evanberkowitz / isle

Lattice Monte-Carlo for carbon nano systems
https://evanberkowitz.github.io/isle/
MIT License
2 stars 4 forks source link

Schur complement solver #19

Open evanberkowitz opened 4 years ago

evanberkowitz commented 4 years ago

Can we implement a Schur complement solver, as in http://arxiv.org/abs/1803.05478 ? We have encountered some numerical instability when Ubeta becomes too large---presumably the Schur solver does not suffer?

(Unfortunately it is only for exp discretization.)