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

Allow changing data type of EQLs Jacobian elements #176

Open
santisoler opened this issue Jul 17, 2020 · 0 comments
Open

Allow changing data type of EQLs Jacobian elements #176

santisoler opened this issue Jul 17, 2020 · 0 comments
Labels
enhancement Idea or request for a new feature

Comments

@santisoler
Copy link
Member

Description of the desired feature

We know that fitting an EQL gridder is a process that requires a fair amount of memory, mainly to store the Jacobian matrix.
One small step into reducing the required memory would be to change the data type used to store the Jacobian elements, which are defined as float64 by default.
Currently, the EQLs offer no option to change that, even though the jacobian method has an optional argument dtype to change it.
Would be nice to have a jac_dtype argument that could be passed to the EQL constructor in order to change the representation of the Jacobian elements.
Although the memory consumption won't drop by a lot, users may reduce it by half storing the elements as float32.
I suspect that using float32 won't significantly compromise the accuracy of the predictions.

Are you willing to help implement and maintain this feature? Yes

@santisoler santisoler added enhancement Idea or request for a new feature help wanted good first issue Good for newcomers (doesn’t require deep knowledge of the project) labels Jul 17, 2020
@santisoler santisoler changed the title Allow changing data type for EQLs Jacobian matrix Allow changing data type of EQLs Jacobian matrix Jul 17, 2020
@santisoler santisoler changed the title Allow changing data type of EQLs Jacobian matrix Allow changing data type of EQLs Jacobian elements Jul 17, 2020
@santisoler santisoler removed the good first issue Good for newcomers (doesn’t require deep knowledge of the project) label Oct 1, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement Idea or request for a new feature
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants