I just did some timing tests, partly to see if using an array might help with
speed. It didn't, but in doing the test I thought of a neater way of testing
than I've done before.
Usually my testing would resemble this:
on oneway
put the milliseconds into t
repeat with a = 1 to 10000
dooneway
end repeat
put the milliseconds - t
end oneway
on otherway
put the milliseconds into t
repeat with a = 1 to 10000
dootherway
end repeat
put the milliseconds - t
end otherway
Then I would type in "oneway" or "otherway" in the message box. It works well
enough, but if I decide to change how many repeats to do I have to go in and
change each place. There's also a lot of same lines, where I'm making note of
the time.
What I did just now is below, and to test the various technique I just type
"timeit technique,repeatcount" in the message box. So for example, to test the
"for each" way of working through a set of items, and I want it to do enough
that I get a good average, I would type "timeit foreach,10000".
The results I got for working through 1000 entries, 10,000 times, was:
for each 1075mS
repeat with 1370mS
count array 4250
global someitems,anarray,t
on fillitems
put empty into someitems
repeat with a = 1 to 1000
put random(100) into item a of someitems
end repeat
put someitems into anarray
split anarray by the itemdelimiter
end fillitems
on timeit what,howmany
put the milliseconds into t
repeat with a = 1 to howmany
do what
end repeat
put the milliseconds - t
end timeit
on foreach
repeat for each item tItem in someitems
end repeat
end foreach
on repeatwith
repeat with i = 1 to the number of items in someitems
end repeat
end repeatwith
on countarray
repeat for each element tElement in anarray
end repeat
end countarray
_______________________________________________
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode
