The MAST-U reference diagnostic is prepared for review
The saved trapped gyro-Landau fluid (TGLF) reference cannot establish acceptance for the retained MAST Upgrade candidate. It rejects a face the candidate serves. Its saved profiles and final mask contain no raw coefficient or rejection-reason evidence that identifies the cause.
A separate diagnostic runner is now prepared. It preserves the existing scenario, solver settings, guards and every face. After stepping, it records inputs, raw reference and fallback coefficients, generic guard features and masks at every accepted state. Its report reconstructs the guard exactly, retains overlapping reasons and identifies the investigated face explicitly.
| preparation evidence | result |
|---|---|
| synthetic smoke and focused checks | 32 focused checks pass in each mode; a two-step synthetic training-equilibrium smoke is finite and reconstructs the raw guard exactly; all 77 source modules pass typing, and lint/format plus batch syntax passC-192 |
| physical rejection cause and reference admissibility | unresolved; compute awaits review |
The smoke uses a training equilibrium and synthetic worker responses. It forbids the physical TGLF solver and produces no physics labels or acceptance metrics. The immutable source package binds original inputs, the consumed candidate, full resolved configuration and dependency versions/bytes. Historical missing dependency bytes remain missing; this preparation does not supply retrospective evidence.
The concrete diagnostic job needs the owner’s approval in OQ-31. It may reproduce the same rejected face. Its purpose is to explain that rejection before reviewing a physically justified remedy. The separately proposed feedback-sampling change still needs its own frozen recipe and compute review.
K-030 remains open. The current model and guard stay selected. No new production fitting, reference compute, synchronization or replacement has occurred.
What’s next. Review the prepared diagnostic request, then inspect its result before deciding whether a reference remedy or sampling experiment should proceed.
Technical details → Verification › reference diagnostic preparation
Decisions → OQ-31