StarPU Handbook - StarPU Introduction
|
#include <cusolverDn.h>
#include <cusolverSp.h>
#include <cusolverRf.h>
Go to the source code of this file.
#define | STARPU_CUSOLVER_REPORT_ERROR(status) |
void | starpu_cusolver_init (void) |
void | starpu_cusolver_shutdown (void) |
cusolverDnHandle_t | starpu_cusolverDn_get_local_handle (void) |
cusolverSpHandle_t | starpu_cusolverSp_get_local_handle (void) |
cusolverRfHandle_t | starpu_cusolverRf_get_local_handle (void) |
void | starpu_cusolver_report_error (const char *func, const char *file, int line, cusolverStatus_t status) |
#define STARPU_CUSOLVER_REPORT_ERROR | ( | status | ) |
Call starpu_cusolver_report_error(), passing the current function, file and line position.
void starpu_cusolver_shutdown | ( | void | ) |
Synchronously deinitialize the CUSOLVER library on every CUDA device.
See CUDA-specificOptimizations
cusolverDnHandle_t starpu_cusolverDn_get_local_handle | ( | void | ) |
Return the CUSOLVER Dense handle to be used to queue CUSOLVER kernels. It is properly initialized and configured for multistream by starpu_cusolver_init().
See CUDA-specificOptimizations
cusolverSpHandle_t starpu_cusolverSp_get_local_handle | ( | void | ) |
Return the CUSOLVER Sparse handle to be used to queue CUSOLVER kernels. It is properly initialized and configured for multistream by starpu_cusolver_init().
See CUDA-specificOptimizations
cusolverRfHandle_t starpu_cusolverRf_get_local_handle | ( | void | ) |
Return the CUSOLVER Refactorization handle to be used to queue CUSOLVER kernels. It is properly initialized and configured for multistream by starpu_cusolver_init().
See CUDA-specificOptimizations
void starpu_cusolver_report_error | ( | const char * | func, |
const char * | file, | ||
int | line, | ||
cusolverStatus_t | status | ||
) |
Report a CUSOLVER error. See CUDASupport for more details.