-
Notifications
You must be signed in to change notification settings - Fork 7
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Error: multiscale level does not fit in preferred memory footprint #217
Comments
Thank you for reporting this! We are currently thinking about making a fix for this to go back to the old behavior. The code is being extra careful to not allocate huge gpu textures or huge amounts of cpu memory. Indeed we will probably remove the error condition, change it to a console warning, and just try to load the lowest multiresolution level. (Not sure why max atlas edge was "undefined" but we will look into that.) |
On a related topic, if I may ... We're trying to use (multi-resolution) Zarr files with Volume Viewer and I was wondering if there's a way for the user to specify the level of the image pyramid that is used for visualization, or even to allow for dynamic display of higher-res images as the user zooms in on the data? Thanks! :) |
You can't explicitly select your multiresolution level right now. The main reason is that we are trying to pick the level automatically to maintain performance within a reasonable memory footprint. (In principle I believe you SHOULD be allowed to explicitly choose this if you know what you're doing). Currently there is a "secret" advanced settings dialog which you can access by clicking ctrl-alt-1 . In there there's a setting to bias the selection of multiresolution level so if you set the bias correctly it should load a higher or lower resolution. The same dialog also has a "maxAtlasSize" setting which lets you basically dial in the texture memory we are using for the gpu volume data. That will also indicate to the viewer that it can pick a higher or lower resolution level. Dynamically changing resolution with zoom is of course a very interesting feature but we haven't started to go down that road yet. |
In a coming build, we have removed the exception error for too-large volumes. They will now put a console error and still attempt to load. Your mileage may vary! Also coming soon, is a "Advanced Settings" setting to let you set the explicit multiresolution level. Unfortunately the advanced settings don't persist yet, but it's a start. Partly noting this here for our internal tracking. It may be a couple weeks still before we issue the next build but the above changes are already on main branch. |
Description
First of all, this is an amazing 3d viewer! Something happened in the last update though, where I can no longer load my OME-Zarrs.
For example, this link used to work:
https://allen-cell-animated.github.io/website-3d-cell-viewer/viewer?url=https%3A%2F%2Fjanelia-flylight-imagery-dev.s3.amazonaws.com%2FFly-eFISH%2FNP01_1_1_SS00790_AstA546_CCHa1_647_1x_LOL.chunked.zarr%2F0
But now it throws an error, although nothing has changed about my hardware or data:
In the console I see this:
I'm wondering if the issue is the undefined
maxAtlasEdge
?Expected Behavior
I would expect at least the lowest scale level to load. The scale level 3 data is definitely small enough to git into my GPU memory.
Reproduction
See link above.
Environment
Firefox 126.0.1, macOS 13.6.1, AMD Radeon Pro Vega II 32 GB.
The text was updated successfully, but these errors were encountered: