a quick, free distributed C/C++ compiler

Velocity, it appears to me, gives the one genuinely fashionable pleasure.
— Aldous Huxley

distcc is a program to distribute builds of C, C++,
Goal C or Goal C++ code throughout a number of machines on a
community. distcc ought to all the time generate the identical outcomes as a
native construct, is straightforward to put in and use, and is often
a lot quicker than an area compile.

distcc doesn’t require all machines to share a filesystem, have
synchronized clocks, or to have the identical libraries or header
recordsdata put in. They will even have totally different processors or
working methods, if cross-compilers are put in.

Simply wished to drop you a line to say that we are actually utilizing distcc at work and
it’s glorious. Now we have a rack of opteron machines that we use for
computational duties and we are actually utilizing them as a compile farm to compile
our ~1MLOC C++ tree (which might take an hour on a single CPU to recompile if
we modify sure header recordsdata). We tried utilizing Solar’s grid engine to do that
(we already use it to schedule our computational jobs), however the mixture
of its polled operation and the overhead of NFS led to little enchancment
(and drastically confused our community). With distcc compile instances are means down
and my productiveness has drastically improved; the 2 greatest options for me are
its low influence on the community and its simplicity.

Thanks rather a lot for a terrific software! — Jeremy Barnes

60-second directions:

1. For every machine, download distcc
   unpack, and do
   
   ./configure && make && sudo make set up
2. On every of the servers, run distccd --daemon, with
   --allow choices to limit entry.
3. Put the names of the servers in your setting:
   export DISTCC_POTENTIAL_HOSTS='localhost pink inexperienced blue'
4. Construct!
   Wrap your construct command within the “pump” script
   and use “distcc” as your C compiler:
   
   cd ~/work/myproject; pump make -j8 CC=distcc

For extra detailed set up directions, see the
INSTALL file.

distcc is developed on GNU/Linux, however has been reported to work
on different methods together with FreeBSD, NetBSD, Darwin, Solaris,
HP-UX, IRIX, Cygwin and BSD/OS.

distcc sends the whole preprocessed supply code throughout the
community for every job, so all it requires of the volunteer
machines is that they be operating the distccd daemon,
and that they’ve an applicable compiler put in.

distcc shouldn’t be itself a compiler, however quite a front-end to the
GNU C/C++ compiler (gcc) and LLVM compiler (clang). (There
is preliminary assist for another compilers however the principle focus is
gcc.) Virtually all gcc choices and options work as regular.

distcc is designed for use with the -j parallel-build
characteristic in GNU Make
or SCons,
or different construct instruments.
Transport recordsdata throughout the community takes
time, however few cycles on the shopper machine. Any recordsdata that may
be constructed remotely are primarily “without spending a dime” by way of shopper
CPU.

distcc has been beneath growth since early 2002. It reliably
and efficiently compiles massive, complicated free and proprietary
software program methods. Applications identified to construct accurately with distcc
embrace
the Linux kernel,
rsync,
KDE,
GNOME (by way of
GARNOME),
Samba and
Wireshark.

distcc is sort of linearly scalable for small numbers of
machines: Constructing Linux 2.4.19 on a single 1700MHz Pentium IV
machine with distcc 0.15 takes 6 minutes, 45 seconds. Utilizing
distcc throughout three such machines on a 100Mbps swap takes solely
2 minutes, 30 seconds: 2.6x quicker. The (unreachable)
theoretical most speedup is 3.0x, so on this case distcc
scales with 89% effectivity.

You do not want a whole lot of machines to profit: a laptop and a single desktop is much faster than a laptop alone.

See Also

41 Years in UX: A Profession Retrospective

There are a number of instruments out there individually that
improve or complement distcc:

• ccache
  caches compiler output to speed up builds.
• dmucs helps distcc choose
  applicable servers.

• ccontrol
  offers centralized management of parallelism, caching and distribution
  even whereas the construct is operating.

• crosstool
  automates constructing cross-compilation toolchains.

• distccWebView is a CGI to indicate which of your
  servers are up and operating.

• tcpbalance
  is a load-balancing TCP proxy written in Erlang that works effectively with
  distcc.