lou@bearcat.rutgers.edu (Lou Steinberg) (12/09/88)
We are trying to time the execution of some functions, and are getting
strange results. This is on an 1109. It appears that there are two
functions that give what we want (cpu time without gc or swap time):
breakdown and timeall. But when we use breakdown on the same computation
several times, we get numbers that vary much more than I would expect
(20% or more). Furthermore, comparing breakdown and timeall it seems
that they give the same time when there are few page faults, but sometimes
(not always) when there are many page faults breakdown gives a higher
cpu time figure than timeall.
Can anyone clarify this mystery for us, or give suggestions how to do the
timing we want?
Thanks
(Please respond by email as I do not keep up regularly with this newsgroup.)
Lou Steinberg
uucp: {pretty much any major site}!rutgers!aramis.rutgers.edu!lou
arpa: lou@aramis.rutgers.edu
--
Lou Steinberg
uucp: {pretty much any major site}!rutgers!aramis.rutgers.edu!lou
arpa: lou@aramis.rutgers.edu