; PARALLEL ;--------------------------------------------------------------- ;! Verb to set or show degree of parallelism ;# VERB POPS ;--------------------------------------------------------------- ;; Copyright (C) 1997 ;; Associated Universities, Inc. Washington DC, USA. ;; ;; This program is free software; you can redistribute it and/or ;; modify it under the terms of the GNU General Public License as ;; published by the Free Software Foundation; either version 2 of ;; the License, or (at your option) any later version. ;; ;; This program is distributed in the hope that it will be useful, ;; but WITHOUT ANY WARRANTY; without even the implied warranty of ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ;; GNU General Public License for more details. ;; ;; You should have received a copy of the GNU General Public ;; License along with this program; if not, write to the Free ;; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ;; MA 02139, USA. ;; ;; Correspondence concerning AIPS should be addressed as follows: ;; Internet email: aipsmail@nrao.edu. ;; Postal address: AIPS Project Office ;; National Radio Astronomy Observatory ;; 520 Edgemont Road ;; Charlottesville, VA 22903-2475 USA ;--------------------------------------------------------------- PARALLEL LLLLLLLLLLLLUUUUUUUUUUUU CCCCCCCCCCCCCCCCCCCCCCCCCCCCC PARALLEL: Verb to set or show degree of parallelism The number of processors to be used for parallelized AIPS programs is given as the immediate argument. ---------------------------------------------------------------- PARALLEL Type: Verb Use: PARALLEL(p) will interpret its immediate argument as the number of processors that will be used by AIPS tasks that have been enabled for parallel processing. This will not have any effect on tasks that have not been specifically compiled to use multiple processors or on tasks that are already running. In general, large problems benefit from having more processors working on them but you may experience substantial slow-downs if you set p (the number of processors too large). In preparing for a long run (which is where parallel processing is most worth while) it is often useful to select a small subset of your data that can be processed in a few minutes and experiment with different values of p to find the fastest option. Task- specific explain files may give more detailed advice. The following rules of thumb may be helpful. - Keep p = 1 unless you are reasonably sure that parallel processing will help with the problem that you are dealing with. - If your machine has n processors limit p to n-1 so that one processor is free to handle the AIPS command-line processor and other functions (most parallel FORTRAN systems suffer badly if they have to share processors with other programs). - Reduce the number of processors if you are sharing a machine with other users. Special cases: PRINTER(0): (or no argument) show the current number of processors that will be used by parallel tasks. Adverbs: The immediate argument is the number of logical processors. No actual adverbs. ----------------------------------------------------------------