Re: GiST -- making my index faster makes is slower - Mailing list pgsql-hackers

From Tom Lane
Subject Re: GiST -- making my index faster makes is slower
Date
Msg-id 3296.1082153803@sss.pgh.pa.us
Whole thread Raw
In response to Re: GiST -- making my index faster makes is slower  (David Blasby <dblasby@refractions.net>)
Responses Re: GiST -- making my index faster makes is slower
Re: GiST -- making my index faster makes is slower
List pgsql-hackers
David Blasby <dblasby@refractions.net> writes:
> Tom Lane wrote:
>> I'd suggest profiling the backend with both key types to get an idea of
>> where the time is going.

> I've been trying to use gprof to do some profiling, but I'm having 
> troubles.  Whats the best way to profile?

It's not hard; the only real gotcha is that on Linux systems you need to
compile with -DLINUX_PROFILE so that the postmaster works around some
Linux brain damage with dropping the profile status at fork().
I usually do (in an already configured and built source tree)
cd src/backendmake cleanmake PROFILE="-pg -DLINUX_PROFILE" allmake install-bin

Don't forget that the backends will drop their gmon.out files in
$PGDATA/data/DBOID/gmon.out.
        regards, tom lane


pgsql-hackers by date:

Previous
From: David Blasby
Date:
Subject: Re: GiST -- making my index faster makes is slower
Next
From: David Blasby
Date:
Subject: Re: GiST -- making my index faster makes is slower