-
Notifications
You must be signed in to change notification settings - Fork 25
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
[python] Partition sparse matrix reads in tiledbsoma.io.to_anndata
#3328
Conversation
tiledbsoma.io.to_anndata
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @bkmartinjr -- this is a tour de force.
Looks good to me 👀 -wise; also, I ran to_anndata
on local disk, S3, and cloud-hosted, comparing main
to this branch, on EC2 Ubuntu m5.4xlarge
, as well as on my M1 Mac. (Several variables there, all test-driven.) (Re Mac: I have a general feeling, more from the R side than the Python side admittedly, that 'threads are a little different' on MacOS so this was worth some interactive/hands-on/eyes-on experimenting.) I saw a speedup at a solid 2X at 100K cells, with all storage backends. I'll try 300K runs later; I won't report back here unless I see something surprising (which I don't expect.)
cc @nguyenv for another pair of eyes.
Partition reads to sparse matrices in the
soma.io.to_anndata
code path. This change is solely for performance, and has no functional impact.#3345 [sc-59595]