Re: Memory problems.

[Marcus G. Daniels]

>>>>> "HKM" == Hoe Kok Meng <address@hidden> writes:

HKM> it crashes with the following error message:
HKM> /src/Swarm/swarm/src/misc/xmalloc.c: 26 xmalloc

If malloc fails on Windows 98, you probably ran out of virtual memory.

HKM> The crash normally occurs after 500K iterations. However, the
HKM> program with 15 ants runs fine if I reduce the number of
HKM> iterations to 400K. I suspect a memory leak as the memory monitor
HKM> I'm running shows rather rapid drop in free memory after the
HKM> first 400K or so iterations.

You can use "xprint (aZone)" or "xfprint (aZone)" to tell you about
what's in a Zone.  Substitute "aZone" with globalZone or scratchZone
or [someObject getZone].  Often it makes sense to have "someObject" be
the ModelSwarm.  You can start making the x[f]print calls at 450K or whatever
if you think there is something special about the dynamics that means
leaking only occurs at that late time.  Or just start auditing right from
the start of the run.  Auditing can either be done by inspection
(e.g. in a Emacs shell buffer that you can search around), or using
shell scripts (finding create/drop pointers with the same value).


                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.