search

SciML / SciMLBase.jl

The Base interface of the SciML ecosystem
https://docs.sciml.ai/SciMLBase/stable
MIT License
118 stars 91 forks source link

autospecialize rather than full specialize by default for the various `NonlinearProblems` #681

Closed oscardssmith closed 2 months ago

oscardssmith commented 2 months ago

This doesn't change anything yet, but is a step in the direction of making NonlinearSolve not specialize on the user given function which will reduce compilation time on first nonlinear solve.

codecov[bot] commented 2 months ago

Codecov Report

Attention: Patch coverage is 66.66667% with 3 lines in your changes are missing coverage. Please review.

Project coverage is 31.67%. Comparing base (9e92fc6) to head (671c208).

Files Patch % Lines
src/problems/basic_problems.jl 66.66% 3 Missing :warning:
Additional details and impacted files ```diff @@ Coverage Diff @@ ## master #681 +/- ## ========================================== - Coverage 31.74% 31.67% -0.07% ========================================== Files 55 55 Lines 4505 4508 +3 ========================================== - Hits 1430 1428 -2 - Misses 3075 3080 +5 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

ChrisRackauckas commented 2 months ago

might as well

oscardssmith commented 2 months ago

are you sure the CI failures here weren't real? They looked plausibly related

avik-pal commented 2 months ago

They do seem related

ChrisRackauckas commented 2 months ago

Revert

oscardssmith commented 2 months ago

https://github.com/SciML/SciMLBase.jl/pull/682

oscardssmith commented 2 months ago

I'm really confused how this was causing problems since from looking at the code, AutoSpecialize should only change things for in place ODEFunctions.