Need help in optimizing my code which solve circuit with matrices -- 2

It is a relatively simple code using Numpy (less than 300 lines) that, given some parameters, build a circuit and solves it at each time step. Since the inputs have a stochastic distribution, I need to solve the circuit in times hundreds of times. At the moment, each simulation is 20min, so it is not possible to do them, I need much more speed. A multi-thread approach has been used, and the code is faster by 4 times. However, I would like to reach much faster speed, maybe by using GPU or other approaches.

I need someone good with optimal resource allocation and code optimization for numerical methods. Thanks.

Make the code at least 10x faster by code optimization and CPU,thread, GPU, other approaches.