| From: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | Alvaro Herrera <alvherre(at)commandprompt(dot)com>, Martijn van Oosterhout <kleptog(at)svana(dot)org>, pgsql-hackers(at)postgreSQL(dot)org |
| Subject: | Re: That EXPLAIN ANALYZE patch still needs work |
| Date: | 2006-06-09 10:11:51 |
| Message-ID: | 1149847912.2691.240.camel@localhost.localdomain |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, 2006-06-08 at 17:21 -0400, Tom Lane wrote:
> Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> > Tried on two machines. The first (Turion64 laptop) gives 44-45 ms for
> > the SELECT, and 50-51 ms for the EXPLAIN ANALYZE.
>
> > The second machine, desktop Celeron 533, gives 197-200 ms for the SELECT
> > and 788-790 for the EXPLAIN ANALYZE. I guess this is the reproduction
> > you were looking for.
>
> Do you have oprofile installed on these? Comparing oprofile results
> might give some more insight where the time is going.
On my office PC the overhead from EA is about 20%.
stracing a backend, I also notice that on a 1.5 million row table we
make only 39158 calls to gettimeofday, on a table of 24591 blocks.
I'm thinking that the instrumentation overhead has been reduced as a
result of the page-at-a-time heap scans we now have?
--
Simon Riggs
EnterpriseDB http://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Agent M | 2006-06-09 11:09:12 | Re: Fabian Pascal and RDBMS deficiencies in fully implementing the relational model |
| Previous Message | Simon Riggs | 2006-06-09 09:23:11 | Re: ADD/DROP INHERITS |