Classification in parallel using MPI
The classify_in_mpi_parallel
module defines functions that
use MPI to save Cayley graph classifications or partial classifications in parallel.
AUTHORS:
Paul Leopardi (2017-10-13)
- boolean_cayley_graphs.classify_in_mpi_parallel.save_class_parts_in_parallel(comm, name_prefix, form, c_len=1, directory=None)[source]
Using MPI, construct a complete list of the partial Cayley graph classifications corresponding to a given bent function or algebraic normal form.
INPUT:
comm
– MPI communicator.name_prefix
– String. Name prefix to use withsave_mangled
to save each class part.form
– A bent function or an algebraic normal form.c_len
– Integer. Default=1. The number of values of c to use in each class part.directory
– string, optional. The directory where the object is to be saved. Default is None, meaning the current directory.
OUTPUT: None.
EFFECT: Uses
name_prefix
to save all partial classifications corresponding tobentf
.
- boolean_cayley_graphs.classify_in_mpi_parallel.save_classifications_in_parallel(comm, name_prefix, list_of_f, start=0, stop=None, directory=None)[source]
Using MPI, construct and save a number of Cayley graph classifications corresponding to a list of bent functions.
INPUT:
comm
– MPI communicator.name_prefix
– String. Name prefix to use withsave_mangled
to save each classification.list_of_f
– List of forms or bent functions.start
– Integer. Default=0. Index of start position in the list.stop
– Integer. Default=None. Index after end position, orNone
if whole remaining list.directory
– string, optional. The directory where the object is to be saved. Default is None, meaning the current directory.
OUTPUT: None.
EFFECT: Uses
name
to save the classifications corresponding tolist_of_f
.