mbtrack2.tracking.parallel module

Module to handle parallel computation

class Mpi(filling_pattern: ndarray[tuple[int, ...], dtype[_ScalarType_co]])[source]

Bases: object

Class which handle parallel computation via the mpi4py module [1].

Parameters

filling_patternbool array of shape (h,)

Filling pattern of the beam, like Beam.filling_pattern

Attributes

commMPI.Intracomm object

MPI intra-comminicator of the processor group, used to manage communication between processors.

rankint

Rank of the processor which run the program

sizeint

Number of processor within the processor group (in fact in the intra-comminicator group)

tableint array of shape (size, 2)

Table of correspondance between the rank of the processor and its associated bunch number

bunch_numint

Return the bunch number corresponding to the current processor

next_bunchint

Return the rank of the next tracked bunch

previous_bunchint

Return the rank of the previous tracked bunch

Methods

write_table(filling_pattern)

Write a table with the rank and the corresponding bunch number for each bunch of the filling pattern

rank_to_bunch(rank)

Return the bunch number corresponding to rank

bunch_to_rank(bunch_num)

Return the rank corresponding to the bunch number bunch_num

share_distributions(beam)

Compute the bunch profiles and share it between the different bunches.

share_means(beam)

Compute the bunch means and share it between the different bunches.

share_stds(beam)

Compute the bunch standard deviations and share it between the different bunches.

References

[1] L. Dalcin, P. Kler, R. Paz, and A. Cosimo, Parallel Distributed Computing using Python, Advances in Water Resources, 34(9):1124-1139, 2011.

__init__(filling_pattern: ndarray[tuple[int, ...], dtype[_ScalarType_co]])[source]
write_table(filling_pattern: ndarray[tuple[int, ...], dtype[_ScalarType_co]])[source]

Write a table with the rank and the corresponding bunch number for each bunch of the filling pattern

Parameters

filling_patternbool array of shape (h,)

Filling pattern of the beam, like Beam.filling_pattern

rank_to_bunch(rank: int) → int[source]

Return the bunch number corresponding to rank

Parameters

rankint

Rank of a processor

Returns

bunch_numint

Bunch number corresponding to the input rank

bunch_to_rank(bunch_num: int) → int[source]

Return the rank corresponding to the bunch number bunch_num

Parameters

bunch_numint

Bunch number

Returns

rankint

Rank of the processor which tracks the input bunch number

property bunch_num: int

Return the bunch number corresponding to the current processor

property next_bunch: int

Return the rank of the next tracked bunch

property previous_bunch: int

Return the rank of the previous tracked bunch

share_distributions(beam: Beam, dimensions: str | list[str] = 'tau', n_bin: int = 75)[source]

Compute the bunch profiles and share it between the different bunches.

Parameters

beam : Beam object dimension : str or list of str, optional

Dimensions in which the binning is done. The default is “tau”.

n_binint or list of int, optional

Number of bins. The default is 75.

share_means(beam: Beam)[source]

Compute the bunch means and share it between the different bunches.

Parameters

beam : Beam object

share_stds(beam: Beam)[source]

Compute the bunch standard deviations and share it between the different bunches.

Parameters

beam : Beam object