Commits · 6fd1403181b00512aa24b90134c68aef961f7827 · TNL / tnl-dev

Oct 15, 2016
- Optimized CUDA traversers for cells and vertices · 6fd14031
  Jakub Klinkovský authored 8 years ago
  
  Entity basis and orientation is passed only to faces and edges.
  6fd14031
- Removed orientation and basis from specialization for vertices · 27a5c94b
  Jakub Klinkovský authored 8 years ago
  
  27a5c94b
- Avoiding unnecessary (de)allocations in LinearSystemAssembler · 216517fa
  Jakub Klinkovský authored 8 years ago
  
  216517fa
Oct 13, 2016
- Merge branch 'cuda-tuning' into develop · d7b7276b
  Tomáš Oberhuber authored 8 years ago
  
  d7b7276b
- Merge branch 'cuda-tuning' of geraldine.fjfi.cvut.cz:/local/projects/tnl/tnl into cuda-tuning · 797e663a
  Tomáš Oberhuber authored 8 years ago
  
  797e663a
- Merge branch 'namespaces' of geraldine.fjfi.cvut.cz:/local/projects/tnl/tnl into namespaces · 6667520a
  Tomáš Oberhuber authored 8 years ago
  
  6667520a
- Fixed solver monitor. · ddaaa19b
  Tomáš Oberhuber authored 8 years ago
  
  ddaaa19b
Oct 12, 2016
- Optimized layers of NeighboutGridEntitiesStorage and fixed tnl-benchmark-simple-heat-eqauation. · dcc60b21
  Tomáš Oberhuber authored 8 years ago
  
  dcc60b21
Oct 09, 2016
- Merge branch 'cuda-tuning' of geraldine.fjfi.cvut.cz:/local/projects/tnl/tnl into cuda-tuning · b14a72a0
  Tomáš Oberhuber authored 8 years ago
  
  b14a72a0
- Optimizing CUDA explicit solver. · 7658b3b3
  Tomáš Oberhuber authored 8 years ago
  
  7658b3b3
- Simplified omp critical region in multireduction · cb2000e0
  Jakub Klinkovský authored 8 years ago
  
  cb2000e0
Oct 08, 2016
- Fixed compilation with older GCC versions · 23daa000
  Jakub Klinkovský authored 8 years ago
  
  23daa000
- Fixed bug in 3D grid traverser · ded4bb8e
  Jakub Klinkovský authored 8 years ago
  
  ded4bb8e
- Fixed isOMPEnabled and added missing HAVE_OPENMP guards · bb6342a0
  Jakub Klinkovský authored 8 years ago
  
  bb6342a0
- Parallelization of the CWYGMRES solver using OpenMP · 1c128e27
  Jakub Klinkovský authored 8 years ago
  
  1c128e27
- Functions in Constants.h can be constexpr functions · 3a0d6bdc
  Jakub Klinkovský authored 8 years ago
  
  3a0d6bdc
- Fixed math functions in reduction-operations.h · 664c84c0
  Jakub Klinkovský authored 8 years ago
  
  664c84c0
- commonReductionOnDevice will be useful also in OpenMP parallelization on host · 9d816749
  Jakub Klinkovský authored 8 years ago
  
  9d816749
- Fixed includes in CudaMultireductionKernel.h · 7110b53a
  Jakub Klinkovský authored 8 years ago
  
  7110b53a
- Fixed fallback to multireduction on host · 89bf3543
  Jakub Klinkovský authored 8 years ago
  
  89bf3543
- Moved multireduction kernel launcher to CudaMultireductionKernel.h · 8172e35e
  Jakub Klinkovský authored 8 years ago
  
  8172e35e
Oct 07, 2016
- Merge branch 'develop' of geraldine.fjfi.cvut.cz:/local/projects/tnl/tnl into develop · 2ded6e73
  Tomáš Oberhuber authored 8 years ago
  
  2ded6e73
- Optimizing CUDA solver. · d8e0c387
  Tomáš Oberhuber authored 8 years ago
  
  d8e0c387
Oct 06, 2016
- Fixed asserts and traverser for faces in 3D grids · 64576ef5
  Jakub Klinkovský authored 8 years ago
  
  64576ef5
- Made the checkCudaDevice macro little bit safer · 01ce8f31
  Jakub Klinkovský authored 8 years ago
  
  01ce8f31
Oct 05, 2016
- Fixed installation of some headers in CMakeLists.txt · 56ec5b01
  Jakub Klinkovský authored 8 years ago
  
  56ec5b01
- Style changes in cuda-reduction_impl.h · 10d56a85
  Jakub Klinkovský authored 8 years ago
  
  10d56a85
- Avoided some compiler warnings · 95f376b0
  Jakub Klinkovský authored 8 years ago
  
  95f376b0
- Added CWYGMRES solver (GMRES using Hauseholder orthogonalization and the... · caae8ef4
  Jakub Klinkovský authored 8 years ago
  
  Added CWYGMRES solver (GMRES using Hauseholder orthogonalization and the compact CWY representation)
  caae8ef4
- Style changes in GMRES - no modifications of the algorithms · ed489170
  Jakub Klinkovský authored 8 years ago
  
  ed489170
Oct 03, 2016
- Optimized vector operations on host - prefetching and unrolling · 4c2bccf2
  Jakub Klinkovský authored 8 years ago
  
  4c2bccf2
Oct 02, 2016
- Debuging heat equation benchmark. · d2a4ec4d
  Tomáš Oberhuber authored 8 years ago
  
  d2a4ec4d
- Fixed formatting of floating-point numbers in solver monitor · 8c7bf210
  Jakub Klinkovský authored 8 years ago
  
  8c7bf210
- Avoided some compiler warnings from Neumann b.c. · c2067017
  Tomáš Oberhuber authored 8 years ago
  
  c2067017
- "Fix" in solver monitor. · e47e725d
  Tomáš Oberhuber authored 8 years ago
  
  e47e725d
- Fixed bug in boundary conditions setter. · 8f726c5c
  Tomáš Oberhuber authored 8 years ago
  
  8f726c5c
- Fixed mesh entity creation in CUDA grid traverser. · 07c0214b
  Tomáš Oberhuber authored 8 years ago
  
  07c0214b
Oct 01, 2016
- Merge branch 'namespaces' into develop · ca30741d
  Tomáš Oberhuber authored 8 years ago
  
  ca30741d
Sep 30, 2016

Merge branch 'develop' into namespaces · 858848a7

Conflicts:
	src/TNL/CMakeLists.txt
	src/TNL/Devices/Cuda.cpp
	src/TNL/Meshes/GridDetails/GridTraverser_impl.h
	src/TNL/Problems/HeatEquationProblem.h
	src/TNL/Solvers/ODE/ExplicitSolver.h
	src/TNL/TimerCPU.h
	src/TNL/TimerRT.h

858848a7

Optimized smart pointers · f2042755

Jakub Klinkovský authored 8 years ago

The original method for detecting changes to the wrapped objects (using
the 'modified' bool flag) is used to avoid unnecessary calls to
std::memcmp.

f2042755