| From: | Mike Mascari <mascarm(at)mascari(dot)com> |
|---|---|
| To: | Rod Taylor <pg(at)rbt(dot)ca> |
| Cc: | Andrew Dunstan <andrew(at)dunslane(dot)net>, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>, pgsql-hackers(at)postgresql(dot)org, Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> |
| Subject: | Re: Call for 7.5 feature completion |
| Date: | 2005-08-26 01:44:41 |
| Message-ID: | 430E7409.6050805@mascari.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Rod Taylor wrote:
> On Thu, 2005-08-25 at 21:27 -0400, Andrew Dunstan wrote:
>
>>Rod Taylor wrote:
>>
>>
>>> * Multi-CPU sorts. Take a large single sort like an index creation
>>> and split the work among multiple CPUs.
>
>>This really implies threading, doesn't it? And presumably it would have
>>many possible uses besides this one for doing parallel work, e.g. maybe
>>the planner could evaluate several alternative plans in parallel.
>
> I don't think threading is needed.
>
> I pictured PostgreSQL spawning one process per CPU explicitly for
> sorting which standard backends could use as required to do batch work.
This is one area where PostgreSQL needs a lot of work to catch up to the
competition. Oracle, DB2, Ingres, even SQL Server Enterprise edition
all have parallel query capabilities. I have an older 8-processor Sun
Enterprise 3500, as an example. It still has use with other vendors'
database products due to their parallel feature set (make -j 9 is nice
too), but behaves like the boat-anchor it is w.r.t. PostgreSQL.
Mike Mascari
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Josh Berkus | 2005-08-26 02:48:28 | Re: Call for 7.5 feature completion |
| Previous Message | Christopher Kings-Lynne | 2005-08-26 01:38:21 | Re: Call for 7.5 feature completion |