test(mitc4): record dual-mode integration verification
T4 operational verification only; no tracked source/test/docs/reference diff.
Source SHA: 6bbbb2ba8c
Evidence: .build/drilling-options-20261020/T4
Commands/exits/raw/XML/report: verification-commands.txt, summary.json,
final-report.json, verification-report.txt, ctest-logs/, failure-evidence/.
Native Debug full: 1035 run = 1024 pass / 10 fail / 1 inconclusive skip;
unit exit1, integration exit0, reference exit0. CMake build0, full CTest
same1035 identities/counts, exit8. Pinned isolated Python164 pass, exit0.
Legacy Debug build0, selected162 unique =155 pass /7 fail /0 skip;
5 known physical failures +2 actual Direct-default-only selection errors
remain raw FAIL. Related reference25, nonlinear supplement2 and official3
all pass. Both modes actual linear/coarse-NL/eigen comparisons pass.
Direct official729node/101frame Release HHT CLI0/comparator0; U101families
pass, max relative RMS0.2282326248391787%. Legacy actual small HHT3frame
CLI/readback pass.10 fresh HDF5 marker/schema/all-frame readbacks pass.
Known10 Direct failures/1skip remain unresolved. Old thermal NT11 numerical
FAIL remains diagnostic (87 failed rows). Formal fine/fine_02/pressure/new
thermal data missing; no current PASS claimed.729node Legacy HHT not rerun
under approved related scope. No performance certification, tolerance/order/
physics/test changes, reference/temp use, push or dev branch changes.
34 formal reference originals unchanged; Legacy source differs only at
private production mode constant. Final binary/DLL hashes rechecked.
Evidence manifest SHA256: f84f5b133c81aab0b9290c504859553d125f719b77c48401918fbb8f387b1aab
This commit is contained in:
KOKO\Mimi