recipes/docker-images/mia-tp4/docs/results/overnight-20260924-0311/ob-e3-k5.summary.json

{
 "base": "http://10.0.0.1:8888",
 "reps": 5,
 "max_tokens": 512,
 "phases": {
  "c1-prose": {
   "start": "2026-09-24T02:47:58.425777Z",
   "end": "2026-09-24T02:49:35.658327Z",
   "decode_tps": {
    "median": 26.65194227698271,
    "min": 26.575666760214414,
    "max": 26.866988565404657,
    "stdev": 0.11652374368652321,
    "n": 5
   },
   "e2e_tps": {
    "median": 26.273383979399163,
    "min": 26.21847606911791,
    "max": 26.49177959003907,
    "stdev": 0.10864417314055967,
    "n": 5
   },
   "ttft_s": {
    "median": 0.3071267190098297,
    "min": 0.2979864460066892,
    "max": 0.3143154710123781,
    "stdev": 0.0064497911328499895,
    "n": 5
   },
   "completion_tokens": [
    512,
    512,
    512,
    512,
    512
   ],
   "distinct_outputs": 4,
   "errors": []
  },
  "c1-code": {
   "start": "2026-09-24T02:49:35.658599Z",
   "end": "2026-09-24T02:50:25.068806Z",
   "decode_tps": {
    "median": 53.399770489486755,
    "min": 53.278380367372165,
    "max": 53.46958858346519,
    "stdev": 0.06472891907392202,
    "n": 5
   },
   "e2e_tps": {
    "median": 51.80581444843791,
    "min": 51.71459092929944,
    "max": 51.92790037684898,
    "stdev": 0.06864111769716208,
    "n": 5
   },
   "ttft_s": {
    "median": 0.3093616570113227,
    "min": 0.30299046699656174,
    "max": 0.3226510109961964,
    "stdev": 0.006721503629573961,
    "n": 5
   },
   "completion_tokens": [
    512,
    512,
    512,
    512,
    512
   ],
   "distinct_outputs": 2,
   "errors": []
  },
  "c1-chat_sampled": {
   "start": "2026-09-24T02:50:25.069108Z",
   "end": "2026-09-24T02:51:52.678695Z",
   "decode_tps": {
    "median": 29.223089630676597,
    "min": 28.11224232664131,
    "max": 30.967130023600127,
    "stdev": 1.1095645506583982,
    "n": 5
   },
   "e2e_tps": {
    "median": 28.81579654573845,
    "min": 27.72029252339613,
    "max": 30.500438671697896,
    "stdev": 1.0806582528330808,
    "n": 5
   },
   "ttft_s": {
    "median": 0.285276081005577,
    "min": 0.28170862901606597,
    "max": 0.2930894090095535,
    "stdev": 0.004355819955353136,
    "n": 5
   },
   "completion_tokens": [
    512,
    512,
    512,
    512,
    512
   ],
   "distinct_outputs": 5,
   "errors": []
  },
  "c1-prose2": {
   "start": "2026-09-24T02:51:52.678828Z",
   "end": "2026-09-24T02:53:36.888613Z",
   "decode_tps": {
    "median": 24.966216879364005,
    "min": 24.64151254373325,
    "max": 25.067353936862673,
    "stdev": 0.17921145080566958,
    "n": 5
   },
   "e2e_tps": {
    "median": 24.656867574784773,
    "min": 24.362057635410686,
    "max": 24.74013355833017,
    "stdev": 0.15472478495489725,
    "n": 5
   },
   "ttft_s": {
    "median": 0.2800692449964117,
    "min": 0.27888751099817455,
    "max": 0.3799264869885519,
    "stdev": 0.040117055307938604,
    "n": 5
   },
   "completion_tokens": [
    512,
    512,
    512,
    512,
    512
   ],
   "distinct_outputs": 4,
   "errors": []
  },
  "c4": {
   "start": "2026-09-24T02:53:36.888970Z",
   "end": "2026-09-24T02:56:21.825116Z",
   "agg_tps": {
    "median": 62.44611395277183,
    "min": 60.91879174840309,
    "max": 62.891919547152035,
    "stdev": 0.7617041927220578,
    "n": 5
   },
   "per_stream_decode_median": 21.280133672723494,
   "waves": [
    {
     "wall_s": 32.56380175301456,
     "tokens": 2048,
     "agg_tps": 62.891919547152035,
     "per_stream_decode": [
      15.886309543544481,
      27.78291831085306,
      15.97984861859736,
      27.2458713503655
     ],
     "errors": []
    },
    {
     "wall_s": 33.618526258011116,
     "tokens": 2048,
     "agg_tps": 60.91879174840309,
     "per_stream_decode": [
      15.381007455007909,
      27.354712815405897,
      15.918167479403714,
      27.87929294709162
     ],
     "errors": []
    },
    {
     "wall_s": 32.79627618700033,
     "tokens": 2048,
     "agg_tps": 62.44611395277183,
     "per_stream_decode": [
      16.11039411250605,
      26.979695481463736,
      15.95906112880281,
      26.355070577147007
     ],
     "errors": []
    },
    {
     "wall_s": 32.65073344198754,
     "tokens": 2048,
     "agg_tps": 62.72447152339781,
     "per_stream_decode": [
      15.844884076963051,
      27.874486780753905,
      16.205196768299977,
      27.92962512823872
     ],
     "errors": []
    },
    {
     "wall_s": 33.304828892985824,
     "tokens": 2048,
     "agg_tps": 61.49258435107348,
     "per_stream_decode": [
      16.09299746242486,
      27.49369408723635,
      15.5330549811367,
      27.041372684741198
     ],
     "errors": []
    }
   ]
  }
 }
}