Recently, to answer the referee’s questions, I implemented the complete LL equation into EPOCH. Here I wanna talk about the issues I encountered in this process.

At first, I was not sure how to implement the derivative of fields into particle pusher. It turned out that it’s super easy. All you need to do is to store the fields felt by the particle from the last time step and calculate the deviation.

How to store the fields from the last time step?

This is the first problem I met. The particle pusher is called every step, so I was stuck by how to get the previous field. At first, I define a plain variable in particle pusher to store fields. However, when I calculate the deviation, the variable is always 0 (cos it is initialized as 0). Jing suggested that I could add an additional particle information and store it. Marvellous. In that way, I can easily calc the difference of fields by letting current fields minus last fields stored in the particle information.

Double free or corruption error.

Quite a dumb mistake. When I added an additional particle property, I forgot to add to the counter of total properties.

Benchmark

I benchmarked my result with Vranic’s, which is quite the same.

Underlying Problem

Howard suggested that time is not aligned in my method. I should get the derivative of fields from updating fields module.