cuquantum.custatevec.apply_matrix_batched_get_workspace_size¶

cuquantum.custatevec.apply_matrix_batched_get_workspace_size(intptr_t handle, int sv_data_type, uint32_t n_index_bits, uint32_t n_svs, _Index sv_stride, int map_type, matrix_indices, intptr_t matrices, int matrix_data_type, int layout, int32_t adjoint, uint32_t n_matrices, uint32_t n_targets, uint32_t n_controls, int compute_type) → size_t[source]¶

Computes the required workspace size for apply_matrix_batched().

Parameters

handle (intptr_t) – The library handle.
sv_data_type (cuquantum.cudaDataType) – The data type of the statevector.
n_index_bits (uint32_t) – The number of index bits.
n_svs (uint32_t) – The number of batched statevectors.
sv_stride (int64_t) – The stride between each state vector in the batch.
map_type (MatrixMapType) – Specify the way to assign matrices.
matrix_indices –
An array of matrix indices to indicate, in order, which matrix is to be applied to the statevectors in the batch. It can be
- an int as the pointer address to the array (on host or device)
- a Python sequence of bits on host
matrices (intptr_t) – The pointer address (as Python int) to the matrices (on either host or device).
matrix_data_type (cuquantum.cudaDataType) – The data type of the matrix.
layout (MatrixLayout) – The memory layout the the matrix.
adjoint (int32_t) – Whether the adjoint of the matrix would be applied.
n_matrices (uint32_t) – The number of matrices.
n_targets (uint32_t) – The length of targets.
n_controls (uint32_t) – The length of controls.
compute_type (cuquantum.ComputeType) – The compute type of matrix multiplication.

Returns

The required workspace size (in bytes).

Return type

size_t