| From: | "Bucky Jordan" <bjordan(at)lumeta(dot)com> |
|---|---|
| To: | "Guillaume Smet" <guillaume(dot)smet(at)gmail(dot)com>, Jérôme BENOIS <benois(at)argia-engineering(dot)fr> |
| Cc: | "Scott Marlowe" <smarlowe(at)g2switchworks(dot)com>, "Dave Dutcher" <dave(at)tridecap(dot)com>, <pgsql-performance(at)postgresql(dot)org>, "Xavier Milliard" <milliard(at)argia(dot)fr> |
| Subject: | Re: High CPU Load |
| Date: | 2006-09-14 22:50:21 |
| Message-ID: | 78ED28FACE63744386D68D8A9D1CF5D4209999@MAIL.corp.lumeta.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
>Hyper threading. It's usually not recommended to enable it on
>PostgreSQL servers. On most servers, you can disable it directly in
>the BIOS.
Maybe for specific usage scenarios, but that's generally not been my experience with relatively recent versions of PG. We ran some tests with pgbench, and averaged 10% or more performance improvement. Now, I agree pgbench isn't the most realistic performance, but we did notice a slight improvement in our application performance too.
Also, here's some benchmarks that were posted earlier by the folks at tweakers.net also showing hyperthreading to be faster:
http://tweakers.net/reviews/646/10
I'm not sure if it's dependent on OS- our tests were on BSD 5.x and PG 7.4 and 8.0/8.1 and were several months ago, so I don't remember many more specifics than that.
So, not saying it's a best practice one way or another, but this is pretty easy to test and you should definitely try it out both ways for your workload.
- Bucky
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Francisco Reyes | 2006-09-14 23:30:52 | Re: Vacuums on large busy databases |
| Previous Message | Guillaume Smet | 2006-09-14 22:24:43 | Re: High CPU Load |