Bad performance on build 2015.2646

Hello,

when compling my game using Corona build 2014.2393, the performances are great, but then, when I updated my corona App to build 2015.2646, performances became horrible, the game is unplayable.

I’m using Graphics 2.0 on both version, since I’ve downgraded to build 2014.2393 from https://developer.coronalabs.com/downloads/archive

What’s the reason that makes my game look laggish on newer Corona build?

I need to use the latest Facebook plugin (which is not supported on build 2014.2393)

Thanks.

Sluggish on what devices, which OS version, etc.  Give all the details.  Also, have you looked at your logs?  Are you getting any warnings, error messages, etc?

Lastly, why 2646?   2678 is the latest.  Try that and see if things improve?

Tip: Don’t get builds from the public release.  Get them from daily builds if you want to stay up-to-date:

https://developer.coronalabs.com/downloads/daily-builds/

PS - There is a small risk of bugs and instability in daily builds, but I find using them is almost always better than using an older public release.

Thanks for your answer. I tried on Galaxy Note 3 with KitKat 4.4.2 and Lollipop 5.0, as well as the iPhone 6 Plus with iOS 8.4

Anyway I tried to use the last daily build, but nothing changes. 

The logs only say “Simulator does not support multitouch events”.

Once again, building with build 2014.2393, the game works fine.

What does the logcat output say for the devices anything interesting?  How about the xCode console output for the apple device?

Also, have you tried any of the demos that come with Corona and compared their performance?

I’m not seeing any performance issues on my Nexus 7 with the latest daily build, but since I don’t know what you’re doing in your app, we’re really comparing apples to oranges.

You may need to get someone to look at your code and help you debug this.

Also, you may want to start at version 2014.2393 and move forward, trying to find what first version of Corona starts to perform badly.  This will help the engineers find the problem.  I know this might be painful, but ‘2014.2393’ is WAY back.  i.e. Lots has changed.

I’m not suggesting you install and build with each version.  Rather, skip ahead quickly till you first see issues, then go backwards till you find the version where the performance first falls apart.

Ex: ???.2393, ???.2493, ???.2593, …

Sluggish on what devices, which OS version, etc.  Give all the details.  Also, have you looked at your logs?  Are you getting any warnings, error messages, etc?

Lastly, why 2646?   2678 is the latest.  Try that and see if things improve?

Tip: Don’t get builds from the public release.  Get them from daily builds if you want to stay up-to-date:

https://developer.coronalabs.com/downloads/daily-builds/

PS - There is a small risk of bugs and instability in daily builds, but I find using them is almost always better than using an older public release.

Thanks for your answer. I tried on Galaxy Note 3 with KitKat 4.4.2 and Lollipop 5.0, as well as the iPhone 6 Plus with iOS 8.4

Anyway I tried to use the last daily build, but nothing changes. 

The logs only say “Simulator does not support multitouch events”.

Once again, building with build 2014.2393, the game works fine.

What does the logcat output say for the devices anything interesting?  How about the xCode console output for the apple device?

Also, have you tried any of the demos that come with Corona and compared their performance?

I’m not seeing any performance issues on my Nexus 7 with the latest daily build, but since I don’t know what you’re doing in your app, we’re really comparing apples to oranges.

You may need to get someone to look at your code and help you debug this.

Also, you may want to start at version 2014.2393 and move forward, trying to find what first version of Corona starts to perform badly.  This will help the engineers find the problem.  I know this might be painful, but ‘2014.2393’ is WAY back.  i.e. Lots has changed.

I’m not suggesting you install and build with each version.  Rather, skip ahead quickly till you first see issues, then go backwards till you find the version where the performance first falls apart.

Ex: ???.2393, ???.2493, ???.2593, …