2.4 Parallel implementation

Most time-consuming codes of Multiwfn have been parallelized by OpenMP technique. If your CPU has multiple cores, you can greatly benefit from parallelization. To enable parallelization, just modify “nthreads” parameter in settings.ini to a proper number. For example, your computer has a CPU with 12-cores physical cores, then commonly you should change “nthreads” to 12.

If Multiwfn crashes during parallel calculation for a very large system, try to enlarge “ompstacksize” in settings.ini (for Windows version) or enlarge the value of the environment variable OMP_STACKSIZE (for Linux or MacOS version).