lib.mpi

Collection of classes and functions to perform MPI operations.

Note

Arrays are assumed to scatter along the first dimension.

Functions

argmax_array	Return index of maximum in input array data along axis.
argmin_array	Return index of minimum in input array data along axis.
average_array	Return weighted average of input array a along axis axis.
bcast_seed	Generate array of seeds.
corrcoef_array	Return weighted correlation matrix of input arrays m (y).
cov_array	Estimate a covariance matrix, given data and weights.
dot_array	Return dot product of input arrays a and b.
front_pad_array	Pad an array in the front with items before this rank.
local_size	Divide global size into local (process) size.
max_array	Return maximum of input array data along axis.
mean_array	Return mean of array data along axis.
min_array	Return minimum of input array data along axis.
prod_array	Return product of input array data along axis.
quantile_array	Return array quantiles.
recv_array	Receive array from process source.
send_array	Send input array data to process dest.
set_common_seed	Set same global np.random and random seed for all MPI processes.
set_independent_seed	Set independent global np.random and random seed for all MPI processes.
shape_array	Return global shape of data array (scattered along the first dimension).
size_array	Return global size of data array.
sort_array	Sort input array data along axis.
std_array	Return weighted standard deviation of input array along axis axis.
sum_array	Return sum of input array data along axis.
var_array	Estimate variance, given data and weights.
weighted_quantile_array	Return weighted array quantiles.

Classes

MPIPool	A processing pool that distributes tasks using MPI.
MPIRandomState	A Random number generator that is invariant against number of ranks, when the total size of random number requested is kept the same.