Memory test for 1.1 release

[pixelzoom]

For #191 ... Memory testing should be done before the sim goes into dev test, for phet and phet-io brands.

[pixelzoom]

Testing 1.1.0-dev.16 on macOS 14.5 + Chrome 126. I see no evidence of memory leaks.

Details...

---

phet brand was tested with https://phet-dev.colorado.edu/html/gas-properties/1.1.0-dev.16/phet/gas-properties_en_phet.html?fuzz. No memory leaks noted. Heap snapsnots were taken at 1-minute intervals, and stabilized at ~110MB.

---

phet-io brand was tested with https://phet-dev.colorado.edu/html/gas-properties/1.1.0-dev.16/phet-io/gas-properties_all_phet-io.html?postMessageOnError&phetioStandalone&fuzz. No memory leaks noted. Heap snapsnots were taken at 1-minute intervals, and stabilized at ~130MB.

[pixelzoom]

Please verify the above memory testing results for dev test phetsims/qa#1100.

OK to close if memory testing looks OK.

[KatieWoe]

Looking at the phet-brand, the memory test looks a bit worse, and seems to be leaking at least a bit.

[KatieWoe]

Same thing for PhET-iO Memory

[pixelzoom]

@KatieWoe said:

... the memory test looks a bit worse, and seems to be leaking at least a bit.

The memory tests in #222 (comment) were done with 1.1.0-dev.16, published 6/17/24.

Dev test phetsims/qa#1100 was done with 1.1.0-dev.18, published 6/24/24.

It's possible that a leak could have snuck in during the 7 days between those versions. I'll do another test with a new dev version.

[pixelzoom]

Links for next memory tests:

https://phet-dev.colorado.edu/html/gas-properties/1.1.0-dev.19/phet/gas-properties_all_phet.html
https://phet-dev.colorado.edu/html/gas-properties/1.1.0-dev.19/phet-io/

Testing on macOS 14.5 + Chrome 126.

[pixelzoom]

Testing 1.1.0-dev.19 on macOS 14.5 + Chrome 126. There may be a very slow memory leak, but nothing to be concerned about in practice.

Details...

---

phet brand was tested with https://phet-dev.colorado.edu/html/gas-properties/1.1.0-dev.19/phet/gas-properties_en_phet.html?fuzz. Heap snapsnots were taken at 1-minute intervals, with 10 minutes between Snapshots 9 & 10. Heap size was observed to go down once (Snapshot 8). There may be a very slow memory leak, but nothing to be concerned about in practice.

---

phet-io brand was tested with https://phet-dev.colorado.edu/html/gas-properties/1.1.0-dev.19/phet-io/gas-properties_all_phet-io.html?postMessageOnError&phetioStandalone&fuzz. Heap snapsnots were taken at 1-minute intervals, with 10 minutes between Snapshots 9 & 10. Heap size was observed to go down once (Snapshot 6). As with phet brand, there may be a very slow memory leak, but nothing to be concerned about in practice.

[pixelzoom]

I'll keep this open for verification in 1.1.0-rc.1.

[pixelzoom]

Please verify memory tests (phet and phet-io brands) for phetsims/qa#1107.

[KatieWoe]

Memory tests look similar to before. Small but non-worrying leak.
PhET Brand:

PhET-iO:

[pixelzoom]

Thanks @KatieWoe. We won't be making any changes that affect memory management going forward, so closing this issue.

[pixelzoom]

Reopening because we have a suspected memory problem on iPad. From #289 (comment):

@samreid noted that heap comparisons indicate leaks with PaintColorProperty (related to StopwatchNode) and SpriteInstance. Hard to believe that this would crash the sim in 10 seconds, but I'll reopen #222 and investigate.

[pixelzoom]

#289 turned out to be caused by WebGL, not memory related.

I did another heap size test, and heap size still stabilizes ~140MB. Re "PaintColorProperty (related to StopwatchNode) and SpriteInstance", I didn't see anything alarming. There may be slow leaks, but tracking them down seems like more trouble than it's worth.

So... re-closing this issue.