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

Add IntegralsArblibExt #199

Merged
merged 5 commits into from
Nov 19, 2023
Merged

Add IntegralsArblibExt #199

merged 5 commits into from
Nov 19, 2023

Conversation

lxvm
Copy link
Collaborator

@lxvm lxvm commented Nov 19, 2023

This pr adds an extension for the recently-released Arblib.jl package (v1) with high-precision integration routines for holomorphic functions. Although the package is somewhat limited (i.e. no vector integrands or batching) I think it would be interesting to have alongside other algorithms.

I am not a user of Arblib.jl so I may have made a mistake in the implementation regarding the handling of the special number types in the package. In particular, the result type is a number containing the error estimate, instead of the integral and error estimate being separate like other Integrals.jl algorithms

Copy link
Member

@ChrisRackauckas ChrisRackauckas left a comment

Choose a reason for hiding this comment

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

Wow I didn't know this existed, that's really cool. Nice addition.

@ChrisRackauckas ChrisRackauckas merged commit 1174432 into SciML:master Nov 19, 2023
5 of 6 checks passed
@ChrisRackauckas
Copy link
Member

In particular, the result type is a number containing the error estimate, instead of the integral and error estimate being separate like other Integrals.jl algorithms

I do wonder if we should split it. But the Arb type generally has the radius as part of it, so it would be weird to not return a valid Arb. Maybe an improvement would be to just define the resid based off of the radius part anyways as a nicety to users, but keep the return val intact. Since that wouldn't be a breaking change I just merged this anyways, but it might be good to get an issue or follow up PR on that.

@lxvm
Copy link
Collaborator Author

lxvm commented Nov 19, 2023

Sure, extracting the radius as the error estimate makes sense, but from what I could test, the only methods that worked were abs_ubound and abs_lbound. I suppose that max(abs(abs(midpoint(ball)) - abs_ubound(ball)), abs(abs(midpoint(ball)) - abs_lbound(ball))) would give the radius

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.

2 participants