[distcc] Synthetic benchmark

[Dan Kegel]

Because the real world is too messy, maybe I'll use a simplified
model of gcc to test my three or four hypotheses about
bottlenecks in distcc on shared clusters of workstations.
In particular, I'll use something that looks like gcc,
but always simply calculates the first million primes
or twiddles its thumbs a million times some other way.
I won't even bother with preprocessing or linking.
Also, I won't use 'make' to drive distcc; I'll just use a shell
script.  So my 'build script' to simulate building five executables
would look like


echo foo > foo.c
for i in 0 1 2 3 4; do
for j in 0 1 2 3 4 5 6 7; do
   distcc gcc -c foo.c &
done
wait
done

or something like that.  That should be just enough to let
me see if distcc has the issues I suspect it does.
- Dan

-- 
My technical stuff: http://kegel.com
My politics: see http://www.misleader.org for examples of why I'm for regime change