Performance improvement

[parkag]

During reconstructions I encountered a potential performance problem.
KCSD2D reconstruction for {16^2 (256) electrodes, 1000 sources} gives quite reasonable reconstruction time (~ 45sec) on an average machine, but that's it.

{32^2 electrodes (1024), 1000 sources} feels very slow (7m40s)

Reconstruction for 128^2 (~16k) electrodes seems impossible. What kind of electrode arrangements (number of electrodes) do we expect in practice?

[Jay-sanjay]

@parkag sir are you still facing the same issue , I would be happy to contribute , would you explain it a bit more precisely