docs(methods): sampler-benchmarks points at the autofit inference search catalogue (search-ext A5 consumers)

wiki/methods/concepts/sampler-benchmarks.md

@@ -5,6 +5,7 @@ topics: [samplers, benchmarks, profiling, jax]
 sources:
   - autofit_workspace_developer/searches_minimal/output/ (local, regenerable)
   - autolens_profiling (local repo, HPC campaign results)
+  - autofit_inference/catalogue/search_catalogue.json (pinned commit 3463aa1, protocol gaussian_x3@1)
 status: drafted
 ---
 
@@ -72,6 +73,43 @@ From the `autolens_profiling` HPC runs ("searches first-class" series):
 Raw outputs, job scripts and result JSON/PNGs live in
 `autolens_profiling` (per-run directories; see that repo's AGENTS.md).
 
+## The search catalogue — every PyAutoFit search, per task (2026-10)
+
+The organism's first benchmark that covers **every search in the
+manifest** rather than a hand-picked few lives in `autofit_inference`:
+`catalogue/search_catalogue.json` (schema
+`autofit-inference-search-catalogue`, version 1), read here at commit
+`3463aa19014b6bd83563018568f0fb4a2376f8eb`.
+
+- **Protocol id `gaussian_x3@1`** — a toy likelihood (three blended or
+  separated 1D Gaussians, datasets `gaussian_x3_blend` /
+  `gaussian_x3_separated`), pre-registered in that repo's
+  `wiki/project/protocol_gaussian_x3.md`. Campaign task
+  `gaussian_x3_search_wave1`.
+- **Grouped by requested task, never compared across tasks**:
+  `point_map` (judged by protocol §4(a): max log posterior within 1 nat
+  of the reference MAP), `posterior` (§4(b): relabelled medians,
+  sigma-ratio band, PPC χ², mode coverage), `evidence` (§4(b)+§4(c):
+  posterior accuracy and ln Z within 1 nat).
+- **Wave 1 ranks nothing** (protocol §9): its rows calibrate the
+  protocol's thresholds. 223 of 520 expected runs landed before the
+  B3 wrap-up ruling (2026-10-08) stopped the pilot; the rest are
+  deferred to wave 2 on RAL. The pilot ran on PyAutoFit `main` before
+  phase A2 (gradients under JAX), so the BFGS / L-BFGS non-finite
+  results it records are not a judgement on those searches.
+- Per-search status at that commit: 13 `measured`, NSS `deferred`
+  (phase A3b), Drawer `unsupported` (a sanity floor that answers no
+  task).
+- The catalogue embeds the search-manifest snapshot it was built
+  against (`catalogue/search_manifest.json`, `search-manifest@1`), so a
+  row is keyed by the same registry name the rosters, the capability
+  matrix and the Brain's samplers faculty use. The faculty lists the
+  catalogue's cells as its `mature (autofit_inference search
+  catalogue)` tier.
+
+Read it as coverage and failure modes per task until a ranking wave
+lands; quote no "fastest search" from wave 1.
+
 ## Benchmark hygiene
 
 - Same seed, same data, same priors across samplers (the minimal tier

wiki/methods/index.md

@@ -25,7 +25,8 @@ should use this wiki.
 - [[initialization-chaining]] — which samplers provide vs require
   starting points; PyAutoFit chained searches.
 - [[sampler-benchmarks]] — the benchmarking-campaign record (MLTracker
-  contract, minimal + A100 results, raw-output locations).
+  contract, minimal + A100 results, raw-output locations) and the
+  per-task search catalogue (`autofit_inference`, protocol `gaussian_x3@1`).
 - [[gradient-optimizer-benchmarks]] — JAX MAP/optimization study on the MGE
   lens likelihood (CPU + A100): diversity × gradients; multi-start Adam scales
   (GIGA-Lens), SVGD reaches truth, CMA-ES collapses.

wiki/methods/log.md

@@ -132,3 +132,13 @@ and pocoMC. Resolved the dynesty source section to existing `Speagle2020`; added
 five canonical entries and compact claim support. Added candidate curation and
 an on-demand public-provenance Insight handoff using existing catch-up machinery.
 Integration is a dated source-roster check; no benchmark verdict, install or run.
+
+
+## 2026-10-15 — Search catalogue pointer on the sampler-benchmarks page
+
+Search-extensibility A5 (consumers). [[sampler-benchmarks]] gains a
+hand-written section pointing at `autofit_inference`'s
+`catalogue/search_catalogue.json` (commit 3463aa1, protocol
+`gaussian_x3@1`): per-task grouping, wave 1 ranks nothing, the
+pilot's coverage and statuses. The index entry names it. No benchmark
+verdict was added.