Hi!
With my last request, why ACL/W 2.0 seems to be soooo slow under NT, I got
the answer that the task switch system might be the problem.
OK! That sounds good, BUT: how do I get a correct time ???????
In WIN32s I have task switching, too, i.e. while calling
(process-pending-events).
Therefore, the time I get with (time <form>) or the profiling tools are
not the real time.
In MCL the time function is so smart to differentiate the time in
time used for GC, time used in the 'cooperative multitasking' (which isn't
very exact either, but better than nothing) and really used time in LISP
code.
Does anybody knows of a patch for the ACL time/profiling function(s) to
get a similiar behaviour?
BTW, the time function in MCL also prints the amount of memory used during
execution. That would be great in ACL/W, too!!
- stefan
_____________________________________________________________________
*** Support bacteria -- it's the only culture some people have! ****
_____________________________________________________________________
Stefan K. Bamberger email: ·····@informatik.uni-wuerzburg.de
Lehrstuhl f"ur Informatik VI voice : ++49 931 7056114
Universit"at W"urzburg / Germany Fax : ++49 931 7056120
_____________________________________________________________________