Skip to content
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

Ensuring mode solver fields are returned in requested precision #1266

Merged
merged 1 commit into from
Nov 27, 2023

Conversation

momchil-flex
Copy link
Collaborator

This was motivated by two things when using the default single precision:

  • Smaller file size for download
  • Smaller chance of OOM which can sometimes happen for very large mode solves in the post-processing (colocating, normalizing, sorting modes).

Copy link
Collaborator

@tylerflex tylerflex left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me. I feel like eventually we can move towards using Enum for things like this, but for now it's ok. Maybe along with #564 we can try to use enums internally as well throughout the codebase. @daquinteroflex

@dbochkov-flexcompute
Copy link
Contributor

Looks good to me as well. When slicing data for GUI I convert everything into single precision no matter what's the input to make storage and downloading more efficient, do you think that's ok to keep it that way? (cc: @xin-flex)

@momchil-flex
Copy link
Collaborator Author

Conversely this makes me wonder if we should just always store the mode fields in single precision? The precision argument was added because in some cases it may be important for the mode solve; but casting the final results seems pretty innocuous?

@momchil-flex
Copy link
Collaborator Author

I think using the "least amount of work" razor we can assume both are fine as they are. :D

@momchil-flex momchil-flex merged commit 708efca into pre/2.5 Nov 27, 2023
14 checks passed
@momchil-flex momchil-flex deleted the momchil/mode_precision branch November 27, 2023 19:10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants