1GFORTRAN(1) GNU GFORTRAN(1)
2
6 gfortran - GNU Fortran compiler
7
9 gfortran [-c|-S|-E]
10 [-g] [-pg] [-Olevel]
11 [-Wwarn...] [-pedantic]
12 [-Idir...] [-Ldir...]
13 [-Dmacro[=defn]...] [-Umacro]
14 [-foption...]
15 [-mmachine-option...]
16 [-o outfile] infile...
17 Only the most useful options are listed here; see below for the
19 remainder.
20
22 The gfortran command supports all the options supported by the gcc
23 command. Only options specific to GNU Fortran are documented here.
24 All GCC and GNU Fortran options are accepted both by gfortran and by
26 gcc (as well as any other drivers built at the same time, such as g++),
27 since adding GNU Fortran to the GCC distribution enables acceptance of
28 GNU Fortran options by all of the relevant drivers.
29 In some cases, options have positive and negative forms; the negative
31 form of -ffoo would be -fno-foo. This manual documents only one of
32 these two forms, whichever one is not the default.
33
35 Here is a summary of all the options specific to GNU Fortran, grouped
36 by type. Explanations are in the following sections.
37 Fortran Language Options
39 -fall-intrinsics -ffree-form -fno-fixed-form -fdollar-ok
40 -fimplicit-none -fmax-identifier-length -std=std -fd-lines-as-code
41 -fd-lines-as-comments -ffixed-line-length-n
42 -ffixed-line-length-none -ffree-line-length-n
43 -ffree-line-length-none -fdefault-double-8 -fdefault-integer-8
44 -fdefault-real-8 -fcray-pointer -fopenmp -fno-range-check
45 -fbackslash -fmodule-private
46 Preprocessing Options
48 -cpp -dD -dI -dM -dN -dU -fworking-directory -imultilib dir
49 -iprefix file -isysroot dir -iquote -isystem dir -nocpp -nostdinc
50 -undef -Aquestion=answer -A-question[=answer] -C -CC -Dmacro[=defn]
51 -Umacro -H -P
52 Error and Warning Options
54 -fmax-errors=n -fsyntax-only -pedantic -pedantic-errors -Wall
55 -Waliasing -Wampersand -Warray-bounds -Wcharacter-truncation
56 -Wconversion -Wimplicit-interface -Wimplicit-procedure
57 -Wline-truncation -Wintrinsics-std -Wsurprising -Wno-tabs
58 -Wunderflow -Wunused-parameter -Wintrinsics-shadow
59 -Wno-align-commons
60 Debugging Options
62 -fdump-parse-tree -ffpe-trap=list -fdump-core -fbacktrace
63 Directory Options
65 -Idir -Jdir -fintrinsic-modules-path dir
66 Link Options
68 -static-libgfortran
69 Runtime Options
71 -fconvert=conversion -fno-range-check -frecord-marker=length
72 -fmax-subrecord-length=length -fsign-zero
73 Code Generation Options
75 -fno-automatic -ff2c -fno-underscoring -fwhole-file
76 -fsecond-underscore -fbounds-check -fcheck-array-temporaries
77 -fmax-array-constructor =n
78 -fcheck=<all|array-temps|bounds|do|mem|pointer|recursion>
79 -fmax-stack-var-size=n -fpack-derived -frepack-arrays
80 -fshort-enums -fexternal-blas -fblas-matmul-limit=n -frecursive
81 -finit-local-zero -finit-integer=n
82 -finit-real=<zero|inf|-inf|nan|snan> -finit-logical=<true|false>
83 -finit-character=n -fno-align-commons -fno-protect-parens
84 Options controlling Fortran dialect
86 The following options control the details of the Fortran dialect
87 accepted by the compiler:
88 -ffree-form
90 -ffixed-form
91 Specify the layout used by the source file. The free form layout
92 was introduced in Fortran 90. Fixed form was traditionally used in
93 older Fortran programs. When neither option is specified, the
94 source form is determined by the file extension.
95 -fall-intrinsics
97 This option causes all intrinsic procedures (including the GNU-
98 specific extensions) to be accepted. This can be useful with
99 -std=f95 to force standard-compliance but get access to the full
100 range of intrinsics available with gfortran. As a consequence,
101 -Wintrinsics-std will be ignored and no user-defined procedure with
102 the same name as any intrinsic will be called except when it is
103 explicitly declared "EXTERNAL".
104 -fd-lines-as-code
106 -fd-lines-as-comments
107 Enable special treatment for lines beginning with "d" or "D" in
108 fixed form sources. If the -fd-lines-as-code option is given they
109 are treated as if the first column contained a blank. If the
110 -fd-lines-as-comments option is given, they are treated as comment
111 lines.
112 -fdefault-double-8
114 Set the "DOUBLE PRECISION" type to an 8 byte wide type. If
115 -fdefault-real-8 is given, "DOUBLE PRECISION" would instead be
116 promoted to 16 bytes if possible, and -fdefault-double-8 can be
117 used to prevent this. The kind of real constants like "1.d0" will
118 not be changed by -fdefault-real-8 though, so also
119 -fdefault-double-8 does not affect it.
120 -fdefault-integer-8
122 Set the default integer and logical types to an 8 byte wide type.
123 Do nothing if this is already the default. This option also
124 affects the kind of integer constants like 42.
125 -fdefault-real-8
127 Set the default real type to an 8 byte wide type. Do nothing if
128 this is already the default. This option also affects the kind of
129 non-double real constants like 1.0, and does promote the default
130 width of "DOUBLE PRECISION" to 16 bytes if possible, unless
131 "-fdefault-double-8" is given, too.
132 -fdollar-ok
134 Allow $ as a valid non-first character in a symbol name. Symbols
135 that start with $ are rejected since it is unclear which rules to
136 apply to implicit typing as different vendors implement different
137 rules. Using $ in "IMPLICIT" statements is also rejected.
138 -fbackslash
140 Change the interpretation of backslashes in string literals from a
141 single backslash character to "C-style" escape characters. The
142 following combinations are expanded "\a", "\b", "\f", "\n", "\r",
143 "\t", "\v", "\\", and "\0" to the ASCII characters alert,
144 backspace, form feed, newline, carriage return, horizontal tab,
145 vertical tab, backslash, and NUL, respectively. Additionally,
146 "\x"nn, "\u"nnnn and "\U"nnnnnnnn (where each n is a hexadecimal
147 digit) are translated into the Unicode characters corresponding to
148 the specified code points. All other combinations of a character
149 preceded by \ are unexpanded.
150 -fmodule-private
152 Set the default accessibility of module entities to "PRIVATE".
153 Use-associated entities will not be accessible unless they are
154 explicitly declared as "PUBLIC".
155 -ffixed-line-length-n
157 Set column after which characters are ignored in typical fixed-form
158 lines in the source file, and through which spaces are assumed (as
159 if padded to that length) after the ends of short fixed-form lines.
160 Popular values for n include 72 (the standard and the default), 80
162 (card image), and 132 (corresponding to "extended-source" options
163 in some popular compilers). n may also be none, meaning that the
164 entire line is meaningful and that continued character constants
165 never have implicit spaces appended to them to fill out the line.
166 -ffixed-line-length-0 means the same thing as
167 -ffixed-line-length-none.
168 -ffree-line-length-n
170 Set column after which characters are ignored in typical free-form
171 lines in the source file. The default value is 132. n may be none,
172 meaning that the entire line is meaningful. -ffree-line-length-0
173 means the same thing as -ffree-line-length-none.
174 -fmax-identifier-length=n
176 Specify the maximum allowed identifier length. Typical values are
177 31 (Fortran 95) and 63 (Fortran 2003 and Fortran 2008).
178 -fimplicit-none
180 Specify that no implicit typing is allowed, unless overridden by
181 explicit "IMPLICIT" statements. This is the equivalent of adding
182 "implicit none" to the start of every procedure.
183 -fcray-pointer
185 Enable the Cray pointer extension, which provides C-like pointer
186 functionality.
187 -fopenmp
189 Enable the OpenMP extensions. This includes OpenMP "!$omp"
190 directives in free form and "c$omp", *$omp and "!$omp" directives
191 in fixed form, "!$" conditional compilation sentinels in free form
192 and "c$", "*$" and "!$" sentinels in fixed form, and when linking
193 arranges for the OpenMP runtime library to be linked in. The
194 option -fopenmp implies -frecursive.
195 -fno-range-check
197 Disable range checking on results of simplification of constant
198 expressions during compilation. For example, GNU Fortran will give
199 an error at compile time when simplifying "a = 1. / 0". With this
200 option, no error will be given and "a" will be assigned the value
201 "+Infinity". If an expression evaluates to a value outside of the
202 relevant range of ["-HUGE()":"HUGE()"], then the expression will be
203 replaced by "-Inf" or "+Inf" as appropriate. Similarly, "DATA
204 i/Z'FFFFFFFF'/" will result in an integer overflow on most systems,
205 but with -fno-range-check the value will "wrap around" and "i" will
206 be initialized to -1 instead.
207 -std=std
209 Specify the standard to which the program is expected to conform,
210 which may be one of f95, f2003, f2008, gnu, or legacy. The default
211 value for std is gnu, which specifies a superset of the Fortran 95
212 standard that includes all of the extensions supported by GNU
213 Fortran, although warnings will be given for obsolete extensions
214 not recommended for use in new code. The legacy value is
215 equivalent but without the warnings for obsolete extensions, and
216 may be useful for old non-standard programs. The f95, f2003 and
217 f2008 values specify strict conformance to the Fortran 95, Fortran
218 2003 and Fortran 2008 standards, respectively; errors are given for
219 all extensions beyond the relevant language standard, and warnings
220 are given for the Fortran 77 features that are permitted but
221 obsolescent in later standards.
222 Enable and customize preprocessing
224 Preprocessor related options. See section Preprocessing and conditional
225 compilation for more detailed information on preprocessing in gfortran.
226 -cpp
228 -nocpp
229 Enable preprocessing. The preprocessor is automatically invoked if
230 the file extension is .fpp, .FPP, .F, .FOR, .FTN, .F90, .F95, .F03
231 or .F08. Use this option to manually enable preprocessing of any
232 kind of Fortran file.
233 To disable preprocessing of files with any of the above listed
235 extensions, use the negative form: -nocpp.
236 The preprocessor is run in traditional mode, be aware that any
238 restrictions of the file-format, e.g. fixed-form line width, apply
239 for preprocessed output as well.
240 -dM Instead of the normal output, generate a list of '#define'
242 directives for all the macros defined during the execution of the
243 preprocessor, including predefined macros. This gives you a way of
244 finding out what is predefined in your version of the preprocessor.
245 Assuming you have no file foo.f90, the command
246 touch foo.f90; gfortran -cpp -dM foo.f90
248 will show all the predefined macros.
250 -dD Like -dM except in two respects: it does not include the predefined
252 macros, and it outputs both the "#define" directives and the result
253 of preprocessing. Both kinds of output go to the standard output
254 file.
255 -dN Like -dD, but emit only the macro names, not their expansions.
257 -dU Like dD except that only macros that are expanded, or whose
259 definedness is tested in preprocessor directives, are output; the
260 output is delayed until the use or test of the macro; and '#undef'
261 directives are also output for macros tested but undefined at the
262 time.
263 -dI Output '#include' directives in addition to the result of
265 preprocessing.
266 -fworking-directory
268 Enable generation of linemarkers in the preprocessor output that
269 will let the compiler know the current working directory at the
270 time of preprocessing. When this option is enabled, the
271 preprocessor will emit, after the initial linemarker, a second
272 linemarker with the current working directory followed by two
273 slashes. GCC will use this directory, when it's present in the
274 preprocessed input, as the directory emitted as the current working
275 directory in some debugging information formats. This option is
276 implicitly enabled if debugging information is enabled, but this
277 can be inhibited with the negated form -fno-working-directory. If
278 the -P flag is present in the command line, this option has no
279 effect, since no "#line" directives are emitted whatsoever.
280 -idirafter dir
282 Search dir for include files, but do it after all directories
283 specified with -I and the standard system directories have been
284 exhausted. dir is treated as a system include directory. If dir
285 begins with "=", then the "=" will be replaced by the sysroot
286 prefix; see --sysroot and -isysroot.
287 -imultilib dir
289 Use dir as a subdirectory of the directory containing target-
290 specific C++ headers.
291 -iprefix prefix
293 Specify prefix as the prefix for subsequent -iwithprefix options.
294 If the prefix represents a directory, you should include the final
295 '/'.
296 -isysroot dir
298 This option is like the --sysroot option, but applies only to
299 header files. See the --sysroot option for more information.
300 -iquote dir
302 Search dir only for header files requested with "#include "file"";
303 they are not searched for "#include <file>", before all directories
304 specified by -I and before the standard system directories. If dir
305 begins with "=", then the "=" will be replaced by the sysroot
306 prefix; see --sysroot and -isysroot.
307 -isystem dir
309 Search dir for header files, after all directories specified by -I
310 but before the standard system directories. Mark it as a system
311 directory, so that it gets the same special treatment as is applied
312 to the standard system directories. If dir begins with "=", then
313 the "=" will be replaced by the sysroot prefix; see --sysroot and
314 -isysroot.
315 -nostdinc
317 Do not search the standard system directories for header files.
318 Only the directories you have specified with -I options (and the
319 directory of the current file, if appropriate) are searched.
320 -undef
322 Do not predefine any system-specific or GCC-specific macros. The
323 standard predefined macros remain defined.
324 -Apredicate=answer
326 Make an assertion with the predicate predicate and answer answer.
327 This form is preferred to the older form -A predicate(answer),
328 which is still supported, because it does not use shell special
329 characters.
330 -A-predicate=answer
332 Cancel an assertion with the predicate predicate and answer answer.
333 -C Do not discard comments. All comments are passed through to the
335 output file, except for comments in processed directives, which are
336 deleted along with the directive.
337 You should be prepared for side effects when using -C; it causes
339 the preprocessor to treat comments as tokens in their own right.
340 For example, comments appearing at the start of what would be a
341 directive line have the effect of turning that line into an
342 ordinary source line, since the first token on the line is no
343 longer a '#'.
344 Warning: this currently handles C-Style comments only. The
346 preprocessor does not yet recognize Fortran-style comments.
347 -CC Do not discard comments, including during macro expansion. This is
349 like -C, except that comments contained within macros are also
350 passed through to the output file where the macro is expanded.
351 In addition to the side-effects of the -C option, the -CC option
353 causes all C++-style comments inside a macro to be converted to
354 C-style comments. This is to prevent later use of that macro from
355 inadvertently commenting out the remainder of the source line. The
356 -CC option is generally used to support lint comments.
357 Warning: this currently handles C- and C++-Style comments only. The
359 preprocessor does not yet recognize Fortran-style comments.
360 -Dname
362 Predefine name as a macro, with definition 1.
363 -Dname=definition
365 The contents of definition are tokenized and processed as if they
366 appeared during translation phase three in a '#define' directive.
367 In particular, the definition will be truncated by embedded newline
368 characters.
369 If you are invoking the preprocessor from a shell or shell-like
371 program you may need to use the shell's quoting syntax to protect
372 characters such as spaces that have a meaning in the shell syntax.
373 If you wish to define a function-like macro on the command line,
375 write its argument list with surrounding parentheses before the
376 equals sign (if any). Parentheses are meaningful to most shells, so
377 you will need to quote the option. With sh and csh,
378 "-D'name(args...)=definition'" works.
379 -D and -U options are processed in the order they are given on the
381 command line. All -imacros file and -include file options are
382 processed after all -D and -U options.
383 -H Print the name of each header file used, in addition to other
385 normal activities. Each name is indented to show how deep in the
386 '#include' stack it is.
387 -P Inhibit generation of linemarkers in the output from the
389 preprocessor. This might be useful when running the preprocessor
390 on something that is not C code, and will be sent to a program
391 which might be confused by the linemarkers.
392 -Uname
394 Cancel any previous definition of name, either built in or provided
395 with a -D option.
396 Options to request or suppress errors and warnings
398 Errors are diagnostic messages that report that the GNU Fortran
399 compiler cannot compile the relevant piece of source code. The
400 compiler will continue to process the program in an attempt to report
401 further errors to aid in debugging, but will not produce any compiled
402 output.
403 Warnings are diagnostic messages that report constructions which are
405 not inherently erroneous but which are risky or suggest there is likely
406 to be a bug in the program. Unless -Werror is specified, they do not
407 prevent compilation of the program.
408 You can request many specific warnings with options beginning -W, for
410 example -Wimplicit to request warnings on implicit declarations. Each
411 of these specific warning options also has a negative form beginning
412 -Wno- to turn off warnings; for example, -Wno-implicit. This manual
413 lists only one of the two forms, whichever is not the default.
414 These options control the amount and kinds of errors and warnings
416 produced by GNU Fortran:
417 -fmax-errors=n
419 Limits the maximum number of error messages to n, at which point
420 GNU Fortran bails out rather than attempting to continue processing
421 the source code. If n is 0, there is no limit on the number of
422 error messages produced.
423 -fsyntax-only
425 Check the code for syntax errors, but don't actually compile it.
426 This will generate module files for each module present in the
427 code, but no other output file.
428 -pedantic
430 Issue warnings for uses of extensions to Fortran 95. -pedantic
431 also applies to C-language constructs where they occur in GNU
432 Fortran source files, such as use of \e in a character constant
433 within a directive like "#include".
434 Valid Fortran 95 programs should compile properly with or without
436 this option. However, without this option, certain GNU extensions
437 and traditional Fortran features are supported as well. With this
438 option, many of them are rejected.
439 Some users try to use -pedantic to check programs for conformance.
441 They soon find that it does not do quite what they want---it finds
442 some nonstandard practices, but not all. However, improvements to
443 GNU Fortran in this area are welcome.
444 This should be used in conjunction with -std=f95, -std=f2003 or
446 -std=f2008.
447 -pedantic-errors
449 Like -pedantic, except that errors are produced rather than
450 warnings.
451 -Wall
453 Enables commonly used warning options pertaining to usage that we
454 recommend avoiding and that we believe are easy to avoid. This
455 currently includes -Waliasing, -Wampersand, -Wsurprising,
456 -Wintrinsics-std, -Wno-tabs, -Wintrinsic-shadow and
457 -Wline-truncation.
458 -Waliasing
460 Warn about possible aliasing of dummy arguments. Specifically, it
461 warns if the same actual argument is associated with a dummy
462 argument with "INTENT(IN)" and a dummy argument with "INTENT(OUT)"
463 in a call with an explicit interface.
464 The following example will trigger the warning.
466 interface
468 subroutine bar(a,b)
469 integer, intent(in) :: a
470 integer, intent(out) :: b
471 end subroutine
472 end interface
473 integer :: a
474 call bar(a,a)
476 -Wampersand
478 Warn about missing ampersand in continued character constants. The
479 warning is given with -Wampersand, -pedantic, -std=f95, -std=f2003
480 and -std=f2008. Note: With no ampersand given in a continued
481 character constant, GNU Fortran assumes continuation at the first
482 non-comment, non-whitespace character after the ampersand that
483 initiated the continuation.
484 -Warray-temporaries
486 Warn about array temporaries generated by the compiler. The
487 information generated by this warning is sometimes useful in
488 optimization, in order to avoid such temporaries.
489 -Wcharacter-truncation
491 Warn when a character assignment will truncate the assigned string.
492 -Wline-truncation
494 Warn when a source code line will be truncated.
495 -Wconversion
497 Warn about implicit conversions between different types.
498 -Wimplicit-interface
500 Warn if a procedure is called without an explicit interface. Note
501 this only checks that an explicit interface is present. It does
502 not check that the declared interfaces are consistent across
503 program units.
504 -Wimplicit-procedure
506 Warn if a procedure is called that has neither an explicit
507 interface nor has been declared as "EXTERNAL".
508 -Wintrinsics-std
510 Warn if gfortran finds a procedure named like an intrinsic not
511 available in the currently selected standard (with -std) and treats
512 it as "EXTERNAL" procedure because of this. -fall-intrinsics can
513 be used to never trigger this behavior and always link to the
514 intrinsic regardless of the selected standard.
515 -Wsurprising
517 Produce a warning when "suspicious" code constructs are
518 encountered. While technically legal these usually indicate that
519 an error has been made.
520 This currently produces a warning under the following
522 circumstances:
523 · An INTEGER SELECT construct has a CASE that can never be
525 matched as its lower value is greater than its upper value.
526 · A LOGICAL SELECT construct has three CASE statements.
528 · A TRANSFER specifies a source that is shorter than the
530 destination.
531 · The type of a function result is declared more than once with
533 the same type. If -pedantic or standard-conforming mode is
534 enabled, this is an error.
535 · A "CHARACTER" variable is declared with negative length.
537 -Wtabs
539 By default, tabs are accepted as whitespace, but tabs are not
540 members of the Fortran Character Set. For continuation lines, a
541 tab followed by a digit between 1 and 9 is supported. -Wno-tabs
542 will cause a warning to be issued if a tab is encountered. Note,
543 -Wno-tabs is active for -pedantic, -std=f95, -std=f2003, -std=f2008
544 and -Wall.
545 -Wunderflow
547 Produce a warning when numerical constant expressions are
548 encountered, which yield an UNDERFLOW during compilation.
549 -Wintrinsic-shadow
551 Warn if a user-defined procedure or module procedure has the same
552 name as an intrinsic; in this case, an explicit interface or
553 "EXTERNAL" or "INTRINSIC" declaration might be needed to get calls
554 later resolved to the desired intrinsic/procedure.
555 -Wunused-parameter
557 Contrary to gcc's meaning of -Wunused-parameter, gfortran's
558 implementation of this option does not warn about unused dummy
559 arguments, but about unused "PARAMETER" values. -Wunused-parameter
560 is not included in -Wall but is implied by -Wall -Wextra.
561 -Walign-commons
563 By default, gfortran warns about any occasion of variables being
564 padded for proper alignment inside a COMMON block. This warning can
565 be turned off via -Wno-align-commons. See also -falign-commons.
566 -Werror
568 Turns all warnings into errors.
569 Some of these have no effect when compiling programs written in
571 Fortran.
572 Options for debugging your program or GNU Fortran
574 GNU Fortran has various special options that are used for debugging
575 either your program or the GNU Fortran compiler.
576 -fdump-parse-tree
578 Output the internal parse tree before starting code generation.
579 Only really useful for debugging the GNU Fortran compiler itself.
580 -ffpe-trap=list
582 Specify a list of IEEE exceptions when a Floating Point Exception
583 (FPE) should be raised. On most systems, this will result in a
584 SIGFPE signal being sent and the program being interrupted,
585 producing a core file useful for debugging. list is a (possibly
586 empty) comma-separated list of the following IEEE exceptions:
587 invalid (invalid floating point operation, such as "SQRT(-1.0)"),
588 zero (division by zero), overflow (overflow in a floating point
589 operation), underflow (underflow in a floating point operation),
590 precision (loss of precision during operation) and denormal
591 (operation produced a denormal value).
592 Some of the routines in the Fortran runtime library, like CPU_TIME,
594 are likely to trigger floating point exceptions when
595 "ffpe-trap=precision" is used. For this reason, the use of
596 "ffpe-trap=precision" is not recommended.
597 -fbacktrace
599 Specify that, when a runtime error is encountered or a deadly
600 signal is emitted (segmentation fault, illegal instruction, bus
601 error or floating-point exception), the Fortran runtime library
602 should output a backtrace of the error. This option only has
603 influence for compilation of the Fortran main program.
604 -fdump-core
606 Request that a core-dump file is written to disk when a runtime
607 error is encountered on systems that support core dumps. This
608 option is only effective for the compilation of the Fortran main
609 program.
610 Options for directory search
612 These options affect how GNU Fortran searches for files specified by
613 the "INCLUDE" directive and where it searches for previously compiled
614 modules.
615 It also affects the search paths used by cpp when used to preprocess
617 Fortran source.
618 -Idir
620 These affect interpretation of the "INCLUDE" directive (as well as
621 of the "#include" directive of the cpp preprocessor).
622 Also note that the general behavior of -I and "INCLUDE" is pretty
624 much the same as of -I with "#include" in the cpp preprocessor,
625 with regard to looking for header.gcc files and other such things.
626 This path is also used to search for .mod files when previously
628 compiled modules are required by a "USE" statement.
629 -Jdir
631 This option specifies where to put .mod files for compiled modules.
632 It is also added to the list of directories to searched by an "USE"
633 statement.
634 The default is the current directory.
636 -fintrinsic-modules-path dir
638 This option specifies the location of pre-compiled intrinsic
639 modules, if they are not in the default location expected by the
640 compiler.
641 Influencing the linking step
643 These options come into play when the compiler links object files into
644 an executable output file. They are meaningless if the compiler is not
645 doing a link step.
646 -static-libgfortran
648 On systems that provide libgfortran as a shared and a static
649 library, this option forces the use of the static version. If no
650 shared version of libgfortran was built when the compiler was
651 configured, this option has no effect.
652 Influencing runtime behavior
654 These options affect the runtime behavior of programs compiled with GNU
655 Fortran.
656 -fconvert=conversion
658 Specify the representation of data for unformatted files. Valid
659 values for conversion are: native, the default; swap, swap between
660 big- and little-endian; big-endian, use big-endian representation
661 for unformatted files; little-endian, use little-endian
662 representation for unformatted files.
663 This option has an effect only when used in the main program. The
665 "CONVERT" specifier and the GFORTRAN_CONVERT_UNIT environment
666 variable override the default specified by -fconvert.
667 -fno-range-check
669 Disable range checking of input values during integer "READ"
670 operations. For example, GNU Fortran will give an error if an
671 input value is outside of the relevant range of
672 ["-HUGE()":"HUGE()"]. In other words, with "INTEGER (kind=4) :: i"
673 , attempting to read -2147483648 will give an error unless
674 -fno-range-check is given.
675 -frecord-marker=length
677 Specify the length of record markers for unformatted files. Valid
678 values for length are 4 and 8. Default is 4. This is different
679 from previous versions of gfortran, which specified a default
680 record marker length of 8 on most systems. If you want to read or
681 write files compatible with earlier versions of gfortran, use
682 -frecord-marker=8.
683 -fmax-subrecord-length=length
685 Specify the maximum length for a subrecord. The maximum permitted
686 value for length is 2147483639, which is also the default. Only
687 really useful for use by the gfortran testsuite.
688 -fsign-zero
690 When enabled, floating point numbers of value zero with the sign
691 bit set are written as negative number in formatted output and
692 treated as negative in the "SIGN" intrinsic. "fno-sign-zero" does
693 not print the negative sign of zero values and regards zero as
694 positive number in the "SIGN" intrinsic for compatibility with F77.
695 Default behavior is to show the negative sign.
696 Options for code generation conventions
698 These machine-independent options control the interface conventions
699 used in code generation.
700 Most of them have both positive and negative forms; the negative form
702 of -ffoo would be -fno-foo. In the table below, only one of the forms
703 is listed---the one which is not the default. You can figure out the
704 other form by either removing no- or adding it.
705 -fno-automatic
707 Treat each program unit (except those marked as RECURSIVE) as if
708 the "SAVE" statement were specified for every local variable and
709 array referenced in it. Does not affect common blocks. (Some
710 Fortran compilers provide this option under the name -static or
711 -save.) The default, which is -fautomatic, uses the stack for
712 local variables smaller than the value given by
713 -fmax-stack-var-size. Use the option -frecursive to use no static
714 memory.
715 -ff2c
717 Generate code designed to be compatible with code generated by g77
718 and f2c.
719 The calling conventions used by g77 (originally implemented in f2c)
721 require functions that return type default "REAL" to actually
722 return the C type "double", and functions that return type
723 "COMPLEX" to return the values via an extra argument in the calling
724 sequence that points to where to store the return value. Under the
725 default GNU calling conventions, such functions simply return their
726 results as they would in GNU C---default "REAL" functions return
727 the C type "float", and "COMPLEX" functions return the GNU C type
728 "complex". Additionally, this option implies the
729 -fsecond-underscore option, unless -fno-second-underscore is
730 explicitly requested.
731 This does not affect the generation of code that interfaces with
733 the libgfortran library.
734 Caution: It is not a good idea to mix Fortran code compiled with
736 -ff2c with code compiled with the default -fno-f2c calling
737 conventions as, calling "COMPLEX" or default "REAL" functions
738 between program parts which were compiled with different calling
739 conventions will break at execution time.
740 Caution: This will break code which passes intrinsic functions of
742 type default "REAL" or "COMPLEX" as actual arguments, as the
743 library implementations use the -fno-f2c calling conventions.
744 -fno-underscoring
746 Do not transform names of entities specified in the Fortran source
747 file by appending underscores to them.
748 With -funderscoring in effect, GNU Fortran appends one underscore
750 to external names with no underscores. This is done to ensure
751 compatibility with code produced by many UNIX Fortran compilers.
752 Caution: The default behavior of GNU Fortran is incompatible with
754 f2c and g77, please use the -ff2c option if you want object files
755 compiled with GNU Fortran to be compatible with object code created
756 with these tools.
757 Use of -fno-underscoring is not recommended unless you are
759 experimenting with issues such as integration of GNU Fortran into
760 existing system environments (vis-@`{a}-vis existing libraries,
761 tools, and so on).
762 For example, with -funderscoring, and assuming other defaults like
764 -fcase-lower and that "j()" and "max_count()" are external
765 functions while "my_var" and "lvar" are local variables, a
766 statement like
767 I = J() + MAX_COUNT (MY_VAR, LVAR)
769 is implemented as something akin to:
771 i = j_() + max_count__(&my_var__, &lvar);
773 With -fno-underscoring, the same statement is implemented as:
775 i = j() + max_count(&my_var, &lvar);
777 Use of -fno-underscoring allows direct specification of user-
779 defined names while debugging and when interfacing GNU Fortran code
780 with other languages.
781 Note that just because the names match does not mean that the
783 interface implemented by GNU Fortran for an external name matches
784 the interface implemented by some other language for that same
785 name. That is, getting code produced by GNU Fortran to link to
786 code produced by some other compiler using this or any other method
787 can be only a small part of the overall solution---getting the code
788 generated by both compilers to agree on issues other than naming
789 can require significant effort, and, unlike naming disagreements,
790 linkers normally cannot detect disagreements in these other areas.
791 Also, note that with -fno-underscoring, the lack of appended
793 underscores introduces the very real possibility that a user-
794 defined external name will conflict with a name in a system
795 library, which could make finding unresolved-reference bugs quite
796 difficult in some cases---they might occur at program run time, and
797 show up only as buggy behavior at run time.
798 In future versions of GNU Fortran we hope to improve naming and
800 linking issues so that debugging always involves using the names as
801 they appear in the source, even if the names as seen by the linker
802 are mangled to prevent accidental linking between procedures with
803 incompatible interfaces.
804 -fwhole-file
806 By default, GNU Fortran parses, resolves and translates each
807 procedure in a file separately. Using this option modifies this
808 such that the whole file is parsed and placed in a single front-end
809 tree. During resolution, in addition to all the usual checks and
810 fixups, references to external procedures that are in the same file
811 effect resolution of that procedure, if not already done, and a
812 check of the interfaces. The dependences are resolved by changing
813 the order in which the file is translated into the backend tree.
814 Thus, a procedure that is referenced is translated before the
815 reference and the duplication of backend tree declarations
816 eliminated.
817 -fsecond-underscore
819 By default, GNU Fortran appends an underscore to external names.
820 If this option is used GNU Fortran appends two underscores to names
821 with underscores and one underscore to external names with no
822 underscores. GNU Fortran also appends two underscores to internal
823 names with underscores to avoid naming collisions with external
824 names.
825 This option has no effect if -fno-underscoring is in effect. It is
827 implied by the -ff2c option.
828 Otherwise, with this option, an external name such as "MAX_COUNT"
830 is implemented as a reference to the link-time external symbol
831 "max_count__", instead of "max_count_". This is required for
832 compatibility with g77 and f2c, and is implied by use of the -ff2c
833 option.
834 -fcheck=<keyword>
836 Enable the generation of run-time checks; the argument shall be a
837 comma-delimited list of the following keywords.
838 all Enable all run-time test of -fcheck.
840 array-temps
842 Warns at run time when for passing an actual argument a
843 temporary array had to be generated. The information generated
844 by this warning is sometimes useful in optimization, in order
845 to avoid such temporaries.
846 Note: The warning is only printed once per location.
848 bounds
850 Enable generation of run-time checks for array subscripts and
851 against the declared minimum and maximum values. It also
852 checks array indices for assumed and deferred shape arrays
853 against the actual allocated bounds and ensures that all string
854 lengths are equal for character array constructors without an
855 explicit typespec.
856 Some checks require that -fcheck=bounds is set for the
858 compilation of the main program.
859 Note: In the future this may also include other forms of
861 checking, e.g., checking substring references.
862 do Enable generation of run-time checks for invalid modification
864 of loop iteration variables.
865 mem Enable generation of run-time checks for memory allocation.
867 Note: This option does not affect explicit allocations using
868 the "ALLOCATE" statement, which will be always checked.
869 pointer
871 Enable generation of run-time checks for pointers and
872 allocatables.
873 recursion
875 Enable generation of run-time checks for recursively called
876 subroutines and functions which are not marked as recursive.
877 See also -frecursive. Note: This check does not work for
878 OpenMP programs and is disabled if used together with
879 -frecursive and -fopenmp.
880 -fbounds-check
882 Deprecated alias for -fcheck=bounds.
883 -fcheck-array-temporaries
885 Deprecated alias for -fcheck=array-temps.
886 -fmax-array-constructor=n
888 This option can be used to increase the upper limit permitted in
889 array constructors. The code below requires this option to expand
890 the array at compile time.
891 C<program test>
893 C<implicit none>
894 C<integer j>
895 C<integer, parameter :: n = 100000>
896 C<integer, parameter :: i(n) = (/ (2*j, j = 1, n) /)>
897 C<print '(10(I0,1X))', i>
898 C<end program test>
899 Caution: This option can lead to long compile times and
901 excessively large object files.
902 The default value for n is 65535.
904 -fmax-stack-var-size=n
906 This option specifies the size in bytes of the largest array that
907 will be put on the stack; if the size is exceeded static memory is
908 used (except in procedures marked as RECURSIVE). Use the option
909 -frecursive to allow for recursive procedures which do not have a
910 RECURSIVE attribute or for parallel programs. Use -fno-automatic to
911 never use the stack.
912 This option currently only affects local arrays declared with
914 constant bounds, and may not apply to all character variables.
915 Future versions of GNU Fortran may improve this behavior.
916 The default value for n is 32768.
918 -fpack-derived
920 This option tells GNU Fortran to pack derived type members as
921 closely as possible. Code compiled with this option is likely to
922 be incompatible with code compiled without this option, and may
923 execute slower.
924 -frepack-arrays
926 In some circumstances GNU Fortran may pass assumed shape array
927 sections via a descriptor describing a noncontiguous area of
928 memory. This option adds code to the function prologue to repack
929 the data into a contiguous block at runtime.
930 This should result in faster accesses to the array. However it can
932 introduce significant overhead to the function call, especially
933 when the passed data is noncontiguous.
934 -fshort-enums
936 This option is provided for interoperability with C code that was
937 compiled with the -fshort-enums option. It will make GNU Fortran
938 choose the smallest "INTEGER" kind a given enumerator set will fit
939 in, and give all its enumerators this kind.
940 -fexternal-blas
942 This option will make gfortran generate calls to BLAS functions for
943 some matrix operations like "MATMUL", instead of using our own
944 algorithms, if the size of the matrices involved is larger than a
945 given limit (see -fblas-matmul-limit). This may be profitable if
946 an optimized vendor BLAS library is available. The BLAS library
947 will have to be specified at link time.
948 -fblas-matmul-limit=n
950 Only significant when -fexternal-blas is in effect. Matrix
951 multiplication of matrices with size larger than (or equal to) n
952 will be performed by calls to BLAS functions, while others will be
953 handled by gfortran internal algorithms. If the matrices involved
954 are not square, the size comparison is performed using the
955 geometric mean of the dimensions of the argument and result
956 matrices.
957 The default value for n is 30.
959 -frecursive
961 Allow indirect recursion by forcing all local arrays to be
962 allocated on the stack. This flag cannot be used together with
963 -fmax-stack-var-size= or -fno-automatic.
964 -finit-local-zero
966 -finit-integer=n
967 -finit-real=<zero|inf|-inf|nan|snan>
968 -finit-logical=<true|false>
969 -finit-character=n
970 The -finit-local-zero option instructs the compiler to initialize
971 local "INTEGER", "REAL", and "COMPLEX" variables to zero, "LOGICAL"
972 variables to false, and "CHARACTER" variables to a string of null
973 bytes. Finer-grained initialization options are provided by the
974 -finit-integer=n, -finit-real=<zero|inf|-inf|nan|snan> (which also
975 initializes the real and imaginary parts of local "COMPLEX"
976 variables), -finit-logical=<true|false>, and -finit-character=n
977 (where n is an ASCII character value) options. These options do
978 not initialize components of derived type variables, nor do they
979 initialize variables that appear in an "EQUIVALENCE" statement.
980 (This limitation may be removed in future releases).
981 Note that the -finit-real=nan option initializes "REAL" and
983 "COMPLEX" variables with a quiet NaN. For a signalling NaN use
984 -finit-real=snan; note, however, that compile-time optimizations
985 may convert them into quiet NaN and that trapping needs to be
986 enabled (e.g. via -ffpe-trap).
987 -falign-commons
989 By default, gfortran enforces proper alignment of all variables in
990 a COMMON block by padding them as needed. On certain platforms this
991 is mandatory, on others it increases performance. If a COMMON block
992 is not declared with consistent data types everywhere, this padding
993 can cause trouble, and -fno-align-commons can be used to disable
994 automatic alignment. The same form of this option should be used
995 for all files that share a COMMON block. To avoid potential
996 alignment issues in COMMON blocks, it is recommended to order
997 objects from largests to smallest.
998 -fno-protect-parens
1000 By default the parentheses in expression are honored for all
1001 optimization levels such that the compiler does not do any re-
1002 association. Using -fno-protect-parens allows the compiler to
1003 reorder REAL and COMPLEX expressions to produce faster code. Note
1004 that for the re-association optimization -fno-signed-zeros and
1005 -fno-trapping-math need to be in effect.
1006
1008 The gfortran compiler currently does not make use of any environment
1009 variables to control its operation above and beyond those that affect
1010 the operation of gcc.
1011
1013 For instructions on reporting bugs, see
1014 <http://bugzilla.redhat.com/bugzilla>.
1015
1017 gpl(7), gfdl(7), fsf-funding(7), cpp(1), gcov(1), gcc(1), as(1), ld(1),
1018 gdb(1), adb(1), dbx(1), sdb(1) and the Info entries for gcc, cpp,
1019 gfortran, as, ld, binutils and gdb.
1020
1022 See the Info entry for gfortran for contributors to GCC and GNU
1023 Fortran.
1024
1026 Copyright (c) 2004, 2005, 2006, 2007, 2008, 2009 Free Software
1027 Foundation, Inc.
1028 Permission is granted to copy, distribute and/or modify this document
1030 under the terms of the GNU Free Documentation License, Version 1.2 or
1031 any later version published by the Free Software Foundation; with the
1032 Invariant Sections being "Funding Free Software", the Front-Cover Texts
1033 being (a) (see below), and with the Back-Cover Texts being (b) (see
1034 below). A copy of the license is included in the gfdl(7) man page.
1035 (a) The FSF's Front-Cover Text is:
1037 A GNU Manual
1039 (b) The FSF's Back-Cover Text is:
1041 You have freedom to copy and modify this GNU Manual, like GNU
1043 software. Copies published by the Free Software Foundation raise
1044 funds for GNU development.
1045gcc-4.5.1 2010-09-24 GFORTRAN(1)