[Beowulf] thermal/power limits

Peter St. John peter.st.john at gmail.com
Mon Aug 12 15:32:36 PDT 2013

I don't think it was "world size", as such, as the number of units that are
within range of hostile units. Imagine simple AI for a stack of cataphracts
in Civ; it considers every possible move, and then what moves the opponent
units within range can make in reaction. Like chess but only a couple ply
deep. In MMOs, you get lag when too many hostile units are within
perception range; they all see you and all your allies, and each computes
the possible reactions of all of them, for all of them. So the call stack
gets big.

Civ I think has always been fine with enough RAM, but you can't have enough
bandwidth if a raid of a hundred players invades a castle with a hundred
liches, all in sight of one courtyard.

That said, it does sound like the problem of cellularized weather
simulation; the pressure and humidity etc at one cell impacts adjacent
cells, but that propogates, so yes, I think the guys who code MPI for that
should make a game with AI for lots of independent mobile units :-)

On Mon, Aug 12, 2013 at 5:38 PM, Nathan Moore <ntmoore at gmail.com> wrote:

> Can anyone on the list comment on the structure of codes like Sid Meier's
> "Civilization" or MS "Age of Empires"?  I lost a week or so to each in
> school and I recall there was a certain "world size," at which the game
> play would completely stall.  I assume because there was an all to all, or
> all to most calculation between game elements that suddenly grew larger
> than cache.
> Strategy games like this don't seem that different than a relaxation-style
> solution to a laplace problem.  It would seem natural to write the game to
> take advantage of multiple cores.
> Is this really how its done?
> --
> - - - - - - -   - - - - - - -   - - - - - - -
> Nathan Moore
> Winona, MN
> - - - - - - -   - - - - - - -   - - - - - - -
> - - - -
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
> To change your subscription (digest mode or unsubscribe) visit
> http://www.beowulf.org/mailman/listinfo/beowulf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.beowulf.org/pipermail/beowulf/attachments/20130812/748baf04/attachment.html>

More information about the Beowulf mailing list