You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In Julia, you can use automatic Differentiation for the ODE solvers. However, this is so far not possible in this code. The reason is that the code has matrix products with the state. There is a work around that, see (https://docs.sciml.ai/DiffEqDocs/stable/basics/faq/#I-get-Dual-number-errors-when-I-solve-my-ODE-with-Rosenbrock-or-SDIRK-methods). However, this requires changing a fair bit of code. Also, you cannot do that with sub-arrays i.e., x[idx]. Hence, it is not prioritized at the moment. It it however stated that the performance decrease for using finite difference is small for BDF solvers which is used in the code.
The text was updated successfully, but these errors were encountered:
In Julia, you can use automatic Differentiation for the ODE solvers. However, this is so far not possible in this code. The reason is that the code has matrix products with the state. There is a work around that, see (https://docs.sciml.ai/DiffEqDocs/stable/basics/faq/#I-get-Dual-number-errors-when-I-solve-my-ODE-with-Rosenbrock-or-SDIRK-methods). However, this requires changing a fair bit of code. Also, you cannot do that with sub-arrays i.e., x[idx]. Hence, it is not prioritized at the moment. It it however stated that the performance decrease for using finite difference is small for BDF solvers which is used in the code.
The text was updated successfully, but these errors were encountered: