Alex:
Just want to comment on the dual core issue. Multiple E+ runs work nicely
on a dual core machine (and would subsequently work with four on a quad
core). I wrote a testing script that monitors each processor on the machine
while constantly running two different E+ instances. When only one runs, my
machine (Lenovo laptop, 2GHz Intel Core Duo, 3GB RAM, XP Professional)
doesn't notice anything is happening. When both processors fill up, the
machine does lag, but not that bad to where it completely halts. The issue
that most people come across when wanting to improve E+ performance is that
it can't easily be applied to run a single simulation on multiple computers.
In a program like Fluent, for instance, the CFD grid can be split, and each
processor can simulate its part, and only need to communicate at the
interfaces. For a building simulation like E+, there isn't really a logical
way to split the data, since each piece of equipment may need to know
information from all sorts of other parts of the simulation.

Two pieces of advice:

If you can get your parametric study to run two simulations (two instances
of EnergyPlus.exe) concurrently, that should nearly double the performance
on a dual core/dual processor machine.

While you are debugging your model, make sure you only run a single design
day, and not any run periods. If the design days come through without
problems, the run period is **likely** to not have any either. (key is the
**'s)

Edwin Lee