Jammy2211 · Oct 8, 2026 · Oct 8, 2026
diff --git a/campaigns.yaml b/campaigns.yaml
@@ -77,10 +77,10 @@ campaigns:
   title: Measurement reliability and profiling tools
   status: active
   updated: '2026-10-08'
-  next: 'timing_noise_audit phase 1 merged (autolens_profiling#402): inventory of 31 timing assertions/gates, 10 UNSAFE-SILENT
-    including the dashboard qualify() flag Pulse relies on (pulse/catalogue.py requires qualified: true). Fix phase (2)
-    will change the profiling-summary contract - Pulse must coordinate. New campaign page wiki/campaigns/measurement_tools.md.
-    Other tasks unchanged and unissued.'
+  next: 'timing_noise_audit fix phases (1) and (2) done (autolens_profiling#404 merged; #405 open, awaiting /prm). Correction
+    2026-10-08: the live Pulse lens feed is dashboard/catalogue.json (profiling-summary@2, registry switched 2026-10-05
+    in 3c7bed2), whose producer hard-codes qualified: false, so the v1 qualify() gap never reached Pulse live and fix (2)
+    needed no Pulse change. Next: fix phase (3) INCONCLUSIVE for go/lever rules. Other tasks unchanged and unissued.'
   evidence: https://github.com/PyAutoLabs/autolens_profiling/blob/main/wiki/campaigns/measurement_tools.md
 - id: autofit
   title: PyAutoFit profiling bootstrap
@@ -139,14 +139,14 @@ tasks:
   priority: high
   path: tasks/timing_noise_audit.md
   updated: '2026-10-08'
-  next: 'Phase 2 MERGED 2026-10-08 as autolens_profiling#404: one abba_overhead_verdict (excess ms vs the 12 ms budget,
-    t-interval, PASS/FAIL/FAIL_GROSS/INCONCLUSIVE) serves the CI test and the fixed-light numba cell; promotion requires
-    PASS; 6/17 published RAL rows re-judge INCONCLUSIVE (recorded, no JSON rewritten). Human decision: the ~16 ms CI
-    fixture cannot resolve 12 ms, so the CI test is the coverage + gross-breakage guard. Inventory now 13 SOUND / 9 FRAGILE /
-    9 UNSAFE-SILENT. Next: fix phase (2) dashboard qualify()/drift wording - a profiling-summary contract change that needs
-    a paired PyAutoPulse reader change (plan as a two-repo task); then (3) INCONCLUSIVE for go/lever rules.'
+  next: 'Fix phase (2) PR OPEN 2026-10-08 as autolens_profiling#405 (awaiting /prm; Heart YELLOW reasons human-acknowledged):
+    dashboard qualify() leaves laptop rows, rows without a load average and hpc rows without a host unqualified with a
+    reason; a within-2x comparison is flat only when both endpoints carry a repeat summary (none do yet), else insufficient;
+    drifted/improved keep status with a single-sample caveat (human decision 2026-10-08). Effect on summary.json: qualified
+    2 -> 2, flat 4 -> 0, insufficient 135 -> 139. Not a Pulse contract change (live feed is v2 catalogue.json, qualified
+    hard-false). Next after merge: fix phase (3) INCONCLUSIVE for go/lever rules, tie sets instead of argmax.'
   issue: https://github.com/PyAutoLabs/autolens_profiling/issues/362
-  evidence: https://github.com/PyAutoLabs/autolens_profiling/pull/404
+  evidence: https://github.com/PyAutoLabs/autolens_profiling/pull/405
 - id: mge_likelihood_breakdown_steps_are_cumulative_an
   campaign: measurement-tools
   title: MGE likelihood_breakdown steps are cumulative and `linear_gaussians` is reported as 0

diff --git a/dashboard.html b/dashboard.html
diff --git a/dashboard.md b/dashboard.md
@@ -1,6 +1,6 @@
 # PyAutoPulse — profiling dashboard
 
-<!-- pulse-campaigns:af6f0d180a436637f5db5be8ad42e33eaec6af7c87278d4294ed46e61fe379d4 -->
+<!-- pulse-campaigns:652e6582f3ead66f68ae1120a79f68f1c18219dbc4ffee221291672133f5fc79 -->
 ## Profiling Check In
 
 <details><summary>Full check-in prompt</summary>
@@ -133,13 +133,13 @@ Reviewed 2026-10-04. No new results since the pin. Over-sampling phase 1: add --
 
 <details><summary>Measurement reliability and profiling tools — active</summary>
 
-Reviewed 2026-10-08. timing_noise_audit phase 1 merged (autolens_profiling#402): inventory of 31 timing assertions/gates, 10 UNSAFE-SILENT including the dashboard qualify() flag Pulse relies on (pulse/catalogue.py requires qualified: true). Fix phase (2) will change the profiling-summary contract - Pulse must coordinate. New campaign page wiki/campaigns/measurement_tools.md. Other tasks unchanged and unissued.
+Reviewed 2026-10-08. timing_noise_audit fix phases (1) and (2) done (autolens_profiling#404 merged; #405 open, awaiting /prm). Correction 2026-10-08: the live Pulse lens feed is dashboard/catalogue.json (profiling-summary@2, registry switched 2026-10-05 in 3c7bed2), whose producer hard-codes qualified: false, so the v1 qualify() gap never reached Pulse live and fix (2) needed no Pulse change. Next: fix phase (3) INCONCLUSIVE for go/lever rules. Other tasks unchanged and unissued.
 
 [Campaign evidence](https://github.com/PyAutoLabs/autolens_profiling/blob/main/wiki/campaigns/measurement_tools.md)
 
 ### Active tasks
 
-- [Audit timing tests and profiling gates for measurement noise](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/timing_noise_audit.md) — active: Phase 2 MERGED 2026-10-08 as autolens_profiling#404: one abba_overhead_verdict (excess ms vs the 12 ms budget, t-interval, PASS/FAIL/FAIL_GROSS/INCONCLUSIVE) serves the CI test and the fixed-light numba cell; promotion requires PASS; 6/17 published RAL rows re-judge INCONCLUSIVE (recorded, no JSON rewritten). Human decision: the ~16 ms CI fixture cannot resolve 12 ms, so the CI test is the coverage + gross-breakage guard. Inventory now 13 SOUND / 9 FRAGILE / 9 UNSAFE-SILENT. Next: fix phase (2) dashboard qualify()/drift wording - a profiling-summary contract change that needs a paired PyAutoPulse reader change (plan as a two-repo task); then (3) INCONCLUSIVE for go/lever rules.
+- [Audit timing tests and profiling gates for measurement noise](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/timing_noise_audit.md) — active: Fix phase (2) PR OPEN 2026-10-08 as autolens_profiling#405 (awaiting /prm; Heart YELLOW reasons human-acknowledged): dashboard qualify() leaves laptop rows, rows without a load average and hpc rows without a host unqualified with a reason; a within-2x comparison is flat only when both endpoints carry a repeat summary (none do yet), else insufficient; drifted/improved keep status with a single-sample caveat (human decision 2026-10-08). Effect on summary.json: qualified 2 -&gt; 2, flat 4 -&gt; 0, insufficient 135 -&gt; 139. Not a Pulse contract change (live feed is v2 catalogue.json, qualified hard-false). Next after merge: fix phase (3) INCONCLUSIVE for go/lever rules, tie sets instead of argmax.
 - [MGE likelihood_breakdown steps are cumulative and `linear_gaussians` is reported as 0](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/mge_likelihood_breakdown_steps_are_cumulative_an.md) — ready: Unchanged and unissued (no GitHub refs, 2026-10-04). Select one bounded step when measurement-tools is prioritised.
 - [A gradient-cost probe: forward vs `value_and_grad` ms/eval and a strict FD check, on any registry cell](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/gradient_cost_probe.md) — ready: Unissued. New context 2026-10-07: Heart unit-test timing flagged PyAutoLens forward-gradient tests +150%/+96% after PyAutoLens#768 rewrote that test file; not a controlled comparison, and this probe is the right instrument. Select one bounded step when prioritised.
 - [Numba breakdown harness: perturb the instance so the operated-matrix memo cannot hide a step](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/numba_breakdown_harness_memo_blind.md) — ready: Unchanged and unissued (PyAutoArray#496 closed 2026-08-27). Select one bounded step when prioritised.

diff --git a/tasks/setup_baseline_campaign.md b/tasks/setup_baseline_campaign.md
@@ -75,3 +75,7 @@ the later collection and review deliverables.
 - Specification merged: https://github.com/PyAutoLabs/autolens_profiling/pull/385 (2026-10-06; issue #384 closed 2026-10-06; PyAutoPulse#17 closed 2026-10-06). `baseline/campaign.json` is status draft, version 1, with `baseline/README.md` readiness contract (https://github.com/PyAutoLabs/autolens_profiling/blob/719459fefb40c7e5be157f652818dbf923830557/baseline/campaign.json, https://github.com/PyAutoLabs/autolens_profiling/blob/719459fefb40c7e5be157f652818dbf923830557/wiki/campaigns/setup_baseline.md).
 - Six unknowns recorded verbatim: no frozen revisions/env lock; no collection authorization/window; hardware/affinity identity unselected; each cell needs complete settings, input hashes, witness + tolerances; compile per-cell builder unavailable; proposed fp64/mixed coverage is not evidence of support.
 - Status stays `needs-decision`: no collection or acceptance authorized; the check-in authorizes none.
+
+## Note 2026-10-08 (from timing_noise_audit fix phase 2)
+
+- When the frozen specification names a reference host and the v2 catalogue starts writing `qualified: true`, reuse autolens_profiling's `build_dashboard.is_reference_host_class` rule (laptop rows never qualify; a missing load average or host is unqualified with a reason) so the v1 and v2 qualification cannot drift apart. Source: autolens_profiling#405.
diff --git a/tasks/timing_noise_audit.md b/tasks/timing_noise_audit.md
@@ -43,6 +43,17 @@ The original task below is retained verbatim. Current scheduling state lives in
 - Human decision: on the ~15–17 ms CI fixture the shared 12 ms budget can only fail through the gross guard; the CI test is documented as the coverage + cached-site-count + gross-breakage guard, and the ms budget is judged on the 225–415 ms production rows.
 - Next: fix phase (2) — `build_dashboard.qualify`/`drift` wording. This changes the `profiling-summary` contract Pulse ingests; plan it as a two-repo task (autolens_profiling producer + PyAutoPulse reader/fixtures) before any producer change lands.
 
+
+## Execution 2026-10-08 — phase 3 (fix phase 2: dashboard qualification + drift wording)
+
+- Human: "continue … --auto" after the plan was presented in the Pulse chat; issue #362 reused (plan comment https://github.com/PyAutoLabs/autolens_profiling/issues/362#issuecomment-6067418030). Mind `active/timing_noise_audit_phase3_qualify_drift.md` (f27988f1); branch `feature/timing-noise-audit-p3-qualify-drift`; Opus executes.
+- **Correction to the phase 1/2 notes above:** Pulse's live `lens` registry reads `dashboard/catalogue.json` (`profiling-summary@2`) since 2026-10-05 (3c7bed2). That producer (`build_catalogue.py`) hard-codes `qualified: false` on every record, and its comparisons list is empty. `build_dashboard.qualify()`/`drift()` feed only the v1 `summary.json` (project dashboard and badge). The P6 gap therefore never reached Pulse live, and fix (2) is a single-repo change: no Pulse reader change was needed.
+- Human decision 2026-10-08: `drifted` / `improved` keep their status and carry a "single-sample endpoint(s)" caveat; only the within-band case changes (`flat` only with repeat summaries on both endpoints, else `insufficient`).
+- Heart readiness was YELLOW (8 organism-wide manifest-drift reasons + "no rehearsal for current source"); the human acknowledged that list for this PR on 2026-10-08.
+- PR [autolens_profiling#405](https://github.com/PyAutoLabs/autolens_profiling/pull/405) opened 2026-10-08. Verified on the branch: 1169 passed / 6 skipped; independent review CLEAN; Pulse v1 `validate()` returns [] before and after. `summary.json` effect: 159 records, qualified 2 → 2 (no record changed), `flat` 4 → 0, `insufficient` 135 → 139, `improved` 6 → 6 with the caveat. Audit counts now 15 SOUND / 8 FRAGILE / 8 UNSAFE-SILENT.
+- Carry-forward: when the v2 catalogue starts qualifying records (setup-baseline reference-host decision), it must reuse the producer's `is_reference_host_class` rule rather than a separate one.
+- Next: /prm the PR (judge tier, human merge); then fix phase (3) — INCONCLUSIVE for go / lever rules, tie sets instead of argmax.
+
 ---
 
 # Audit timing tests and profiling gates for measurement noise