From: | Gavin Sherry <swm(at)linuxworld(dot)com(dot)au> |
---|---|
To: | Andrew Dunstan <andrew(at)dunslane(dot)net> |
Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Qingqing Zhou <zhouqq(at)cs(dot)toronto(dot)edu>, pgsql-patches(at)postgresql(dot)org |
Subject: | Re: remove BufferBlockPointers for speed and space |
Date: | 2005-08-11 12:47:15 |
Message-ID: | Pine.LNX.4.58.0508112239500.2131@linuxworld.com.au |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-patches |
On Thu, 11 Aug 2005, Andrew Dunstan wrote:
>
>
> Gavin Sherry wrote:
>
> >>Or more than one hardware architecture (which you didn't even say what
> >>you tested...)
> >>
> >>
> >
> >Well, he tested on SunOS (!) and Linux -- I presume that's two
> >architectures.
> >
> >
>
> Sun still calls Solaris SunOs - try doing uname -s on a Solaris box (or
> look at a buildfarm solaris build info)
True. But my previous experience in university environments is that SunOS
usually refers to SunOS 2.6 -- and the performance indicates old hardware.
The thing is, compilser optimised versions of the test reveal very little
difference in performance. This may be because the compiler is very good
at optimising sequential annd predictable access to the array. Instead, we
should mimic what we see in the real world: random access.
Gavin
From | Date | Subject | |
---|---|---|---|
Next Message | Andrew Dunstan | 2005-08-11 12:53:08 | Re: remove BufferBlockPointers for speed and space |
Previous Message | Andrew Dunstan | 2005-08-11 12:32:43 | Re: remove BufferBlockPointers for speed and space |