Determining the total Mechanical Loss

I need to be able to extract from my model the total mechanical loss from all sources, including the internal friction, the thermoelastic loss, radiative cooling, etc, as a function of frequency. I am currently performing a frequency sweep and fitting in MatLab, the frequency response of the total stress energy. At resonance the loss is the peak width divided by the resonant frequency. This works well enough but:

- it is slow because the MatLab livelink seems to be limited to using 1 core of the server, and
- it is limited to results at the resonant frequencies.

Is there a faster way? Is there an example model of such a calculation?

Also, what is the method for adding in loss mechanisms not prebuilt into COMSOL? An example model would be most welcome.