Fast correlation function calculator. A high-performance pair-counting toolkit

Context. A novel high-performance exact pair-counting toolkit called fast correlation function calculator (FCFC) is presented. Aims. With the rapid growth of modern cosmological datasets, evaluation functions with observational and simulation catalogues has become a challenge. High-efficiency codes are thus in great demand. Methods. We introduce different data structures algorithms that can be used for problems, perform comprehensive benchmarks to identify most efficient real-world applications. then describe three levels parallelisms by FCFC, SIMD, OpenMP, MPI, run extensive tests investigate scalabilities. Finally, we compare efficiency FCFC alternative codes. Results. The histogram update implemented shown outperform methods. does not benefit greatly from SIMD because bottleneck our algorithm mainly cache latency. Nevertheless, scales well numbers OpenMP threads MPI processes, even though speedups may degraded over few thousand total. found faster than (if all) other public