From: | Steve Atkins <steve(at)blighty(dot)com> |
---|---|
To: | POSTGRES-PERFORMANCE <pgsql-performance(at)postgresql(dot)org> |
Subject: | Re: how much mem to give postgres? |
Date: | 2004-10-20 17:50:39 |
Message-ID: | 20041020175038.GA1339@gp.word-to-the-wise.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
On Wed, Oct 20, 2004 at 03:07:00PM +0100, Matt Clark wrote:
> You turn it off in the BIOS. There is no 'other half', the processor is
> just pretending to have two cores by shuffling registers around, which
> gives maybe a 5-10% performance gain in certain multithreaded
> situations.
> <opinion>A hack to overcome marchitactural limitations due
> to the overly long pipeline in the Prescott core.</opinion>. Really of
> most use for desktop interactivity rather than actual throughput.
<OT>
Hyperthreading is actually an excellent architectural feature that
can give significant performance gains when implemented well and used
for an appropriate workload under a decently HT aware OS.
IMO, typical RDBMS streams are not an obviously appropriate workload,
Intel didn't implement it particularly well and I don't think there
are any OSes that support it particularly well.
</OT>
But don't write off using it in the future, when it's been improved
at both the OS and the silicon levels.
Cheers,
Steve
From | Date | Subject | |
---|---|---|---|
Next Message | Matt Clark | 2004-10-20 18:16:18 | Re: how much mem to give postgres? |
Previous Message | Manfred Spraul | 2004-10-20 17:39:13 | Re: futex results with dbt-3 |