Okay, that's true too, but the virtual machine does more-or-less guarantee
that the number of cycles you get on a particular instance type is
consistent (with the exception of micro-instances, but we wouldn't go there
anyway). Nothing short of writing our own dedicated operating system is
going to guarantee absolute uniformity on *any* type of platform, but I
think we all agree it's a big improvement over the current status quo.

Also, I chose AWS as my example because I really like it and it's the
easiest way to easily spin up a few dozen machines for just a few hours at a
time, but I'm not emotionally vested in it or anything. If we can find a
university who's willing to rent out a consistent bunch of virtual hardware
for a day, that's good too. For me, it's more about holding a big tournament
where the post-mortem doesn't read like an autopsy.

Cheers,
Adrian

On Tue, Sep 28, 2010 at 11:44 AM, Dave Dyer <[email protected]> wrote:

>
> * Completely predictable and uniform hardware across each competitor.
>
>
> This isn't a realistic assumption either.  The virtual environment may be
> standardized, but
> the physical hardware and collateral environment among all the containers
> is not.
>
> _______________________________________________
> Computer-go mailing list
> [email protected]
> http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
>
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go

Reply via email to