add Lee-Ping's MTSVVVR custom integrator

[dacase]

Jason: I've added Lee-Ping's MTSVVVR custom integrator to runmd.py. It seems to be working, but
has not had a lot of testing. I use the following arguments to run it:

--gamma_ln 1.0 --nrespa 4 --dt 2.0 --ntp

I modified what Lee-Ping had done by making MonteCarloBarostat a slow force, which seems right to me.

This is not a very elegant implementation: the integrator is just defined within the runmd.py file. And I may be importing more things than I need to from simtk. But thought you might want to look at it. Plus, this is my first time at creating a gitub-style pull request!

...dac

[swails]

I modified what Lee-Ping had done by making MonteCarloBarostat a slow force, which seems right to me.

I'm not exactly sure what this does. The barostat only acts every 25 steps (by default -- you can change this). And it gets the step number from the context, I think. Which means "steps" refers to outer-time steps. It may count function calls, but that seems unlikely to me. I agree with the assessment that it should be a slow force, but I'm not sure exactly what this implies for the Barostat. But where did you set the force group of the Barostat? I didn't see that in the commit (maybe I just have to look harder).

On the whole, looks good -- go ahead and click merge :)