diff options
author | Cedric Nugteren <web@cedricnugteren.nl> | 2017-03-10 21:24:35 +0100 |
---|---|---|
committer | Cedric Nugteren <web@cedricnugteren.nl> | 2017-03-10 21:24:35 +0100 |
commit | 49e04c7fce8fed45559e143137cef3a1a36328cc (patch) | |
tree | f73a5c280f12cc5e38f6d4fd4e853b8b8e1aa432 /include/clblast.h | |
parent | de3500ed18ddb39261ffa270f460909571276462 (diff) |
Added API and test infrastructure for the batched GEMM routine
Diffstat (limited to 'include/clblast.h')
-rw-r--r-- | include/clblast.h | 12 |
1 files changed, 12 insertions, 0 deletions
diff --git a/include/clblast.h b/include/clblast.h index a1f14471..2520d601 100644 --- a/include/clblast.h +++ b/include/clblast.h @@ -619,6 +619,18 @@ StatusCode AxpyBatched(const size_t n, const size_t batch_count, cl_command_queue* queue, cl_event* event = nullptr); +// Batched version of GEMM: SGEMMBATCHED/DGEMMBATCHED/CGEMMBATCHED/ZGEMMBATCHED/HGEMMBATCHED +template <typename T> +StatusCode GemmBatched(const Layout layout, const Transpose a_transpose, const Transpose b_transpose, + const size_t m, const size_t n, const size_t k, + const T *alphas, + const cl_mem a_buffer, const size_t *a_offsets, const size_t a_ld, + const cl_mem b_buffer, const size_t *b_offsets, const size_t b_ld, + const T *betas, + cl_mem c_buffer, const size_t *c_offsets, const size_t c_ld, + const size_t batch_count, + cl_command_queue* queue, cl_event* event = nullptr); + // ================================================================================================= // CLBlast stores binaries of compiled kernels into a cache in case the same kernel is used later on |