Dear List,
Many of my CSPs eat *very* much memory. Of course, recomputation helps.
Still, it would be even better if less memory would be used in the
first place. So, I try to analyse how I can rewrite my CSPs..
Most of my data structure does not change at all during the search
process. Therefore, I came up with the idea to leave the data structure
outside of the script (in the top-level space) and only put the
variables inside. I only need some linking from the data structure to
the actual variables, because the meaning of these variables (and their
constraints) depends on this data structure. Then I can see the data
structure from inside the script, and seemingly apply constraints on
this data structure (which actually constrains variables local to the
script). Of course, such an approach makes the definition of the CSP a
bit less convenient. Therefore my question: would it actually be worth
the effort (a little inital test says otherwise).
I did some initial profiling. To my big surprise I learned that my data
structure on its own actually takes much memory than cloning a space
which contains this data structure. Can this actually happen (or does
this only confirm that I am a total newbie in profiling ;-) )? Which
brings me to my question: What is actually copied when a space is
cloned?
Thank you very much!
Best,
Torsten
--
Torsten Anders
Sonic Arts Research Centre
Queen's University Belfast (UK)
www.torsten-anders.de
_________________________________________________________________________________
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users
