{
  "protocolVersion": "osai-ctxfit/1",
  "createdAt": "2026-10-20T07:47:08.658Z",
  "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": "qwen3:4b",
    "digest": "359d7dd4bcdab3d86b87d73ac27966f4dbb9f5efdfcc75d34a8764a09474fae7",
    "weightsBytes": 2497293931,
    "family": "qwen3",
    "parameterSize": "4.0B",
    "quantization": "Q4_K_M",
    "trainedContextTokens": 262144
  },
  "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": 3547332608,
        "freeBytes": 7190085632
      },
      "loadDurationNs": 4803678216,
      "ps": {
        "sizeBytes": 3178149969,
        "sizeVramBytes": 3178149969,
        "hostBytes": 0,
        "contextLength": 4096
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 361425000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 74920000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 359087000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5902000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 358577000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5873000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 360084000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5910000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 178.22978832427796,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 8192,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 4243587072,
        "freeBytes": 6493831168
      },
      "loadDurationNs": 1978700839,
      "ps": {
        "sizeBytes": 3873366343,
        "sizeVramBytes": 3873366343,
        "hostBytes": 0,
        "contextLength": 8192
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 360277000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 52942000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 358522000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5879000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 358604000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5982000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 359284000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5920000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 178.46984417351732,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 16384,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 5468323840,
        "freeBytes": 5269094400
      },
      "loadDurationNs": 1941236018,
      "ps": {
        "sizeBytes": 5098103111,
        "sizeVramBytes": 5098103111,
        "hostBytes": 0,
        "contextLength": 16384
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 361938000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 45451000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 357941000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 6004000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 358324000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5893000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 357873000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5952000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 178.80041682847173,
        "fullyInVram": true
      }
    },
    {
      "numCtx": 32768,
      "gpuBefore": {
        "usedBytes": 93323264,
        "freeBytes": 10644094976,
        "utilizationPercent": 0
      },
      "gpuAfterLoad": {
        "usedBytes": 7917797376,
        "freeBytes": 2819620864
      },
      "loadDurationNs": 1969118815,
      "ps": {
        "sizeBytes": 7547576647,
        "sizeVramBytes": 7547576647,
        "hostBytes": 0,
        "contextLength": 32768
      },
      "passes": [
        {
          "warmup": true,
          "evalCount": 64,
          "evalDurationNs": 360830000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 44735000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 357867000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5906000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 359294000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 6029000
        },
        {
          "warmup": false,
          "evalCount": 64,
          "evalDurationNs": 358102000,
          "promptEvalCount": 35,
          "promptEvalDurationNs": 5944000
        }
      ],
      "derived": {
        "measuredPasses": 3,
        "generationTokensPerSecond": 178.72002948880487,
        "fullyInVram": true
      }
    }
  ]
}
