[distcc] SMP volunteers + compression

[Martin Pool]

Some rough numbers for LZO vs gzip on .i files, on a 550MHz PIII:

mbp/3 tmp$ time gzip -2 < distcc.i > distcc.i.2.gz

real	0m0.020s
user	0m0.020s
sys	0m0.000s
mbp/3 tmp$ time gzip -3 < distcc.i > distcc.i.3.gz

real	0m0.028s
user	0m0.010s
sys	0m0.010s
mbp/3 tmp$ time gzip -1 < distcc.i > distcc.i.1.gz

real	0m0.027s
user	0m0.020s
sys	0m0.000s
mbp/3 tmp$ time lzop -c1F < distcc.i > distcc.i.1.lzo

real	0m0.008s
user	0m0.000s
sys	0m0.000s
mbp/3 tmp$ ls -l distcc.i*
-rw-rw-r--    1 mbp      mbp         80977 Jun 12 09:31 distcc.i
-rw-rw-r--    1 mbp      mbp         20217 Jun 12 09:33 distcc.i.1.gz
-rw-rw-r--    1 mbp      mbp         19408 Jun 12 09:32 distcc.i.2.gz
-rw-rw-r--    1 mbp      mbp         18858 Jun 12 09:32 distcc.i.3.gz
-rw-rw-r--    1 mbp      mbp         26397 Jun 12 09:38 distcc.i.1.lzo

So the compression is a bit worse than gzip, but it's somewhat
faster.  It's hard to measure for such a small file on this machine,
and there will be some overheads.  There's a fair amount of variation
in time from one run to another.

-- 
Martin