index = list.index;
len = list.length;
while (index length) {
It's a problem with len/gth, you overlooked the 'length' in the
while() when you renamed it to 'len'.
Below is the reworked testing class. It creates the array in the
beginning, then
weird results!
I do get very different results from the IDE, but I am testing using
Flash 8 and Flash Player IDE is 8
could that (it should) explain the difference with results seen in
the browser (which has Flash Player 9) ?
Cedric
index = list.index;
len
Hey Cedric,
What results are you getting? Which one is how much faster in what player?
It would surprise me if there were different results for FP8 and FP9,
because it runs in the old VM and I doubt they reworked it since it's
there for backwards compatibility anyway. It's probably more of a
Hey Cedric,
What results are you getting? Which one is how much faster in what
player?
F8 IDE (FP 8 IDE):
flatten: 1576
flatten2: 1554
flatten: 1451
flatten2: 1593
flatten: 1390
flatten2: 1437
flatten: 1526
flatten2: 1439
flatten: 1313
flatten2: 1528
flatten: 1368
Thanks, Cedric.
I assume those were results compiled by the Flash IDE, so we can
conclude recursion *is* the fastest. It's also the most elegant
solution, not requirements in this case, but nice to have anyway.
I still don't understand why Danny got contrary results, but I like
recursion better
On 7/26/06, Mark Winterhalder [EMAIL PROTECTED] wrote:
I still don't understand why Danny got contrary results, but I like
recursion better so let's just leave it like that :)
OK, let's not -- I had another look, and found a ! was missing in the
recursive method. I've updated the test class
OK, my final conclusion:
http://www.snafoo.org/tmp/flatten.swf
Iterative *is* faster than recursion. Run the test and examine the
test class below. (Use an empty .fla, import it, call TestArray.main()
)
Also, we learn that whenever you have to get the item at the same
position of an array
Sweet.
The only other method that uses recursion is eql.
-Steven
___
Flashcoders@chattyfig.figleaf.com
To change your subscription options or search the archive:
http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
Brought to you by Fig Leaf
On 7/25/06, Mark Winterhalder [EMAIL PROTECTED] wrote:
On 7/25/06, Steven Sacks | BLITZ [EMAIL PROTECTED] wrote:
It is iterating. It still needs to iterate through all nested arrays
using the same method, which means recursion.
Yes, but function calls aren't the only way to do recursion.
On 7/25/06, Mark Winterhalder [EMAIL PROTECTED] wrote:
Hmm... now that I'm having coffee and slowly waking up, I'm getting
serious doubts about how to do depth-first without function calls. I'm
probably wrong, sorry.
Interesting problem, though. Got me thinking.
So, here's my proposal for a
On 7/25/06, Mark Winterhalder [EMAIL PROTECTED] wrote:
On 7/25/06, Steven Sacks | BLITZ [EMAIL PROTECTED] wrote:
It is iterating. It still needs to iterate through all nested arrays
using the same method, which means recursion.
Yes, but function calls aren't the only way to do
Ha - you beat me to it...
public static function flatten( inArray : Array ) : Array {
var outArray : Array = [];
var list = inArray;
list.index = 0;
var item : Object;
var index : Number;
var length
On 7/25/06, Danny Kodicek [EMAIL PROTECTED] wrote:
Just out of interest, I tried a speed test using the three methods - mine
was the fastest by a very long way:
Thanks -- I was hoping somebody would test it :)
This seemed a weird result, so I took a look for other differences. It turns
out
BTW, the reason I didn't have a detailed look were the shortened,
non-descriptive variable names. This is most likely not necessary. It
makes sense for properties, but variables usually are stored as
registers (if the compiler chooses to use function2 instead of the old
function tag, which
On 7/25/06, Mark Winterhalder [EMAIL PROTECTED] wrote:
This seems to be extreme. Any chance something is going wrong with the
test? Did you wait a few frames before doing the first test to ensure
initialization has finished? Maybe switch the order to see if the
first one always is the slowest?
First off, thanks to everyone helping on this thread. I originally was
using object == typeof(this[a]). Glad to see instanceof Array was
fastest.
What I'm trying to do here is write new methods for Array (and next up
is String) that other languages have but Flash does not. Some of these
Oh, I just noticed that Mark's was fastest. But, I can't get it to work
(freezes flash) There seems to be an error in the final while loop.
It's a single = not a double ==. Is that a typo?
___
Flashcoders@chattyfig.figleaf.com
To change your
Ok, I got Mark's to work however it does not work when turned into a
prototype and I set list = this. Why it doesn't work may be the result
of another problem, though, and I'm not sure yet what it is, but here's
the behavior.
I created a movie with 3 buttons. An new array is instantiated.
On 7/25/06, Steven Sacks | BLITZ [EMAIL PROTECTED] wrote:
Oh, I just noticed that Mark's was fastest. But, I can't get it to work
(freezes flash) There seems to be an error in the final while loop.
It's a single = not a double ==. Is that a typo?
No, not a typo. I know it's confusing, but
On 7/25/06, Steven Sacks | BLITZ [EMAIL PROTECTED] wrote:
Mark, please check your script to try and find out what's wrong with it.
I'd like to run some more accurate comparisons.
Ah, yes, just noticed I forgot to look into the freezing problem you
mentioned in your last mail (also, I noticed I
One thing I'm wondering is if a decrementing loop and then a reverse()
at the end is faster than an incrementing loop.
I'd be astounded if it was!
I'm working on compacting Danny's script at the moment.
Sounds like an odd idea to me: Mark's was way faster and was essentially the
same idea
Oh, I just noticed that Mark's was fastest.
Sorry, replied to an easlier mail before noticing the thread continued...
Danny
___
Flashcoders@chattyfig.figleaf.com
To change your subscription options or search the archive:
Could you post (or send me) the test script, please?
Sure, here you go.
---
function randomArray(p) {
var a = [];
for (var i = 0; i p; i++) {
if (random(2) random(5)) {
a.push(i);
} else {
(;);
output:
a,b,c,d,e,d,e,d,e,d,e,f,g,h,,i,j
Bernard
-Oorspronkelijk bericht-
Van: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Namens
Danny Kodicek
Verzonden: dinsdag 25 juli 2006 21:51
Aan: Flashcoders mailing list
Onderwerp: Re: [Flashcoders] Attention Recursion Speed Experts
] Attention Recursion Speed Experts
I don't know if I understand this correctly, but is it making a array
flat?
Then this is the fastest way I think:
var foo : Array = [a, b, c, [d, e,[d, e,[d, e,[d,
e, f, [g, [h]], [[], i], j];
var fooString:String = foo.join(;);
fooString = fooString.split
I am assuming this is for AS2 - right?
If you want speed that probably means you are dealing with a lot of data(?)
1- What is the typical recursion level?
2- What is the typical number of items?
3- What is the typical size of sub arrays?
In general making a function call is not fast at all.
] Attention Recursion Speed Experts
I am assuming this is for AS2 - right?
If you want speed that probably means you are dealing with a lot of
data(?)
1- What is the typical recursion level?
2- What is the typical number of items?
3- What is the typical size of sub arrays?
In general
On 7/25/06, Steven Sacks | BLITZ [EMAIL PROTECTED] wrote:
It is iterating. It still needs to iterate through all nested arrays
using the same method, which means recursion.
Yes, but function calls aren't the only way to do recursion.
What you're doing is essentially like a depth-first
28 matches
Mail list logo
