{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T06:58:27.607Z",
  "tool": "scripts/lab/context-fit-sweep.mjs",
  "runtime": {
    "name": "ollama",
    "version": "0.35.1",
    "endpoint": "loopback"
  },
  "system": {
    "cpu": {
      "model": "Intel(R) Core(TM) i7-14700K",
      "logicalCores": 28
    },
    "memory": {
      "totalBytes": 34018738176
    },
    "os": {
      "platform": "win32",
      "release": "10.0.26200"
    },
    "gpu": {
      "model": "NVIDIA GeForce RTX 4070 Ti",
      "totalVramBytes": 12878610432,
      "driverVersion": "617.42",
      "provider": "nvidia-smi"
    }
  },
  "model": {
    "identifier": "llama3.1:8b",
    "digest": "46e0c10c039e019119339687c3c1757cc81b9da49709a3b3924863ba87ca666e",
    "weightsBytes": 4920753328,
    "family": "llama",
    "parameterSize": "8.0B",
    "quantization": "Q4_K_M",
    "trainedContextTokens": 131072
  },
  "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": 1825570816,
        "freeBytes": 11053039616,
        "utilizationPercent": 22
      },
      "gpuAfterLoad": {
        "usedBytes": 7269777408,
        "freeBytes": 5608833024
      },
      "loadDurationNs": 3553175100,
      "ps": {
        "sizeBytes": 5271715839,
        "sizeVramBytes": 5271715839,
        "hostBytes": 0,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 823692000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 54681000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 801637000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13459000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 801758000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13035000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 805363000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13033000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 79.82458547342216,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 1789919232,
        "freeBytes": 11088691200,
        "utilizationPercent": 26
      },
      "gpuAfterLoad": {
        "usedBytes": 7939817472,
        "freeBytes": 4938792960
      },
      "loadDurationNs": 2307724700,
      "ps": {
        "sizeBytes": 5934426357,
        "sizeVramBytes": 5934426357,
        "hostBytes": 0,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 810869000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 45869000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 825985000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13259000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 808429000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13679000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 804268000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13820000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 79.1658884082585,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 1843396608,
        "freeBytes": 11035213824,
        "utilizationPercent": 18
      },
      "gpuAfterLoad": {
        "usedBytes": 9050259456,
        "freeBytes": 3828350976
      },
      "loadDurationNs": 2662966200,
      "ps": {
        "sizeBytes": 7024945397,
        "sizeVramBytes": 7024945397,
        "hostBytes": 0,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 831258000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 50640000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 840370000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13774000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 816769000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13809000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 845890000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 12841000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 76.15693087568572,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 1808793600,
        "freeBytes": 11069816832,
        "utilizationPercent": 16
      },
      "gpuAfterLoad": {
        "usedBytes": 11171528704,
        "freeBytes": 1707081728
      },
      "loadDurationNs": 2320303600,
      "ps": {
        "sizeBytes": 9059172351,
        "sizeVramBytes": 9059172351,
        "hostBytes": 0,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 831009000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 46198000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 843646000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13146000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 838492000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13722000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 833653000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13813000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 76.3275022301942,
        "fullyInVram": true
      }
    }
  ]
}
