Quest v35 slightly different memory footprint>

#1

Hi there, are you aware of the Quest v35 update having a slightly different memory footprint?
I am getting out of memory crashes on models (both cloud and local) that were working fine a couple hours before.
Was on the pre-release version of v35 (via the Test Channel) all working fine, updated to the public release version and now the larger models crash the app.

Also getting a problem occasionally, if it crashes, the headset gats caught in a black no-mans land, sometimes the 3 progress dots are displayed, sometimes not. I can hear button interactions but nothing is displayed and I cannot get back to Home without restarting the headset.

I am on the current app lab release.

#2

Hi! Sorry, it looks like we’ll need to investigate, but I have no idea what could be going wrong at the moment. Did you try with a range of models? Do smaller ones still work fine, whereas only larger ones no longer fit? (The various different error modes you report are definitely matching our experience with out-of-memory errors.)

#3

Yes, this is what I was getting at, those models were definitely knocking on the door of memory errors, in fact I could only get them to work via the cloud rather than direct from Sketchup, so I assume there is some form of compression or block size difference between cloud and local.
Smaller models did work.
I’ve just gone through and optimised further an significantly reduced the file size and face / edge count.
Just uploading now to see if that makes a difference. Will let you know

#4

Ok, that seems to have done the trick. So for the two models that I was having trouble with, I reduced them from about 220Mb to about 155Mb optimising textures and component geometry complexity (then UL’ing to cloud page)
They were loading at around 220Mb previously and it seems that approx 155 - 160Mb is the week spot.
Ill just have to be less lazy around optimisation :wink: