perf+exp: 轮次34——复用cum_xf进库(22项指标逐位不变)、批量认证压最坏尾部(80→37ms)、FP16/BF16实测无收益

- 撤回一条自己的假收益：先前"合并几何查询→24.5ms"不成立。①第一版把 xf= 传给含5行的 gaps，
  只有第0行走完宽相位，认证点全部退回截顶40.0（少干活）；②真实顺序 特征→推理→候选→认证，
  认证点依赖网络输出，故"特征间隙与认证点合并一次调用"实现上做不到
- 可实现并已落地：gm_groups(..., xf=) 复用 features 已算的 cum_xf，B=1 交错实测 P50 −3.4ms；
  256题重跑 22 项指标与轮次33 base 逐项相同
- 新增 --cert_batch（默认0不变）：K条候选并成1次 gaps 调用。合成夹具(B=2)断言选择/轨迹/证书逐位一致，
  变异检查（换回错误索引）2项测试立即失败；256题重跑同样 22 项指标逐位相同
  实测交换：典型 33.5→37.6ms，最坏（前面候选被拒）80.2→37.1ms，P99 64.7→48.3ms
- FP16/BF16 实测不比 FP32 快（推理 7.68→8.31→8.58ms），整周期不变而 P99 更差，动作偏差 0.005°/0.038°
  → 小批量下不是算力受限，半精度不划算，推理保持 FP32
- 仍不达 20ms：下一刀是流水线复用（下次规划的特征间隙＝本次已认证的末点），可整块省掉特征 ~13ms
- 测试 32→34 项

docs/experiment_round34_cert_batch_check.json

@@ -0,0 +1,71 @@
+{
+  "config": {
+    "ckpt": "ml/runs/base_dag2_dag1/best.pt",
+    "samples": "ml/data/samples",
+    "tasks": null,
+    "split": "val",
+    "n": 256,
+    "K": 8,
+    "commit": 2,
+    "fb": 2.2,
+    "ki": 0.0,
+    "fb_scale": 0.0,
+    "fb_near": 0.2,
+    "fb_fallback": 0.0,
+    "kdecay": 0.9,
+    "kimax": 0.5,
+    "fb_every": 1,
+    "safety_try": 0,
+    "sel_w": 0.1,
+    "dls_safety": 1,
+    "fb_clear": 0.0,
+    "fb_cmin": 0.3,
+    "sub_refresh": 0,
+    "fb_block_gain": 0.5,
+    "dump": null,
+    "out": "docs/experiment_round34_cert_batch_check.json",
+    "trace": null,
+    "select": "endpoint",
+    "replan": "empty",
+    "safety": 1,
+    "partial": 1,
+    "inc_cap": 0,
+    "gap_margin": 0.0,
+    "cert_batch": 1,
+    "max_sec": null,
+    "seed": 0,
+    "dls": "analytic",
+    "device": "cuda"
+  },
+  "batch_size": 256,
+  "batch_seed_rule": "same seed per batch",
+  "checkpoint_sha256": "f7f686fd8ee92fb6948fbcc4ee954636065ad87bb54f661c4f8f575bc16c2ce1",
+  "tasks_sha256": "249bd93457825cda98c59e64d9a096d3473c69aeb3553de1dea6d9e56be71772",
+  "kin_fp": "2ebcbf17",
+  "act_scale": 1.0,
+  "metrics": {
+    "n": 256,
+    "n_ref": 256,
+    "success": 0.57421875,
+    "collision_free_success": 0.57421875,
+    "time_ok_1p3": 0.390625,
+    "collision_rate": 0.0,
+    "collision_steps_mean": 0.0,
+    "travel_ratio_median": 2.4449527263641357,
+    "time_mean_s": 2.3076188564300537,
+    "brakes_mean": 19.25,
+    "start_fail": 0.0,
+    "goal_ratio_2s": 0.20997506380081177,
+    "err_fail_med": 229.68746948242188,
+    "err_success_med": 0.18899640440940857,
+    "dls_share": 0.625,
+    "time_budget": "teacher×2(≥3 s)",
+    "speed_ratio_p50": 1.6600004434585571,
+    "speed_ratio_p95": 1.6600035429000854,
+    "speed_ratio_max": 1.660003662109375,
+    "over_vmax_rate": 1.0,
+    "accel_ratio_p95": 1.5835107564926147,
+    "accel_ratio_max": 1.6600004434585571
+  },
+  "simulation_wall_s": 306.9490996999957
+}