[MUD-Dev] Re: TECH: Distributed Muds

Brian Hook bwh at wksoftware.com
Wed Apr 25 14:42:03 New Zealand Standard Time 2001

At 01:02 PM 4/25/01 -0400, Derek Snider wrote:

> A special list could be set up for objects that rot, and objects
> with programs to avoid traversing the entire global object list.

That would seem to make the most sense.  A global registry where
objects can register what events they respond to.  Of course, you
still have to deal with pathological worst case scenarios.

> On a side note, from tests I have done a program can traverse a
> linked list of 1,000,000 links in 0.16 seconds on a PII 333Mhz
> machine running a mud, webserver, ftp server, etc (90% idle).

The traversal operation itself isn't the problem, since in a best case
contrived scenario you're going to be accessing nearly sequential
chunks of data (depending on what you mean by "traversal" -- if it's
just pure traversal, then it could end up being a completely linear
iteration of memory which would be fast, but if it's "traversal and
examination" and the node data blocks aren't contiguous with the
node's pointers then you'll have some cache thrashing).

And as you state, the problem is the vmem system, not the actual
traversal itself.

Brian Hook

MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the MUD-Dev mailing list