cusparse-0.3.0.0: FFI bindings to the CUDA Sparse BLAS library
Copyright[2017] Trevor L. McDonell
LicenseBSD3
MaintainerTrevor L. McDonell <tmcdonell@cse.unsw.edu.au>
Stabilityexperimental
Portabilitynon-portable (GHC extensions)
Safe HaskellNone
LanguageHaskell2010

Foreign.CUDA.BLAS.Sparse.Convert

Description

For more information see the cuSPARSE function reference:

http://docs.nvidia.com/cuda/cusparse/index.html#cusparse-format-conversion-reference

Synopsis

Documentation

data Action Source #

This type indicates whether the operation is performed only on indices (Symbolic) or on data and indices (Numeric).

http://docs.nvidia.com/cuda/cusparse/index.html#cusparseactiont

Constructors

Symbolic 
Numeric 

data Hybrid Source #

An opaque structure holding the matrix in hybrid (HYB) format.

http://docs.nvidia.com/cuda/cusparse/index.html#cusparsehybmatt

data Algorithm_csr2csc Source #

Indices the algorithm to use for CSR to CSC matrix conversion

Algorithm 1 requires extra storage proportional to the number of nonzero values nnz. It is in general faster than algorithm 2 and the result is deterministic.

Algorithm 2 requires extra storage proportional to the number of rows m. It is non-deterministic, and does not ensure always the same ordering of CSC column indices and values. It is faster than algorithm 1 for regular matrices.

Constructors

Csr2cscAlg1 
Csr2cscAlg2