ComputeType#

enum nvmath.sparse.ComputeType(value)[source]#

The subset of cudaDataType_t that is supported by the generic sparse APIs.

Member Type:: int

Valid values are as follows:

CUDA_R_32F = <ComputeType.CUDA_R_32F: 0>#

CUDA_C_32F = <ComputeType.CUDA_C_32F: 4>#

CUDA_R_64F = <ComputeType.CUDA_R_64F: 1>#

CUDA_C_64F = <ComputeType.CUDA_C_64F: 5>#