Function cusparseGather 

pub unsafe extern "C" fn cusparseGather(
    handle: cusparseHandle_t,
    vecY: cusparseConstDnVecDescr_t,
    vecX: cusparseSpVecDescr_t,
) -> cusparseStatus_t

The function gathers the elements of the dense vector vecY into the sparse vector vecX

In other words,

for i=0 to nnz-1
X_values[i] = Y[X_indices[i]]

cusparseGather supports the following index type for representing the sparse vector vecX:

• 32-bit indices (cusparseIndexType_t::CUSPARSE_INDEX_32I)
• 64-bit indices (cusparseIndexType_t::CUSPARSE_INDEX_64I)

cusparseGather supports the following data types:

X/Y
cudaDataType_t::CUDA_R_16F
cudaDataType_t::CUDA_R_16BF
cudaDataType_t::CUDA_R_32F
cudaDataType_t::CUDA_R_64F
cudaDataType_t::CUDA_C_16F [DEPRECATED]
cudaDataType_t::CUDA_C_16BF [DEPRECATED]
cudaDataType_t::CUDA_C_32F
cudaDataType_t::CUDA_C_64F

cusparseGather has the following constraints:

• The arrays representing the sparse vector vecX must be aligned to 16 bytes

cusparseGather has the following properties:

• The routine requires no extra storage
• The routine supports asynchronous execution
• Provides deterministic (bit-wise) results for each run if the the sparse vector vecX indices are distinct
• The routine allows indices of vecX to be unsorted

cusparseGather supports the following optimizations:

• CUDA graph capture
• Hardware Memory Compression

Please visit cuSPARSE Library Samples - cusparseGather for a code example.

Parameters

• handle: Handle to the cuSPARSE library context.
• vecY: Dense vector Y.
• vecX: Sparse vector X.