FWIW, in summer I have done a little bit of testing on one of our
dual-cpu machines; among this I have been running OSDB (open source
database benchmark), 32 simulated clients, against Postgres (7.2.1)/Linux
(2.4.18), once bootet with maxcpus=1 and once with maxcpus=2; if I
remember correctly I saw something between 80-90% performance improvement
on the IR benchmark with the second cpu activated.
Note the run was completely cpu-bound, neither harddisk nor memory was the
bottleneck, so you may see less of an improvement if other parts of your
system are the limit; but Postgres itself appears to make use of the
available cpus quite nicely.
Regards
--
Helge Bahmann <bahmann@math.tu-freiberg.de> /| \__
The past: Smart users in front of dumb terminals /_|____\
_/\ | __)
$ ./configure \\ \|__/__|
checking whether build environment is sane... yes \\/___/ |
checking for AIX... no (we already did this) |