Open GoogleCodeExporter opened 8 years ago
Unless this affects performance a lot...
Original comment by rany...@gmail.com
on 22 Nov 2010 at 6:22
One reason against doing this is the increased verbosity in the specification
of the propensity functions, Although it may make it more readable,
particularly in large models.
with regards to performance, I'd have to see.
Also, I think that rates being a list makes the code much simpler cause we can
use the positional information. Dictionaries do not maintain the original order
Original comment by fccoelho
on 22 Nov 2010 at 6:32
You can always use the sorted keys to get the same order, but when do you use
this information in the code?
Original comment by rany...@gmail.com
on 22 Nov 2010 at 6:38
In the GSSA method in the Model Class.
if you want to try and modify it, go ahead.
Original comment by fccoelho
on 22 Nov 2010 at 7:15
I just modified it and it works, I haven't checked the effect on runtime yet.
Original comment by rany...@gmail.com
on 22 Nov 2010 at 7:32
Attachments:
Thanks,
I'll test them side by side and see which one is faster. I guess it'll be a
matter of which is faster: an array or a dictionary lookup.
Original comment by fccoelho
on 23 Nov 2010 at 9:35
Original comment by fccoelho
on 23 Nov 2010 at 9:36
I tested it once yesterday.
The tuple took 27 seconds to run,
a dictionnary with numerical keys took 35 seconds,
a dictionnary with string keys took 47 seconds.
These numbers are just from one time each, but it's almost double the time...
Oh well = )
Original comment by rany...@gmail.com
on 23 Nov 2010 at 10:13
Original issue reported on code.google.com by
rany...@gmail.com
on 22 Nov 2010 at 6:21