Turn on floating point exceptions in tests

We have some issues with floating point exceptions. Usually we have floating point exceptions turned off because frankly we don't care. However, some simulations, which we want to run in situ with, do care a lot about bad floating point math. We should do our best to make sure we don't bring down such a simulation. We probably at some point need to turn on floating point exceptions in one or all dashboards to check for these conditions.