search

mbakker7 / ttim

MIT License
34 stars 23 forks source link

attempt fix for numba #41

Closed dbrakenhoff closed 1 year ago

dbrakenhoff commented 1 year ago

Attempt to get CI to work again...

mbakker7 commented 1 year ago

Thanks for diving into this problem. I think you isolated it nicely. Strange that it didn't happen before. The solution seems easy: We need to divide the times over the intervals only once and not twice. TTim already does this for the model.potential and model.disvec functions by calling the invlapcomp function. The only place that it doesn't happen is the Element.discharge function. This must have been an oversight when invlapcomp was created.

I propose to close this PR and open a new one where we use the invlapcomp function inside the Element.discharge function. I will modify the Issue accordingly.

dbrakenhoff commented 1 year ago

Sounds good 👍. Closing this PR.