Using GHC GHC, using using GHC GHC can work in one of three “modes”: ghc ––interactive interactive mode ghci Interactive mode, which is also available as ghci. Interactive mode is described in more detail in . ghc ––make make mode In this mode, GHC will build a multi-module Haskell program automatically, figuring out dependencies for itself. If you have a straightforward Haskell program, this is likely to be much easier, and faster, than using make. ghc -E -C -S -c This is the traditional batch-compiler mode, in which GHC can compile source files one at a time, or link objects together into an executable. Options overview GHC's behaviour is controlled by options, which for historical reasons are also sometimes referred to as command-line flags or arguments. Options can be specified in three ways: Command-line arguments structure, command-line command-linearguments argumentscommand-line An invocation of GHC takes the following form: ghc [argument...] Command-line arguments are either options or file names. Command-line options begin with -. They may not be grouped: is different from . Options need not precede filenames: e.g., ghc *.o -o foo. All options are processed and then applied to all files; you cannot, for example, invoke ghc -c -O1 Foo.hs -O2 Bar.hs to apply different optimisation levels to the files Foo.hs and Bar.hs. Command line options in source files source-file options Sometimes it is useful to make the connection between a source file and the command-line options it requires quite tight. For instance, if a Haskell source file uses GHC extensions, it will always need to be compiled with the option. Rather than maintaining the list of per-file options in a Makefile, it is possible to do this directly in the source file using the OPTIONS pragma OPTIONS pragma: {-# OPTIONS -fglasgow-exts #-} module X where ... OPTIONS pragmas are only looked for at the top of your source files, upto the first (non-literate,non-empty) line not containing OPTIONS. Multiple OPTIONS pragmas are recognised. Note that your command shell does not get to the source file options, they are just included literally in the array of command-line arguments the compiler driver maintains internally, so you'll be desperately disappointed if you try to glob etc. inside OPTIONS. NOTE: the contents of OPTIONS are prepended to the command-line options, so you do have the ability to override OPTIONS settings via the command line. It is not recommended to move all the contents of your Makefiles into your source files, but in some circumstances, the OPTIONS pragma is the Right Thing. (If you use and have OPTION flags in your module, the OPTIONS will get put into the generated .hc file). Setting options in GHCi Options may also be modified from within GHCi, using the :set command. See for more details. Static vs. Dynamic options staticoptions dynamicoptions Each of GHC's command line options is classified as either static or dynamic. A static flag may only be specified on the command line, whereas a dynamic flag may also be given in an OPTIONS pragma in a source file or set from the GHCi command-line with :set. As a rule of thumb, all the language options are dynamic, as are the warning options and the debugging options. The rest are static, with the notable exceptions of , , , , and . The flag reference tables () lists the status of each flag. Meaningful file suffixes suffixes, file file suffixes for GHC File names with “meaningful” suffixes (e.g., .lhs or .o) cause the “right thing” to happen to those files. .lhs lhs suffix A “literate Haskell” module. .hs A not-so-literate Haskell module. .hi A Haskell interface file, probably compiler-generated. .hc Intermediate C file produced by the Haskell compiler. .c A C file not produced by the Haskell compiler. .s An assembly-language source file, usually produced by the compiler. .o An object file, produced by an assembler. Files with other suffixes (or without suffixes) are passed straight to the linker. Help and verbosity options help options verbosity options Cause GHC to spew a long usage message to standard output and then exit. The option makes GHC verbose: it reports its version number and shows (on stderr) exactly how it invokes each phase of the compilation system. Moreover, it passes the flag to most phases; each reports its version number (and possibly some other information). Please, oh please, use the option when reporting bugs! Knowing that you ran the right bits in the right order is always the first thing we want to verify. n To provide more control over the compiler's verbosity, the flag takes an optional numeric argument. Specifying on its own is equivalent to , and the other levels have the following meanings: Disable all non-essential messages (this is the default). Minimal verbosity: print one line per compilation (this is the default when or is on). Print the name of each compilation phase as it is executed. (equivalent to ). The same as , except that in addition the full command line (if appropriate) for each compilation phase is also printed. The same as except that the intermediate program representation after each compilation phase is also printed (excluding preprocessed and C/assembly files). Print a one-line string including GHC's version number. Print GHC's numeric version number only. Print the path to GHC's library directory. This is the top of the directory tree containing GHC's libraries, interfaces, and include files (usually something like /usr/local/lib/ghc-5.04 on Unix). This is the value of $libdirlibdir in the package configuration file (see ). Using <command>ghc</command> <option>––make</option> separate compilation When given the option, GHC will build a multi-module Haskell program by following dependencies from a single root module (usually Main). For example, if your Main module is in a file called Main.hs, you could compile and link the program like this: ghc ––make Main.hs The command line must contain one source file or module name; GHC will figure out all the modules in the program by following the imports from this initial module. It will then attempt to compile each module which is out of date, and finally if the top module is Main, the program will also be linked into an executable. The main advantages to using ghc ––make over traditional Makefiles are: GHC doesn't have to be restarted for each compilation, which means it can cache information between compilations. Compiling a muli-module program with ghc ––make can be up to twice as fast as running ghc individually on each source file. You don't have to write a Makefile. Makefilesavoiding GHC re-calculates the dependencies each time it is invoked, so the dependencies never get out of sync with the source. Any of the command-line options described in the rest of this chapter can be used with , but note that any options you give on the command line will apply to all the source files compiled, so if you want any options to apply to a single source file only, you'll need to use an OPTIONS pragma (see ). If the program needs to be linked with additional objects (say, some auxilliary C code), these can be specified on the command line as usual. Note that GHC can only follow dependencies if it has the source file available, so if your program includes a module for which there is no source file, even if you have an object and an interface file for the module, then GHC will complain. The exception to this rule is for package modules, which may or may not have source files. The source files for the program don't all need to be in the same directory; the option can be used to add directories to the search path (see ). GHC without <option>––make</option> Without , GHC will compile one or more source files given on the command line. The first phase to run is determined by each input-file suffix, and the last phase is determined by a flag. If no relevant flag is present, then go all the way through linking. This table summarises: Phase of the compilation system Suffix saying “start here” Flag saying “stop after” (suffix of) output file literate pre-processor .lhs - .hs C pre-processor (opt.) .hs (with ) .hspp Haskell compiler .hs , .hc, .s C compiler (opt.) .hc or .c .s assembler .s .o linker other - a.out Thus, a common invocation would be: ghc -c Foo.hs Note: What the Haskell compiler proper produces depends on whether a native-code generatornative-code generator is used (producing assembly language) or not (producing C). See for more details. Note: C pre-processing is optional, the flag turns it on. See for more details. Note: The option -E option runs just the pre-processing passes of the compiler, dumping the result in a file. Note that this differs from the previous behaviour of dumping the file to standard output. Re-directing the compilation output(s) output-directing options redirecting compilation output GHC's compiled output normally goes into a .hc, .o, etc., file, depending on the last-run compilation phase. The option -o option re-directs the output of that last-run phase to file foo. Note: this “feature” can be counterintuitive: ghc -C -o foo.o foo.hs will put the intermediate C code in the file foo.o, name notwithstanding! Note: on Windows, if the result is an executable file, the extension ".exe" is added if the specified filename does not already have an extension. Thus ghc -o foo Main.hs will compile and link the module Main.hs, and put the resulting executable in foo.exe (not foo). The option isn't of much use if you have several input files… Non-interface output files are normally put in the same directory as their corresponding input file came from. You may specify that they be put in another directory using the -odir <dir> option (the “Oh, dear” option). For example: % ghc -c parse/Foo.hs parse/Bar.hs gurgle/Bumble.hs -odir `arch` The output files, Foo.o, Bar.o, and Bumble.o would be put into a subdirectory named after the architecture of the executing machine (sun4, mips, etc). The directory must already exist; it won't be created. Note that the option does not affect where the interface files are put. In the above example, they would still be put in parse/Foo.hi, parse/Bar.hi, and gurgle/Bumble.hi. file The interface output may be directed to another file bar2/Wurble.iface with the option (not recommended). WARNING: if you redirect the interface file somewhere that GHC can't find it, then the recompilation checker may get confused (at the least, you won't get any recompilation avoidance). We recommend using a combination of and options instead, if possible. To avoid generating an interface at all, you could use this option to redirect the interface into the bit bucket: -ohi /dev/null, for example. directory Redirects all generated interface files into directory, instead of the default which is to place the interface file in the same directory as the source file. suffix suffix suffix EXOTICA: The suffix will change the .o file suffix for object files to whatever you specify. We use this when compiling libraries, so that objects for the profiling versions of the libraries don't clobber the normal ones. Similarly, the suffix will change the .hi file suffix for non-system interface files (see ). Finally, the option suffix will change the .hc file suffix for compiler-generated intermediate C files. The / game is useful if you want to compile a program with both GHC and HBC (say) in the same directory. Let HBC use the standard .hi/.o suffixes; add to your make rule for GHC compiling… Keeping Intermediate Files intermediate files, saving .hc files, saving .s files, saving The following options are useful for keeping certain intermediate files around, when normally GHC would throw these away after compilation: Keep intermediate .hc files when doing .hs-to-.o compilations via C (NOTE: .hc files aren't generated when using the native code generator, you may need to use to force them to be produced). Keep intermediate .s files. Keep intermediate .raw-s files. These are the direct output from the C compiler, before GHC does “assembly mangling” to produce the .s file. Again, these are not produced when using the native code generator. temporary files keeping Instructs the GHC driver not to delete any of its temporary files, which it normally keeps in /tmp (or possibly elsewhere; see ). Running GHC with will show you what temporary files were generated along the way. Redirecting temporary files temporary files redirecting If you have trouble because of running out of space in /tmp (or wherever your installation thinks temporary files should go), you may use the -tmpdir <dir> option option to specify an alternate directory. For example, says to put temporary files in the current working directory. Alternatively, use your TMPDIR environment variable.TMPDIR environment variable Set it to the name of the directory where temporary files should be put. GCC and other programs will honour the TMPDIR variable as well. Even better idea: Set the DEFAULT_TMPDIR make variable when building GHC, and never worry about TMPDIR again. (see the build documentation). Warnings and sanity-checking sanity-checking options warnings GHC has a number of options that select which types of non-fatal error messages, otherwise known as warnings, can be generated during compilation. By default, you get a standard set of warnings which are generally likely to indicate bugs in your program. These are: , , , , and . The following flags are simple ways to select standard “packages” of warnings: : -W option Provides the standard warnings plus , , , , and . : Turns off all warnings, including the standard ones. : Turns on all warning options. The full set of warning options is described below. To turn off any warning, simply give the corresponding option on the command line. : deprecations Causes a warning to be emitted when a deprecated function or type is used. Entities can be marked as deprecated using a pragma, see . : duplicate exports, warning export lists, duplicates Have the compiler warn about duplicate entries in export lists. This is useful information if you maintain large export lists, and want to avoid the continued export of a definition after you've deleted (one) mention of it in the export list. This option is on by default. : shadowing interface files Causes the compiler to emit a warning when a module or interface file in the current directory is shadowing one with the same module name in a library or other directory. : incomplete patterns, warning patterns, incomplete Similarly for incomplete patterns, the function g below will fail when applied to non-empty lists, so the compiler will emit a warning about this when is enabled. g [] = 2 This option isn't enabled be default because it can be a bit noisy, and it doesn't always indicate a bug in the program. However, it's generally considered good practice to cover all the cases in your functions. : Turns on warnings for various harmless but untidy things. This currently includes: importing a type with (..) when the export is abstract, and listing duplicate class assertions in a qualified type. : missing fields, warning fields, missing This option is on by default, and warns you whenever the construction of a labelled field constructor isn't complete, missing initializers for one or more fields. While not an error (the missing fields are initialised with bottoms), it is often an indication of a programmer error. : missing methods, warning methods, missing This option is on by default, and warns you whenever an instance declaration is missing one or more methods, and the corresponding class declaration has no default declaration for them. : type signatures, missing If you would like GHC to check that every top-level function/value has a type signature, use the option. This option is off by default. : shadowing, warning This option causes a warning to be emitted whenever an inner-scope value has the same name as an outer-scope value, i.e. the inner value shadows the outer one. This can catch typographical errors that turn into hard-to-find bugs, e.g., in the inadvertent cyclic definition let x = ... x ... in. Consequently, this option does will complain about cyclic recursive definitions. : overlapping patterns, warning patterns, overlapping By default, the compiler will warn you if a set of patterns are overlapping, i.e., f :: String -> Int f [] = 0 f (_:xs) = 1 f "2" = 2 where the last pattern match in f won't ever be reached, as the second pattern overlaps it. More often than not, redundant patterns is a programmer mistake/error, so this option is enabled by default. : Causes the compiler to warn about lambda-bound patterns that can fail, eg. \(x:xs)->.... Normally, these aren't treated as incomplete patterns by . : defaulting mechanism, warning Have the compiler warn/inform you where in your source the Haskell defaulting mechanism for numeric types kicks in. This is useful information when converting code from a context that assumed one default into one with another, e.g., the `default default' for Haskell 1.4 caused the otherwise unconstrained value 1 to be given the type Int, whereas Haskell 98 defaults it to Integer. This may lead to differences in performance and behaviour, hence the usefulness of being non-silent about this. This warning is off by default. : unused binds, warning binds, unused Report any function definitions (and local bindings) which are unused. For top-level functions, the warning is only given if the binding is not exported. : unused imports, warning imports, unused Report any objects that are explicitly imported but never used. : unused matches, warning matches, unused Report all unused variables which arise from pattern matches, including patterns consisting of a single variable. For instance f x y = [] would report x and y as unused. To eliminate the warning, all unused variables can be replaced with wildcards. If you're feeling really paranoid, the option is a good choice. It turns on heavyweight intra-pass sanity-checking within GHC. (It checks GHC's sanity, not yours.) &separate; &packages; Optimisation (code improvement) optimisation improvement, code The options specify convenient “packages” of optimisation flags; the options described later on specify individual optimisations to be turned on/off; the options specify machine-specific optimisations to be turned on/off. <option>-O*</option>: convenient “packages” of optimisation flags. There are many options that affect the quality of code produced by GHC. Most people only have a general goal, something like “Compile quickly” or “Make my program run like greased lightning.” The following “packages” of optimisations (or lack thereof) should suffice. Once you choose a “package,” stick with it—don't chop and change. Modules' interfaces will change with a shift to a new option, and you may have to recompile a large chunk of all importing modules before your program can again be run safely (see ). No -type option specified: -O* not specified This is taken to mean: “Please compile quickly; I'm not over-bothered about compiled-code quality.” So, for example: ghc -c Foo.hs : Means “turn off all optimisation”, reverting to the same settings as if no options had been specified. Saying can be useful if eg. make has inserted a on the command line already. or : -O option -O1 option optimisenormally Means: “Generate good-quality code without taking too long about it.” Thus, for example: ghc -c -O Main.lhs : -O2 option optimiseaggressively Means: “Apply every non-dangerous optimisation, even if it means significantly longer compile times.” The avoided “dangerous” optimisations are those that can make runtime or space worse if you're unlucky. They are normally turned on or off individually. At the moment, is unlikely to produce better code than . : -Ofile <file> option optimising, customised (NOTE: not supported yet in GHC 5.x. Please ask if you're interested in this.) For those who need absolute control over exactly what options are used (e.g., compiler writers, sometimes :-), a list of options can be put in a file and then slurped in with . In that file, comments are of the #-to-end-of-line variety; blank lines and most whitespace is ignored. Please ask if you are baffled and would like an example of ! We don't use a flag for day-to-day work. We use to get respectable speed; e.g., when we want to measure something. When we want to go for broke, we tend to use (and we go for lots of coffee breaks). The easiest way to see what (etc.) “really mean” is to run with , then stand back in amazement. <option>-f*</option>: platform-independent flags -f* options (GHC) -fno-* options (GHC) These flags turn on and off individual optimisations. They are normally set via the options described above, and as such, you shouldn't need to set any of them explicitly (indeed, doing so could lead to unexpected results). However, there are one or two that may be of interest: : When this option is given, intermediate floating point values can have a greater precision/range than the final type. Generally this is a good thing, but some programs may rely on the exact precision/range of Float/Double values and should not use this option for their compilation. : Causes GHC to ignore uses of the function Exception.assert in source code (in other words, rewriting Exception.assert p e to e (see ). This flag is turned on by . Turns off the strictness analyser; sometimes it eats too many cycles. Turns off the CPR (constructed product result) analysis; it is somewhat experimental. : strict constructor fields constructor fields, strict This option causes all constructor fields which are marked strict (i.e. “!”) to be unboxed or unpacked if possible. For example: data T = T !Float !Float will create a constructor T containing two unboxed floats if the flag is given. This may not always be an optimisation: if the T constructor is scrutinised and the floats passed to a non-strict function for example, they will have to be reboxed (this is done automatically by the compiler). This option should only be used in conjunction with , in order to expose unfoldings to the compiler so the reboxing can be removed as often as possible. For example: f :: T -> Float f (T f1 f2) = f1 + f2 The compiler will avoid reboxing f1 and f2 by inlining + on floats, but only when is on. Any single-constructor data is eligible for unpacking; for example data T = T !(Int,Int) will store the two Ints directly in the T constructor, by flattening the pair. Multi-level unpacking is also supported: data T = T !S data S = S !Int !Int will store two unboxed Int#s directly in the T constructor. Switches on an experimental "optimisation". Switching it on makes the compiler a little keener to inline a function that returns a constructor, if the context is that of a thunk. x = plusInt a b If we inlined plusInt we might get an opportunity to use update-in-place for the thunk 'x'. : inlining, controlling unfolding, controlling (Default: 45) Governs the maximum size that GHC will allow a function unfolding to be. (An unfolding has a “size” that reflects the cost in terms of “code bloat” of expanding that unfolding at at a call site. A bigger function would be assigned a bigger cost.) Consequences: (a) nothing larger than this will be inlined (unless it has an INLINE pragma); (b) nothing larger than this will be spewed into an interface file. Increasing this figure is more likely to result in longer compile times than faster code. The next option is more useful: : inlining, controlling unfolding, controlling (Default: 8) This is the magic cut-off figure for unfolding: below this size, a function definition will be unfolded at the call-site, any bigger and it won't. The size computed for a function depends on two things: the actual size of the expression minus any discounts that apply (see ). &phases; Using Concurrent Haskell Concurrent Haskell—use GHC supports Concurrent Haskell by default, without requiring a special option or libraries compiled in a certain way. To get access to the support libraries for Concurrent Haskell, just import Control.Concurrent (details are in the accompanying library documentation). RTS options are provided for modifying the behaviour of the threaded runtime system. See . Concurrent Haskell is described in more detail in the documentation for the Control.Concurrent module. Using Parallel Haskell Parallel Haskell—use [You won't be able to execute parallel Haskell programs unless PVM3 (Parallel Virtual Machine, version 3) is installed at your site.] To compile a Haskell program for parallel execution under PVM, use the option,-parallel option both when compiling and linking. You will probably want to import Parallel into your Haskell modules. To run your parallel program, once PVM is going, just invoke it “as normal”. The main extra RTS option is , to say how many PVM “processors” your program to run on. (For more details of all relevant RTS options, please see .) In truth, running Parallel Haskell programs and getting information out of them (e.g., parallelism profiles) is a battle with the vagaries of PVM, detailed in the following sections. Dummy's guide to using PVM PVM, how to use Parallel Haskell—PVM use Before you can run a parallel program under PVM, you must set the required environment variables (PVM's idea, not ours); something like, probably in your .cshrc or equivalent: setenv PVM_ROOT /wherever/you/put/it setenv PVM_ARCH `$PVM_ROOT/lib/pvmgetarch` setenv PVM_DPATH $PVM_ROOT/lib/pvmd Creating and/or controlling your “parallel machine” is a purely-PVM business; nothing specific to Parallel Haskell. The following paragraphs describe how to configure your parallel machine interactively. If you use parallel Haskell regularly on the same machine configuration it is a good idea to maintain a file with all machine names and to make the environment variable PVM_HOST_FILE point to this file. Then you can avoid the interactive operations described below by just saying pvm $PVM_HOST_FILE You use the pvmpvm command command to start PVM on your machine. You can then do various things to control/monitor your “parallel machine;” the most useful being: ControlD exit pvm, leaving it running halt kill off this “parallel machine” & exit add <host> add <host> as a processor delete <host> delete <host> reset kill what's going, but leave PVM up conf list the current configuration ps report processes' status pstat <pid> status of a particular process The PVM documentation can tell you much, much more about pvm! Parallelism profiles parallelism profiles profiles, parallelism visualisation tools With Parallel Haskell programs, we usually don't care about the results—only with “how parallel” it was! We want pretty pictures. Parallelism profiles (à la hbcpp) can be generated with the -qP RTS option (concurrent, parallel) RTS option. The per-processor profiling info is dumped into files named <full-path><program>.gr. These are then munged into a PostScript picture, which you can then display. For example, to run your program a.out on 8 processors, then view the parallelism profile, do: $ ./a.out +RTS -qP -qp8 $ grs2gr *.???.gr > # combine the 8 .gr files into one $ gr2ps -O # cvt to .ps; output in $ ghostview -seascape # look at it! The scripts for processing the parallelism profiles are distributed in ghc/utils/parallel/. Other useful info about running parallel programs The “garbage-collection statistics” RTS options can be useful for seeing what parallel programs are doing. If you do either -Sstderr RTS option or , then you'll get mutator, garbage-collection, etc., times on standard error. The standard error of all PE's other than the `main thread' appears in /tmp/pvml.nnn, courtesy of PVM. Whether doing or not, a handy way to watch what's happening overall is: tail -f /tmp/pvml.nnn. RTS options for Concurrent/Parallel Haskell RTS options, concurrent RTS options, parallel Concurrent Haskell—RTS options Parallel Haskell—RTS options Besides the usual runtime system (RTS) options (), there are a few options particularly for concurrent/parallel execution. : -qp<N> RTS option (PARALLEL ONLY) Use <N> PVM processors to run this program; the default is 2. : -C<us> RTS option Sets the context switch interval to <s> seconds. A context switch will occur at the next heap block allocation after the timer expires (a heap block allocation occurs every 4k of allocation). With or , context switches will occur as often as possible (at every heap block allocation). By default, context switches occur every 20ms milliseconds. Note that GHC's internal timer ticks every 20ms, and the context switch timer is always a multiple of this timer, so 20ms is the maximum granularity available for timed context switches. : -q RTS option (PARALLEL ONLY) Produce a quasi-parallel profile of thread activity, in the file <program>.qp. In the style of hbcpp, this profile records the movement of threads between the green (runnable) and red (blocked) queues. If you specify the verbose suboption (), the green queue is split into green (for the currently running thread only) and amber (for other runnable threads). We do not recommend that you use the verbose suboption if you are planning to use the hbcpp profiling tools or if you are context switching at every heap check (with ). --> : -qt<num> RTS option (PARALLEL ONLY) Limit the thread pool size, i.e. the number of concurrent threads per processor to <num>. The default is 32. Each thread requires slightly over 1K words in the heap for thread state and stack objects. (For 32-bit machines, this translates to 4K bytes, and for 64-bit machines, 8K bytes.) : -qe<num> RTS option (parallel) (PARALLEL ONLY) Limit the spark pool size i.e. the number of pending sparks per processor to <num>. The default is 100. A larger number may be appropriate if your program generates large amounts of parallelism initially. : -qQ<num> RTS option (parallel) (PARALLEL ONLY) Set the size of packets transmitted between processors to <num>. The default is 1024 words. A larger number may be appropriate if your machine has a high communication cost relative to computation speed. : -qh<num> RTS option (parallel) (PARALLEL ONLY) Select a packing scheme. Set the number of non-root thunks to pack in one packet to <num>-1 (0 means infinity). By default GUM uses full-subgraph packing, i.e. the entire subgraph with the requested closure as root is transmitted (provided it fits into one packet). Choosing a smaller value reduces the amount of pre-fetching of work done in GUM. This can be advantageous for improving data locality but it can also worsen the balance of the load in the system. : -qg<num> RTS option (parallel) (PARALLEL ONLY) Select a globalisation scheme. This option affects the generation of global addresses when transferring data. Global addresses are globally unique identifiers required to maintain sharing in the distributed graph structure. Currently this is a binary option. With <num>=0 full globalisation is used (default). This means a global address is generated for every closure that is transmitted. With <num>=1 a thunk-only globalisation scheme is used, which generated global address only for thunks. The latter case may lose sharing of data but has a reduced overhead in packing graph structures and maintaining internal tables of global addresses. Platform-specific Flags -m* options platform-specific options machine-specific options Some flags only make sense for particular target platforms. : (SPARC machines)-mv8 option (SPARC only) Means to pass the like-named option to GCC; it says to use the Version 8 SPARC instructions, notably integer multiply and divide. The similiar GCC options for SPARC also work, actually. : (iX86 machines)-monly-N-regs option (iX86 only) GHC tries to “steal” four registers from GCC, for performance reasons; it almost always works. However, when GCC is compiling some modules with four stolen registers, it will crash, probably saying: Foo.hc:533: fixed or forbidden register was spilled. This may be due to a compiler bug or to impossible asm statements or clauses. Just give some registers back with . Try `3' first, then `2'. If `2' doesn't work, please report the bug to us. &runtime; Generating External Core Files intermediate code generation GHC can dump its optimized intermediate code (said to be in “Core” format) to a file as a side-effect of compilation. Core files, which are given the suffix .hcr, can be read and processed by non-GHC back-end tools. The Core format is formally described in An External Representation for the GHC Core Language, and sample tools (in Haskell) for manipulating Core files are available in the GHC source distribution directory /fptools/ghc/utils/ext-core. Note that the format of .hcr files is different (though similar) to the Core output format generated for debugging purposes (). Generate .hcr files. &debug; &flags;