5 Jun
2012
5 Jun
'12
3:44 p.m.
On Tue, Jun 05, 2012 at 05:24:49PM +0200, Kai Hendry wrote:
On 5 June 2012 17:16, Phil Stewart
<phil.stewart(a)lichp.co.uk> wrote:
I believe that PAE slows things down considerably. PAE is just a
horrible hack based on segmented memory (yes, remember that from the
DOS days?).
Hugo.
--
=== Hugo Mills: hugo@... carfax.org.uk | darksatanic.net | lug.org.uk ===
PGP key: 515C238D from wwwkeys.eu.pgp.net or http://www.carfax.org.uk
--- I am an opera lover from planet Zog. Take me to your lieder. ---
An in-memory datastore like Memcache or Redis can
quite happily chew through
as much RAM as you can throw at it if you have a big enough data set. I
believe Redis can be configured to swap, but performance would take a
nose-dive.
Could such proggies not spawn another process in order to use more
memory though? Instead of swapping say, to slow disk IIUC.
That's what I'm wondering. I assumed PAE kernels might aid processes
to do this automagically too.