Memory allocation

Delaney, Timothy C (Timothy) tdelaney at avaya.com
Wed Sep 24 00:34:26 CEST 2003


> From: Batista, Facundo [mailto:FBatista at uniFON.com.ar]
> 
> The detail is that I *already have* the serialized object in disk, so:
> 
> 	1. Load the pickled object from disk (already must)
> 	2. Unpickle the object to memory (already must)
> 	3. Keep the "backup" pickled object in memory
> 	4. ...
> 	5. Pickle the object (already must)
> 	6. Compare to the backup object
> 	7. If different, write to disk (already must)
> 
> So, the only overhead in the change of process is 
> 
> 	1) keep the backup pickled object in memory: can do this if the
> memory space overhead is not too big. 
> 	2) compare the serialized objects: if is to prevent a 
> disk access,
> makes the system faster.

So why not just try it? Set up a number of test cases, using various amounts of data, then see what happens.

It would take a lot less time to just set this up and try it than to ask questions that have no right answer.

Tim Delaney





More information about the Python-list mailing list