General optimization

[SirPurrcival]

Functions we can probably get the most performance out of:

• get_data
• lfp approximation (<- was pretty dirty, sorry about that)
• graphing? Not sure

[SirPurrcival]

https://wiki.python.org/moin/PythonSpeed/PerformanceTips

[SirPurrcival]

lfp approximation cut down by a lot (from minutes to seconds), probably still room for improvement.

[SirPurrcival]

I timed the runtime of the script on my computer with: Nscale = 0.1 Kscale = .15 Sscale = 1. Rscale = Nscale * .3

Time required for setup: 0.25959324836730957 Time required to populate network: 1.4485700130462646 Time required for connection setup: 4.657220363616943 Time required for simulation: 68.16679549217224 Time required for fetching data: 68.60523128509521 Time required for graphing grid: 74.39840173721313 Time required for layer estimation procedure: 83.97682785987854 Time required for plotting and final time: 84.18127965927124

Note, those are times relative to the beginning of the script, not between each step. I don't know if we can improve the simulation performance somehow, maybe some nest settings about better/more parallelization that I don't know about. Connection setup might be improved, depending on how nest handles it. A few notes

• Fetching data does not seem an issue anymore
• Graphing can probably be improved by doing better preprocessing (raster function)
• LFP estimation is already down by quite a lot but there might be a better way of doing it. I indicated the greatest bottleneck in the code

I'll get you a coffee if you manage to bring down processing time further :)