Gauntlet #340: fix first-run baseline self-compare; pathfinder ramp start

First full run exposed two issues: (1) write_run promotes baseline.json before
the results scene loads it, so a first run compared against itself and showed
'= same' instead of FIRST RUN -- detect via the _promoted_baseline flag;
(2) pathfinder_rush started at 5 queries/tick, but queries burst on one 100 ms
tick frame, so the very first hold window failed (p95 38.9 ms) and scored 0 --
start the ramp at 1 query/tick so the sustainable level is actually found.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

tests/benchmarks/gauntlet/gauntlet_main.py

@@ -47,7 +47,12 @@ def trial_meta():
 
 # ------------------------------------------------------------------ results
 def build_results_scene(record):
-    baseline = baseline_io.load_baseline()
+    # If this run was just promoted to baseline.json, it IS the baseline --
+    # show FIRST RUN rather than comparing the run against itself.
+    if record.get("_promoted_baseline"):
+        baseline = None
+    else:
+        baseline = baseline_io.load_baseline()
     cmp = baseline_io.compare(record, baseline)
     meta = trial_meta()
 

tests/benchmarks/gauntlet/trials/pathfinder_rush.py

@@ -20,7 +20,9 @@ class PathfinderRush(Trial):
     unit = "queries/tick"
     accent = (76, 194, 110)
     description = "A* find_path queries across a static maze"
-    base_load = 5
+    # Queries are issued in a burst on the 100 ms sim tick, so even a handful
+    # spike a single frame; start the ramp at 1 so the sustainable level is found.
+    base_load = 1
     growth = 1.6
 
     def setup(self, scene, ui):