1 <chapter id="using-ghc">
2 <title>Using GHC</title>
4 <indexterm><primary>GHC, using</primary></indexterm>
5 <indexterm><primary>using GHC</primary></indexterm>
7 <para>GHC can work in one of three “modes”:</para>
11 <term><cmdsynopsis><command>ghc</command>
12 <arg choice=plain>––interactive</arg>
14 <indexterm><primary>interactive mode</primary>
16 <indexterm><primary>ghci</primary>
19 <para>Interactive mode, which is also available as
20 <command>ghci</command>. Interactive mode is described in
21 more detail in <xref linkend="ghci">.</para>
26 <term><cmdsynopsis><command>ghc</command>
27 <arg choice=plain>––make</arg>
29 <indexterm><primary>make mode</primary>
31 <indexterm><primary><option>––make</option></primary>
34 <para>In this mode, GHC will build a multi-module Haskell
35 program automatically, figuring out dependencies for itself.
36 If you have a straightforward Haskell program, this is likely
37 to be much easier, and faster, than using
38 <command>make</command>.</para>
44 <command>ghc</command>
52 <indexterm><primary><option>-E</option></primary></indexterm>
53 <indexterm><primary><option>-C</option></primary></indexterm>
54 <indexterm><primary><option>-S</option></primary></indexterm>
55 <indexterm><primary><option>-c</option></primary></indexterm>
57 <para>This is the traditional batch-compiler mode, in which
58 GHC can compile source files one at a time, or link objects
59 together into an executable.</para>
65 <title>Options overview</title>
67 <para>GHC's behaviour is controlled by
68 <firstterm>options</firstterm>, which for historical reasons are
69 also sometimes referred to as command-line flags or arguments.
70 Options can be specified in three ways:</para>
73 <title>Command-line arguments</title>
75 <indexterm><primary>structure, command-line</primary></indexterm>
76 <indexterm><primary>command-line</primary><secondary>arguments</secondary></indexterm>
77 <indexterm><primary>arguments</primary><secondary>command-line</secondary></indexterm>
79 <para>An invocation of GHC takes the following form:</para>
85 <para>Command-line arguments are either options or file names.</para>
87 <para>Command-line options begin with <literal>-</literal>.
88 They may <emphasis>not</emphasis> be grouped:
89 <option>-vO</option> is different from <option>-v -O</option>.
90 Options need not precede filenames: e.g., <literal>ghc *.o -o
91 foo</literal>. All options are processed and then applied to
92 all files; you cannot, for example, invoke <literal>ghc -c -O1
93 Foo.hs -O2 Bar.hs</literal> to apply different optimisation
94 levels to the files <filename>Foo.hs</filename> and
95 <filename>Bar.hs</filename>.</para>
98 <Sect2 id="source-file-options">
99 <title>Command line options in source files</title>
101 <indexterm><primary>source-file options</primary></indexterm>
103 <para>Sometimes it is useful to make the connection between a
104 source file and the command-line options it requires quite
105 tight. For instance, if a Haskell source file uses GHC
106 extensions, it will always need to be compiled with the
107 <option>-fglasgow-exts</option> option. Rather than maintaining
108 the list of per-file options in a <filename>Makefile</filename>,
109 it is possible to do this directly in the source file using the
110 <literal>OPTIONS</literal> pragma <indexterm><primary>OPTIONS
111 pragma</primary></indexterm>:</para>
114 {-# OPTIONS -fglasgow-exts #-}
119 <para><literal>OPTIONS</literal> pragmas are only looked for at
120 the top of your source files, upto the first
121 (non-literate,non-empty) line not containing
122 <literal>OPTIONS</literal>. Multiple <literal>OPTIONS</literal>
123 pragmas are recognised. Note that your command shell does not
124 get to the source file options, they are just included literally
125 in the array of command-line arguments the compiler driver
126 maintains internally, so you'll be desperately disappointed if
127 you try to glob etc. inside <literal>OPTIONS</literal>.</para>
129 <para>NOTE: the contents of OPTIONS are prepended to the
130 command-line options, so you <emphasis>do</emphasis> have the
131 ability to override OPTIONS settings via the command
134 <para>It is not recommended to move all the contents of your
135 Makefiles into your source files, but in some circumstances, the
136 <literal>OPTIONS</literal> pragma is the Right Thing. (If you
137 use <option>-keep-hc-file-too</option> and have OPTION flags in
138 your module, the OPTIONS will get put into the generated .hc
143 <title>Setting options in GHCi</title>
145 <para>Options may also be modified from within GHCi, using the
146 <literal>:set</literal> command. See <xref linkend="ghci-set">
147 for more details.</para>
151 <sect1 id="static-dynamic-flags">
152 <title>Static vs. Dynamic options</title>
153 <indexterm><primary>static</primary><secondary>options</secondary>
155 <indexterm><primary>dynamic</primary><secondary>options</secondary>
158 <para>Each of GHC's command line options is classified as either
159 <firstterm>static</firstterm> or <firstterm>dynamic</firstterm>.
160 A static flag may only be specified on the command line, whereas a
161 dynamic flag may also be given in an <literal>OPTIONS</literal>
162 pragma in a source file or set from the GHCi command-line with
163 <literal>:set</literal>.</para>
165 <para>As a rule of thumb, all the language options are dynamic, as
166 are the warning options and the debugging options. The rest are
167 static, with the notable exceptions of <option>-v</option>,
168 <option>-cpp</option>, <option>-fasm</option>,
169 <option>-fvia-C</option>, and <option>-#include</option>.
171 The flag reference tables (<xref linkend="flag-reference">) lists
172 the status of each flag.</para>
175 <sect1 id="file-suffixes">
176 <title>Meaningful file suffixes</title>
178 <indexterm><primary>suffixes, file</primary></indexterm>
179 <indexterm><primary>file suffixes for GHC</primary></indexterm>
181 <para>File names with “meaningful” suffixes (e.g.,
182 <filename>.lhs</filename> or <filename>.o</filename>) cause the
183 “right thing” to happen to those files.</para>
188 <term><filename>.lhs</filename></term>
189 <indexterm><primary><literal>lhs</literal> suffix</primary></indexterm>
191 <para>A “literate Haskell” module.</para>
196 <term><filename>.hs</filename></term>
198 <para>A not-so-literate Haskell module.</para>
203 <term><filename>.hi</filename></term>
205 <para>A Haskell interface file, probably
206 compiler-generated.</para>
211 <term><filename>.hc</filename></term>
213 <para>Intermediate C file produced by the Haskell
219 <term><filename>.c</filename></term>
221 <para>A C file not produced by the Haskell
227 <term><filename>.s</filename></term>
229 <para>An assembly-language source file, usually produced by
235 <term><filename>.o</filename></term>
237 <para>An object file, produced by an assembler.</para>
242 <para>Files with other suffixes (or without suffixes) are passed
243 straight to the linker.</para>
247 <sect1 id="options-help">
248 <title>Help and verbosity options</title>
250 <IndexTerm><Primary>help options</Primary></IndexTerm>
251 <IndexTerm><Primary>verbosity options</Primary></IndexTerm>
255 <term><option>––help</option></term>
256 <term><option>-?</option></term>
257 <indexterm><primary><option>-?</option></primary></indexterm>
258 <indexterm><primary><option>––help</option></primary></indexterm>
260 <para>Cause GHC to spew a long usage message to standard
261 output and then exit.</para>
266 <term><option>-v</option></term>
267 <indexterm><primary><option>-v</option></primary></indexterm>
269 <para>The <option>-v</option> option makes GHC
270 <emphasis>verbose</emphasis>: it reports its version number
271 and shows (on stderr) exactly how it invokes each phase of
272 the compilation system. Moreover, it passes the
273 <option>-v</option> flag to most phases; each reports its
274 version number (and possibly some other information).</para>
276 <para>Please, oh please, use the <option>-v</option> option
277 when reporting bugs! Knowing that you ran the right bits in
278 the right order is always the first thing we want to
284 <term><option>-v</option><replaceable>n</replaceable></term>
285 <indexterm><primary><option>-v</option></primary></indexterm>
287 <para>To provide more control over the compiler's verbosity,
288 the <option>-v</option> flag takes an optional numeric
289 argument. Specifying <option>-v</option> on its own is
290 equivalent to <option>-v3</option>, and the other levels
291 have the following meanings:</para>
295 <term><option>-v0</option></term>
297 <para>Disable all non-essential messages (this is the
303 <term><option>-v1</option></term>
305 <para>Minimal verbosity: print one line per
306 compilation (this is the default when
307 <option>––make</option> or
308 <option>––interactive</option> is on).</para>
313 <term><option>-v2</option></term>
315 <para>Print the name of each compilation phase as it
316 is executed. (equivalent to
317 <option>-dshow-passes</option>).</para>
322 <term><option>-v3</option></term>
324 <para>The same as <option>-v2</option>, except that in
325 addition the full command line (if appropriate) for
326 each compilation phase is also printed.</para>
331 <term><option>-v4</option></term>
333 <para>The same as <option>-v3</option> except that the
334 intermediate program representation after each
335 compilation phase is also printed (excluding
336 preprocessed and C/assembly files).</para>
344 <term><option>––version</option></term>
345 <indexterm><primary><option>––version</option></primary></indexterm>
347 <para>Print a one-line string including GHC's version number.</para>
352 <term><option>––numeric-version</option></term>
353 <indexterm><primary><option>––numeric-version</option></primary></indexterm>
355 <para>Print GHC's numeric version number only.</para>
360 <term><option>––print-libdir</option></term>
361 <indexterm><primary><option>––print-libdir</option></primary></indexterm>
363 <para>Print the path to GHC's library directory. This is
364 the top of the directory tree containing GHC's libraries,
365 interfaces, and include files (usually something like
366 <literal>/usr/local/lib/ghc-5.04</literal> on Unix). This
368 <literal>$libdir</literal><indexterm><primary><literal>libdir</literal></primary>
369 </indexterm>in the package configuration file (see <xref
370 linkend="packages">).</para>
377 <sect1 id="make-mode">
378 <title>Using <command>ghc</command> <option>––make</option></title>
380 <indexterm><primary><option>––make</option></primary>
382 <indexterm><primary>separate compilation</primary>
385 <para>When given the <option>––make</option> option, GHC will
386 build a multi-module Haskell program by following dependencies
387 from a single root module (usually <literal>Main</literal>). For
388 example, if your <literal>Main</literal> module is in a file
389 called <filename>Main.hs</filename>, you could compile and link
390 the program like this:</para>
393 ghc ––make Main.hs
396 <para>The command line must contain one source file or module
397 name; GHC will figure out all the modules in the program by
398 following the imports from this initial module. It will then
399 attempt to compile each module which is out of date, and finally
400 if the top module is <literal>Main</literal>, the program
401 will also be linked into an executable.</para>
403 <para>The main advantages to using <literal>ghc ––make</literal>
404 over traditional <literal>Makefile</literal>s are:</para>
408 <para>GHC doesn't have to be restarted for each compilation,
409 which means it can cache information between compilations.
410 Compiling a muli-module program with <literal>ghc
411 ––make</literal> can be up to twice as fast as running
412 <literal>ghc</literal> individually on each source
416 <para>You don't have to write a
417 <literal>Makefile</literal>.</para>
419 <indexterm><primary><literal>Makefile</literal>s</primary><secondary>avoiding</secondary>
422 <para>GHC re-calculates the dependencies each time it is
423 invoked, so the dependencies never get out of sync with the
428 <para>Any of the command-line options described in the rest of
429 this chapter can be used with <option>––make</option>, but note
430 that any options you give on the command line will apply to all
431 the source files compiled, so if you want any options to apply to
432 a single source file only, you'll need to use an
433 <literal>OPTIONS</literal> pragma (see <xref
434 linkend="source-file-options">).</para>
436 <para>If the program needs to be linked with additional objects
437 (say, some auxilliary C code), these can be specified on the
438 command line as usual.</para>
440 <para>Note that GHC can only follow dependencies if it has the
441 source file available, so if your program includes a module for
442 which there is no source file, even if you have an object and an
443 interface file for the module, then GHC will complain. The
444 exception to this rule is for package modules, which may or may
445 not have source files.</para>
447 <para>The source files for the program don't all need to be in the
448 same directory; the <option>-i</option> option can be used to add
449 directories to the search path (see <xref
450 linkend="options-finding-imports">).</para>
454 <Sect1 id="options-order">
455 <title>GHC without <option>––make</option></title>
457 <para>Without <option>––make</option>, GHC will compile one or
458 more source files given on the command line.</para>
460 <para>The first phase to run is determined by each input-file
461 suffix, and the last phase is determined by a flag. If no
462 relevant flag is present, then go all the way through linking.
463 This table summarises:</para>
467 <colspec align="left">
468 <colspec align="left">
469 <colspec align="left">
470 <colspec align="left">
474 <entry>Phase of the compilation system</entry>
475 <entry>Suffix saying “start here”</entry>
476 <entry>Flag saying “stop after”</entry>
477 <entry>(suffix of) output file</entry>
482 <entry>literate pre-processor</entry>
483 <entry><literal>.lhs</literal></entry>
485 <entry><literal>.hs</literal></entry>
489 <entry>C pre-processor (opt.)
491 <entry><literal>.hs</literal> (with
492 <option>-cpp</option>)</entry>
493 <entry><option>-E</option></entry>
494 <entry><literal>.hspp</literal></entry>
498 <entry>Haskell compiler</entry>
499 <entry><literal>.hs</literal></entry>
500 <entry><option>-C</option>, <option>-S</option></entry>
501 <entry><literal>.hc</literal>, <literal>.s</literal></entry>
505 <entry>C compiler (opt.)</entry>
506 <entry><literal>.hc</literal> or <literal>.c</literal></entry>
507 <entry><option>-S</option></entry>
508 <entry><literal>.s</literal></entry>
512 <entry>assembler</entry>
513 <entry><literal>.s</literal></entry>
514 <entry><option>-c</option></entry>
515 <entry><literal>.o</literal></entry>
519 <entry>linker</entry>
520 <entry><replaceable>other</replaceable></entry>
522 <entry><filename>a.out</filename></entry>
528 <indexterm><primary><option>-C</option></primary></indexterm>
529 <indexterm><primary><option>-E</option></primary></indexterm>
530 <indexterm><primary><option>-S</option></primary></indexterm>
531 <indexterm><primary><option>-c</option></primary></indexterm>
533 <para>Thus, a common invocation would be: <literal>ghc -c
534 Foo.hs</literal></para>
536 <para>Note: What the Haskell compiler proper produces depends on
537 whether a native-code generator<indexterm><primary>native-code
538 generator</primary></indexterm> is used (producing assembly
539 language) or not (producing C). See <xref
540 linkend="options-codegen"> for more details.</para>
542 <para>Note: C pre-processing is optional, the
543 <option>-ccp</option><indexterm><primary><option>-cpp</option></primary>
544 </indexterm>flag turns it on. See <xref
545 linkend="c-pre-processor"> for more details.</para>
547 <para>Note: The option <option>-E</option><IndexTerm><Primary>-E
548 option</Primary></IndexTerm> runs just the pre-processing passes
549 of the compiler, dumping the result in a file. Note that this
550 differs from the previous behaviour of dumping the file to
551 standard output.</para>
554 <sect1 id="options-output">
555 <title>Re-directing the compilation output(s)</title>
557 <indexterm><primary>output-directing options</primary></indexterm>
558 <indexterm><primary>redirecting compilation output</primary></indexterm>
563 <term><option>-o</option></term>
564 <indexterm><primary><option>-o</option></primary></indexterm>
566 <para>GHC's compiled output normally goes into a
567 <filename>.hc</filename>, <filename>.o</filename>, etc.,
568 file, depending on the last-run compilation phase. The
569 option <option>-o foo</option><IndexTerm><Primary>-o
570 option</Primary></IndexTerm> re-directs the output of that
571 last-run phase to file <filename>foo</filename>.</para>
573 <para>Note: this “feature” can be
574 counterintuitive: <command>ghc -C -o foo.o foo.hs</command>
575 will put the intermediate C code in the file
576 <filename>foo.o</filename>, name notwithstanding!</para>
581 <term><option>-odir</option></term>
582 <indexterm><primary><option>-odir</option></primary></indexterm>
584 <para>The <option>-o</option> option isn't of much use if
585 you have <emphasis>several</emphasis> input files…
586 Non-interface output files are normally put in the same
587 directory as their corresponding input file came from. You
588 may specify that they be put in another directory using the
589 <option>-odir <dir></option><IndexTerm><Primary>-odir
590 <dir> option</Primary></IndexTerm> (the “Oh,
591 dear” option). For example:</para>
594 % ghc -c parse/Foo.hs parse/Bar.hs gurgle/Bumble.hs -odir `arch`
597 <para>The output files, <filename>Foo.o</filename>,
598 <filename>Bar.o</filename>, and
599 <filename>Bumble.o</filename> would be put into a
600 subdirectory named after the architecture of the executing
601 machine (<filename>sun4</filename>,
602 <filename>mips</filename>, etc). The directory must already
603 exist; it won't be created.</para>
605 <para>Note that the <option>-odir</option> option does
606 <emphasis>not</emphasis> affect where the interface files
607 are put. In the above example, they would still be put in
608 <filename>parse/Foo.hi</filename>,
609 <filename>parse/Bar.hi</filename>, and
610 <filename>gurgle/Bumble.hi</filename>.</para>
615 <term><option>-ohi</option> <replaceable>file</replaceable></term>
616 <indexterm><primary><option>-ohi</option></primary>
619 <para>The interface output may be directed to another file
620 <filename>bar2/Wurble.iface</filename> with the option
621 <option>-ohi bar2/Wurble.iface</option> (not
624 <para>WARNING: if you redirect the interface file somewhere
625 that GHC can't find it, then the recompilation checker may
626 get confused (at the least, you won't get any recompilation
627 avoidance). We recommend using a combination of
628 <option>-hidir</option> and <option>-hisuf</option> options
629 instead, if possible.</para>
631 <para>To avoid generating an interface at all, you could use
632 this option to redirect the interface into the bit bucket:
633 <literal>-ohi /dev/null</literal>, for example.</para>
638 <term><option>-hidir</option> <replaceable>directory</replaceable></term>
639 <indexterm><primary><option>-hidir</option></primary>
642 <para>Redirects all generated interface files into
643 <replaceable>directory</replaceable>, instead of the default
644 which is to place the interface file in the same directory
645 as the source file.</para>
650 <term><option>-osuf</option> <replaceable>suffix</replaceable></term>
651 <term><option>-hisuf</option> <replaceable>suffix</replaceable></term>
652 <term><option>-hcsuf</option> <replaceable>suffix</replaceable></term>
653 <indexterm><primary><option>-osuf</option></primary></indexterm>
654 <indexterm><primary><option>-hisuf</option></primary></indexterm>
655 <indexterm><primary><option>-hcsuf</option></primary></indexterm>
657 <para>EXOTICA: The <option>-osuf</option>
658 <replaceable>suffix</replaceable> will change the
659 <literal>.o</literal> file suffix for object files to
660 whatever you specify. We use this when compiling libraries,
661 so that objects for the profiling versions of the libraries
662 don't clobber the normal ones.</para>
664 <para>Similarly, the <option>-hisuf</option>
665 <replaceable>suffix</replaceable> will change the
666 <literal>.hi</literal> file suffix for non-system interface
667 files (see <XRef LinkEnd="hi-options">).</para>
669 <para>Finally, the option <option>-hcsuf</option>
670 <replaceable>suffix</replaceable> will change the
671 <literal>.hc</literal> file suffix for compiler-generated
672 intermediate C files.</para>
674 <para>The <option>-hisuf</option>/<option>-osuf</option>
675 game is useful if you want to compile a program with both
676 GHC and HBC (say) in the same directory. Let HBC use the
677 standard <filename>.hi</filename>/<filename>.o</filename>
678 suffixes; add <option>-hisuf g_hi -osuf
679 g_o</option> to your <command>make</command> rule for
680 GHC compiling…</para>
685 <sect2 id="keeping-intermediates">
686 <title>Keeping Intermediate Files</title>
687 <indexterm><primary>intermediate files, saving</primary>
689 <indexterm><primary><literal>.hc</literal> files, saving</primary>
691 <indexterm><primary><literal>.s</literal> files, saving</primary>
695 <para>The following options are useful for keeping certain
696 intermediate files around, when normally GHC would throw these
697 away after compilation:</para>
701 <term><option>-keep-hc-files</option></term>
703 <primary><option>-keep-hc-files</option></primary>
706 <para>Keep intermediate <literal>.hc</literal> files when
707 doing <literal>.hs</literal>-to-<literal>.o</literal>
708 compilations via C (NOTE: <literal>.hc</literal> files
709 aren't generated when using the native code generator, you
710 may need to use <option>-fvia-C</option> to force them
711 to be produced).</para>
716 <term><option>-keep-s-files</option></term>
718 <primary><option>-keep-s-files</option></primary>
721 <para>Keep intermediate <literal>.s</literal> files.</para>
726 <term><option>-keep-raw-s-files</option></term>
728 <primary><option>-keep-raw-s-files</option></primary>
731 <para>Keep intermediate <literal>.raw-s</literal> files.
732 These are the direct output from the C compiler, before
733 GHC does “assembly mangling” to produce the
734 <literal>.s</literal> file. Again, these are not produced
735 when using the native code generator.</para>
740 <term><option>-keep-tmp-files</option></term>
742 <primary><option>-keep-tmp-files</option></primary>
745 <primary>temporary files</primary>
746 <secondary>keeping</secondary>
749 <para>Instructs the GHC driver not to delete any of its
750 temporary files, which it normally keeps in
751 <literal>/tmp</literal> (or possibly elsewhere; see <xref
752 linkend="temp-files">). Running GHC with
753 <option>-v</option> will show you what temporary files
754 were generated along the way.</para>
760 <sect2 id="temp-files">
761 <title>Redirecting temporary files</title>
764 <primary>temporary files</primary>
765 <secondary>redirecting</secondary>
770 <term><option>-tmpdir</option></term>
771 <indexterm><primary><option>-tmpdir</option></primary></indexterm>
773 <para>If you have trouble because of running out of space
774 in <filename>/tmp</filename> (or wherever your
775 installation thinks temporary files should go), you may
776 use the <option>-tmpdir
777 <dir></option><IndexTerm><Primary>-tmpdir
778 <dir> option</Primary></IndexTerm> option to specify
779 an alternate directory. For example, <option>-tmpdir
780 .</option> says to put temporary files in the current
781 working directory.</para>
783 <para>Alternatively, use your <Constant>TMPDIR</Constant>
784 environment variable.<IndexTerm><Primary>TMPDIR
785 environment variable</Primary></IndexTerm> Set it to the
786 name of the directory where temporary files should be put.
787 GCC and other programs will honour the
788 <Constant>TMPDIR</Constant> variable as well.</para>
790 <para>Even better idea: Set the
791 <Constant>DEFAULT_TMPDIR</Constant> make variable when
792 building GHC, and never worry about
793 <Constant>TMPDIR</Constant> again. (see the build
794 documentation).</para>
802 <sect1 id="options-sanity">
803 <title>Warnings and sanity-checking</title>
805 <indexterm><primary>sanity-checking options</primary></indexterm>
806 <indexterm><primary>warnings</primary></indexterm>
809 <para>GHC has a number of options that select which types of
810 non-fatal error messages, otherwise known as warnings, can be
811 generated during compilation. By default, you get a standard set
812 of warnings which are generally likely to indicate bugs in your
814 <option>-fwarn-overlpapping-patterns</option>,
815 <option>-fwarn-deprecations</option>,
816 <option>-fwarn-duplicate-exports</option>,
817 <option>-fwarn-missing-fields</option>, and
818 <option>-fwarn-missing-methods</option>. The following flags are
819 simple ways to select standard “packages” of warnings:
825 <term><option>-W</option>:</term>
827 <IndexTerm><Primary>-W option</Primary></IndexTerm>
828 <para>Provides the standard warnings plus
829 <option>-fwarn-incomplete-patterns</option>,
830 <option>-fwarn-unused-matches</option>,
831 <option>-fwarn-unused-imports</option>,
832 <option>-fwarn-misc</option>, and
833 <option>-fwarn-unused-binds</option>.</para>
838 <term><option>-w</option>:</term>
840 <IndexTerm><Primary><option>-w</option></Primary></IndexTerm>
841 <para>Turns off all warnings, including the standard ones.</para>
846 <term><option>-Wall</option>:</term>
848 <indexterm><primary><option>-Wall</option></primary></indexterm>
849 <para>Turns on all warning options.</para>
855 <para>The full set of warning options is described below. To turn
856 off any warning, simply give the corresponding
857 <option>-fno-warn-...</option> option on the command line.</para>
862 <term><option>-fwarn-deprecations</option>:</term>
864 <indexterm><primary><option>-fwarn-deprecations</option></primary>
866 <indexterm><primary>deprecations</primary></indexterm>
867 <para>Causes a warning to be emitted when a deprecated
868 function or type is used. Entities can be marked as
869 deprecated using a pragma, see <xref
870 linkend="deprecated-pragma">.</para>
875 <term><option>-fwarn-duplicate-exports</option>:</term>
877 <indexterm><primary><option>-fwarn-duplicate-exports</option></primary></indexterm>
878 <indexterm><primary>duplicate exports, warning</primary></indexterm>
879 <indexterm><primary>export lists, duplicates</primary></indexterm>
881 <para>Have the compiler warn about duplicate entries in
882 export lists. This is useful information if you maintain
883 large export lists, and want to avoid the continued export
884 of a definition after you've deleted (one) mention of it in
885 the export list.</para>
887 <para>This option is on by default.</para>
892 <term><option>-fwarn-hi-shadowing</option>:</term>
894 <indexterm><primary><option>-fwarn-hi-shadowing</option></primary></indexterm>
895 <indexterm><primary>shadowing</primary>
896 <secondary>interface files</secondary></indexterm>
898 <para>Causes the compiler to emit a warning when a module or
899 interface file in the current directory is shadowing one
900 with the same module name in a library or other
906 <term><option>-fwarn-incomplete-patterns</option>:</term>
908 <indexterm><primary><option>-fwarn-incomplete-patterns</option></primary></indexterm>
909 <indexterm><primary>incomplete patterns, warning</primary></indexterm>
910 <indexterm><primary>patterns, incomplete</primary></indexterm>
912 <para>Similarly for incomplete patterns, the function
913 <function>g</function> below will fail when applied to
914 non-empty lists, so the compiler will emit a warning about
915 this when <option>-fwarn-incomplete-patterns</option> is
922 <para>This option isn't enabled be default because it can be
923 a bit noisy, and it doesn't always indicate a bug in the
924 program. However, it's generally considered good practice
925 to cover all the cases in your functions.</para>
930 <term><option>-fwarn-misc</option>:</term>
931 <indexterm><primary><option>-fwarn-misc</option></primary></indexterm>
933 <para>Turns on warnings for various harmless but untidy
934 things. This currently includes: importing a type with
935 <literal>(..)</literal> when the export is abstract, and
936 listing duplicate class assertions in a qualified type.</para>
941 <term><option>-fwarn-missing-fields</option>:</term>
943 <indexterm><primary><option>-fwarn-missing-fields</option></primary></indexterm>
944 <indexterm><primary>missing fields, warning</primary></indexterm>
945 <indexterm><primary>fields, missing</primary></indexterm>
947 <para>This option is on by default, and warns you whenever
948 the construction of a labelled field constructor isn't
949 complete, missing initializers for one or more fields. While
950 not an error (the missing fields are initialised with
951 bottoms), it is often an indication of a programmer error.</para>
956 <term><option>-fwarn-missing-methods</option>:</term>
958 <indexterm><primary><option>-fwarn-missing-methods</option></primary></indexterm>
959 <indexterm><primary>missing methods, warning</primary></indexterm>
960 <indexterm><primary>methods, missing</primary></indexterm>
962 <para>This option is on by default, and warns you whenever
963 an instance declaration is missing one or more methods, and
964 the corresponding class declaration has no default
965 declaration for them.</para>
970 <term><option>-fwarn-missing-signatures</option>:</term>
972 <indexterm><primary><option>-fwarn-missing-signatures</option></primary></indexterm>
973 <indexterm><primary>type signatures, missing</primary></indexterm>
975 <para>If you would like GHC to check that every top-level
976 function/value has a type signature, use the
977 <option>-fwarn-missing-signatures</option> option. This
978 option is off by default.</para>
983 <term><option>-fwarn-name-shadowing</option>:</term>
985 <indexterm><primary><option>-fwarn-name-shadowing</option></primary></indexterm>
986 <indexterm><primary>shadowing, warning</primary></indexterm>
988 <para>This option causes a warning to be emitted whenever an
989 inner-scope value has the same name as an outer-scope value,
990 i.e. the inner value shadows the outer one. This can catch
991 typographical errors that turn into hard-to-find bugs, e.g.,
992 in the inadvertent cyclic definition <literal>let x = ... x
993 ... in</literal>.</para>
995 <para>Consequently, this option does
996 <emphasis>will</emphasis> complain about cyclic recursive
1002 <term><option>-fwarn-overlapping-patterns</option>:</term>
1003 <indexterm><primary><option>-fwarn-overlapping-patterns</option></primary></indexterm>
1004 <indexterm><primary>overlapping patterns, warning</primary></indexterm>
1005 <indexterm><primary>patterns, overlapping</primary></indexterm>
1007 <para>By default, the compiler will warn you if a set of
1008 patterns are overlapping, i.e.,</para>
1011 f :: String -> Int
1017 <para>where the last pattern match in <Function>f</Function>
1018 won't ever be reached, as the second pattern overlaps
1019 it. More often than not, redundant patterns is a programmer
1020 mistake/error, so this option is enabled by default.</para>
1025 <term><option>-fwarn-simple-patterns</option>:</term>
1027 <indexterm><primary><option>-fwarn-simple-patterns</option></primary>
1029 <para>Causes the compiler to warn about lambda-bound
1030 patterns that can fail, eg. <literal>\(x:xs)->...</literal>.
1031 Normally, these aren't treated as incomplete patterns by
1032 <option>-fwarn-incomplete-patterns</option>.</para>
1037 <term><option>-fwarn-type-defaults</option>:</term>
1039 <indexterm><primary><option>-fwarn-type-defaults</option></primary></indexterm>
1040 <indexterm><primary>defaulting mechanism, warning</primary></indexterm>
1041 <para>Have the compiler warn/inform you where in your source
1042 the Haskell defaulting mechanism for numeric types kicks
1043 in. This is useful information when converting code from a
1044 context that assumed one default into one with another,
1045 e.g., the `default default' for Haskell 1.4 caused the
1046 otherwise unconstrained value <Constant>1</Constant> to be
1047 given the type <literal>Int</literal>, whereas Haskell 98
1048 defaults it to <literal>Integer</literal>. This may lead to
1049 differences in performance and behaviour, hence the
1050 usefulness of being non-silent about this.</para>
1052 <para>This warning is off by default.</para>
1057 <term><option>-fwarn-unused-binds</option>:</term>
1059 <indexterm><primary><option>-fwarn-unused-binds</option></primary></indexterm>
1060 <indexterm><primary>unused binds, warning</primary></indexterm>
1061 <indexterm><primary>binds, unused</primary></indexterm>
1062 <para>Report any function definitions (and local bindings)
1063 which are unused. For top-level functions, the warning is
1064 only given if the binding is not exported.</para>
1069 <term><option>-fwarn-unused-imports</option>:</term>
1071 <indexterm><primary><option>-fwarn-unused-imports</option></primary></indexterm>
1072 <indexterm><primary>unused imports, warning</primary></indexterm>
1073 <indexterm><primary>imports, unused</primary></indexterm>
1075 <para>Report any objects that are explicitly imported but
1081 <term><option>-fwarn-unused-matches</option>:</term>
1083 <indexterm><primary><option>-fwarn-unused-matches</option></primary></indexterm>
1084 <indexterm><primary>unused matches, warning</primary></indexterm>
1085 <indexterm><primary>matches, unused</primary></indexterm>
1087 <para>Report all unused variables which arise from pattern
1088 matches, including patterns consisting of a single variable.
1089 For instance <literal>f x y = []</literal> would report
1090 <VarName>x</VarName> and <VarName>y</VarName> as unused. To
1091 eliminate the warning, all unused variables can be replaced
1092 with wildcards.</para>
1098 <para>If you're feeling really paranoid, the
1099 <option>-dcore-lint</option>
1100 option<indexterm><primary><option>-dcore-lint</option></primary></indexterm>
1101 is a good choice. It turns on heavyweight intra-pass
1102 sanity-checking within GHC. (It checks GHC's sanity, not
1110 <sect1 id="options-optimise">
1111 <title>Optimisation (code improvement)</title>
1113 <indexterm><primary>optimisation</primary></indexterm>
1114 <indexterm><primary>improvement, code</primary></indexterm>
1116 <para>The <option>-O*</option> options specify convenient
1117 “packages” of optimisation flags; the
1118 <option>-f*</option> options described later on specify
1119 <emphasis>individual</emphasis> optimisations to be turned on/off;
1120 the <option>-m*</option> options specify
1121 <emphasis>machine-specific</emphasis> optimisations to be turned
1124 <sect2 id="optimise-pkgs">
1125 <title><option>-O*</option>: convenient “packages” of optimisation flags.</title>
1127 <para>There are <emphasis>many</emphasis> options that affect
1128 the quality of code produced by GHC. Most people only have a
1129 general goal, something like “Compile quickly” or
1130 “Make my program run like greased lightning.” The
1131 following “packages” of optimisations (or lack
1132 thereof) should suffice.</para>
1134 <para>Once you choose a <option>-O*</option>
1135 “package,” stick with it—don't chop and
1136 change. Modules' interfaces <emphasis>will</emphasis> change
1137 with a shift to a new <option>-O*</option> option, and you may
1138 have to recompile a large chunk of all importing modules before
1139 your program can again be run safely (see <XRef
1140 LinkEnd="recomp">).</para>
1145 <term>No <option>-O*</option>-type option specified:</term>
1146 <indexterm><primary>-O* not specified</primary></indexterm>
1148 <para>This is taken to mean: “Please compile
1149 quickly; I'm not over-bothered about compiled-code
1150 quality.” So, for example: <command>ghc -c
1151 Foo.hs</command></para>
1156 <term><option>-O0</option>:</term>
1157 <indexterm><primary><option>-O0</option></primary></indexterm>
1159 <para>Means “turn off all optimisation”,
1160 reverting to the same settings as if no
1161 <option>-O</option> options had been specified. Saying
1162 <option>-O0</option> can be useful if
1163 eg. <command>make</command> has inserted a
1164 <option>-O</option> on the command line already.</para>
1169 <term><option>-O</option> or <option>-O1</option>:</term>
1170 <indexterm><primary>-O option</primary></indexterm>
1171 <indexterm><primary>-O1 option</primary></indexterm>
1172 <indexterm><primary>optimise</primary><secondary>normally</secondary></indexterm>
1174 <para>Means: “Generate good-quality code without
1175 taking too long about it.” Thus, for example:
1176 <command>ghc -c -O Main.lhs</command></para>
1181 <term><option>-O2</option>:</term>
1182 <indexterm><primary>-O2 option</primary></indexterm>
1183 <indexterm><primary>optimise</primary><secondary>aggressively</secondary></indexterm>
1185 <para>Means: “Apply every non-dangerous
1186 optimisation, even if it means significantly longer
1187 compile times.”</para>
1189 <para>The avoided “dangerous” optimisations
1190 are those that can make runtime or space
1191 <emphasis>worse</emphasis> if you're unlucky. They are
1192 normally turned on or off individually.</para>
1194 <para>At the moment, <option>-O2</option> is
1195 <emphasis>unlikely</emphasis> to produce better code than
1196 <option>-O</option>.</para>
1201 <term><option>-Ofile <file></option>:</term>
1202 <indexterm><primary>-Ofile <file> option</primary></indexterm>
1203 <indexterm><primary>optimising, customised</primary></indexterm>
1205 <para>(NOTE: not supported yet in GHC 5.x. Please ask if
1206 you're interested in this.)</para>
1208 <para>For those who need <emphasis>absolute</emphasis>
1209 control over <emphasis>exactly</emphasis> what options are
1210 used (e.g., compiler writers, sometimes :-), a list of
1211 options can be put in a file and then slurped in with
1212 <option>-Ofile</option>.</para>
1214 <para>In that file, comments are of the
1215 <literal>#</literal>-to-end-of-line variety; blank
1216 lines and most whitespace is ignored.</para>
1218 <para>Please ask if you are baffled and would like an
1219 example of <option>-Ofile</option>!</para>
1224 <para>We don't use a <option>-O*</option> flag for day-to-day
1225 work. We use <option>-O</option> to get respectable speed;
1226 e.g., when we want to measure something. When we want to go for
1227 broke, we tend to use <option>-O -fvia-C</option> (and we go for
1228 lots of coffee breaks).</para>
1230 <para>The easiest way to see what <option>-O</option> (etc.)
1231 “really mean” is to run with <option>-v</option>,
1232 then stand back in amazement.</para>
1235 <sect2 id="options-f">
1236 <title><option>-f*</option>: platform-independent flags</title>
1238 <indexterm><primary>-f* options (GHC)</primary></indexterm>
1239 <indexterm><primary>-fno-* options (GHC)</primary></indexterm>
1241 <para>These flags turn on and off individual optimisations.
1242 They are normally set via the <option>-O</option> options
1243 described above, and as such, you shouldn't need to set any of
1244 them explicitly (indeed, doing so could lead to unexpected
1245 results). However, there are one or two that may be of
1250 <term><option>-fexcess-precision</option>:</term>
1252 <indexterm><primary><option>-fexcess-precision</option></primary></indexterm>
1253 <para>When this option is given, intermediate floating
1254 point values can have a <emphasis>greater</emphasis>
1255 precision/range than the final type. Generally this is a
1256 good thing, but some programs may rely on the exact
1258 <literal>Float</literal>/<literal>Double</literal> values
1259 and should not use this option for their compilation.</para>
1264 <term><option>-fignore-asserts</option>:</term>
1266 <indexterm><primary><option>-fignore-asserts</option></primary></indexterm>
1267 <para>Causes GHC to ignore uses of the function
1268 <literal>Exception.assert</literal> in source code (in
1269 other words, rewriting <literal>Exception.assert p
1270 e</literal> to <literal>e</literal> (see <xref
1271 linkend="sec-assertions">). This flag is turned on by
1272 <option>-O</option>.
1278 <term><option>-fno-strictness</option></term>
1279 <indexterm><primary><option>-fno-strictness</option></primary>
1282 <para>Turns off the strictness analyser; sometimes it eats
1283 too many cycles.</para>
1288 <term><option>-fno-cpr-analyse</option></term>
1289 <indexterm><primary><option>-fno-cpr-analyse</option></primary>
1292 <para>Turns off the CPR (constructed product result)
1293 analysis; it is somewhat experimental.</para>
1298 <term><option>-funbox-strict-fields</option>:</term>
1300 <indexterm><primary><option>-funbox-strict-fields</option></primary></indexterm>
1301 <indexterm><primary>strict constructor fields</primary></indexterm>
1302 <indexterm><primary>constructor fields, strict</primary></indexterm>
1304 <para>This option causes all constructor fields which are
1305 marked strict (i.e. “!”) to be unboxed or
1306 unpacked if possible. For example:</para>
1309 data T = T !Float !Float
1312 <para>will create a constructor <literal>T</literal>
1313 containing two unboxed floats if the
1314 <option>-funbox-strict-fields</option> flag is given.
1315 This may not always be an optimisation: if the
1316 <Function>T</Function> constructor is scrutinised and the
1317 floats passed to a non-strict function for example, they
1318 will have to be reboxed (this is done automatically by the
1321 <para>This option should only be used in conjunction with
1322 <option>-O</option>, in order to expose unfoldings to the
1323 compiler so the reboxing can be removed as often as
1324 possible. For example:</para>
1328 f (T f1 f2) = f1 + f2
1331 <para>The compiler will avoid reboxing
1332 <Function>f1</Function> and <Function>f2</Function> by
1333 inlining <Function>+</Function> on floats, but only when
1334 <option>-O</option> is on.</para>
1336 <para>Any single-constructor data is eligible for
1337 unpacking; for example</para>
1340 data T = T !(Int,Int)
1343 <para>will store the two <literal>Int</literal>s directly
1344 in the <Function>T</Function> constructor, by flattening
1345 the pair. Multi-level unpacking is also supported:</para>
1349 data S = S !Int !Int
1352 <para>will store two unboxed <literal>Int#</literal>s
1353 directly in the <Function>T</Function> constructor.</para>
1358 <term><option>-funfolding-update-in-place<n></option></term>
1359 <indexterm><primary><option>-funfolding-update-in-place</option></primary></indexterm>
1361 <para>Switches on an experimental "optimisation".
1362 Switching it on makes the compiler a little keener to
1363 inline a function that returns a constructor, if the
1364 context is that of a thunk.
1368 If we inlined plusInt we might get an opportunity to use
1369 update-in-place for the thunk 'x'.</para>
1374 <term><option>-funfolding-creation-threshold<n></option>:</term>
1376 <indexterm><primary><option>-funfolding-creation-threshold</option></primary></indexterm>
1377 <indexterm><primary>inlining, controlling</primary></indexterm>
1378 <indexterm><primary>unfolding, controlling</primary></indexterm>
1380 <para>(Default: 45) Governs the maximum size that GHC will
1381 allow a function unfolding to be. (An unfolding has a
1382 “size” that reflects the cost in terms of
1383 “code bloat” of expanding that unfolding at
1384 at a call site. A bigger function would be assigned a
1385 bigger cost.) </para>
1387 <para> Consequences: (a) nothing larger than this will be
1388 inlined (unless it has an INLINE pragma); (b) nothing
1389 larger than this will be spewed into an interface
1393 <para> Increasing this figure is more likely to result in longer
1394 compile times than faster code. The next option is more
1400 <term><option>-funfolding-use-threshold<n></option>:</term>
1402 <indexterm><primary><option>-funfolding-use-threshold</option></primary></indexterm>
1403 <indexterm><primary>inlining, controlling</primary></indexterm>
1404 <indexterm><primary>unfolding, controlling</primary></indexterm>
1406 <para>(Default: 8) This is the magic cut-off figure for
1407 unfolding: below this size, a function definition will be
1408 unfolded at the call-site, any bigger and it won't. The
1409 size computed for a function depends on two things: the
1410 actual size of the expression minus any discounts that
1411 apply (see <option>-funfolding-con-discount</option>).</para>
1422 <sect1 id="sec-using-concurrent">
1423 <title>Using Concurrent Haskell</title>
1425 <indexterm><primary>Concurrent Haskell—use</primary></indexterm>
1428 GHC supports Concurrent Haskell by default, without requiring a
1429 special option or libraries compiled in a certain way. To get access
1430 to the support libraries for Concurrent Haskell, just import
1431 <literal>Control.Concurrent</literal> (details are in the accompanying
1432 library documentation).</para>
1435 RTS options are provided for modifying the behaviour of the threaded
1436 runtime system. See <XRef LinkEnd="parallel-rts-opts">.
1440 Concurrent Haskell is described in more detail in the documentation
1441 for the <literal>Control.Concurrent</literal> module.
1446 <Sect1 id="sec-using-parallel">
1447 <title>Using Parallel Haskell</title>
1450 <indexterm><primary>Parallel Haskell—use</primary></indexterm>
1454 [You won't be able to execute parallel Haskell programs unless PVM3
1455 (Parallel Virtual Machine, version 3) is installed at your site.]
1459 To compile a Haskell program for parallel execution under PVM, use the
1460 <Option>-parallel</Option> option,<IndexTerm><Primary>-parallel
1461 option</Primary></IndexTerm> both when compiling <Emphasis>and
1462 linking</Emphasis>. You will probably want to <Literal>import
1463 Parallel</Literal> into your Haskell modules.
1467 To run your parallel program, once PVM is going, just invoke it
1468 “as normal”. The main extra RTS option is
1469 <Option>-qp<n></Option>, to say how many PVM
1470 “processors” your program to run on. (For more details of
1471 all relevant RTS options, please see <XRef
1472 LinkEnd="parallel-rts-opts">.)
1476 In truth, running Parallel Haskell programs and getting information
1477 out of them (e.g., parallelism profiles) is a battle with the vagaries of
1478 PVM, detailed in the following sections.
1481 <Sect2 id="pvm-dummies">
1482 <Title>Dummy's guide to using PVM</Title>
1485 <indexterm><primary>PVM, how to use</primary></indexterm>
1486 <indexterm><primary>Parallel Haskell—PVM use</primary></indexterm>
1487 Before you can run a parallel program under PVM, you must set the
1488 required environment variables (PVM's idea, not ours); something like,
1489 probably in your <filename>.cshrc</filename> or equivalent:
1492 setenv PVM_ROOT /wherever/you/put/it
1493 setenv PVM_ARCH `$PVM_ROOT/lib/pvmgetarch`
1494 setenv PVM_DPATH $PVM_ROOT/lib/pvmd
1500 Creating and/or controlling your “parallel machine” is a purely-PVM
1501 business; nothing specific to Parallel Haskell. The following paragraphs
1502 describe how to configure your parallel machine interactively.
1506 If you use parallel Haskell regularly on the same machine configuration it
1507 is a good idea to maintain a file with all machine names and to make the
1508 environment variable PVM_HOST_FILE point to this file. Then you can avoid
1509 the interactive operations described below by just saying
1517 You use the <Command>pvm</Command><IndexTerm><Primary>pvm command</Primary></IndexTerm> command to start PVM on your
1518 machine. You can then do various things to control/monitor your
1519 “parallel machine;” the most useful being:
1525 <ColSpec Align="Left">
1529 <entry><KeyCombo><KeyCap>Control</KeyCap><KeyCap>D</KeyCap></KeyCombo></entry>
1530 <entry>exit <command>pvm</command>, leaving it running</entry>
1534 <entry><command>halt</command></entry>
1535 <entry>kill off this “parallel machine” & exit</entry>
1539 <entry><command>add <host></command></entry>
1540 <entry>add <command><host></command> as a processor</entry>
1544 <entry><command>delete <host></command></entry>
1545 <entry>delete <command><host></command></entry>
1549 <entry><command>reset</command></entry>
1550 <entry>kill what's going, but leave PVM up</entry>
1554 <entry><command>conf</command></entry>
1555 <entry>list the current configuration</entry>
1559 <entry><command>ps</command></entry>
1560 <entry>report processes' status</entry>
1564 <entry><command>pstat <pid></command></entry>
1565 <entry>status of a particular process</entry>
1574 The PVM documentation can tell you much, much more about <command>pvm</command>!
1579 <Sect2 id="par-profiles">
1580 <Title>Parallelism profiles</Title>
1583 <indexterm><primary>parallelism profiles</primary></indexterm>
1584 <indexterm><primary>profiles, parallelism</primary></indexterm>
1585 <indexterm><primary>visualisation tools</primary></indexterm>
1589 With Parallel Haskell programs, we usually don't care about the
1590 results—only with “how parallel” it was! We want pretty pictures.
1594 Parallelism profiles (à la <Command>hbcpp</Command>) can be generated with the
1595 <Option>-qP</Option><IndexTerm><Primary>-qP RTS option (concurrent, parallel)</Primary></IndexTerm> RTS option. The
1596 per-processor profiling info is dumped into files named
1597 <Filename><full-path><program>.gr</Filename>. These are then munged into a PostScript picture,
1598 which you can then display. For example, to run your program
1599 <Filename>a.out</Filename> on 8 processors, then view the parallelism profile, do:
1605 <prompt>$</prompt> ./a.out +RTS -qP -qp8
1606 <prompt>$</prompt> grs2gr *.???.gr > temp.gr # combine the 8 .gr files into one
1607 <prompt>$</prompt> gr2ps -O temp.gr # cvt to .ps; output in temp.ps
1608 <prompt>$</prompt> ghostview -seascape temp.ps # look at it!
1614 The scripts for processing the parallelism profiles are distributed
1615 in <filename>ghc/utils/parallel/</filename>.
1621 <Title>Other useful info about running parallel programs</Title>
1624 The “garbage-collection statistics” RTS options can be useful for
1625 seeing what parallel programs are doing. If you do either
1626 <Option>+RTS -Sstderr</Option><IndexTerm><Primary>-Sstderr RTS option</Primary></IndexTerm> or <Option>+RTS -sstderr</Option>, then
1627 you'll get mutator, garbage-collection, etc., times on standard
1628 error. The standard error of all PE's other than the `main thread'
1629 appears in <filename>/tmp/pvml.nnn</filename>, courtesy of PVM.
1633 Whether doing <option>+RTS -Sstderr</option> or not, a handy way to watch
1634 what's happening overall is: <command>tail -f /tmp/pvml.nnn</command>.
1639 <Sect2 id="parallel-rts-opts">
1640 <title>RTS options for Concurrent/Parallel Haskell
1644 <indexterm><primary>RTS options, concurrent</primary></indexterm>
1645 <indexterm><primary>RTS options, parallel</primary></indexterm>
1646 <indexterm><primary>Concurrent Haskell—RTS options</primary></indexterm>
1647 <indexterm><primary>Parallel Haskell—RTS options</primary></indexterm>
1651 Besides the usual runtime system (RTS) options
1652 (<XRef LinkEnd="runtime-control">), there are a few options particularly
1653 for concurrent/parallel execution.
1660 <Term><Option>-qp<N></Option>:</Term>
1663 <IndexTerm><Primary>-qp<N> RTS option</Primary></IndexTerm>
1664 (PARALLEL ONLY) Use <Literal><N></Literal> PVM processors to run this program;
1670 <term><option>-C[<us>]</option>:</term>
1673 <indexterm><primary>-C<us> RTS option</primary></indexterm> Sets
1674 the context switch interval to <literal><s></literal> seconds.
1675 A context switch will occur at the next heap block allocation after
1676 the timer expires (a heap block allocation occurs every 4k of
1677 allocation). With <option>-C0</option> or <option>-C</option>,
1678 context switches will occur as often as possible (at every heap block
1679 allocation). By default, context switches occur every 20ms
1680 milliseconds. Note that GHC's internal timer ticks every 20ms, and
1681 the context switch timer is always a multiple of this timer, so 20ms
1682 is the maximum granularity available for timed context switches.
1687 <term><option>-q[v]</option>:</term>
1690 <indexterm><primary>-q RTS option</primary></indexterm>
1691 (PARALLEL ONLY) Produce a quasi-parallel profile of thread activity,
1692 in the file <FIlename><program>.qp</FIlename>. In the style of <command>hbcpp</command>, this profile
1693 records the movement of threads between the green (runnable) and red
1694 (blocked) queues. If you specify the verbose suboption (<option>-qv</option>), the
1695 green queue is split into green (for the currently running thread
1696 only) and amber (for other runnable threads). We do not recommend
1697 that you use the verbose suboption if you are planning to use the
1698 <Command>hbcpp</Command> profiling tools or if you are context switching at every heap
1699 check (with <Option>-C</Option>).
1705 <Term><Option>-qt<num></Option>:</Term>
1708 <IndexTerm><Primary>-qt<num> RTS option</Primary></IndexTerm>
1709 (PARALLEL ONLY) Limit the thread pool size, i.e. the number of concurrent
1710 threads per processor to <Literal><num></Literal>. The default is
1711 32. Each thread requires slightly over 1K <Emphasis>words</Emphasis> in
1712 the heap for thread state and stack objects. (For 32-bit machines, this
1713 translates to 4K bytes, and for 64-bit machines, 8K bytes.)
1719 <Term><Option>-d</Option>:</Term>
1722 <IndexTerm><Primary>-d RTS option (parallel)</Primary></IndexTerm>
1723 (PARALLEL ONLY) Turn on debugging. It pops up one xterm (or GDB, or
1724 something…) per PVM processor. We use the standard <Command>debugger</Command>
1725 script that comes with PVM3, but we sometimes meddle with the
1726 <Command>debugger2</Command> script. We include ours in the GHC distribution,
1727 in <Filename>ghc/utils/pvm/</Filename>.
1733 <Term><Option>-qe<num></Option>:</Term>
1736 <IndexTerm><Primary>-qe<num> RTS option
1737 (parallel)</Primary></IndexTerm> (PARALLEL ONLY) Limit the spark pool size
1738 i.e. the number of pending sparks per processor to
1739 <Literal><num></Literal>. The default is 100. A larger number may be
1740 appropriate if your program generates large amounts of parallelism
1746 <Term><Option>-qQ<num></Option>:</Term>
1749 <IndexTerm><Primary>-qQ<num> RTS option (parallel)</Primary></IndexTerm>
1750 (PARALLEL ONLY) Set the size of packets transmitted between processors
1751 to <Literal><num></Literal>. The default is 1024 words. A larger number may be
1752 appropriate if your machine has a high communication cost relative to
1758 <Term><Option>-qh<num></Option>:</Term>
1761 <IndexTerm><Primary>-qh<num> RTS option (parallel)</Primary></IndexTerm>
1762 (PARALLEL ONLY) Select a packing scheme. Set the number of non-root thunks to pack in one packet to
1763 <num>-1 (0 means infinity). By default GUM uses full-subgraph
1764 packing, i.e. the entire subgraph with the requested closure as root is
1765 transmitted (provided it fits into one packet). Choosing a smaller value
1766 reduces the amount of pre-fetching of work done in GUM. This can be
1767 advantageous for improving data locality but it can also worsen the balance
1768 of the load in the system.
1773 <Term><Option>-qg<num></Option>:</Term>
1776 <IndexTerm><Primary>-qg<num> RTS option
1777 (parallel)</Primary></IndexTerm> (PARALLEL ONLY) Select a globalisation
1778 scheme. This option affects the
1779 generation of global addresses when transferring data. Global addresses are
1780 globally unique identifiers required to maintain sharing in the distributed
1781 graph structure. Currently this is a binary option. With <num>=0 full globalisation is used
1782 (default). This means a global address is generated for every closure that
1783 is transmitted. With <num>=1 a thunk-only globalisation scheme is
1784 used, which generated global address only for thunks. The latter case may
1785 lose sharing of data but has a reduced overhead in packing graph structures
1786 and maintaining internal tables of global addresses.
1797 <sect1 id="options-platform">
1798 <title>Platform-specific Flags</title>
1800 <indexterm><primary>-m* options</primary></indexterm>
1801 <indexterm><primary>platform-specific options</primary></indexterm>
1802 <indexterm><primary>machine-specific options</primary></indexterm>
1804 <para>Some flags only make sense for particular target
1810 <term><option>-mv8</option>:</term>
1812 <para>(SPARC machines)<indexterm><primary>-mv8 option (SPARC
1813 only)</primary></indexterm> Means to pass the like-named
1814 option to GCC; it says to use the Version 8 SPARC
1815 instructions, notably integer multiply and divide. The
1816 similiar <option>-m*</option> GCC options for SPARC also
1817 work, actually.</para>
1822 <term><option>-monly-[32]-regs</option>:</term>
1824 <para>(iX86 machines)<indexterm><primary>-monly-N-regs
1825 option (iX86 only)</primary></indexterm> GHC tries to
1826 “steal” four registers from GCC, for performance
1827 reasons; it almost always works. However, when GCC is
1828 compiling some modules with four stolen registers, it will
1829 crash, probably saying:
1832 Foo.hc:533: fixed or forbidden register was spilled.
1833 This may be due to a compiler bug or to impossible asm
1834 statements or clauses.
1837 Just give some registers back with
1838 <option>-monly-N-regs</option>. Try `3' first, then `2'.
1839 If `2' doesn't work, please report the bug to us.</para>
1848 <sect1 id="ext-core">
1849 <title>Generating External Core Files</title>
1851 <indexterm><primary>intermediate code generation</primary></indexterm>
1853 <para>GHC can dump its optimized intermediate code (said to be in “Core” format)
1854 to a file as a side-effect of compilation. Core files, which are given the suffix
1855 <filename>.hcr</filename>, can be read and processed by non-GHC back-end
1856 tools. The Core format is formally described in <ulink url="http://www.haskell.org/ghc/docs/papers/core.ps.gz"
1857 <citetitle>An External Representation for the GHC Core Language</citetitle></ulink>,
1858 and sample tools (in Haskell)
1859 for manipulating Core files are available in the GHC source distribution
1860 directory <literal>/fptools/ghc/utils/ext-core</literal>.
1861 Note that the format of <literal>.hcr</literal>
1862 files is <emphasis>different</emphasis> (though similar) to the Core output format generated
1863 for debugging purposes (<xref linkend="options-debugging">).</para>
1868 <term><option>-fext-core</option></term>
1870 <primary><option>-fext-core</option></primary>
1873 <para>Generate <literal>.hcr</literal> files.</para>
1887 ;;; Local Variables: ***
1889 ;;; sgml-parent-document: ("users_guide.sgml" "book" "chapter") ***