|
Doxygen
1.9.8
|
Wrapper class around calls GPU accelerated linear algebra libraries. More...
#include <tsgAcceleratedDataStructures.hpp>
Public Member Functions | |
| GpuEngine ()=default | |
| Default construction, does not use any GPU acceleration, will set handles on-the-fly. | |
| GpuEngine (GpuEngine const &)=delete | |
| Do not copy, even if no GPU backend is enabled. | |
| GpuEngine & | operator= (GpuEngine const &)=delete |
| Do not copy, even if no GPU backend is enabled. | |
| GpuEngine (GpuEngine &&)=default | |
| Moving the engine is allowed, unique pointers will be moved. | |
| GpuEngine & | operator= (GpuEngine &&)=default |
| Moving the engine is allowed, unique pointers will be moved. | |
| void | setCuBlasHandle (void *handle) |
| Manually sets the cuBlas handle, handle must be a valid cublasHandle_t associated with this CUDA device. | |
| void | setCuSparseHandle (void *handle) |
| Manually sets the cuSparse handle, handle must be a valid cusparseHandle_t associated with this CUDA device. | |
| void | setCuSolverDnHandle (void *handle) |
| Manually sets the cuSparse handle, handle must be a valid cusolverDnHandle_t associated with this CUDA device. | |
Public Attributes | |
| std::unique_ptr< int, HandleDeleter< AccHandle::Cublas > > | cublas_handle |
| Holds the cuBlas handle. | |
| std::unique_ptr< int, HandleDeleter< AccHandle::Cusparse > > | cusparse_handle |
| Holds the cuSparse handle. | |
| std::unique_ptr< int, HandleDeleter< AccHandle::Cusolver > > | cusolver_handle |
| Holds the cuSolver handle. | |
| bool | called_magma_init = false |
| Avoids an empty engine when no acceleration is enabled, allows for default constructor/move/copy, skips extraneous calls to MAGMA init. | |
Wrapper class around calls GPU accelerated linear algebra libraries.
The class also manages the required handles and queues and holds the context of the active GPU device.