You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A given notebook session can eventually run out of memory at which point the only way to fix it appears to be reloading the page.
My expectation is that when the output of a cell is cleared, the memory should be freed.
To repro, repeatedly execute a cell that outputs a viewer widget until it stops outputting.
At this point the console output for the page will show:
WebAssembly module validated with warning: failed to allocate executable memory for module
Uncaught (in promise) InternalError: out of memory
It would be nice if we could also catch this and show the error in the notebook. Instead all we get is an empty white canvas and the newly introduced timeout warning.
The text was updated successfully, but these errors were encountered:
It's new. The old notebook approach added an iframe to the page. When the previous iframe is removed from the DOM, it seems like the right cleanup happens.
Digging into the memory inspector and taking snapshots after each cell execution, you can see we leak approximately 100MB per execution. Something is holding a reference that keeps it from being GC'd.
A given notebook session can eventually run out of memory at which point the only way to fix it appears to be reloading the page.
My expectation is that when the output of a cell is cleared, the memory should be freed.
To repro, repeatedly execute a cell that outputs a viewer widget until it stops outputting.
At this point the console output for the page will show:
It would be nice if we could also catch this and show the error in the notebook. Instead all we get is an empty white canvas and the newly introduced timeout warning.
The text was updated successfully, but these errors were encountered: