Fix calls to Mersenne Twister for CBMC functions

[LSchwiebert]

The previous patch for issue #498 didn't address the problem for functions that were in the CBMC branch, so the same problem of inconsistent results was recurring for simulations that ran on gcc versus icc.

This patch updates those files. Also did a search and there don't seem to be any other function calls that might have this same problem, so hopefully this patch resolves the issue completely.