{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T07:51:06.881Z",
  "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": "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": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 5640290304,
        "freeBytes": 5097127936
      },
      "loadDurationNs": 7566965818,
      "ps": {
        "sizeBytes": 5271715839,
        "sizeVramBytes": 5271715839,
        "hostBytes": 0,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 542022000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 128603000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 540575000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8645000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 540302000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8763000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 539752000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8880000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 118.45227298806965,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 6302990336,
        "freeBytes": 4434427904
      },
      "loadDurationNs": 2521447220,
      "ps": {
        "sizeBytes": 5934426357,
        "sizeVramBytes": 5934426357,
        "hostBytes": 0,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 543908000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 49540000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 540816000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8648000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 539181000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8653000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 541633000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8705000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 118.33969409189078,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 7393509376,
        "freeBytes": 3343908864
      },
      "loadDurationNs": 2513155278,
      "ps": {
        "sizeBytes": 7024945397,
        "sizeVramBytes": 7024945397,
        "hostBytes": 0,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 542000000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 47618000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 538937000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8610000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 540641000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8673000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 540089000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 8662000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 118.49898812973416,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 9201254400,
        "freeBytes": 1536163840
      },
      "loadDurationNs": 3303483680,
      "ps": {
        "sizeBytes": 9451654347,
        "sizeVramBytes": 8832145162,
        "hostBytes": 619509185,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 860264000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 79038000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 859035000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 13822000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 858766000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14264000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 854317000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 14205000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 74.52554013549674,
        "fullyInVram": false
      }
    }
  ]
}
