Skip to content

Check the boundary chipf against its definition - #11

Merged
jurasic-pf merged 1 commit into
mainfrom
chipf-boundary-invariant
Sep 8, 2026
Merged

jurasic-pf merged 1 commit into
mainfrom
chipf-boundary-invariant

Conversation

@jurasic-pf

@jurasic-pf jurasic-pf commented Sep 8, 2026 •

Copy link
Copy Markdown
Contributor

The V&V job has been failing since proximafusion/vmecpp#736 changed the boundary
chipf extrapolation

image

VMEC++ extrapolates the boundary chipf like iotaf, following PARVMEC's
add_fluxes.f90, since proximafusion/vmecpp#736. The Fortran references keep
the 8.52 rule 2*chips(ns) - chips(ns1), so the last full-grid point mismatches
at 5.1e-3 on dshape_beta1_mn7_ns51 and 4.3e-3 on heliotron_beta5_mn7_ns51
while every interior point agrees to 2.8e-16.

The reference value at that point is replaced by iotaf * phipf. phipf is
constant without aphi, which no configuration in the suite uses, so the
identity is exact: the residual is 2.2e-16 and 4.4e-16 on the two
configurations, against 6.7e-3 and 1.5e-2 for the references there.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@jurasic-pf
jurasic-pf force-pushed the chipf-boundary-invariant branch from 7febd28 to 90c41a0 Compare September 8, 2026 11:35
@jurasic-pf jurasic-pf changed the title Verify the boundary chipf against its definition instead of the reference Check the boundary chipf against its definition Sep 8, 2026
@jurasic-pf
jurasic-pf merged commit 8b9c41d into main Sep 8, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant