On-the-fly clustering for exascale molecular dynamics simulations.

The image processing method for developing connected components has been established. We will now transition to a new implementation focused on graph processing.

As previously mentioned, the algorithm developed during the PhD is efficient and well-optimized. However, it has a limitation: it consolidates information into a central MPI node. To address this, we aimed to create an alternative algorithm for two main reasons. First, we sought a simpler algorithm that could be adapted for GPU implementation in the near future. Second, for large simulations, we wanted to ensure robustness, avoiding scenarios where all information is gathered on a single node.

The alternative implementation we have recently finalized maintains the same starting point as before. Regarding the local MPI process, it operates similarly to the previous algorithm.