CPU not calculated correctly in ThreadCpuTimer

Naming is hard, and the initial implementation seems to have stumbled on a method naming confusion with the effect that it was only possible to hit a cpu time out shortly after a new thread is created (which is true when the tests run, and thus the test was passing). Essentially it was grabbing an elapsed time for the timer and, treating it as the total time for the thread.

Fixed locally, will submit PR soon...