On Mon, 2007-03-05 at 11:10 +0200, Hannu Krosing wrote:
> > My proposal for a fix: ensure that when relations larger (much larger?)
> > than buffer cache are scanned, they are mapped to a single page in the
> > shared buffer cache.
>
> How will this approach play together with synchronized scan patches ?
>
Thanks for considering my patch in this discussion. I will test by
turning shared_buffers down as low as I can, and see if that makes a big
difference.
> Or should synchronized scan rely on systems cache only ?
>
I don't know what the performance impact of that will be; still good
compared to reading from disk, but I assume much more CPU time.
Regards,Jeff Davis