Commits · 4ff930e3597b5e4d4e89c6ea5d56d0708e84f26f · Christian Heinigk / dune-codegen

Jan 17, 2017
- Make permutation and vectorization possible · 4ff930e3
  René Heß authored 8 years ago
  
  4ff930e3
- Cleanup permutation in sumfactorization · 6023ba18
  René Heß authored 8 years ago
  
  6023ba18
Jan 16, 2017
- Permutation of directions is working, needs cleanup · ea33081e
  René Heß authored 8 years ago
  
  ea33081e
Jan 13, 2017
- Fix documentation typo · 2600d421
  René Heß authored 8 years ago
  
  2600d421
- Cleanup and improve documentation · fc427d44
  René Heß authored 8 years ago
  
  fc427d44
- Refactor trialfunction evaluation to be reusable for apply function · a37601db
  Dominic Kempf authored 8 years ago
  
  a37601db
Jan 12, 2017
- Use tensors is SF · a7382743
  René Heß authored 8 years ago
  
  a7382743
- Do jacobian_apply time measurements · bbd821f3
  Dominic Kempf authored 8 years ago
  
  bbd821f3
- Update documentation of sf · 57fa6878
  René Heß authored 8 years ago
  
  57fa6878
Jan 09, 2017
- Finegrainde intrumentation levels · 004f2c5e
  René Heß authored 8 years ago
  
  004f2c5e
Dec 22, 2016
- Pytest · 3ea2e80b
  René Heß authored 8 years ago
  
  3ea2e80b
Dec 21, 2016
- Detect faces in shape generation regardless of the size being 1 · fafe98af
  Dominic Kempf authored 8 years ago
  
  fafe98af
Dec 20, 2016
- Finalize broadcasting of input coefficients in vectorized case · 73abd9a0
  Dominic Kempf authored 8 years ago
  
  73abd9a0
- Remove large parts of loopy vectorization checker · 27383643
  Dominic Kempf authored 8 years ago
  
  It is not (yet!) suitable for our needs.
  27383643
Dec 19, 2016
- Clean up code and use old version for gradvec · d3ed7056
  René Heß authored 8 years ago
  
  Note: sumfact_poisson_fastdg_gradvec_nonquadvec does not work!
  d3ed7056
- Avoid direct vector acces for gradvec · d21c936f
  René Heß authored 8 years ago
  
  At the moment loopy tries to acces a vector class variable and throws a not implemented error. Probably needs some changes in the horizontal add code.
  d21c936f
- Stage 1 works if we do not vectorize · 24e28d88
  René Heß authored 8 years ago
  
  24e28d88
- Small cleanup · 4673a1e5
  René Heß authored 8 years ago
  
  4673a1e5
- Accumulation possible for fastdg · dc799b48
  René Heß authored 8 years ago
  
  dc799b48
- Change size to actual size · a5ae37b7
  Dominic Kempf authored 8 years ago
  
  a5ae37b7
- Towards fastdg accumulation · de423230
  René Heß authored 8 years ago
  
  de423230
Dec 15, 2016
- Isuue horizontal_add instructions! · 4572ed8f
  Dominic Kempf authored 8 years ago
  
  4572ed8f
- Set padded entries to zero in gradvec vectorization · 6a7d917a
  Dominic Kempf authored 8 years ago
  
  6a7d917a
- Eliminate trivial reductions · 227304ca
  Dominic Kempf authored 8 years ago
  
  For some reason, the patch refuses to schedule the kernel for sumfact_poisson_dg_symdiff.
  227304ca
- Correctly treat restriction in the jacobian_skeleton case · 0f0d69d0
  Dominic Kempf authored 8 years ago
  
  0f0d69d0
- Have index handling information be covered by memoization · c13b3424
  Dominic Kempf authored 8 years ago
  
  c13b3424
Dec 14, 2016
- Correctly treat restriction in the jacobian_skeleton case · 06941da4
  Dominic Kempf authored 8 years ago
  
  06941da4
- Have index handling information be covered by memoization · 9ecbbea3
  Dominic Kempf authored 8 years ago
  
  9ecbbea3
Dec 13, 2016
- Address all the comments · b2887ab0
  Dominic Kempf authored 8 years ago
  
  b2887ab0
Dec 12, 2016
- Fix sumfactorized jacobians on facets · ca547479
  Dominic Kempf authored 8 years ago
  
  ca547479
Dec 09, 2016
- Fix accumulation: Correctly choose s/n · 3ed8f693
  Dominic Kempf authored 8 years ago
  
  3ed8f693
- Fix on sumfact output buffer · c56531c8
  Dominic Kempf authored 8 years ago
  
  c56531c8
- small fixes · a16c7a17
  Dominic Kempf authored 8 years ago
  
  a16c7a17
- Fix facetheta matrix generation · e8cda866
  Dominic Kempf authored 8 years ago
  
  e8cda866
- Implement the evaluation of basis for jacobians · 4b137861
  Dominic Kempf authored 8 years ago
  
  4b137861
- Shrink the quadrature loop on intersection by one iname · 25807a97
  Dominic Kempf authored 8 years ago
  
  25807a97
- Move some implementation work to the AMatrix class · 1b3115e4
  Dominic Kempf authored 8 years ago
  
  1b3115e4
- Put amatrix sequence construction into a function! · 4d1f3cd2
  Dominic Kempf authored 8 years ago
  
  4d1f3cd2
Dec 07, 2016
- Last fixes to combined vectorization · 6f0041b5
  Dominic Kempf authored 8 years ago
  
  6f0041b5
Dec 06, 2016
- Combine vectorization strategies · e3b1849d
  Dominic Kempf authored 8 years ago
  
  Only 6 orders of magnitude to go!
  e3b1849d