Skip to content

2nd order sigma stencil

Ask Hjorth Larsen requested to merge 2nd-order-sigma-stencil into master

Evaluate sigma=|grad n|^2 using second-order stencil.

With this change, S22 binding curves with PAW/libvdwxc will be well converged for h=0.14 whereas results would still be subject to small but annoying energy shifts when using the first-order stencil.

Update references of tests (approximately 26 were failing) to match new values.

A few values changed in ways that may be worth thinking about; updates to those tests are included in separate commits whereas the remainder (uninteresting updates) are in a single commit.

Merge request reports