Average body mass in IPC-toolkit vs average lumped nodal mass in IPC

[halehOssadat]

I have a question about the arguments in adaptive barrier stiffness function called ipc::initial_barrier_stiffness, it has one argument called const double average_mass which indicates as average mass of all bodies while in IPC function called suggestKappa(kappa) and according to Supplement of IPC paper, it should be average of lumped nodal mass. Does it the same as IPC?
And what types of CCD IPC-toolkit does use for ipc::compute_collision_free_stepsize, it seems that it is TIGHT_INCLUSION while it seems that IPC uses FLOATING_POINT_ROOT_FINDER as default. Can I use this with IPC-toolkit?

[zfergus]

The average_mass indeed refers to the "average lumped nodal mass" as described in the supplemental document of our original IPC paper.

By default, the IPC Toolkit uses the Tight-Inclusion CCD algorithm. To use the same CCD as in the original IPC codebase you can set the CMake option IPC_TOOLKIT_WITH_CORRECT_CCD to OFF.

[halehOssadat]

Thanks for your clarification, I assume that ipc::compute_collision_free_stepsize doesn't compute the CFL-INSPIRED CULLING OF CCD. I also want to implement this part using IPC-toolkit, but I am not sure what is the equivalent functions for sh.build(result, searchDir, alpha, result.avgEdgeLen / 3.0); and what is the best way to do it using IPC-toolkit?

[zfergus]

I added the CFL-inspired culling of CCD in PR #50. I will merge it in once I test it. Feel free to use the branch if you want to try it out.

[zfergus]

See this file for example usage.