{
  "evidenceVersion": "local-ollama-speed-v1",
  "capturedAt": "2026-06-01T23:18:26.783Z",
  "hardwareSlug": "rtx-3080-16gb-mobile",
  "modelTag": "gemma2:9b",
  "catalogSlug": "gemma-2-9b-it",
  "quant": "Q4_K_M",
  "contextSize": 4096,
  "numPredict": 256,
  "prompt": "Write a detailed explanation of how transformer attention works.",
  "scenario": "single-stream",
  "runtime": "ollama-api",
  "command": "POST http://localhost:11434/api/generate model=gemma2:9b temperature=0 top_p=1 seed=42 num_ctx=4096 num_predict=256",
  "env": {
    "os": "Microsoft Windows [Version 10.0.26200.8457]",
    "platform": "win32",
    "arch": "x64",
    "cpu": "AMD Ryzen 9 5900HX with Radeon Graphics        ",
    "cpuThreads": 16,
    "ramGb": 31.9,
    "gpu": {
      "vendor": "nvidia",
      "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
      "driver": "571.96",
      "memoryTotalMb": 16384,
      "memoryUsedMb": 1881,
      "tempC": 55,
      "smClockMhz": 210,
      "powerDrawW": 17.09
    },
    "ollamaVersionCli": "ollama version is 0.24.0",
    "ollamaApiVersion": {
      "version": "0.24.0"
    }
  },
  "ollamaTagDetails": {
    "name": "gemma2:9b",
    "model": "gemma2:9b",
    "modified_at": "2026-06-02T02:17:51.5387577+03:00",
    "size": 5443152417,
    "digest": "ff02c3702f322b9e075e9568332d96c0a7028002f1a5a056e0a6784320a4db0b",
    "details": {
      "parent_model": "",
      "format": "gguf",
      "family": "gemma2",
      "families": [
        "gemma2"
      ],
      "parameter_size": "9.2B",
      "quantization_level": "Q4_0"
    }
  },
  "notes": null,
  "logHashSha256": "5f40fb7a71f0b4a7a1990a92cf4930173fc486ac127ad6ea3a6f8ec0744f48e3",
  "runsCaptured": 5,
  "decodeStatsAllRuns": {
    "min": 67.8918871555153,
    "max": 69.4380909646044,
    "median": 68.38466029492595,
    "p5": 67.91602225725454,
    "p95": 69.40091975698765,
    "mean": 68.59588720115555,
    "n": 5
  },
  "decodeStatsSteadyRuns": {
    "min": 67.8918871555153,
    "max": 69.25223492652067,
    "median": 68.19861147956871,
    "p5": 67.90998848181974,
    "p95": 69.12209873178146,
    "mean": 68.38533626029334,
    "n": 4
  },
  "runs": [
    {
      "runIndex": 1,
      "startedAt": "2026-06-01T23:17:55.747Z",
      "finishedAt": "2026-06-01T23:18:05.231Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run1.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run1.log",
      "decodeTokPerSec": 69.4380909646044,
      "prefillTokPerSec": 744.7767629649956,
      "totalMs": 9402.044,
      "loadMs": 4504.9167,
      "promptEvalMs": 25.511,
      "evalMs": 3686.7373,
      "promptTokens": 19,
      "responseTokens": 256,
      "responseHashSha256": "088a5606afc6deb52dcf14475ae426e3a6714495f5f0a04806539cea8b24f6dc",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 1881,
        "tempC": 55,
        "smClockMhz": 210,
        "powerDrawW": 17.11
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 62,
        "smClockMhz": 1500,
        "powerDrawW": 114.67
      }
    },
    {
      "runIndex": 2,
      "startedAt": "2026-06-01T23:18:05.311Z",
      "finishedAt": "2026-06-01T23:18:10.604Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run2.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run2.log",
      "decodeTokPerSec": 68.01256266421149,
      "prefillTokPerSec": 891.6545824006157,
      "totalMs": 5201.5814,
      "loadMs": 337.0043,
      "promptEvalMs": 21.3087,
      "evalMs": 3764.0105,
      "promptTokens": 19,
      "responseTokens": 256,
      "responseHashSha256": "5b7069acaacf301e5c6e17772a8f558dfb08bb9c48005e9996e054181e1532fe",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 62,
        "smClockMhz": 1770,
        "powerDrawW": 109.16
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 63,
        "smClockMhz": 1485,
        "powerDrawW": 113.42
      }
    },
    {
      "runIndex": 3,
      "startedAt": "2026-06-01T23:18:10.685Z",
      "finishedAt": "2026-06-01T23:18:16.015Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run3.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run3.log",
      "decodeTokPerSec": 68.38466029492595,
      "prefillTokPerSec": 1046.0135871659638,
      "totalMs": 5240.0768,
      "loadMs": 339.1463,
      "promptEvalMs": 18.1642,
      "evalMs": 3743.5296,
      "promptTokens": 19,
      "responseTokens": 256,
      "responseHashSha256": "5b7069acaacf301e5c6e17772a8f558dfb08bb9c48005e9996e054181e1532fe",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 63,
        "smClockMhz": 1485,
        "powerDrawW": 106.94
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 65,
        "smClockMhz": 1470,
        "powerDrawW": 114.47
      }
    },
    {
      "runIndex": 4,
      "startedAt": "2026-06-01T23:18:16.091Z",
      "finishedAt": "2026-06-01T23:18:21.405Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run4.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run4.log",
      "decodeTokPerSec": 67.8918871555153,
      "prefillTokPerSec": 1099.91895334028,
      "totalMs": 5233.2092,
      "loadMs": 338.6419,
      "promptEvalMs": 17.274,
      "evalMs": 3770.7009,
      "promptTokens": 19,
      "responseTokens": 256,
      "responseHashSha256": "5b7069acaacf301e5c6e17772a8f558dfb08bb9c48005e9996e054181e1532fe",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 63,
        "smClockMhz": 1470,
        "powerDrawW": 108.1
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 66,
        "smClockMhz": 1425,
        "powerDrawW": 114.57
      }
    },
    {
      "runIndex": 5,
      "startedAt": "2026-06-01T23:18:21.483Z",
      "finishedAt": "2026-06-01T23:18:26.782Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run5.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/gemma2-9b/run5.log",
      "decodeTokPerSec": 69.25223492652067,
      "prefillTokPerSec": 889.7797092761876,
      "totalMs": 5219.0817,
      "loadMs": 350.501,
      "promptEvalMs": 21.3536,
      "evalMs": 3696.6316,
      "promptTokens": 19,
      "responseTokens": 256,
      "responseHashSha256": "5b7069acaacf301e5c6e17772a8f558dfb08bb9c48005e9996e054181e1532fe",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 66,
        "smClockMhz": 1425,
        "powerDrawW": 108.1
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9164,
        "tempC": 68,
        "smClockMhz": 1665,
        "powerDrawW": 113.16
      }
    }
  ],
  "errors": [],
  "uploadEligible": true
}