An estimation of scalability when using a k-d tree as the data structure for neuron touch detection

Sammanfattning: To deepen our understanding of the brain and its function, a lot of effort has been put into the subject of whole brain simulation. One prerequisite problem that needs to be solved, in order to do these simulations, involves finding synapse connections between neurons to figure out the pathway of signals in the brain. Since the amount of data needed to represent a brain is quite large, consideration need to be put into the efficiency of the method of finding these connections. An algorithm using a k-d tree as a base was produced and benchmarked in order to estimate the scale of this issue. While the scalability in terms of Big O was quite satisfactory, the case was far from the same in absolute terms. It is concluded that the problem is memory bound and that this should be taken into consideration before time complexity.