28 Jan
2010
28 Jan
'10
4:09 p.m.
bah, was meant for the list.
---------- Forwarded message ----------
From: Paul Davis <paul(a)linuxaudiosystems.com>
Date: Thu, Jan 28, 2010 at 3:08 PM
Subject: Re: [LAD] hard realtime performance synth
To: David McClanahan <david.mcclanahan(a)gmail.com>
On Thu, Jan 28, 2010 at 3:07 PM, David McClanahan
<david.mcclanahan(a)gmail.com> wrote:
Ok, this may a partial fair point, but I don't
think its impossible that
with a machine with 100+ MB of RAM that one could screw off virtual
memory(paging) especially since I'm not really after having a soundfont
library available so much as having something that calculates a sample at a
time.
The fair part may be the "3 levels of cache" which I assume amounts to a
buffering delay.
no. it refers to your processors L1, L2 and (maybe) L3 cache. where
the next instruction (and data) is when its time to execute it is
going to make a HUGE difference to long it takes. modern general
purpose processors are not built with the idea of predictable,
deterministic execution in mind.