shehzan10 · GitHub

* Added SparseArray and SparseArrayBase wrapper classes
* Added functions to handle/get etc
* Reorder storage enum. CSR=1, CSC=2, COO=3
* Cleanup af_create_sparse_array_from_dense function
* Add function to convert storage types
  * Currently backend limitation only accepts dense
  * In C-API implementation is complete
* Internal API changes
  * SparseArray getRows/Columns -> getRowIdx/getColIdx
  * Rename dense2storage to sparseConvertDenseToStorage
* CSC not supported.
* BLAS is implemented but not working

@shehzan10 shehzan10 changed the title WIP: FEAT Add Sparse operations to CUDA Backend WIP: FEAT Add Sparse operations

Jun 1, 2016
* Add C++ API as functions
* Remove num_rows and num_cols C API
* Rename get_rows/cols to get_row/col_idx
* Rename get_num_values to get_num_nonzero
- Was causing performance issues on intel and amd devices
- Added comments
- Fixed typos and bugs
- Fixed terminology
- Added necessary tests

@pavanky pavanky changed the title WIP: FEAT Add Sparse operations Adding Sparse operations for all backends

Sep 2, 2016

@pavanky

@shehzan10

Read the original on github.com ↗