Getting started

In this short guide we will install Spike and show you a small calculation that you can do with it.

Installing spike

1. Clone the repository

   git clone https://github.com/chegerland/Spike.git

2. Make a build folder and build

   mkdir -p build && cd build
   cmake .. && make

Your first calculation

Spike is both a library and a bunch of executables. You can use every part of the library to build your own programs, e.g. take a look at the following program:

#include <fstream>
#include <iostream>
 
#include "Spike/Spike.h"
 
using namespace Spike;
 
int main(int argc, char *argv[]) {
 
  // define time frame and neuron
  const TimeFrame tf(0.0, 30.0, 1e-3);
  LIFAC lifac(3.1, 0.0, 0.5, 1e-1);
 
  // define arrays for trajectory
  std::vector<double> v(tf.get_size());
  std::vector<double> a(tf.get_size());
 
  // calculate a trajectory
  lifac.get_voltage_curve(tf, v, a);
 
  // write trajectory to file
  std::ofstream file;
  file.open("trajectory.csv");
  file << "# Trajectory for the LIFAC\n";
  file << "# t, v, a\n";
 
  for (size_t i = 0; i < v.size(); i++) {
    file << tf.get_time(i) << "," << v[i] << "," << a[i] << std::endl;
  }
 
  file.close();
 
  return 0;
}

Here we calculate the trajectory of a leaky integrate-and-fire neuron with spike-frequency adaptation and save it to a file. This executable may also be found in app/spike_trajectory.cpp.

Working with the Spike apps

In app/ you can find programs that I have written for my master thesis. Most of them are intended to be used on a computing cluster using MPI. Every simulation takes a .ini file as an input argument, runs the simulation and produces an according .csv output file. As an example, if we want to calculate the linear susceptibility of a leaky integrate-and-fire neuron, we would need an input file like this and put it in data/lif.ini

[TimeFrame]
t_0 = 0.0
t_end = 1000.0
dt = 5e-4
 
[Neuron]
type = LIF
mu = 2.0
D = 1e-1
 
[Simulation]
c = 1.0
N_neurons = 10000

We have defined all parameters and may use the executable bin/spike_suscept in parallel by running

mpirun -np 4 bin/spike_suscept -f data/lif.ini

This uses four processes in parallel and the output file data/lif_suscept_lin.csv contains two columns: the first one being the frequencies and the second one being the complex value of the linear susceptibility at that frequency.