After a year of trying they've finally got round to sorting out our distributed simulation across multiple PCs problem. Now we can use other computers cores to make our models run faster.
We're got 9 quad cores and a duel core, 38 cores for those not great at maths. Previously we could only use a maximum of 4, now we can go 9.5 times faster (not quite as it takes a while to copy across to each new machine).
One of our models takes an hour an a half to run, I reckon we can get it down to just over 10 mins. Brilliant.
The best part about it is my other team members won't know who's stealing their cores. Ha, ha!