1OCAMLOPT(1) General Commands Manual OCAMLOPT(1)
2
6 ocamlopt - The OCaml native-code compiler
7
10 ocamlopt [ options ] filename ...
11 ocamlopt.opt (same options)
13
16 The OCaml high-performance native-code compiler ocamlopt(1) compiles
17 OCaml source files to native code object files and link these object
18 files to produce standalone executables.
19 The ocamlopt(1) command has a command-line interface very close to that
21 of ocamlc(1). It accepts the same types of arguments and processes
22 them sequentially, after all options have been processed:
23 Arguments ending in .mli are taken to be source files for compilation
25 unit interfaces. Interfaces specify the names exported by compilation
26 units: they declare value names with their types, define public data
27 types, declare abstract data types, and so on. From the file x.mli, the
28 ocamlopt(1) compiler produces a compiled interface in the file x.cmi.
29 The interface produced is identical to that produced by the bytecode
30 compiler ocamlc(1).
31 Arguments ending in .ml are taken to be source files for compilation
33 unit implementations. Implementations provide definitions for the names
34 exported by the unit, and also contain expressions to be evaluated for
35 their side-effects. From the file x.ml, the ocamlopt(1) compiler pro‐
36 duces two files: x.o, containing native object code, and x.cmx, con‐
37 taining extra information for linking and optimization of the clients
38 of the unit. The compiled implementation should always be referred to
39 under the name x.cmx (when given a .o file, ocamlopt(1) assumes that it
40 contains code compiled from C, not from OCaml).
41 The implementation is checked against the interface file x.mli (if it
43 exists) as described in the manual for ocamlc(1).
44 Arguments ending in .cmx are taken to be compiled object code. These
46 files are linked together, along with the object files obtained by com‐
47 piling .ml arguments (if any), and the OCaml standard library, to pro‐
48 duce a native-code executable program. The order in which .cmx and .ml
49 arguments are presented on the command line is relevant: compilation
50 units are initialized in that order at run-time, and it is a link-time
51 error to use a component of a unit before having initialized it. Hence,
52 a given x.cmx file must come before all .cmx files that refer to the
53 unit x.
54 Arguments ending in .cmxa are taken to be libraries of object code.
56 Such a library packs in two files lib.cmxa and lib.a a set of object
57 files (.cmx/.o files). Libraries are build with ocamlopt -a (see the
58 description of the -a option below). The object files contained in the
59 library are linked as regular .cmx files (see above), in the order
60 specified when the library was built. The only difference is that if an
61 object file contained in a library is not referenced anywhere in the
62 program, then it is not linked in.
63 Arguments ending in .c are passed to the C compiler, which generates a
65 .o object file. This object file is linked with the program.
66 Arguments ending in .o or .a are assumed to be C object files and
68 libraries. They are linked with the program.
69 The output of the linking phase is a regular Unix executable file. It
71 does not need ocamlrun(1) to run.
72 ocamlopt.opt is the same compiler as ocamlopt, but compiled with itself
74 instead of with the bytecode compiler ocamlc(1). Thus, it behaves
75 exactly like ocamlopt, but compiles faster. ocamlopt.opt is not avail‐
76 able in all installations of OCaml.
77
80 The following command-line options are recognized by ocamlopt(1).
81 -a Build a library (.cmxa/.a file) with the object files (.cmx/.o
83 files) given on the command line, instead of linking them into
84 an executable file. The name of the library must be set with the
85 -o option.
86 If -cclib or -ccopt options are passed on the command line,
88 these options are stored in the resulting .cmxa library. Then,
89 linking with this library automatically adds back the
90 -cclib and -ccopt options as if they had been provided on the
91 command line, unless the -noautolink option is given. Addition‐
92 ally, a substring $CAMLORIGIN inside a -ccopt options will be
93 replaced by the full path to the .cma library, excluding the
94 filename.
95 -absname
97 Show absolute filenames in error messages.
98 -annot Dump detailed information about the compilation (types, bind‐
100 ings, tail-calls, etc). The information for file src.ml is put
101 into file src.annot. In case of a type error, dump all the
102 information inferred by the type-checker before the error. The
103 src.annot file can be used with the emacs commands given in
104 emacs/caml-types.el to display types and other annotations
105 interactively.
106 -bin-annot
108 Dump detailed information about the compilation (types, bind‐
109 ings, tail-calls, etc) in binary format. The information for
110 file src.ml is put into file src.cmt. In case of a type error,
111 dump all the information inferred by the type-checker before the
112 error. The annotation files produced by -bin-annot contain more
113 information and are much more compact than the files produced by
114 -annot.
115 -c Compile only. Suppress the linking phase of the compilation.
117 Source code files are turned into compiled files, but no exe‐
118 cutable file is produced. This option is useful to compile mod‐
119 ules separately.
120 -cc ccomp
122 Use ccomp as the C linker called to build the final executable
123 and as the C compiler for compiling .c source files.
124 -cclib -llibname
126 Pass the -llibname option to the linker. This causes the given C
127 library to be linked with the program.
128 -ccopt option
130 Pass the given option to the C compiler and linker. For
131 instance, -ccopt -Ldir causes the C linker to search for C
132 libraries in directory dir.
133 -color mode
135 Enable or disable colors in compiler messages (especially warn‐
136 ings and errors). The following modes are supported:
137 auto use heuristics to enable colors only if the output supports
139 them (an ANSI-compatible tty terminal);
140 always enable colors unconditionally;
142 never disable color output.
144 The default setting is auto, and the current heuristic checks
146 that the "TERM" environment variable exists and is not empty or
147 "dumb", and that isatty(stderr) holds.
148 The environment variable "OCAML_COLOR" is considered if -color
150 is not provided. Its values are auto/always/never as above.
151 -compact
154 Optimize the produced code for space rather than for time. This
155 results in smaller but slightly slower programs. The default is
156 to optimize for speed.
157 -config
159 Print the version number of ocamlopt(1) and a detailed summary
160 of its configuration, then exit.
161 -depend ocamldep-args
163 Compute dependencies, as ocamldep would do.
164 -for-pack module-path
166 Generate an object file (.cmx and .o files) that can later be
167 included as a sub-module (with the given access path) of a com‐
168 pilation unit constructed with -pack. For instance, ocam‐
169 lopt -for-pack P -c A.ml will generate a.cmx and a.o files that
170 can later be used with ocamlopt -pack -o P.cmx a.cmx.
171 -g Add debugging information while compiling and linking. This
173 option is required in order to produce stack backtraces when the
174 program terminates on an uncaught exception (see ocamlrun(1)).
175 -i Cause the compiler to print all defined names (with their
177 inferred types or their definitions) when compiling an implemen‐
178 tation (.ml file). No compiled files (.cmo and .cmi files) are
179 produced. This can be useful to check the types inferred by the
180 compiler. Also, since the output follows the syntax of inter‐
181 faces, it can help in writing an explicit interface (.mli file)
182 for a file: just redirect the standard output of the compiler to
183 a .mli file, and edit that file to remove all declarations of
184 unexported names.
185 -I directory
187 Add the given directory to the list of directories searched for
188 compiled interface files (.cmi), compiled object code files
189 (.cmx), and libraries (.cmxa). By default, the current directory
190 is searched first, then the standard library directory. Directo‐
191 ries added with -I are searched after the current directory, in
192 the order in which they were given on the command line, but
193 before the standard library directory. See also option -nost‐
194 dlib.
195 If the given directory starts with +, it is taken relative to
197 the standard library directory. For instance, -I +compiler-libs
198 adds the subdirectory compiler-libs of the standard library to
199 the search path.
200 -impl filename
202 Compile the file filename as an implementation file, even if its
203 extension is not .ml.
204 -inline n
206 Set aggressiveness of inlining to n, where n is a positive inte‐
207 ger. Specifying -inline 0 prevents all functions from being
208 inlined, except those whose body is smaller than the call site.
209 Thus, inlining causes no expansion in code size. The default
210 aggressiveness, -inline 1, allows slightly larger functions to
211 be inlined, resulting in a slight expansion in code size. Higher
212 values for the -inline option cause larger and larger functions
213 to become candidate for inlining, but can result in a serious
214 increase in code size.
215 -intf filename
217 Compile the file filename as an interface file, even if its
218 extension is not .mli.
219 -intf-suffix string
221 Recognize file names ending with string as interface files
222 (instead of the default .mli).
223 -keep-docs
225 Keep documentation strings in generated .cmi files.
226 -keep-locs
228 Keep locations in generated .cmi files.
229 -labels
231 Labels are not ignored in types, labels may be used in applica‐
232 tions, and labelled parameters can be given in any order. This
233 is the default.
234 -linkall
236 Force all modules contained in libraries to be linked in. If
237 this flag is not given, unreferenced modules are not linked in.
238 When building a library (-a flag), setting the -linkall flag
239 forces all subsequent links of programs involving that library
240 to link all the modules contained in the library. When compil‐
241 ing a module (option -c), setting the -linkall option ensures
242 that this module will always be linked if it is put in a library
243 and this library is linked.
244 -linscan
246 Use linear scan register allocation. Compiling with this allo‐
247 cator is faster than with the usual graph coloring allocator,
248 sometimes quite drastically so for long functions and modules.
249 On the other hand, the generated code can be a bit slower.
250 -no-alias-deps
252 Do not record dependencies for module aliases.
253 -no-app-funct
255 Deactivates the applicative behaviour of functors. With this
256 option, each functor application generates new types in its
257 result and applying the same functor twice to the same argument
258 yields two incompatible structures.
259 -noassert
261 Do not compile assertion checks. Note that the special form
262 assert false is always compiled because it is typed specially.
263 This flag has no effect when linking already-compiled files.
264 -noautolink
266 When linking .cmxa libraries, ignore -cclib and -ccopt options
267 potentially contained in the libraries (if these options were
268 given when building the libraries). This can be useful if a
269 library contains incorrect specifications of C libraries or C
270 options; in this case, during linking, set -noautolink and pass
271 the correct C libraries and options on the command line.
272 -nodynlink
274 Allow the compiler to use some optimizations that are valid only
275 for code that is never dynlinked.
276 -nostdlib
278 Do not automatically add the standard library directory the list
279 of directories searched for compiled interface files (.cmi),
280 compiled object code files (.cmx), and libraries (.cmxa). See
281 also option -I.
282 -nolabels
284 Ignore non-optional labels in types. Labels cannot be used in
285 applications, and parameter order becomes strict.
286 -o exec-file
288 Specify the name of the output file produced by the linker. The
289 default output name is a.out, in keeping with the Unix tradi‐
290 tion. If the -a option is given, specify the name of the library
291 produced. If the -pack option is given, specify the name of the
292 packed object file produced. If the -output-obj option is
293 given, specify the name of the output file produced. If the
294 -shared option is given, specify the name of plugin file pro‐
295 duced. This can also be used when compiling an interface or
296 implementation file, without linking, in which case it sets the
297 name of the cmi or cmo file, and also sets the module name to
298 the file name up to the first dot.
299 -opaque
301 When compiling a .mli interface file, this has the same effect
302 as the -opaque option of the bytecode compiler. When compiling a
303 .ml implementation file, this produces a .cmx file without
304 cross-module optimization information, which reduces recompila‐
305 tion on module change.
306 -open module
308 Opens the given module before processing the interface or imple‐
309 mentation files. If several -open options are given, they are
310 processed in order, just as if the statements open! module1;;
311 ... open! moduleN;; were added at the top of each file.
312 -output-obj
314 Cause the linker to produce a C object file instead of an exe‐
315 cutable file. This is useful to wrap OCaml code as a C library,
316 callable from any C program. The name of the output object file
317 must be set with the -o option. This option can also be used to
318 produce a compiled shared/dynamic library (.so extension).
319 -p Generate extra code to write profile information when the pro‐
321 gram is executed. The profile information can then be examined
322 with the analysis program gprof(1). The -p option must be given
323 both at compile-time and at link-time. Linking object files not
324 compiled with -p is possible, but results in less precise pro‐
325 filing.
326 See the gprof(1) man page for more information about the pro‐
328 files.
329 Full support for gprof(1) is only available for certain plat‐
331 forms (currently: Intel x86/Linux and Alpha/Digital Unix). On
332 other platforms, the -p option will result in a less precise
333 profile (no call graph information, only a time profile).
334 -pack Build an object file (.cmx and .o files) and its associated com‐
336 piled interface (.cmi) that combines the .cmx object files given
337 on the command line, making them appear as sub-modules of the
338 output .cmx file. The name of the output .cmx file must be
339 given with the -o option. For instance, ocam‐
340 lopt -pack -o P.cmx A.cmx B.cmx C.cmx generates compiled files
341 P.cmx, P.o and P.cmi describing a compilation unit having three
342 sub-modules A, B and C, corresponding to the contents of the
343 object files A.cmx, B.cmx and C.cmx. These contents can be ref‐
344 erenced as P.A, P.B and P.C in the remainder of the program.
345 The .cmx object files being combined must have been compiled
347 with the appropriate -for-pack option. In the example above,
348 A.cmx, B.cmx and C.cmx must have been compiled with ocam‐
349 lopt -for-pack P.
350 Multiple levels of packing can be achieved by combining -pack
352 with -for-pack. See The OCaml user's manual, chapter "Native-
353 code compilation" for more details.
354 -plugin plugin
356 Dynamically load the code of the given plugin (a .cmo, .cma or
357 .cmxs file) in the compiler. The plugin must exist in the same
358 kind of code as the compiler (ocamlopt.byte must load bytecode
359 plugins, while ocamlopt.opt must load native code plugins), and
360 extension adaptation is done automatically for .cma files (to
361 .cmxs files if the compiler is compiled in native code).
362 -pp command
364 Cause the compiler to call the given command as a preprocessor
365 for each source file. The output of command is redirected to an
366 intermediate file, which is compiled. If there are no compila‐
367 tion errors, the intermediate file is deleted afterwards.
368 -ppx command
370 After parsing, pipe the abstract syntax tree through the pre‐
371 processor command. The module Ast_mapper(3) implements the
372 external interface of a preprocessor.
373 -principal
375 Check information path during type-checking, to make sure that
376 all types are derived in a principal way. All programs accepted
377 in -principal mode are also accepted in default mode with equiv‐
378 alent types, but different binary signatures.
379 -rectypes
381 Allow arbitrary recursive types during type-checking. By
382 default, only recursive types where the recursion goes through
383 an object type are supported. Note that once you have created an
384 interface using this flag, you must use it again for all depen‐
385 dencies.
386 -runtime-variant suffix
388 Add suffix to the name of the runtime library that will be used
389 by the program. If OCaml was configured with option
390 -with-debug-runtime, then the d suffix is supported and gives a
391 debug version of the runtime.
392 -S Keep the assembly code produced during the compilation. The
394 assembly code for the source file x.ml is saved in the file x.s.
395 -safe-string
397 Enforce the separation between types string and bytes, thereby
398 making strings read-only. This is the default.
399 -shared
401 Build a plugin (usually .cmxs) that can be dynamically loaded
402 with the Dynlink module. The name of the plugin must be set with
403 the -o option. A plugin can include a number of OCaml modules
404 and libraries, and extra native objects (.o, .a files). Build‐
405 ing native plugins is only supported for some operating system.
406 Under some systems (currently, only Linux AMD 64), all the OCaml
407 code linked in a plugin must have been compiled without the
408 -nodynlink flag. Some constraints might also apply to the way
409 the extra native objects have been compiled (under Linux AMD 64,
410 they must contain only position-independent code).
411 -short-paths
413 When a type is visible under several module-paths, use the
414 shortest one when printing the type's name in inferred inter‐
415 faces and error and warning messages.
416 -strict-sequence
418 The left-hand part of a sequence must have type unit.
419 -unboxed-types
421 When a type is unboxable (i.e. a record with a single argument
422 or a concrete datatype with a single constructor of one argu‐
423 ment) it will be unboxed unless annotated with [@@ocaml.boxed].
424 -no-unboxed-types
426 When a type is unboxable it will be boxed unless annotated with
427 [@@ocaml.unboxed]. This is the default.
428 -unsafe
430 Turn bound checking off for array and string accesses (the
431 v.(i)ands.[i] constructs). Programs compiled with -unsafe are
432 therefore faster, but unsafe: anything can happen if the program
433 accesses an array or string outside of its bounds. Additionally,
434 turn off the check for zero divisor in integer division and mod‐
435 ulus operations. With -unsafe, an integer division (or modulus)
436 by zero can halt the program or continue with an unspecified
437 result instead of raising a Division_by_zero exception.
438 -unsafe-string
440 Identify the types string and bytes, thereby making strings
441 writable. This is intended for compatibility with old source
442 code and should not be used with new software.
443 -v Print the version number of the compiler and the location of the
445 standard library directory, then exit.
446 -verbose
448 Print all external commands before they are executed, in partic‐
449 ular invocations of the assembler, C compiler, and linker.
450 -version or -vnum
452 Print the version number of the compiler in short form (e.g.
453 "3.11.0"), then exit.
454 -w warning-list
456 Enable, disable, or mark as fatal the warnings specified by the
457 argument warning-list. See ocamlc(1) for the syntax of warning-
458 list.
459 -warn-error warning-list
461 Mark as fatal the warnings specified in the argument warn‐
462 ing-list. The compiler will stop with an error when one of
463 these warnings is emitted. The warning-list has the same mean‐
464 ing as for the -w option: a + sign (or an uppercase letter)
465 marks the corresponding warnings as fatal, a - sign (or a lower‐
466 case letter) turns them back into non-fatal warnings, and a @
467 sign both enables and marks as fatal the corresponding warnings.
468 Note: it is not recommended to use the -warn-error option in
470 production code, because it will almost certainly prevent com‐
471 piling your program with later versions of OCaml when they add
472 new warnings or modify existing warnings.
473 The default setting is -warn-error -a+31 (only warning 31 is
475 fatal).
476 -warn-help
478 Show the description of all available warning numbers.
479 -where Print the location of the standard library, then exit.
481 - file Process file as a file name, even if it starts with a dash (-)
483 character.
484 -help or --help
486 Display a short usage summary and exit.
487
490 The IA32 code generator (Intel Pentium, AMD Athlon) supports the fol‐
491 lowing additional option:
492 -ffast-math
494 Use the IA32 instructions to compute trigonometric and exponen‐
495 tial functions, instead of calling the corresponding library
496 routines. The functions affected are: atan, atan2, cos, log,
497 log10, sin, sqrt and tan. The resulting code runs faster, but
498 the range of supported arguments and the precision of the result
499 can be reduced. In particular, trigonometric operations cos,
500 sin, tan have their range reduced to [-2^64, 2^64].
501
504 The AMD64 code generator (64-bit versions of Intel Pentium and AMD
505 Athlon) supports the following additional options:
506 -fPIC Generate position-independent machine code. This is the
508 default.
509 -fno-PIC
511 Generate position-dependent machine code.
512
515 The ARM code generator supports the following additional options:
516 -farch=armv4|armv5|armv5te|armv6|armv6t2|armv7
518 Select the ARM target architecture
519 -ffpu=soft|vfpv2|vfpv3-d16|vfpv3
521 Select the floating-point hardware
522 -fPIC Generate position-independent machine code.
524 -fno-PIC
526 Generate position-dependent machine code. This is the default.
527 -fthumb
529 Enable Thumb/Thumb-2 code generation
530 -fno-thumb
532 Disable Thumb/Thumb-2 code generation
533 The default values for target architecture, floating-point hardware and
535 thumb usage were selected at configure-time when building ocamlopt
536 itself. This configuration can be inspected using ocamlopt -config.
537 Target architecture depends on the "model" setting, while floating-
538 point hardware and thumb support are determined from the ABI setting in
539 "system" ( linux_eabiorlinux_eabihf).
540
543 ocamlc(1).
544 The OCaml user's manual, chapter "Native-code compilation".
545 OCAMLOPT(1)