{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T07:52:08.417Z",
  "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": "qwen3:14b",
    "digest": "bdbd181c33f2ed1b31c972991882db3cf4d192569092138a7d29e973cd9debe8",
    "weightsBytes": 9276198565,
    "family": "qwen3",
    "parameterSize": "14.8B",
    "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": 9140436992,
        "freeBytes": 1596981248
      },
      "loadDurationNs": 13264533302,
      "ps": {
        "sizeBytes": 10153550150,
        "sizeVramBytes": 8757675294,
        "hostBytes": 1395874856,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 6105193000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 329303000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 2867018000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 46323000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 2944328000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 47475000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 3017766000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 46190000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 21.736708681913157,
        "fullyInVram": false
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9257877504,
        "freeBytes": 1479540736
      },
      "loadDurationNs": 3739214724,
      "ps": {
        "sizeBytes": 10820517886,
        "sizeVramBytes": 8876499926,
        "hostBytes": 1944017960,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 3802764000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 281206000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 3785603000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 59698000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 3770220000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 60141000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 3891046000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 62562000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 16.906157354587894,
        "fullyInVram": false
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9060745216,
        "freeBytes": 1676673024
      },
      "loadDurationNs": 3754610281,
      "ps": {
        "sizeBytes": 12181412247,
        "sizeVramBytes": 8679315209,
        "hostBytes": 3502097038,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 6052549000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 433011000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 6033225000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 96140000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 6339903000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 96927000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 6040204000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 102747000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 10.59566862311273,
        "fullyInVram": false
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9075425280,
        "freeBytes": 1661992960
      },
      "loadDurationNs": 4234202017,
      "ps": {
        "sizeBytes": 14960421763,
        "sizeVramBytes": 8691845693,
        "hostBytes": 6268576070,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 8773207000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 626475000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 8742663000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 154494000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 8750443000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 143524000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 8684528000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 140122000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 7.320423994382489,
        "fullyInVram": false
      }
    }
  ]
}
