Potential GSoC proposal: Reduce the speed gap between 'ghc -c' and 'ghc --make'

[Evan Laforge]

> Questions:
>
> Would implementing this optimisation be a worthwhile/realistic GSoC project?
> What are other potential ways to bring 'ghc -c' performance up to par
> with 'ghc --make'?

I implemented a ghc server that runs several persistent ghcs, and
distributes compiles among them.  It seemed to work, but the speed
gains weren't what I'd hoped for.  But perhaps I was doing something
dumb that hurt performance.  I'd be happy to upload the source if
you're interested.