{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T07:49:06.216Z",
  "tool": "scripts/lab/context-fit-sweep.mjs",
  "runtime": {
    "name": "ollama",
    "version": "0.32.5",
    "endpoint": "loopback"
  },
  "system": {
    "cpu": {
      "model": "12th Gen Intel(R) Core(TM) i9-12900K",
      "logicalCores": 24
    },
    "memory": {
      "totalBytes": 32816553984
    },
    "os": {
      "platform": "linux",
      "release": "7.0.0-34-generic"
    },
    "gpu": {
      "model": "NVIDIA GeForce RTX 3080",
      "totalVramBytes": 10737418240,
      "driverVersion": "595.91.07",
      "provider": "nvidia-smi"
    }
  },
  "model": {
    "identifier": "qwen3:8b-q8_0",
    "digest": "e56358ca25dd14db6853a9f68a92d717aaa6f0a94250a72d1a0f3d86a9f30130",
    "weightsBytes": 8851089538,
    "family": "qwen3",
    "parameterSize": "8.2B",
    "quantization": "Q8_0",
    "trainedContextTokens": 40960
  },
  "settings": {
    "contexts": [
      4096,
      8192,
      16384,
      32768
    ],
    "skippedContexts": [],
    "passes": 4,
    "warmups": 1,
    "numPredict": 64,
    "temperature": 0,
    "seed": 42,
    "prompt": "fixed short prompt, see PROMPT in scripts/lab/context-fit-sweep.mjs",
    "settleSeconds": 8
  },
  "steps": [
    {
      "numCtx": 4096,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9276751872,
        "freeBytes": 1460666368
      },
      "loadDurationNs": 12622210536,
      "ps": {
        "sizeBytes": 8892742368,
        "sizeVramBytes": 8892742368,
        "hostBytes": 0,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 878617000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 2968445000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 828567000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13376000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 830795000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13380000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 829130000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13410000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 77.18934304632566,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9201254400,
        "freeBytes": 1536163840
      },
      "loadDurationNs": 3802588924,
      "ps": {
        "sizeBytes": 10234049329,
        "sizeVramBytes": 8819488849,
        "hostBytes": 1414560480,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 2281038000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 188659000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 2239501000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 37177000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 2242749000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 36275000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 2243004000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 38298000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 28.536407774565948,
        "fullyInVram": false
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9209643008,
        "freeBytes": 1527775232
      },
      "loadDurationNs": 3774279256,
      "ps": {
        "sizeBytes": 11457737522,
        "sizeVramBytes": 8829838295,
        "hostBytes": 2627899227,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 4144334000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 293288000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 4203162999,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 68334000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 4088954000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 67657000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 4079542000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 65949000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 15.651924673156019,
        "fullyInVram": false
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9194962944,
        "freeBytes": 1542455296
      },
      "loadDurationNs": 4009868110,
      "ps": {
        "sizeBytes": 13987951409,
        "sizeVramBytes": 8814434713,
        "hostBytes": 5173516696,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 7016852000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 478374000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 6930097000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 118339000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 6864161000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 113641000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 6842108000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 108505000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 9.323790627871345,
        "fullyInVram": false
      }
    }
  ]
}
