1. 13 Jul, 2020 1 commit
  2. 10 Jul, 2020 4 commits
  3. 09 Jul, 2020 2 commits
  4. 08 Jul, 2020 1 commit
  5. 07 Jul, 2020 5 commits
  6. 06 Jul, 2020 2 commits
  7. 02 Jul, 2020 1 commit
    • Martin Hafskjold Thoresen's avatar
      Testing and debugging · dc144031
      Martin Hafskjold Thoresen authored
      We're slowly getting closer towards what's going on. Currently we're looking
      at the FD hessian for pressure tets, which does not seem to be right,
      implying that the forces are not right either.
  8. 01 Jul, 2020 2 commits
  9. 30 Jun, 2020 4 commits
  10. 29 Jun, 2020 6 commits
  11. 26 Jun, 2020 4 commits
  12. 25 Jun, 2020 2 commits
    • Martin Hafskjold Thoresen's avatar
      Draw some book info · 0c1410fc
      Martin Hafskjold Thoresen authored
      This is mostly information that we already have, but it's better structure wise.
    • Martin Hafskjold Thoresen's avatar
      Clean up bookkeeping and newtons method. · 02e43b13
      Martin Hafskjold Thoresen authored
      Now we're further embracing the bookkeeping concept,
      and `ForceData`, which previously was the state of a "simulation"
      has been replaced with `Books`, that now contain all interesting information about
      the simulation. This is good so that we can run multiple runs of a sim without
      doing any extra bookkeeping if we want to.
      Next up is to add GUI for the books.
  13. 24 Jun, 2020 1 commit
  14. 23 Jun, 2020 4 commits
  15. 22 Jun, 2020 1 commit
    • Martin Hafskjold Thoresen's avatar
      Try out aligned dirs with z · f8b9f096
      Martin Hafskjold Thoresen authored
      We try to align the no-stretch axis on all faces with the z axis.
      Expected behaviour is that the donut is shrinking along z
      but not along y or x.
      In reality, Wolfe is complaining about no good step size found.
       1. Double check wolfe (although we know that the step sizes are okay,
          and the energy calculation is seemingly okay).
       2. Visualize the attempted direciton. If this looks plausible, we suspect
          that the problem is in the optimizer and not in the consitutive model.