Linear Optics and Durable Tracking

简体中文 · Previous: Engine connection · Next: Equilibrium chain

These are Studio–Engine v0.1 routes and are enabled for a strict current studio-project/0.1 snapshot. A v0.2 project uses the later equilibrium and specialized screening workflows; it does not enable generic v0.1 tracking.

Evaluate linear optics

Prerequisites are a current v0.1 draft that passes structure/closure and compiles to an Engine snapshot, at least one positive-length non-marker element, and a selected target.

1. Open protocol/v0_1/fixtures/minimum-electron-ring.canonical.studio-project.json.
2. Select a tested execution target.
3. Choose Calculate → Evaluate Optics (Ctrl+R).
4. Studio validates the exact project, opens a session, evaluates, and fetches the Twiss artifact.
5. Inspect Qx, Qy, chromaticities, beta functions, and dispersion.

Before display, Studio checks project/revision identity, artifact hash, row count/order, finite values, positive beta, and coverage of every authoritative occurrence. A result finishing after a project edit is not promoted to the new revision.

This is qualified 4D linear-optics screening, not proof of a coupled machine, nonlinear acceptance, tolerances, or engineering feasibility.

Submit durable CPU tracking

Choose Calculate → Submit CPU Tracking Job…:

Field	Default	Rule
Turns	100	1–1,000,000
Macro-particles	1,000	1–100,000
Random seed	8,675,309	0 to 2^53−1
Checkpoint interval	10 turns	0 to the total turns

Total work cannot exceed 50,000,000 particle-turns. The defaults make a quick example. For a repeatable comparison, keep the seed and change one design variable at a time.

Studio journals the exact request before contact. If the response is ambiguous, the same request can be safely replayed rather than submitting a different job.

Manage jobs in Tasks

Tasks shows Job, Target, Phase, Progress, Access, and Result. Use:

• Refresh target for the selected principal-scoped registry;
• Reconnect selected with the OS-stored token;
• Cancel selected for cooperative cancellation while continuing to observe;
• Open artifact to reverify and display;
• Export artifact… for result JSON, evidence manifest, and checksum sidecar.

Accepted jobs can be observed concurrently. Only the short, not-yet-durably-recorded submission window is serialized.

Reading a tracking result

The page reports turns × particles, survived/lost counts, loss-state categories, final mean and standard deviation for x, px, y, py, zeta, delta, seed/backend/aperture status, project revision, job ID, and artifact hash.

It answers what happened to this finite ensemble in this ideal model. It is not machine DA, Touschek lifetime, or collective-effect evidence. Without modeled aperture, survival cannot prove vacuum-chamber clearance.

Old jobs remain bound to their original revision after an edit. Compare runs only when turns, particles, seed, Engine source, and physics boundary are consistent.