I am running CMU CL, version 17f, on a Sparc 1000. I have noticed that
(get-internal-run-time) always returns 0. This means that the profiling
utilities say every function takes 0.00000 seconds to run.
Does anyone know what could be wrong? Is the current value of
internal-time-units-per-second [which = 100] bad? Or is the problem that
the machine has more than one processor? I should also say that the system
was heavily loaded by other people when I was doing this.
-------
As a quick fix, I have redefined the function as follows:
(defun get-internal-run-time ()
(get-internal-real-time)) ; the real-time one does seem to work
This makes the profiling code give values that mean "something".
Thanks.