The sim really stutters when the mouse changes from hand to arrow and back

[samreid]

Win8/Chrome, The sim really stutters when the mouse changes from hand to and back

[samreid]

This was also observed by an independent group in Fluid Pressure and Flow, and we have promoted it to a general scenery issue: phetsims/scenery#275

[samreid]

Same issue on OSX10.9.2/Chrome37. Pretty severe.

[ariel-phet]

This seems like a high priority to fix as a scenery bug

[samreid]

@jonathanolson can you help look into this?

[jonathanolson]

Didn't find any references to this in the Chromium bug tracker. Will try investigating workarounds.

[jonathanolson]

Please see comments in phetsims/scenery#275, it's somewhat written with ESPB in mind (large number of nodes).

[samreid]

@jonathanolson said:

I'd highly recommend combining many line/path elements into a single path (see GridNode, GaugeNode), and minimizing the number of nodes (it's reflowing the invisible ones also). I see ToggleButtonDeprecated with 4 copies of things. The Scene when loaded in the 1st tab has ~1320 nodes in the tree (including all screens), so I'd recommend testing with the add/remove child for screens instead of visibility toggles to see if that helps.

I tried commenting out the ticks in GaugeNode and GridNode, and getting rid of the ToggleButtonNode, but the reflow time is still very long in Win8/Chrome (difficult to tell if it helped much just by looking).

[samreid]

Additionally setting screenDisplayStrategy: 'setChildren' seems to have significantly reduced the problem. Switching screens doesn't seem like it takes so long that it would cause problems. Maybe this is a good tradeoff.

[samreid]

Here is the current size of the scene graph:

1305 lines in the original version.
580 lines when using 'setChildren' instead of 'setVisible'
546 lines when using 'setChildren' and no gridnode
538 lines when using 'setChildren' and no gaugenode ticks
532 lines when using 'setChildren' and no ToggleNodeDeprecated.
456 lines when using all changes above

[samreid]

Converting AttachDetachToggleButtons to use RadioButtons leaves the node count unchanged at 344.

[samreid]

I switched from DeprecatedToggleButtons in the SceneSelectionPanel in #243 and now the node count is down around 578

[samreid]

Hmmmmmm... If I put ?standalone, then the node count drops to 372. Are we really using 206 nodes in the home screen + navbar?

EDIT: Home screen => home screen + navbar.

[samreid]

After the above commit, I am down to 542 nodes.

[samreid]

After consolidating GridNod lines above, I am down to 510 nodes, and the delay on Win8/Chrome is looking significantly better.

[samreid]

By the way, @jonathanolson said joist/scenery02has consolidated many joist nodes, which may reduce the number further (nearly half of the nodes are now joist).

[samreid]

Checking with ?standalone reveals almost no delay, so perhaps joist is the next place to optimize. I also saw many duplicated nodes with the new RadioButtonGroup, perhaps that is another place we can optimize.

[samreid]

I tried getting @jonathanolson's changes to NavigationBarScreenButton from joist/ohtwo and this reduced the number of nodes from 509 to 344. In my opinion, that is very significant. I'll look into merging that commit into master.

[samreid]

Well, we have gone from 1305 lines and a huge delay to 344 lines and a very short delay, so I think this issue is ready to be closed. We will need to test the new navbar buttons, switching screens, the grid and the speedometer to make sure no new problems were introduced in these changes.

[samreid]

After rendering some things in WebGL, we are down to 303 nodes. However, this latter reduction did not seem to reduce the "hitch" time significantly.

[samreid]

In my opinion, this problem is sufficiently reduced for this sim, and acceptable for publication. Closing.

[samreid]

This got worse in Chrome 38, see phetsims/scenery#275 and noted in #294

[bryo5363]

I did not notice any changes in the simulation performance in Safari 9537.85.10.17.1

[samreid]

This problem seems chrome specific.

[samreid]

Not quite as bad on OSX 10.10 + Chrome Version 38.0.2125.104 on my MacBook Air (11-inch, Mid 2012)

[samreid]

Also not much we can do about this until we are focusing more on WebGL.

[ariel-phet]

Agreed, wait until we have more webgl support.

[jessegreenberg]

I am not seeing this happen at all, though I only have access to Windows 10. The general issue phetsims/scenery#275 has been closed because starting May 2015 developers began to report that this is no longer an issue. @samreid can this issue be closed?

[samreid]

Yes, thanks! Closing.