Erlang & Hyperthreading

[Thomas Lindgren]

--- Ryan Rawson <ryanobjc@REDACTED> wrote:

> In my circumstance, I run a mnesia database on every
> node.  Each node
> answers questions from its local database.  So
> running N nodes on a
> N-CPU/SMP system ends up with N copies of the
> database on 1 machine.

This is one case where using multiple nodes easily
gets complicated or expensive.

> That isn't the end of the world, since practically
> any Unix/Linux
> application on a 32 machine can't use more than 1.5
> GB RAM, but the
> issue I'd be worried about is the additional
> communications overhead.

OK, this again depends on your architecture. In some
cases, a given job can be broken down into pieces that
can be passed around the nodes with relative ease. In
other cases, e.g., e-commerce or so I'm told, you
might be able to spread/scale the data and database
over many nodes. In yet other cases, well ... bummer,
man.

It might also be worth measuring the actual
communications overheads for your specific case. While
waiting for the multithreaded VM, hey? :-)

Best,
Thomas


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com