Spheral++ Activity
Brought to you by:
mikeowen
Activity for Spheral++
-
J. Michael Owen committed [992f09]
Further work getting ready to test JC.
-
Initial implementation of the Johnson-Cook damage model -- untested.
-
Fixing a comment.
-
Renaming 2D tensile rod in preparation for adding JC option.
-
Fixing path to renamed 1D tensile rod test.
-
Adding the option to make D1 and/or D2 constant.
-
Starting to test the JC damage model.
-
Adding boundary conditions to JC.
-
Checkpoint on unit test for JC damage.
-
Renaming tensile rod 1D test since we're going to test more than one kind of damage now.
-
Hacking the mesh silo writer to also explicitly write out the point data too, until the problems with Visit and vardef'ing the points are resolved.
-
Adding the state evolution for the melt energy. Currently only the StrengthModel can provide that -- we should generalize to the EquationOfState as well.
-
Adding JC damage option to 2D tensile rod test.
-
Enforcing minimum failure strain.
-
Forgot to update the effective damage, which is what is actually applied.
-
Initial python bindings for JohnsonCookDamage.
-
Removing debug print.
-
Fixing the runstack error that is screwing us up on LC TOSS systems.
-
Adding a section on building with WSL on Windows 10.
-
default->OpenMP
-
A few minor fixes for Steinberg-Guinan -- doesn't seem to fix the test case though.
-
OpenMP->default
-
Cody Raskin committed [1f1df4]
Fixes to hydrostatic solver to remove debug prints and RPRPS to protect against fault when only 2 procs
-
Fixing our flags for rubble damage behavior.
-
Updating our bit-perfect restart tests -- no sign of any problem in the code.
-
Aha, seem to have our Visit silo file internal paths straightened out.
-
Checkpoint cleaning up viz files.
-
Getting rid of a debug statement.
-
A few changes in this commit:
-
Gah, typo in compile guard for MPI.
-
Experimenting with a differet rule for rolling fully damaged material over to the reference state.
-
Turning off the damage gradient switch by default, since I think for things with surfaces it's doing more harm than good right now. We'll revisit this one. Also updated test reference data so everything is passing right now.
-
Getting interface to compile -- untested.
-
Fixing Hypre linking.
-
Merge default<->default
-
Gah, more head merging.
-
Working on Jens' fractal stuff.
-
Jens Villumsen committed [870106]
If You choose to have no output, for whatever reason, set BaseDirectory equal to the null string and all output, except cerr, with disappear into a black hole.
-
Downselecting to just the cell based viz method now that it supports point mesh mapping as well.
-
Fixing (actually a workaround) for point mesh defvars in siloMeshDump.
-
Merging.
-
Making some progress.
-
Load balance algorithm may generate boxes with no points. Empty box removal included.
-
Extended Manual
-
update Manual
-
Fixed a linking problem.
-
Adding the interface build target.
-
Checkpoint on Hades.
-
Reactivating masking by damage as a possibility, but the default damage threshold makes this a noop by default.
-
Checkpoint work on SEGV.
-
More fiddling with how to write for Hades.
-
Fixing corruption writing out vector<string> to silo.
-
Getting materials straightened out.
-
Removing the rollover to initial H in the presence of damage. This seems to exacerbate some instabilities in SPH.
-
Gah, punting and having dropViz update the connectivityMap before calling the cell visualization.
-
Visit displays again, but hades still SEGVs.
-
Getting DBWrite with arrays squared away.
-
Gah, still working on this.
-
Merging default->default.
-
Bugfix.
-
Improved computational cube finder for isolated boundary conditions.
-
Under special circumstances particles could be deleted by mistake. Bug fixed.
-
More fiddling with Hades output.
-
Merging with Jens' latest.
-
Gah, hades expects lots of coordinates written out.
-
More iterations on writing for Hades.
-
Paul Tsuji committed [9ca7a6]
Merging with default.
-
Adding manual for FractalGravity
-
Updated makefile.in
-
Adding some more Silo bindings.
-
Protecting from NAN sound speeds.
-
Checkpoint.
-
Found a compile bug -- Jens must not actually be using this method?
-
Minor stuff, something is still wrong with hypre install.
-
Checkpoint working on hades output.
-
Getting Fractal to build in Spheral again. I think I'm not building the right set of files though -- gotta check with Jens.
-
Got fftw and hypre bulding again. Putting hypre in it's own prefix install path though since it spews a lot of files with short collision-prone names about.
-
Gah, still not installing.
-
Checkpoint on Hades work.
-
Working on build system for Jens' fractal stuff. Guess we need to get hypre installed as well.
-
Working on the hypre third party target rules.
-
Fixing domain shuffling bug.
-
Filling out more parallel info.
-
Adding variable annotations.
-
Correcting some material writing.
-
Fixing a few contract nits for restarting runs and such.
-
Cleaning up Jens' fractal packages a bit.
-
Checkpoint.
-
Reverting back to explicit returnType in TrampolineGenerator. I have other plans for type None.
-
Constructor arguments rather than overriding attributes.
-
Checkpoint -- working on a new Hades visualization format.
-
default->pybind11
-
default->pybind11
-
Comment change.
-
Checkpoint -- fixed some output coordinates and such.
-
Checkpoint.
-
Updating the use of distributed boundary information.
-
Changing the effective strain to scale with damage, and also how aggressively we rescale properties of damaged nodes.
-
Checkpoint.
-
Checkpoint.
1 >