Skip to content

Handle collision-kernel zero endpoints#132

Open
krystophny wants to merge 14 commits into
mainfrom
fix/flux-pumping/collop-zero-endpoint
Open

Handle collision-kernel zero endpoints#132
krystophny wants to merge 14 commits into
mainfrom
fix/flux-pumping/collop-zero-endpoint

Conversation

@krystophny

Copy link
Copy Markdown
Member

Summary

  • use the analytic identity for the l = 0 recurrence factor
  • use the exact zero limit for the l > 0, x = 0 inner kernel
  • avoid invoking CQUAD for a callback that is identically zero on that endpoint interval
  • add a halting-mode regression for the collision-operator zero endpoint

Stack dependency: #130. This PR remains pointed at main and must not merge before its dependency.

The change preserves the collision operator at every nonendpoint value. It does not change the quadrature algorithm, tolerances, discretization, convergence criteria, memory layout, ABI, or compiler semantics. At the endpoint, it evaluates the analytic limits: the l = 0 recurrence multiplier is one, and the regular l > 0 kernel is zero.

Closes #131.

Verification

Failing before (377226441848b64f39e97462937797e10342bab5):

0% tests passed, 1 tests failed out of 1
IEEE_INVALID_FLAG
FAIL: zero endpoint raised IEEE invalid

Passing after (69fbb86):

100% tests passed, 0 tests failed out of 1
Static analysis passed (335 Fortran files checked)
Build passed
100% tests passed, 0 tests failed out of 13
Lint passed

The focused regression completed without an IEEE invalid flag. git diff --check also passed.

isw_qflux_NA=1 with in_file_pert='none' assembles an identically zero
non-axisymmetric source. That zero vector is the Arnoldi seed, so its
normalization divides by zero (0/0 -> NaN) and zhseqr then fails to
converge at every step, spinning without qflux output. Reject the
configuration in check() with a clear message instead.
Add an optional right-hand side to the non-axisymmetric solve: the
drive of a single-helicity perturbation given by a radial corrugation
field (delta B^s/B_0^phi)_mn and an electrostatic potential harmonic
Phi_mn. The vpar corrugation piece enters the A1/A2 force channels
with velocity weights a3m and x^2 (sigma-odd), the ExB piece with
weights x^-1 and a1m (sigma-even); the two new velocity moments are
computed in compute_sources. Input via new &ntv_input variables
(isw_hel_drive, m_theta_hel, hel_brad_re/im, hel_phim_re/im), default
off. Band profiles, moments and dimensional prefactors follow a
CAS-checked derivation (see issue #118).

Off by default: golden-record QL summary.h5 is reproduced to 1e-15
against a main build.
Review findings on the drive commit: fold the x^(3k-4) weight into a
single power so the x=0 quadrature node stays finite; skip the moment
quadrature on the precomputed-matrix-element path and reject
isw_hel_drive at run time when the moments are unavailable; reject the
unvalidated relativistic branch, reconstruction mode and unset species
data; deallocate asource_hel in collop_unload.
The zero-source guard added upstream rejects isw_qflux_NA=1 with
in_file_pert='none'. With the misalignment drive, that configuration is
valid: the drive supplies the non-axisymmetric source. Relax the guard
so it fires only when the drive is off or all its amplitudes are zero.
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.

Collision precompute raises IEEE invalid at the zero endpoint

1 participant