89 ghost_densemat *w,
const char *transw,
void *alpha,
void *beta,
int reduce,
int printerror);
Functions for global mathematical operations.
ghost_implementation
Possible implementations of a CPU function.
Definition: types.h:323
ghost_implementation impl
Definition: tsmm.h:28
Header file for type definitions.
int xcols
The first configured block size N.
Definition: tsmm.h:23
int vcols
The second configure block size K.
Definition: tsmm.h:27
ghost_densemat_storage xstor
Definition: tsmm.h:30
ghost_error(* ghost_tsmm_kernel)(ghost_densemat *, ghost_densemat *, ghost_densemat *, void *, void *)
A tsmm kernel function.
Definition: tsmm.h:38
int ghost_tsmm_perf_GBs(double *perf, double time, void *varg)
ghost_error
Error return type.
Definition: error.h:23
ghost_alignment
Possible alignments of data access.
Definition: types.h:315
ghost_error ghost_tsmm(ghost_densemat *x, ghost_densemat *v, ghost_densemat *w, void *alpha, void *beta)
Multiply a distributed dense tall skinny matrix with a redundant dense matrix.
Definition: tsmm.cpp:107
ghost_alignment alignment
Definition: tsmm.h:29
int ghost_tsmm_perf_GFs(double *perf, double time, void *varg)
ghost_datatype
Available primitive data types.
Definition: types.h:248
ghost_datatype dt
The data type of the densemats.
Definition: tsmm.h:19
ghost_densemat_storage
Densemat storage orders.
Definition: densemat.h:94
ghost_error ghost_tsmm_valid(ghost_densemat *x, ghost_densemat *v, const char *transv, ghost_densemat *w, const char *transw, void *alpha, void *beta, int reduce, int printerror)
Check whether TSMM can be applied instead of GEMM with the given arguments.
Definition: tsmm.cpp:52
int multipleof
Definition: tsmm.h:32
int unroll
Definition: tsmm.h:31
A dense vector/matrix.
Definition: densemat.h:226