Closed twant closed 5 years ago
Yes a discussion sounds good. First off, these numbers all look great, and we probably don't need to do anything else here. But for learning's sake:
Did you by change save the snapshot data, or look into comparing earlier snapshots to later ones? I only mention because it can be very valuable to compare from a fuzz of about 5 minutes, to one a while later, to see what junk has been accumulated. From there looking for Molarity-specific types whose numbers seem to be growing.
A brief discussion of any questions seems good before a close.
TODO: fuzz for 15 min and compare snapshots 2 & 3 (not on startup) to look for anything piling up.
@zepumph Interesting development in this...when I went back to more closely examine the heap snapshots, I got very different test results. I took snapshots every 5 min, and each time it went up by at least 10MB. I've saved a couple of the snapshots.
I'm currently taking a look to see if I can figure out, but wanted to flag that we should probably discuss on Tuesday!
False alarm, was looking at the a11y view in this process! Just tested again and saw no evidence of a memory leak. Closing.
Both tests were conducted over a period of 4 hours, taking snapshots every 15 - 60 mins.
PhET-iO:
Regular sim:
@zepumph can we discuss these results in our next meeting?