Commits · 4cd830a47fd001d0c1683d9aca1c253719e782d3 · Kaushik Kulkarni / viennacl-dev · GitLab

Aug 06, 2014
- CG: Completed implementation of improved (pipelined) OpenCL kernels. · 4cd830a4
  Karl Rupp authored Aug 06, 2014
  
  4cd830a4
Aug 05, 2014
- CG: Added improved kernels for OpenCL when using compressed_matrix<> · 41c0c5a6
  Karl Rupp authored Aug 05, 2014
  41c0c5a6
Aug 04, 2014
- Pipelined CG: Added implementation for host-based execution. · 812e3918
  Karl Rupp authored Aug 04, 2014
  812e3918
- Merge branch 'wip-sliced-ell-matrix'. · 246370b0
  Karl Rupp authored Aug 04, 2014
  246370b0
- CUDA: Added option for specifying the CUDA arch through CMake. · d82ddd78
  Karl Rupp authored Aug 04, 2014
  d82ddd78
- sliced_ell_matrix: Support for Sliced ELL-matrix format (sliced_ell_matrix). · 1f9db9c9
  Karl Rupp authored Aug 04, 2014
  1f9db9c9
Aug 03, 2014
- Matrix: Added support for the operations A = trans(B), A += trans(B), A -= trans(B) · 0f8d1b6e
  Karl Rupp authored Aug 03, 2014
  0f8d1b6e
Aug 02, 2014
- VS 2012: Fixed compilation problem in compressed_compressed_matrix<> · ec267fcb
  Karl Rupp authored Aug 02, 2014
  ec267fcb
Aug 01, 2014
- Scheduler: Added missing implementations for element_pow() · 25076ce0
  Karl Rupp authored Aug 01, 2014
  
  25076ce0
- Tests: Reduced round-off errors in element_pow() test. · 2b8e01d7
  Karl Rupp authored Aug 01, 2014
  2b8e01d7
- VS2012: Fixed weird compilation problem with std::min() · 5c0ee23f
  Karl Rupp authored Aug 01, 2014
  5c0ee23f
- vector_operations: create temporary scalars with correct context · c3b474a1
  Toby Smithe authored Aug 01, 2014
  
  c3b474a1
Jul 27, 2014
- Device-Specific / GEMM : Now letting the compiler unroll the loop. · 5fae4e32
  Philippe Tillet authored Jul 27, 2014
  5fae4e32
- Device-Specific / Template-Base : Added a default warp size for CPUs · d2e5d704
  Philippe Tillet authored Jul 27, 2014
  
  d2e5d704
Jul 26, 2014
- Device-specific / GEMM : Fixed bug when simd-width > 1 for FETCH_GLOBAL_CONTIGUOUS · 98d4d325
  Philippe Tillet authored Jul 26, 2014
  
  98d4d325
- Device-specific / GEMM : Implemented the new FETCH_GLOBAL_CONTIGUOUS policy.... · c1c04eca
  Philippe Tillet authored Jul 26, 2014
  c1c04eca
Jul 24, 2014
- Device-specific : Added accessor to template::parameters(); rename... · 527352be
  Philippe Tillet authored Jul 24, 2014
  527352be
- CMake: Fixed non-matching conditions in if() and endif() block. · ca0c4d33
  Karl Rupp authored Jul 24, 2014
  
  ca0c4d33
Jul 22, 2014
- CMake, OpenMP, MinGW: Fixed missing OpenMP linkage under MinGW for libviennacl. · 801c95e8
  Karl Rupp authored Jul 22, 2014
  801c95e8
- Device-specific: Fixed error codes definition · 14391c58
  Philippe Tillet authored Jul 22, 2014
  14391c58
- Device-Specific : Further stabilization of the row-wise reduction template · a76dc2d0
  Philippe Tillet authored Jul 22, 2014
  
  a76dc2d0
- Device-specific : Fixed silly bug in row-wise reduction template · 266ff46b
  Philippe Tillet authored Jul 22, 2014
  
  266ff46b
- Core: Replaced matrix/vector_{base}::alignment by viennacl::dense_padding_size. · cb6e54b1
  Philippe Tillet authored Jul 22, 2014
  cb6e54b1
Jul 21, 2014
- Device-specific: template_base::check_invalid() now returns an error code... · b09a4895
  Philippe Tillet authored Jul 22, 2014
  b09a4895
Jul 20, 2014
- OCL : Changed format of add_program debug output · cfe957e4
  Philippe Tillet authored Jul 20, 2014
  cfe957e4
- OCL : Added maxwell tag · 9cb3ad09
  Philippe Tillet authored Jul 20, 2014
  
  9cb3ad09
- Device-specific - Built-in database: More robust database handling · dba96f55
  Philippe Tillet authored Jul 20, 2014
  dba96f55
- Device-specific - Row-wise reduction: Fixed-up the previous commit so that it doesn't break slices · 028d276f
  Philippe Tillet authored Jul 20, 2014
  028d276f
- Device-specific - Row-wise reduction : Now enqueueing the internal sizes to avoid deadlocks · 32e964e2
  Philippe Tillet authored Jul 20, 2014
  
  32e964e2
- Device-specific - Row-wise-reduction : Now detects deadlock case in check_valid() · b0a08822
  Philippe Tillet authored Jul 20, 2014
  
  b0a08822
- Now print build info on VIENNACL_BUILD_INFO and not VIENNACL_DEBUG_BUILD · 9bcbaa3d
  Philippe Tillet authored Jul 20, 2014
  
  9bcbaa3d
Jul 19, 2014
- Device-Specific / Matrix Product Template : Corrected static initializer for rC · 317fe44b
  Philippe Tillet authored Jul 19, 2014
  
  317fe44b
- OCL : Fixed bug in build log output · 4049de32
  Philippe Tillet authored Jul 19, 2014
  4049de32
Jul 18, 2014
- OCL: Added beignet vendor id · d89b83ae
  Philippe Tillet authored Jul 18, 2014
  
  d89b83ae
Jul 17, 2014
- Device-specific : Further improvements in the validity check of the GEMM template · 259fed35
  Philippe Tillet authored Jul 17, 2014
  
  259fed35
- Device-specific : Added missing includes in template_base (it was miraculously... · 9857b62a
  Philippe Tillet authored Jul 17, 2014
  9857b62a
- Scheduler: Fixed access to uninitialized memory when creating a temporary object. · 97333ae7
  Karl Rupp authored Jul 17, 2014
  97333ae7
Jul 15, 2014
- Device-specific : Added fallback accelerator/default profiles; Uniformization... · 3c3ec964
  Philippe Tillet authored Jul 15, 2014
  3c3ec964
Jul 14, 2014
- Linalg (OpenCL) : Fixed the generator-related test failures · 3b4cbf27
  Philippe Tillet authored Jul 14, 2014
  
  3b4cbf27
- OCL: Put vendor_ids into an enum · c2c12f7b
  Philippe Tillet authored Jul 14, 2014
  
  c2c12f7b