ngspice/src/maths
Keith OHara 0638aaa160 Continue Newton iteration until currents settle.
For each type of device, a *convTest function determines if the
current through that device is converged within tolerance,
and sets CKTnoncon if the current is not yet converged.

ASCRconvTest() erroneously subtracted old current from old,
rather than old from new, when evaluating convergence.

Also, since at least 3f5, the calling function NIconvTest assumed
that the worker functions indicated non-convergence through their
return value, so was ignoring the reports of current nonconvergence.
2026-04-06 00:01:12 +02:00
..
KLU bring the KLU_malloc() in right order, ticket #717 2024-12-06 22:48:56 +01:00
cmaths Make fft scaling independent from rounding behaviour for odd data. 2026-03-20 13:37:06 +01:00
dense add *.h to the source files 2022-05-12 17:24:26 +02:00
deriv remove all .cvsignore files 2012-10-26 18:30:14 +02:00
fft fft window functions back to correct scaling - no need need for post scaling step 2024-01-24 23:16:44 +01:00
misc Replace all BOOLEAN, BOOL, _Bool by bool 2024-12-15 10:25:28 +01:00
ni Continue Newton iteration until currents settle. 2026-04-06 00:01:12 +02:00
poly Remove VS compiler warning 2024-07-05 13:53:04 +02:00
sparse Reformat spoutput.c 2025-05-24 11:22:47 +02:00
Makefile.am KLU Integration from scratch #4, changed files 2023-08-16 11:14:10 +02:00