[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]

Re: [zzdev] Re: [zzdev] Re: Two hashtables per cell? No way!

On Tue, 17 Oct 2000, Jack Park wrote:
> From: Tuomas Lukka <lukka@xxxxxxxxxx>
> > On Sat, 14 Oct 2000, Benjamin Fallenstein wrote:
> <snip>
> > > Then, of course, we'd need to make ZZDimSpace abstract, and create a new
> > > ZZLocalDimSpace subclass, which stores the contents in hashtables. But
> > > that's not a big deal, really.
> >
> > Hmm... currently everything is always loaded in memory anyway so we can
> > easily do that.
> >
> Question: does loading everything in memory imply that it is or will be
> possible to build models far too large for memory? 

Currently impossible. 

> Does this mean there is
> or needs to be a virtual memory model for gzz?

There are two models, neither of which has yet been implemented.

The first is Ted's slice model where the space is divided into slices,
some number of which are explicitly brought in / swapped out by the user.

The second (which we computer scientist types seem to prefer) is a flat
space, of which relevant parts are loaded in memory, without requiring the
user to do anything but to browse the space as usual.

We are planning to try both.