[distcc] Re: issue fix proposal for distcc and intel C++ 7.1

[Martin Pool]

On  9 Jul 2003, Ben Elliston <bje at wasabisystems.com> wrote:
> Markus Werle <numerical.simulation at web.de> writes:
> 
> > > As a broader example, people in scientific computing use things like
> > > OpenPBS for distributing work.  Perhaps some of them would be better
> > > served by something based on distcc?
> 
> > I don't know.
> 
> I doubt it.  The time required to compile your average scientific code
> is negligible compared to its runtime.

I meant for distributing the actual work, not the compilation of it.

You can conceptually split distcc into "pretending to be a C compiler"
and "distributing jobs to selected machines across the network".
Roughly half the code complexity is in each one.  Certainly some
people want the second but not the first, for example for scientific
computing.

Would any of them want to pull that part out of distcc and replace the
first part, and would it be economical?  I don't know.  Some people
have hypothetically suggested it, but I have not yet seen a particular
case that would justify it.

--
Martin