lapjv | Go implementation of the LAPJV algorithm | Learning library

In R. Jonker and A. Volgenant, A Shortest Augmenting Path Algorithm for Dense and Sparse Linear Assignment Problems (doi: 10.1007/BF02278710), the authors show that an implementation of their LAPJV algorithm, adapted to sparse graphs and referred to as LAPJVsp, performs well on various problems.

A Pascal implementation of LAPJVsp is currently available here. The augmenting row reduction step of the algorithm is mostly unchanged, and differs from the code provided in the published paper only by using a compressed sparse row representation of the biadjacency matrix of the graph, whose row indices, column indices, and weights are referred to as first, kk, and cc respectively:

I've read that Illegal instruction (core dumped) error may be because of some machine issues or version issues.

However, my case is weird because I have two scripts - one dummy script to test the function and another one is my actual script. The weird thing is, it works for my dummy script (minimum cost and everything calculated correctly, the script executed with no errors, etc). But when I tried to execute the function in my actual script it throws a Illegal instruction (core dumped) error.

I've checked that in both cases I fed in a square matrix of class 'numpy.ndarray' and the elements are of class 'numpy.float64' for both. The only difference is that the dummy matrix is a 4x4 square matrix while the actual square matrix is of size 200 by 200. I tried reducing the size to roughly 50 by 50 just to test and see if thats the issue but it still throws the same error.

Any suggestions would be really appreciated, thanks!