Even on smallish matrices our implementation takes almost the same time when running under GAP compared to when running under HPC-GAP with one thread. This is pretty nice.
On the other hand, using 4 threads almost doubles the CPU time needed to complete one run of the algorithm.
Even on smallish matrices our implementation takes almost the same time when running under GAP compared to when running under HPC-GAP with one thread. This is pretty nice.
On the other hand, using 4 threads almost doubles the CPU time needed to complete one run of the algorithm.