See title. The files generated in DataSimulation, in the runs I've done remotely are 130-150MB at 100 cars with a dt=0.02. Admittedly, the dt is very small (and cranking it up does significantly improve things, dt=0.5 nets 2MB), but it is still prohibitively large. Maybe we can cut down on some data somewhere? Or perhaps JSON is more efficient?
I also think the memory usage whilst running gets prohibitive because of this, causing the whole program to choke more than with a renderer.
Perhaps it would be better to save the data points of each car in a separate file, and the run of a simulation in a folder rather than a file.
See title. The files generated in DataSimulation, in the runs I've done remotely are 130-150MB at 100 cars with a dt=0.02. Admittedly, the dt is very small (and cranking it up does significantly improve things, dt=0.5 nets 2MB), but it is still prohibitively large. Maybe we can cut down on some data somewhere? Or perhaps JSON is more efficient? I also think the memory usage whilst running gets prohibitive because of this, causing the whole program to choke more than with a renderer. Perhaps it would be better to save the data points of each car in a separate file, and the run of a simulation in a folder rather than a file.