Skip to content

Early interp (density)#322

Draft
ShawnL00 wants to merge 2 commits into
inducer:early-interpfrom
ShawnL00:early-interp-sl
Draft

Early interp (density)#322
ShawnL00 wants to merge 2 commits into
inducer:early-interpfrom
ShawnL00:early-interp-sl

Conversation

@ShawnL00

@ShawnL00 ShawnL00 commented Jun 21, 2026

Copy link
Copy Markdown
Contributor

WIP:

This PR adds need-driven interpolation for layer-potential densities.

Plain densities continue to be interpolated as whole units from stage1 to quad_stage2. For densities coupled with geometry factors, this PR adds explicit markers:

  • sym.distribute_interpolation(expr) allows interpolation to be pushed into selected density expressions.
  • sym.geo_density(geometry, density) is a helper for common geometry * unknown densities.
  • sym.interpolation_unit(expr) marks quantities that must remain atomic under interpolation.

For example, this lets expressions such as normal * sigma be written so that the normal is formed using the stage2 geometry parameterization and interpolated to quad_stage2, while the unknown sigma is interpolated from stage1 to quad_stage2. At the same time, QWeight/L2-scaled quantities such as u / sqrt_jac_q_weight remain atomic.

tangential_to_xyz(...) transforms variables from a local tangential basis to the standard Cartesian basis, so the resulting Cartesian vector should be interpolated as an atomic quantity.

@ShawnL00 ShawnL00 changed the title Early interp Early interp (density) Jun 21, 2026
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