{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T07:48:00.409Z",
  "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",
    "digest": "500a1f067a9f782620b40bee6f7b0c89e17ae61f686b92c24933e4ca4b2b8b41",
    "weightsBytes": 5225388164,
    "family": "qwen3",
    "parameterSize": "8.2B",
    "quantization": "Q4_K_M",
    "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": 5961154560,
        "freeBytes": 4776263680
      },
      "loadDurationNs": 8032750767,
      "ps": {
        "sizeBytes": 5578204118,
        "sizeVramBytes": 5578204118,
        "hostBytes": 0,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 593482000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 202207000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 549693000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8910000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 550736000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8913000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 550037000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8886000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 116.35580879104496,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 6678380544,
        "freeBytes": 4059037696
      },
      "loadDurationNs": 2473724289,
      "ps": {
        "sizeBytes": 6295440588,
        "sizeVramBytes": 6295440588,
        "hostBytes": 0,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 571911000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 88268000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 552093000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8838000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 551342000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 9492000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 550110000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8917000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 116.08040018717965,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 7785676800,
        "freeBytes": 2951741440
      },
      "loadDurationNs": 2547952013,
      "ps": {
        "sizeBytes": 7402726358,
        "sizeVramBytes": 7402726358,
        "hostBytes": 0,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 572482000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 104706000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 550841000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8902000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 550051000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8962000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 550237000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8945000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 116.31351581227726,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9295626240,
        "freeBytes": 1441792000
      },
      "loadDurationNs": 3282877459,
      "ps": {
        "sizeBytes": 10362248231,
        "sizeVramBytes": 8914101862,
        "hostBytes": 1448146369,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 1745503000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 166047000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1723678000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 28290000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1730136000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 28114000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 1733025000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 32124000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 36.99131166567253,
        "fullyInVram": false
      }
    }
  ]
}
