| From: | Simon Riggs <simon(at)2ndQuadrant(dot)com> |
|---|---|
| To: | Robert Haas <robertmhaas(at)gmail(dot)com> |
| Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: our buffer replacement strategy is kind of lame |
| Date: | 2011-08-12 11:57:41 |
| Message-ID: | CA+U5nMKEE+59N176Xe-onSOyBW0Z-QUDHG3-+s=L0zGWnk2Xmw@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Fri, Aug 12, 2011 at 11:53 AM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
> I've not been reading "the literature", given the problems we had in
> 2004/5 regarding patents in this area. I also think that since we rely
> on the underlying filesystem for cacheing that we don't have exactly
> the same problem as other systems.
Reading the link you gave, I'm unimpressed by ClockPro. The
measurements made are all in low numbers of MB and the results show
that Clock is roughly same or better for large caches, but one might
read them multiple ways.
The zone of interest for us is above 1GB and the issues we care about
are contention, so even if we think ClockPro has a chance of improving
things we really don't have any measurements that support the cases we
care about.
--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Marko Kreen | 2011-08-12 12:02:03 | Re: sha1, sha2 functions into core? |
| Previous Message | Jean-Baptiste Quenot | 2011-08-12 11:55:06 | Re: plpython crash |