|
file | data_tune.h [code] |
| A class not used currently but was helpful in storing tuning data.
|
|
file | device_mem_pointer_matrix_cl.h [code] |
| Contains a class declaration for an object which represents an OpenCL device memory allocation for Matrix Container.
|
|
file | device_mem_pointer_matrix_cu.h [code] |
| Contains a class declaration for an object which represents an CUDA device memory allocation for Matrix Container.
|
|
file | matrix_col_iterator.inl |
|
file | MemAllocationUpdater.inl |
|
file | MemAllocationUpdater_CU.inl |
|
file | operator_macros_cucl.inl |
| Contains macro defintions for user functions using CUDA, OpenCL and CPU/OpenMP backend.
|
|
file | operator_macros_multi.inl |
| Contains macro defintions with support for different user functions for each back-end. Not used currently in the example code. Still experimental.
|
|
file | sparse_matrix_cuda.inl |
|
file | thread_management.h [code] |
| Contains classes that help with thread management. Currently only Pthreads. This was used earlier when using multiple CUDA devices.
|
|
file | thread_pool.h [code] |
| Contains class definition for thread pooling that was used for multi-GPU CUDA execution to reduce host thread creation overhead.
|
|