2. Future Work
Preamble
This topic is meant to collect feature requests.
New features
- 3Dvisualize (or successor) handle output when Ngrain not equals to 1
- Spectral solver: Layer for breaking boundary conditions in all directions
- get rid of 3 dim data but rather use linear structure (more a general idea, not needed anyway)
- Logical array out of geom? material.config? to know which points to ignore
- Constitutive_none. Nothing, just F as an output No stress, no dPdF, memory efficient as possible
- make Abaqus restart aware, done for standard, expl. not testet
- do the fourier transform and convolution on a 8time higher resolved grid Notes-higherResolvedFourierSpace.pdf
Open questions
- Spectral solver: Check velocity gradient calculation for missing correction (compare to J2 test)
- Version numbering: read this german text or check out semver.org *
DAMASK G MM.mm.ff.pppp
^^^^^^--------------------------- Name
^------------------------- Generation
^^---------------------- Major
^^------------------- Minor
^^---------------- Bugfix
^^^^----------- Patch
- random is random makes bug hunting difficult. Rather used a fixed seed per default?
Missing tests
- MSC.Marc: Element library
- MSC.Marc: Restart
- Phenopoerlaw: Single slip analytic results
- addCauchy
- Spectral: Cutback
- Fiber / gauss (mtex)
- F prescribed: Test it for spectral solver
- write data to HDF5 format
-
mesh_FEasCP
: string comparison and general use. maybe use ORDERPACK.
Pre and Post processing scripts
- Migrate from optparse to newer argparse (requires Python > 2.7)
Full Fortran 2008 support
- use %im and %re for accessing complex numbers
- use mold allocation (i.e. getting shape of arrays)
- use the newunit option
PDF output