Commits · 465f90be29c3a14f3b82fb4b8047a32a953fc87d · Christian Heinigk / dune-codegen

Apr 07, 2017
- Stage 3 vertical vectorization · 465f90be
  Dominic Kempf authored 8 years ago
  
  465f90be
- Generic shape/index thing for vertical/horizontal vercotrization · afb785f5
  Dominic Kempf authored 8 years ago
  
  afb785f5
- Fix quadrature shape for pure vertical vecotrization · 834d1f11
  Dominic Kempf authored 8 years ago
  
  834d1f11
- Some bugfixes for verticality · cf2b8c63
  Dominic Kempf authored 8 years ago
  
  cf2b8c63
- Working vertical mass matrix example · a3ce6477
  Dominic Kempf authored 8 years ago
  
  a3ce6477
Apr 06, 2017
- fix fastdg · ed879613
  Dominic Kempf authored 8 years ago
  
  ed879613
- more fixups · 85789057
  Dominic Kempf authored 8 years ago
  
  85789057
- another fixup · 3cc4ec54
  Dominic Kempf authored 8 years ago
  
  3cc4ec54
- fixups · 42b8b04a
  Dominic Kempf authored 8 years ago
  
  42b8b04a
- Partial solution to the problem of reduced dimensionality in the quadrature loop · 1f00d0e1
  Dominic Kempf authored 8 years ago
  
  1f00d0e1
- Use basis_size/quadrature_size · 6eaa6ac8
  Dominic Kempf authored 8 years ago
  
  6eaa6ac8
- fixup · f6884593
  Dominic Kempf authored 8 years ago
  
  f6884593
- Add slicing option · 800e6af0
  Dominic Kempf authored 8 years ago
  
  800e6af0
- WIP · d27a0a68
  Dominic Kempf authored 8 years ago
  
  d27a0a68
- First implementation of vertical vectorization · 6cb3a57e
  Dominic Kempf authored 8 years ago
  
  6cb3a57e
- First implementation of slicing of basis tabulation matrices along the quadrature axis · 31f14b8d
  Dominic Kempf authored 8 years ago
  
  31f14b8d
- Hvae BasisTabulationMatrixArray refer to the underlying objects directly · bac1cc58
  Dominic Kempf authored 8 years ago
  
  bac1cc58
Apr 05, 2017
- Do not store rows/cols AND transpose on BasisTabulationMatrix · 4a7f6ce0
  Dominic Kempf authored 8 years ago
  
  Instead, store quadrature_size, basis_size and transpose and have properties giving rows and cols.
  4a7f6ce0
- Remove the last explicit references to 4 as the vector width · 8e920583
  Dominic Kempf authored 8 years ago
  
  8e920583
- Refactor collect_rotate to in the future allow vertical vectorization · 869d6fff
  Dominic Kempf authored 8 years ago
  
  869d6fff
- Use padding on theta large matrices and get trid of explicit 4s · a8eadc59
  Dominic Kempf authored 8 years ago
  
  a8eadc59
- Fixup fastdg · 60f5ea19
  Dominic Kempf authored 8 years ago
  
  60f5ea19
- Refactor treatment of vectorized sumfact kernels · ea799b09
  Dominic Kempf authored 8 years ago
  
  ea799b09
Apr 04, 2017
- Small cleanup · f0ef08f8
  Dominic Kempf authored 8 years ago
  
  f0ef08f8
Apr 03, 2017
- First improvement of vectorization strategy · 2c9cdfc6
  Dominic Kempf authored 8 years ago
  
  2c9cdfc6
- Avoid listing SumfactKernel arguments several times · f84650ac
  Dominic Kempf authored 8 years ago
  
  f84650ac
- Move the output generation on the sumfact node · a84d4eda
  Dominic Kempf authored 8 years ago
  
  a84d4eda
- AMatrix -> BasisTabulationMatrix · afd6ad29
  Dominic Kempf authored 8 years ago
  
  afd6ad29
- AMatrix -> BasisTabulation · c1f06168
  Dominic Kempf authored 8 years ago
  
  c1f06168
- Move sumfact node to own module · d02cc280
  Dominic Kempf authored 8 years ago
  
  d02cc280
- sumfact.sumfact -> sumfact.accumulation · 5a2cdb03
  Dominic Kempf authored 8 years ago
  
  5a2cdb03
- Remove underscore from names in sumfact permutation · c3866d3d
  Dominic Kempf authored 8 years ago
  
  c3866d3d
Mar 31, 2017
- pep8 · 492e275d
  Dominic Kempf authored 8 years ago
  
  492e275d
- Lots of cosmetics · bcf54425
  Dominic Kempf authored 8 years ago
  
  bcf54425
- Simplify insn_dep propagation · a1f83023
  Dominic Kempf authored 8 years ago
  
  a1f83023
- Move the direct_output · 70190e27
  Dominic Kempf authored 8 years ago
  
  70190e27
- Remove unused code · dddbdab9
  Dominic Kempf authored 8 years ago
  
  dddbdab9
- temp construction in stage 2 · 70f49d91
  Dominic Kempf authored 8 years ago
  
  70f49d91
- More clear implementation of shape magic · e16d273d
  Dominic Kempf authored 8 years ago
  
  e16d273d
- Refactor output shape · 4b2db15f
  Dominic Kempf authored 8 years ago
  
  4b2db15f