On Friday 24 September 2004 7:32 am, Tom Lane wrote:
> David Helgason <david@uti.is> writes:
> > I'm calling one stored procedure with a prepared statement on the
> > server with 6 arrays of around 1200 elements each as parameters.
> > The parameters are around 220K in total.
>
> Exactly how are you fetching, building, or otherwise providing the
> arrays? Which PG version is this exactly?
>
> > Any suggestions how I go about finding the bottleneck here? What
> > tools do other people use for profiling on Linux.
>
> Rebuild with profiling enabled (make clean; make PROFILE="-pg
> -DLINUX_PROFILE") and then use gprof to produce a report from the
> trace file that the backend drops when it exits.
>
> If that sounds out of your league, send along a self-contained test
> case and I'll be glad to take a look.
>
> > This might sound like a "then don't do that" situation.
>
> I'm betting on some O(N^2) behavior in the array code, but it'll be
> difficult to pinpoint without profile results.
Possibly the bug we discussed early last year (IIRC added to todo but
not fixed):
http://archives.postgresql.org/pgsql-performance/2003-01/msg00235.php
Cheers,
Steve
