Skip to content

v0.4.0

Latest
Latest

Choose a tag to compare

View all tags
@github-actions github-actions released this 03 Oct 14:38
· 13 commits to main since this release
v0.4.0
218fb84
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Breaking

  • ParametricModel::fit takes an Interpolation in place of the polynomial degree: Interpolation::Polynomial { degree } for the previous fit, Interpolation::PiecewiseLinear for Triverio's (#86)
  • 2D FDFD ports: with H along z, S's reflections are the tangential E's, as the 3D ports' and mode expansions', and so minus those before 0.4.0; a mode's backward amplitude follows (#90)
  • Boundaries3d has a new public field, real_stretch: struct literals need it, or ..Boundaries3d::pml(cells) (#84)

Added

  • components and netlists, the foundation of circuits (#71)
  • the circuit solve: one sparse system per netlist, checked against Filipsson's sub-network growth (#74)
  • read and write Touchstone files, Version 1 and 2.0 (#72)
  • the circuit adjoint: every parameter's gradient from one transposed solve (#75)
  • optimization at circuit level through genoxide: a splitter, a ring at critical coupling and a fit (#76)
  • compact models by vector fitting, over parameters, and the measured fidelity (#78)
  • 3D FDFD ports: the grid's own full-vector port modes, one-way mode sources and a reciprocal S-matrix (#77)
  • the first components: waveguide, bend, couplers, MMI, Y-branch, rings and MZI (#79)
  • (studio) a component library and a chip view (#80)
  • validate against measured Mach-Zehnder interferometers (Dwivedi 2015) (#85)
  • (material) AlN's index (Rigler 2015), AlGaN films (Rigler 2013), and InGaP beyond Tanaka's range (Ferrini 2002) (#87)
  • [breaking] compact models checked against their papers: Triverio's piecewise-linear model with an exact uniform stability test (#86)
  • [breaking] QMR preconditioned by ILU(0) on Shin and Fan's operator, with PMLs stretched as much as they absorb (#84)

Fixed

  • a 3D direct solve reaches round-off on any machine, finishing by QMR on an inaccurate factorization (#82)
  • 0.4 follow-ups: MMI port polarization, provenance -0.000, parallel spectra, roadmap (#83)
  • [breaking] 2D reflections with H along z by the tangential E's convention, as the 3D ports' (#90)

Documentation

  • bring the README up to date with 0.3.3 and the 0.4 work on main (#88)
  • bring the site, getting started, the studio's README and AGENTS.md up to date (#89)
  • animate the studio in the README, recorded by a script (#91)
  • describe the crate as 0.4.0 has it, with FDTD, inverse design, layout and PDKs as planned (#92)
Assets 23
Loading