some loading time tests (not very scientific!)

Sparkyidr

so, I just built a mock-up of an app we are working on.
The mock up is 4 scenes
1. a Simple splash screen, 1 image, 2 actors
2. an options scene. 4 actors with very very few rules
3. a little animation type scene
4. a results scene with 2 actors, each with one rule.

so yeah, a pretty simple project. Did an ad-hoc build to see how the loading times are.

Between scenes the loads are really really good (I assume down to there being few rules, and no sounds)

I set up a simple splash scene to see if it made the initial loads any better rather than going into a scene with loads going on.

Here are my findings.

Initially it loaded to my scene1 in around 2 seconds. wow...awesome.

I quit

next load : 4 seconds
next load : 12 seconds
next load : 12 seconds
next load : 11 (12?) seconds

I then loaded up another app, and quit

Then went back to my app
next load : 5 seconds
next load : 9 seconds
next load : 7 seconds
next load : 12 seconds

not sure what this says.....but thought I would share my findings. Deffo seems like something is happening to make the load times extend, the more you load the app??

This is on an iPhone - 32gig 3GS.

Vmlweb

I keep wondering weather GS Long Loading times are something to do with the analytics system that GS games all have. Maybe it takes a long to send the info to the GS Servers

Vmlweb

I just did a little test that confirms what I just said

iCannon Loading On 3MegaBit Wifi - 11 seconds
iCannon Loading On Cellar Network - 25 seconds

This is on a 16Gig iPhone 3GS

design219

Whoa, this is a little startling. The head chefs, as far I have seen, have not weighed in on this load time issue, and I'd sure like to here them make a statement of some sort.

Others on this forum have mentioned the possibility of memory not being released, or actors not being destroyed when changing scenes and various other possibilities.

Thanks for sharing your test results. Let us know if you learn more.

design219

I just timed by game (Food Fight!) loading after a restart. 6 times the load was exactly the same, 18 seconds with playing the game on some of the loads. I'm on an iPhone 3gs.

I've also wondered if having more scenes helps with initial load time.

Sparkyidr

the number of scenes doesn't seem to affect that initial load time too much.
The only thing I have noticed is that if the initial scene has a lot of compressed audio (ogg) that affected my load time a lot.

It would be nice to know what is actually happening on that initial load.
Like I said in the original post, sometimes, this new app I am working on loads up in 2 seconds...that would be ACE if I could keep it around that.but other times it takes 12 seconds. Seems like a massive difference for loading the same app?

synthesis

@ Sparkyidr
Do you know what level of published app size bloating occurs when adding scenes cause or is the scene overhead minimal and the published app size is more relevant to the library content and actor counts within the scene?

In other words...how much end (published) file size increase occurs if you add a blank unused scene? Anyone know the load time for a "blank" scene?

Sparkyidr

I had always assumed app size was more determined by content rather than the number of scenes.
example:
Donk lite v1.0 was around 2meg.
Donk lite v1.1, which contains 8 more ogg files, and 1 extra 30 second m4a is 7.4 meg
but no extra scenes.

load time seems pretty much the same between v1 and v1.1

I will do some testing later though with a bunch of empty scenes to see how that affects the size.

synthesis

Cool!!! YOU DA MAN! (or woman if the case may be

BTW:
A 5.4 MB bump for 9 audio files seems high.
Our music track for one scene is 54 seconds and the file size is around 800KB and sounds fine.

Unless all 9 of your audio clips were around a minute long...it sounds like too much.

You may need to optimize some...unless you want all that HI-FI quality in there.

Sparkyidr

@synth...last time I checked I was da "man" ;o)

Yeah, the audio was all nice quality stuff (am a studio engineer/producer in the day job so quality is important to me)

I got my numbers wrong though
1.0 was 1.8meg
1.1 was 6.4 meg
so a bit less of a difference than I first reported. )

synthesis

If you "check" too often...some might say you're playing with it :P

re: audio file sizes:
Sure quality in important...but in game design...optimization is more important.
Optimize and you can get more samples on there and make the game richer and deeper.

Just a thought.

Sparkyidr

normally I would agree with you. But the "game" in question was actually a soundboard. )