PyLops MPI API

The Application Programming Interface (API) of PyLops MPI enables distributed and parallel processing of large-scale linear algebra operations, inversions and computations.

DistributedArray

Partition(value)	Enum class
DistributedArray(global_shape[, base_comm, ...])	Distributed Numpy Arrays
StackedDistributedArray(distarrays[, base_comm])	Stacked DistributedArrays

Linear operators

Templates

MPILinearOperator([Op, shape, dtype, base_comm])	Common interface for performing matrix-vector products in distributed fashion.
asmpilinearoperator(Op)	Return Op as a MPI LinearOperator.
MPIStackedLinearOperator([shape, dtype, ...])	Common interface for performing matrix-vector products in distributed fashion for StackedLinearOperators.

Basic Operators

MPIBlockDiag(ops[, base_comm, dtype])	MPI Block-diagonal operator.
MPIStackedBlockDiag(ops[, base_comm, dtype])	MPI Stacked BlockDiag Operator
MPIVStack(ops[, base_comm, dtype])	MPI VStack Operator
MPIStackedVStack(ops[, base_comm, dtype])	MPI Stacked VStack Operator
MPIHStack(ops[, base_comm, dtype])	MPI HStack Operator

Derivatives

MPIFirstDerivative(dims[, sampling, kind, ...])	MPI First Derivative
MPISecondDerivative(dims[, sampling, kind, ...])	MPI Second Derivative
MPILaplacian(dims[, axes, weights, ...])	MPI Laplacian
MPIGradient(dims[, sampling, edge, kind, ...])	MPI Gradient

Signal Processing

MPIFredholm1(G[, nz, saveGt, usematmul, ...])

Fredholm integral of first kind.

Wave-Equation processing

MPIMDC(G, nt, nv, nfreq[, dt, dr, twosided, ...])

Multi-dimensional convolution.

Solvers

Basic

CG(Op[, callbacks])	Conjugate gradient
CGLS(Op[, callbacks])	Conjugate gradient least squares

cg(Op, y, x0[, niter, tol, show, itershow, ...])	Conjugate gradient
cgls(Op, y, x0[, niter, damp, tol, show, ...])	Conjugate gradient least squares

Utils

local_split(global_shape, base_comm, ...)

To get the local shape from the global shape

dottest(Op, u, v[, nr, nc, rtol, atol, ...])

Dot test.