscoring, and assuming other defaults like
406 -fcase-lower and that j() and max_count() are external functions
407 while my_var and lvar are local variables, a statement like
408 I = J() + MAX_COUNT (MY_VAR, LVAR)
410 is implemented as something akin to:
412 i = j_() + max_count__(&my_var__, &lvar);
414 With -fno-underscoring, the same statement is implemented as:
416 i = j() + max_count(&my_var, &lvar);
418 Use of -fno-underscoring allows direct specification of user-
420 defined names while debugging and when interfacing gfortran code
421 with other languages.
422 Note that just because the names match does not mean that the
424 interface implemented by gfortran for an external name matches the
425 interface implemented by some other language for that same name.
426 That is, getting code produced by gfortran to link to code produced
427 by some other compiler using this or any other method can be only a
428 small part of the overall solution---getting the code generated by
429 both compilers to agree on issues other than naming can require
430 significant effort, and, unlike naming disagreements, linkers nor‐
431 mally cannot detect disagreements in these other areas.
432 Also, note that with -fno-underscoring, the lack of appended under‐
434 scores introduces the very real possibility that a user-defined
435 external name will conflict with a name in a system library, which
436 could make finding unresolved-reference bugs quite difficult in
437 some cases---they might occur at program run time, and show up only
438 as buggy behavior at run time.
439 In future versions of gfortran we hope to improve naming and link‐
441 ing issues so that debugging always involves using the names as
442 they appear in the source, even if the names as seen by the linker
443 are mangled to prevent accidental linking between procedures with
444 incompatible interfaces.
445 -fsecond-underscore
447 By default, gfortran appends an underscore to external names. If
448 this option is used gfortran appends two underscores to names with
449 underscores and one underscore to external names with no under‐
450 scores. (gfortran also appends two underscores to internal names
451 with underscores to avoid naming collisions with external names.
452 This option has no effect if -fno-underscoring is in effect. It is
454 implied by the -ff2c option.
455 Otherwise, with this option, an external name such as MAX_COUNT is
457 implemented as a reference to the link-time external symbol
458 max_count__, instead of max_count_. This is required for compati‐
459 bility with g77 and f2c, and is implied by use of the -ff2c option.
460 -fbounds-check
462 Enable generation of run-time checks for array subscripts and
463 against the declared minimum and maximum values. It also checks
464 array indices for assumed and deferred shape arrays against the
465 actual allocated bounds.
466 In the future this may also include other forms of checking, eg.
468 checking substring references.
469 -fmax-stack-var-size=n
471 This option specifies the size in bytes of the largest array that
472 will be put on the stack.
473 This option currently only affects local arrays declared with con‐
475 stant bounds, and may not apply to all character variables. Future
476 versions of gfortran may improve this behavior.
477 The default value for n is 32768.
479 -fpackderived
481 This option tells gfortran to pack derived type members as closely
482 as possible. Code compiled with this option is likely to be incom‐
483 patible with code compiled without this option, and may execute
484 slower.
485 -frepack-arrays
487 In some circumstances gfortran may pass assumed shape array sec‐
488 tions via a descriptor describing a discontiguous area of memory.
489 This option adds code to the function prologue to repack the data
490 into a contiguous block at runtime.
491 This should result in faster accesses to the array. However it can
493 introduce significant overhead to the function call, especially
494 when the passed data is discontiguous.
495 -fshort-enums
497 This option is provided for interoperability with C code that was
498 compiled with the -fshort-enums option. It will make gfortran
499 choose the smallest "INTEGER" kind a given enumerator set will fit
500 in, and give all its enumerators this kind.
501
503 GNU Fortran 95 currently does not make use of any environment variables
504 to control its operation above and beyond those that affect the opera‐
505 tion of gcc.
506
508 For instructions on reporting bugs, see <http://gcc.gnu.org/bugs.html>.
509
511 gpl(7), gfdl(7), fsf-funding(7), cpp(1), gcov(1), gcc(1), as(1), ld(1),
512 gdb(1), adb(1), dbx(1), sdb(1) and the Info entries for gcc, cpp, gfor‐
513 tran, as, ld, binutils and gdb.
514
516 See the Info entry for gfortran for contributors to GCC and GFORTRAN.
517
519 Copyright (c) 2004, 2005, 2006 Free Software Foundation, Inc.
520 Permission is granted to copy, distribute and/or modify this document
522 under the terms of the GNU Free Documentation License, Version 1.2 or
523 any later version published by the Free Software Foundation; with the
524 Invariant Sections being "GNU General Public License" and "Funding Free
525 Software", the Front-Cover texts being (a) (see below), and with the
526 Back-Cover Texts being (b) (see below). A copy of the license is
527 included in the gfdl(7) man page.
528 (a) The FSF's Front-Cover Text is:
530 A GNU Manual
532 (b) The FSF's Back-Cover Text is:
534 You have freedom to copy and modify this GNU Manual, like GNU
536 software. Copies published by the Free Software Foundation raise
537 funds for GNU development.
538gcc-4.1.2 2007-09-25 GFORTRAN(1)