Description of the problem: I got my study running online and it works all fine on the development machine (mac) on Chrome and Firefox. I tried it on a couple of other windows computers as well. In chrome after a dozens of trials the study snaps:
It seems that’s because there is not enough memory, but there is no other data in the browser other than the data produced by the study that far.
Also, on another windows computer, when number ‘1’ is typed in a made up text box, it appears onscreen as ‘num_1’. When number ‘10’ is typed in, it appears onscreen as ‘num_1num_0’.
Thanks in advance for any help with addressing these issues!
I’m experiencing a similar issue.
With an online experiment a have a memory leak which increase the consumption of RAM to the sky (up to swap, on a 64GB ram machine).
I tried different OS and browser (chrome, explorer, firefox). Windows is not very good in managing the memory, in general, and then it crashes. Mac or linux are able to deal without the crash, but they still have the leak.
Hi again,
After a few attempts to trial my study again on Windows Chrome and Firefox, the study eventually worked for me. So today I tried running my study online on Prolific Academic. Out of 40 participants who attempted the study, more than 30 messaged me to tell that the study crashed halfway through completion, some participants tried a couple of times with same result, and in some cases, the study crashed after participants completed 95% of it. In none of these cases any data is saved as participants need to complete the study fully for the data to be saved, i.e., until they receive the completion code. (Not to mention this is very frustrating and irritating for participants as they spend a lot of time to no avail).
Some participants provided some print screens with the issue so you can investigate:
Any thoughts for how can this problem be addressed or what other alternatives do I have? I really need to run the study online since I have to prescreen for certain characteristics and I don’t know how I could run the study and save the data from participants if the study keeps on crashing for participants particularly on windows browsers.
Sorry, I hadn’t seen this post and it seems like it’s a persistent issue. Not sure why Chrome is crashing like this on some machines (and not others). Possibly the crash depends on the version (e.g. if there was a version with a now-fixed error in it) but maybe it’s an issue with a memory leak in the graphics that only shows up on some computers (depending on graphics memory). Graphics memory is separate to standard RAM, which would explain the lack of any problem showing up in the memory listing for the browser.
The difficulty for me and @apitiot is getting a machine where this is happening so we can debug
I’m also having the same issue. My experiment: (https://run.pavlovia.org/Jordan_Gallant/accenttyping/html/)
also uses audio files which my also be straining the RAM used by the browser. I’ve had most of my participants report the same error. The issue seems to occur as the result of a resizing window event (as per console log below)
@unagi_pie that looks like something completely different :-/ Also something that we need to fix somehow, but not the same issue and warratns a separate thread with full info about the browsers on which this occurs and whether it’s constant or sporadic. Thanks
I ran into the same situation today and find this post. Did anybody figure out how to solve this?
When I run the experiment on piloting mode it works perfectly, but when I switch it to running then this situation shows up.
