Performance difference between CasADi 3.5.5 and 3.6.4

[nickbianco]

Performance difference between CasADi 3.5.5 and 3.6.4

[jaeandersson]

No idea what might have cause the Jacobian nonzeros to blow up like that. Maybe @jgillis knows. Easiest way to debug is to do a git-bisect on the main branch to find the offending commit.

[nickbianco]

Sounds reasonable enough. I don't bandwidth for this at th moment, but when I do I can try to find the problematic commit.

[nickbianco]

I build our stack against CasADi 3.6.0 and still see the Jacobian nonzeros blowing up. Will continue bisecting from there.

[nickbianco]

@jaeandersson git-bisect identified 428c262 as the offending commit, but the problem was on our end. The syntax changed for the Jacobian sparsity methods and we didn't update properly so our sparsity detection was being ignored completely.

Closing.