Chaoming Wang

This is a very good comparison. This demonstrates that we need more accurate Exponential numerical methods for integrating such complex dynamics. Welcome contributions. For the question, ``can this be a...

Thanks for this information. I have googled and did not find any information about what are ``cnexp`` and ``derivimplicit`` methods. But i found two tutorials which may be useful in...

Dear @CloudyDory , thank you for your report. Actually, BrainPy has a very easy way to given any inputs. For your model, you can give inputs using `brainpy.DSRunner`: ```python inputs...

This is a very good question. I have also considered it for a long time. However, current BrainPy also supports the monitor every few steps. Here is an example: ```python...

So many thanks for your report! It's an excellent feedback for us.

Unit checking is on the way. @Routhleck

Great! This requirement needs to explicitly cast all parameters to ``brainpy.math.float_``. For example, for a HH neuron model, its parameter ``gNa`` should be reinterpreted as ``gNa = bm.asarray(gNa, bm.float_)``. Ideally,...

One more thing that needs to be taken care of is that the coefficients of runge kutta methods should also be cast into `brainpy.math.float_` type.

yes, changes should be made in the brainpy framework. Note that ``dt`` should also be cast in the integrators.

The preallocation can be disabled with the setting of ``brainpy.math.disable_gpu_memory_preallocation()``.