class brainpy.math.sparse.csrmv(data, indices, indptr, vector, *, shape, transpose=False)[source]#

Product of CSR sparse matrix and a dense vector using cuSPARSE algorithm.

This function supports JAX transformations, including jit(), grad(), vmap() and pmap().

  • data (ndarray, float) – An array of shape (nse,).

  • indices (ndarray) – An array of shape (nse,).

  • indptr (ndarray) – An array of shape (shape[0] + 1,) and dtype indices.dtype.

  • vector (ndarray) – An array of shape (shape[0] if transpose else shape[1],) and dtype data.dtype.

  • shape (tuple of int) – A length-2 tuple representing the matrix shape.

  • transpose (bool) – A boolean specifying whether to transpose the sparse matrix before computing.

  • method (str) –

    The method used to compute Matrix-Vector Multiplication. Default is taichi. The candidate methods are:

    • None: default using Taichi kernel.

    • cusparse: using cuSPARSE library.

    • scalar:

    • vector:

    • adaptive:


y – The array of shape (shape[1] if transpose else shape[0],) representing the matrix vector product.

Return type: