cocos2d for iPhone

A fast, easy to use, free, and community supported 2D game engine

1. 

   zenkimoto
   Member

   I am new to iPhone development. If Riq is reading, I wanted to thank him for all his hard work.

   I have dabbled in other game engines before in the past with the more traditional draw/update game loop. I'm curious about the performance of the cocos2d iPhone engine with regards to timers. Is the standard practice to have a lot of timers controlling different game objects (and in this case, I would assume that it would slow down performance)? Or is it best if there is only one timer that updates objects? What's the maximum number of timers that should be used? And what about event handling such as collision detection, should that be on a separate timer or fall under the main step timer?

   I realize that's a lot of questions. I would appreciate any ideas anyone can provide.

   I would assume that the best practice is to have one timer (i.e. step:) updating all the objects and handling events. This would be more similar to the traditional draw/update.

   Posted 2 years ago #
2. 

   riq
   Key Master

   The timers are optimized, but it is better if you use just 1 timer for your game instead of several timers.

   As you said, just schedule 1 timer (step: or update:, or the name that you prefer) and handle all your updates there.

   There is a work-in-progress in branches/branch-0.9-scheduler which is a big refactoring of the scheduler/cocosnode#schedule, but at the moment the performance is not as fast as trunk.

   Posted 2 years ago #
3. 

   Tim
   Member

   I'd like to also add that if you start using multiple timers then it becomes a management problem IMO cause then you're having to make sure to stop/start the various timers as your game logic plays out. For my 3 current apps I use a primary game loop, but for the ones that use accelerometer input I'm using another timer for that method since the updates are more frequent than the overall game loop. YMMV, but this seems to work for me.

   Posted 2 years ago #
4. 

   zenkimoto
   Member

   Thanks for the quick responses.

   So my next question is, should the game loop timer update positions of game objects rather than using actions? If I move an object from say from (0, 0) to some (x, y) using an action, how do I know when it's there when using an action? Do I check the change in time (for the duration of the action) or do I check the position of the object at each iteration of the game loop?

   Is it best to use actions for effects purposes and to manually control game objects in the game loop (or by using a physics engine)?

   Posted 2 years ago #
5. 

   Tim
   Member

   You've got multiple questions there that all depend on what you're doing and want to achieve. I've used the game loop to call a method to update the movement of an object as well as to check boundaries (like what you've mentioned about managing it's coordinates). But I've also used it to kick off and stop actions (e.g. having a running counter that's updated each time the loop is called and once a certain value was hit, initiate wave of baddies).

   Lam posted a while back about creating methods to handle your own actions so as to avoid the allocation/deallocation process for when running actions and the subsequent impact to FPS. This would also be in line with what you're asking in terms of knowing where an object is. There is no particular right or wrong IMO but rather it's driven by what you want to achieve. I just started using Chipmunk and found that it was great in that it got me away from having to deal with stopping and starting actions based upon impact let alone dong the cgpoint/cgrect check for everything. So again, I think it depends on the scenario and what you'd like to achieve cause you could find that you might need to do all three you mentioned.

   Posted 2 years ago #
6. 

   riq
   Key Master

   @zenkimoto: You can use the CCSequence + CCCallFunc actions to know when a certain action is finished.

   For example:

   id move = [CCMoveBy action...];
    id call = [CCCallFunc actionWithTarget:self selector:@selector(done)];
    id seq = [CCSequence actions:move, call, nil];
    [sprite runAction:seq];
   
   -(void) done
   {
      // sprite finished the movement
   }

   Posted 2 years ago #

RSS feed for this topic

Reply

You must log in to post.