Tapping screen on an iPad2 slows down the simulation

[veillette]

While investigating phetsims/plinko-probability#26, we found that tapping plinko-probability slows down the animation. On the second tab, playing the simulation in continuous mode on iPad2, and tapping the blue background slows down the rendering from 45fps to 20fps.

Here is the link to the dev version.
http://www.colorado.edu/physics/phet/dev/html/plinko-probability/1.0.0-dev.5/plinko-probability_en.html

This is mostly apparent in simulations that have a running animation. Using Rutherford scattering as an example, then one can slow down the animation by simply tapping the simulation repeatedly (say 5 times a second) anywhere on the screen.

Tapping the play button repetitively in plinko is an expected user interaction.
The browser found the issue using safari but we can reproduce it using chrome ( 51.0.2704 ) as well.

[veillette]

I have tried to reproduce on a MondoPad (a giant touchscreen, running windows 7) using chrome 51.0.2704 but to no avail. The MondoPad runs plinko at 60fps no matter how much tapping we do (we tried with 30 fingers of tapping). For reference, tapping down with 5 fingers repetitively bring the iPad to 2fps.

[jessegreenberg]

I just tried to observe this issue in rutherford-scattering and plinko-probability, but I could not reproduce in iOS 9.3.1 with an iPad2. I tried tapping repetitively with 1-5 fingers.

[pixelzoom]

6/30/16 dev meeting: Suggested to try built version with all query parameter options off. Also try with and without other tabs open in the browser.

[jessegreenberg]

I just tested this plinko-probability 'Lab' screen in requirejs. The frame rate hovers around 26 fps. When I repeatedly tap the blue background with 5 fingers, the fps drops to 16.

[memo330179]

It would also be worth looking into inputs. Pressing a button on any simulation on the iPad causes the fps to drop. If a button is pressed repeatedly then the fps can drop to below 10. This is true for most simulations, but most simulations don't involve much button pressing by design. For Plinko being able to press a button repeatedly is necessary for the simulation to work properly.

[jonathanolson]

I'd recommend checking what is hit tested, and how expensive it is to hit test. You can run popupDebug and look for the "hits" text in the output to see what is hit tested. Remember that hit-testing Path is much more expensive than hit-testing the optimized versions (Circle,Rectangle/etc.)

[veillette]

Since (1) IPad2 is no longer a supported platform, (2) Plinko-probability has been published., it is safe to tag this issue as stale and close it.