{
  "evidenceVersion": "local-ollama-speed-v1",
  "capturedAt": "2026-06-02T01:27:40.117Z",
  "hardwareSlug": "rtx-3080-16gb-mobile",
  "modelTag": "phi4-reasoning:14b",
  "catalogSlug": "phi-4-reasoning-14b",
  "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=phi4-reasoning:14b 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": 1798,
      "tempC": 55,
      "smClockMhz": 210,
      "powerDrawW": 14.22
    },
    "ollamaVersionCli": "ollama version is 0.24.0",
    "ollamaApiVersion": {
      "version": "0.24.0"
    }
  },
  "ollamaTagDetails": {
    "name": "phi4-reasoning:14b",
    "model": "phi4-reasoning:14b",
    "modified_at": "2026-06-02T04:26:49.6436845+03:00",
    "size": 11117503233,
    "digest": "47e2630ccbcd361abcc9f90d8e8a5ada92583e4d5f9359192130372b42132315",
    "details": {
      "parent_model": "",
      "format": "gguf",
      "family": "phi3",
      "families": [
        "phi3"
      ],
      "parameter_size": "14.7B",
      "quantization_level": "Q4_K_M"
    }
  },
  "notes": null,
  "logHashSha256": "9d3324def3dfa9bd9d53369a0370ce780f0cda0814c9763fb1b14084211855d9",
  "runsCaptured": 5,
  "decodeStatsAllRuns": {
    "min": 39.51834424559342,
    "max": 41.02482052923044,
    "median": 40.45755279690886,
    "p5": 39.698882263623204,
    "p95": 40.93058009430802,
    "mean": 40.39507405241868,
    "n": 5
  },
  "decodeStatsSteadyRuns": {
    "min": 39.51834424559342,
    "max": 40.55361835461836,
    "median": 40.4392935663256,
    "p5": 39.65374775911576,
    "p95": 40.539208520961935,
    "mean": 40.237637433215745,
    "n": 4
  },
  "runs": [
    {
      "runIndex": 1,
      "startedAt": "2026-06-02T01:26:53.990Z",
      "finishedAt": "2026-06-02T01:27:07.101Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run1.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run1.log",
      "decodeTokPerSec": 41.02482052923044,
      "prefillTokPerSec": 1307.9339380329966,
      "totalMs": 13024.9653,
      "loadMs": 5095.802,
      "promptEvalMs": 185.7892,
      "evalMs": 6240.1248,
      "promptTokens": 243,
      "responseTokens": 256,
      "responseHashSha256": "569631803e0009572251c36aa1d97701f8f9c70c99171c9d611ffed75815365d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 1798,
        "tempC": 55,
        "smClockMhz": 210,
        "powerDrawW": 14.16
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 63,
        "smClockMhz": 1095,
        "powerDrawW": 114.54
      }
    },
    {
      "runIndex": 2,
      "startedAt": "2026-06-02T01:27:07.192Z",
      "finishedAt": "2026-06-02T01:27:15.288Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run2.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run2.log",
      "decodeTokPerSec": 40.421034335742355,
      "prefillTokPerSec": 12266.222457787535,
      "totalMs": 8008.7669,
      "loadMs": 190.0208,
      "promptEvalMs": 19.8105,
      "evalMs": 6333.3362,
      "promptTokens": 243,
      "responseTokens": 256,
      "responseHashSha256": "569631803e0009572251c36aa1d97701f8f9c70c99171c9d611ffed75815365d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 63,
        "smClockMhz": 1095,
        "powerDrawW": 107.52
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 65,
        "smClockMhz": 1035,
        "powerDrawW": 114.35
      }
    },
    {
      "runIndex": 3,
      "startedAt": "2026-06-02T01:27:15.365Z",
      "finishedAt": "2026-06-02T01:27:23.523Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run3.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run3.log",
      "decodeTokPerSec": 39.51834424559342,
      "prefillTokPerSec": 16243.315508021391,
      "totalMs": 8062.4281,
      "loadMs": 210.0418,
      "promptEvalMs": 14.96,
      "evalMs": 6478.0042,
      "promptTokens": 243,
      "responseTokens": 256,
      "responseHashSha256": "569631803e0009572251c36aa1d97701f8f9c70c99171c9d611ffed75815365d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 65,
        "smClockMhz": 1035,
        "powerDrawW": 107.1
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 67,
        "smClockMhz": 1065,
        "powerDrawW": 114.54
      }
    },
    {
      "runIndex": 4,
      "startedAt": "2026-06-02T01:27:23.599Z",
      "finishedAt": "2026-06-02T01:27:31.786Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run4.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run4.log",
      "decodeTokPerSec": 40.45755279690886,
      "prefillTokPerSec": 8422.263890670003,
      "totalMs": 8104.6019,
      "loadMs": 198.3249,
      "promptEvalMs": 28.8521,
      "evalMs": 6327.6195,
      "promptTokens": 243,
      "responseTokens": 256,
      "responseHashSha256": "569631803e0009572251c36aa1d97701f8f9c70c99171c9d611ffed75815365d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 66,
        "smClockMhz": 1065,
        "powerDrawW": 109.1
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 69,
        "smClockMhz": 1275,
        "powerDrawW": 114.64
      }
    },
    {
      "runIndex": 5,
      "startedAt": "2026-06-02T01:27:31.864Z",
      "finishedAt": "2026-06-02T01:27:40.116Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run5.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-02/phi4-reasoning-14b/run5.log",
      "decodeTokPerSec": 40.55361835461836,
      "prefillTokPerSec": 14292.099938832163,
      "totalMs": 8156.3115,
      "loadMs": 210.0012,
      "promptEvalMs": 17.0024,
      "evalMs": 6312.6303,
      "promptTokens": 243,
      "responseTokens": 256,
      "responseHashSha256": "569631803e0009572251c36aa1d97701f8f9c70c99171c9d611ffed75815365d",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 69,
        "smClockMhz": 1275,
        "powerDrawW": 107.67
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 13435,
        "tempC": 71,
        "smClockMhz": 1065,
        "powerDrawW": 114.47
      }
    }
  ],
  "errors": [],
  "uploadEligible": true
}