Core Parallel Datastructures
- class NTPolySwig.Matrix_ps(*args)
A Module For Performing Distributed Sparse Matrix Operations.
- Conjugate()
Compute the complex conjugate of a matrix
- FillDense()
This routine will fill a dense matrix so that every element has a given a value of 1. This is useful as a starting point for further filtering or mapping operations.
- FillDistributedPermutation(lb, permuterows=True)
Fill the matrix based on a permutation.:type lb:
Permutation- Parameters:
lb – the permutation.:type permuterows: boolean, optional
permuterows – true if this is a row permutation matrix.
- FillFromTripletList(*args)
- Overload 1:
Fill in the matrix based on the contents of triplet lists.:type triplet_list:
TripletList_r
- Parameters:
triplet_list – list of values. Need to be absolute coordinates.
- Overload 2:
Fill in the matrix based on the contents of triplet lists.:type triplet_list:
TripletList_c
- Parameters:
triplet_list – list of values. Need to be absolute coordinates.
- FillIdentity()
Fills this matrix as the identity matrix.
- Gemm(matA, matB, memory_pool, alpha=1.0, beta=0.0, threshold=0.0)
this := alpha*matA*matB+ beta*this (GEMM):type matA:
Matrix_ps- Parameters:
matA – first mat.:type matB:
Matrix_psmatB – second mat.:type memory_pool:
PMatrixMemoryPoolmemory_pool – memory pool for intermediates.:type alpha: float, optional
alpha – scaling factor.:type beta: float, optional
beta – scaling factor.:type threshold: float, optional
threshold – for flushing small values.
- GetActualDimension()
get the actual dimension of the matrix.
- GetLogicalDimension()
the logical dimension is scaled so each process has an even slice.
- GetMatrixBlock(*args)
- Overload 1:
Extract an arbitrary block of a matrix into a triplet list. Block is defined by the row/column start/end values. This is slower than GetTripletList, because communication is required. Data is returned with absolute coordinates. :type triplet_list:
TripletList_r
- Parameters:
triplet_list – the list to fill. :type start_row: int
start_row – the starting row for data to store on this process. :type end_row: int
end_row – the ending row for data to store on this process. :type start_column: int
start_column – the starting col for data to store on this process :type end_column: int
end_column – the ending col for data to store on this process
- Overload 2:
Extract an arbitrary block of a matrix into a triplet list. Block is defined by the row/column start/end values. This is slower than GetTripletList, because communication is required. Data is returned with absolute coordinates. :type triplet_list:
TripletList_c
- Parameters:
triplet_list – the list to fill. :type start_row: int
start_row – the starting row for data to store on this process. :type end_row: int
end_row – the ending row for data to store on this process. :type start_column: int
start_column – the starting col for data to store on this process :type end_column: int
end_column – the ending col for data to store on this process
- GetMatrixSlice(submatrix, start_row, end_row, start_column, end_column)
Copy an arbitrary slice from a matrix into a new smaller matrix. NTPoly only works with square matrices, so if the number of rows and columns is different the matrix is resized to the maximum size. :type submatrix:
Matrix_ps- Parameters:
submatrix – the slice to fill. :type start_row: int
start_row – the starting row to include in this matrix. :type end_row: int
end_row – the ending row to include in this matrix. :type start_column: int
start_column – the starting column to include in this matrix. :type end_column: int
end_column – the last column to include in this matrix.
- GetSize()
Get the total number of non-zero entries in the matrix
- GetTripletList(*args)
- Overload 1:
Extracts a triplet list of the data that is stored on this process. Data is returned with absolute coordinates. :type triplet_list:
TripletList_r
- Parameters:
triplet_list – the list to fill.
- Overload 2:
Extracts a triplet list of the data that is stored on this process. Data is returned with absolute coordinates. :type triplet_list:
TripletList_c
- Parameters:
triplet_list – the list to fill.
- Increment(matB, alpha=1.0, threshold=0.0)
this = alpha*MatB + this (AXPY):type matB:
Matrix_ps- Parameters:
matB – the matrix to add.:type alpha: float, optional
alpha – scaling factor.:type threshold: float, optional
threshold – for flushing small values.
- IsIdentity()
Determine if this is the identity matrix.
- Norm()
compute the norm of a matrix.
- PairwiseMultiply(matA, matB)
Elementwise multiplication. C_ij = A_ij * B_ij. Also known as a Hadamard:type matA:
Matrix_ps- Parameters:
matA – first mat.:type matB:
Matrix_psmatB – second mat.
- Resize(new_size)
Change the size of a matrix. If the new size is smaller, then values outside that range are deleted. IF the new size is bigger, zero padding is applied. Warning: this requires a full data redistribution.
- Scale(constant)
scale the matrix by a constatn. constant the value to scale by.
- Trace()
compute the trace of a matrix.
- Transpose(matA)
Transpose a sparse matrix.
- WriteToBinary(file_name)
Write the matrix to a custom binary format.:type file_name: string
- Parameters:
file_name – file to write to.
- WriteToMatrixMarket(file_name)
Write the matrix to the matrix market format.:type file_name: string
- Parameters:
file_name – file to write to.
- property thisown
The membership flag
- class NTPolySwig.PMatrixMemoryPool(Matrix)
A memory pool datatype that can be reused for matrix matrix multiplication this is to prevent excessive alloc/dealloc.
- property thisown
The membership flag
- class NTPolySwig.ProcessGrid(*args)
A datatype which stores a process grid and all its communicators.
- GetMyColumn()
Get the column of the current process.
- GetMyRow()
Get the row of the current process.
- GetMySlice()
Get the slice of the current process.
- GetNumColumns()
Get the number of columns in this grid.
- GetNumRows()
Get the number of rows in this grid.
- GetNumSlices()
Get the number of slices in this grid.
- WriteInfo()
Write out some basic information about this process grid to the log.
- property thisown
The membership flag