recipes/docker-images/mia-tp4/docs/results/overnight-20260924-0311/ob-f2-final-reuse.summary.json

{
 "base": "http://10.0.0.1:8888",
 "reps": 5,
 "max_tokens": 512,
 "phases": {
  "c1-prose": {
   "start": "2026-09-24T06:13:07.456951Z",
   "end": "2026-09-24T06:14:23.613837Z",
   "decode_tps": {
    "median": 34.14652405666999,
    "min": 33.85988724492192,
    "max": 34.52475906528059,
    "stdev": 0.2501525493349125,
    "n": 5
   },
   "e2e_tps": {
    "median": 33.60508109290937,
    "min": 33.32614655198231,
    "max": 33.940932149480815,
    "stdev": 0.24448188130552415,
    "n": 5
   },
   "ttft_s": {
    "median": 0.27170867199311033,
    "min": 0.26574828498996794,
    "max": 0.2913744800025597,
    "stdev": 0.009468896702495939,
    "n": 5
   },
   "completion_tokens": [
    512,
    512,
    512,
    512,
    512
   ],
   "distinct_outputs": 3,
   "errors": []
  },
  "c1-code": {
   "start": "2026-09-24T06:14:23.614195Z",
   "end": "2026-09-24T06:15:06.080114Z",
   "decode_tps": {
    "median": 62.24104837155688,
    "min": 61.46504315335072,
    "max": 62.85357491360117,
    "stdev": 0.5375414964436913,
    "n": 5
   },
   "e2e_tps": {
    "median": 60.363319584036006,
    "min": 59.54126223094562,
    "max": 60.83859443499137,
    "stdev": 0.4785512899536748,
    "n": 5
   },
   "ttft_s": {
    "median": 0.2854100909898989,
    "min": 0.27195627099717967,
    "max": 0.3009387729980517,
    "stdev": 0.009287248547467045,
    "n": 5
   },
   "completion_tokens": [
    512,
    512,
    512,
    512,
    512
   ],
   "distinct_outputs": 1,
   "errors": []
  },
  "c1-chat_sampled": {
   "start": "2026-09-24T06:15:06.080389Z",
   "end": "2026-09-24T06:16:16.192505Z",
   "decode_tps": {
    "median": 37.20330727837108,
    "min": 36.248058082039485,
    "max": 37.763801676830276,
    "stdev": 0.5304518189123368,
    "n": 5
   },
   "e2e_tps": {
    "median": 36.58426333185751,
    "min": 35.680963039867294,
    "max": 37.133553600396496,
    "stdev": 0.5076432023434098,
    "n": 5
   },
   "ttft_s": {
    "median": 0.2520815610187128,
    "min": 0.24551763600902632,
    "max": 0.2597504220029805,
    "stdev": 0.0048272690688732165,
    "n": 5
   },
   "completion_tokens": [
    512,
    512,
    512,
    512,
    512
   ],
   "distinct_outputs": 5,
   "errors": []
  },
  "c1-prose2": {
   "start": "2026-09-24T06:16:16.192701Z",
   "end": "2026-09-24T06:17:36.240490Z",
   "decode_tps": {
    "median": 32.55656103431811,
    "min": 31.997337897613313,
    "max": 32.659714023506226,
    "stdev": 0.24978272656672335,
    "n": 5
   },
   "e2e_tps": {
    "median": 32.10533596989721,
    "min": 31.56760094132209,
    "max": 32.19531328660531,
    "stdev": 0.2373185390593803,
    "n": 5
   },
   "ttft_s": {
    "median": 0.2517438460199628,
    "min": 0.24810012700618245,
    "max": 0.2567485769977793,
    "stdev": 0.0032167999252467367,
    "n": 5
   },
   "completion_tokens": [
    512,
    512,
    512,
    512,
    512
   ],
   "distinct_outputs": 4,
   "errors": []
  },
  "c4": {
   "start": "2026-09-24T06:17:36.240735Z",
   "end": "2026-09-24T06:19:52.427243Z",
   "agg_tps": {
    "median": 75.79343951270069,
    "min": 73.36471371141046,
    "max": 76.60329488892638,
    "stdev": 1.243935341968832,
    "n": 5
   },
   "per_stream_decode_median": 24.312903080764492,
   "waves": [
    {
     "wall_s": 26.890230299002724,
     "tokens": 2048,
     "agg_tps": 76.16148977630564,
     "per_stream_decode": [
      19.260773789147976,
      28.777115672636146,
      19.78929918090842,
      29.966202353618012
     ],
     "errors": []
    },
    {
     "wall_s": 26.735142437013565,
     "tokens": 2048,
     "agg_tps": 76.60329488892638,
     "per_stream_decode": [
      19.37607595673969,
      32.29823161107042,
      19.76544782545669,
      32.298844416036566
     ],
     "errors": []
    },
    {
     "wall_s": 27.915327360999072,
     "tokens": 2048,
     "agg_tps": 73.36471371141046,
     "per_stream_decode": [
      18.5472707733846,
      33.242378311279985,
      19.070615912900447,
      29.635868061268702
     ],
     "errors": []
    },
    {
     "wall_s": 27.623401710006874,
     "tokens": 2048,
     "agg_tps": 74.14003610055347,
     "per_stream_decode": [
      18.74881061590412,
      33.23032340382344,
      18.963252856430834,
      29.60866152452925
     ],
     "errors": []
    },
    {
     "wall_s": 27.02080830698833,
     "tokens": 2048,
     "agg_tps": 75.79343951270069,
     "per_stream_decode": [
      19.259836858334495,
      33.25823915382336,
      19.848690488892842,
      29.501133611971337
     ],
     "errors": []
    }
   ]
  }
 }
}