{
  "evidenceVersion": "local-ollama-speed-v1",
  "capturedAt": "2026-06-01T21:58:08.575Z",
  "hardwareSlug": "rtx-3080-16gb-mobile",
  "modelTag": "codegemma:7b",
  "catalogSlug": "codegemma-7b",
  "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=codegemma:7b 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": 61,
      "smClockMhz": 210,
      "powerDrawW": 17.74
    },
    "ollamaVersionCli": "ollama version is 0.24.0",
    "ollamaApiVersion": {
      "version": "0.24.0"
    }
  },
  "ollamaTagDetails": {
    "name": "codegemma:7b",
    "model": "codegemma:7b",
    "modified_at": "2026-06-02T00:57:38.3372053+03:00",
    "size": 5011852809,
    "digest": "0c96700aaada572ce9bb6999d1fda9b53e9e6cef5d74fda1e066a1ba811b93f3",
    "details": {
      "parent_model": "",
      "format": "gguf",
      "family": "gemma",
      "families": [
        "gemma"
      ],
      "parameter_size": "9B",
      "quantization_level": "Q4_0"
    }
  },
  "notes": null,
  "logHashSha256": "c5cdb6776f73591dd29f2b4d3615f208644103c6dca8bed337780a5f02295014",
  "runsCaptured": 5,
  "decodeStatsAllRuns": {
    "min": 78.99625628718398,
    "max": 81.3337024273694,
    "median": 80.36630459344288,
    "p5": 79.23113367256342,
    "p95": 81.21574420543564,
    "mean": 80.3221635679556,
    "n": 5
  },
  "decodeStatsSteadyRuns": {
    "min": 78.99625628718398,
    "max": 81.3337024273694,
    "median": 80.55510795557173,
    "p5": 79.20176353312281,
    "p95": 81.24523376091908,
    "mean": 80.36004365642421,
    "n": 4
  },
  "runs": [
    {
      "runIndex": 1,
      "startedAt": "2026-06-01T21:57:42.509Z",
      "finishedAt": "2026-06-01T21:57:50.823Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run1.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run1.log",
      "decodeTokPerSec": 80.17064321408117,
      "prefillTokPerSec": 963.1475662464935,
      "totalMs": 8232.1991,
      "loadMs": 4308.6355,
      "promptEvalMs": 33.2244,
      "evalMs": 3193.1888,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "b66a792bb50054f7442094e05680003366bd2ffadcf02146e71c9118a785471b",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 1881,
        "tempC": 61,
        "smClockMhz": 210,
        "powerDrawW": 17.77
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 67,
        "smClockMhz": 1290,
        "powerDrawW": 114.01
      }
    },
    {
      "runIndex": 2,
      "startedAt": "2026-06-01T21:57:50.898Z",
      "finishedAt": "2026-06-01T21:57:55.209Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run2.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run2.log",
      "decodeTokPerSec": 81.3337024273694,
      "prefillTokPerSec": 1976.016104531252,
      "totalMs": 4220.1715,
      "loadMs": 338.8249,
      "promptEvalMs": 16.1942,
      "evalMs": 3147.5267,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "2e69dca24becc4ec965c528af1582f8662a1cece3c70e4aefadc8ce52207d13b",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 67,
        "smClockMhz": 1290,
        "powerDrawW": 107.45
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 68,
        "smClockMhz": 1245,
        "powerDrawW": 112.62
      }
    },
    {
      "runIndex": 3,
      "startedAt": "2026-06-01T21:57:55.275Z",
      "finishedAt": "2026-06-01T21:57:59.651Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run3.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run3.log",
      "decodeTokPerSec": 80.74391131770058,
      "prefillTokPerSec": 1824.1101762546457,
      "totalMs": 4281.4376,
      "loadMs": 367.6022,
      "promptEvalMs": 17.5428,
      "evalMs": 3170.5177,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "2e69dca24becc4ec965c528af1582f8662a1cece3c70e4aefadc8ce52207d13b",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 68,
        "smClockMhz": 1245,
        "powerDrawW": 106.46
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 69,
        "smClockMhz": 1200,
        "powerDrawW": 114.47
      }
    },
    {
      "runIndex": 4,
      "startedAt": "2026-06-01T21:57:59.735Z",
      "finishedAt": "2026-06-01T21:58:04.152Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run4.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run4.log",
      "decodeTokPerSec": 78.99625628718398,
      "prefillTokPerSec": 1747.383020897609,
      "totalMs": 4337.5437,
      "loadMs": 414.0019,
      "promptEvalMs": 18.3131,
      "evalMs": 3240.6599,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "2e69dca24becc4ec965c528af1582f8662a1cece3c70e4aefadc8ce52207d13b",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 69,
        "smClockMhz": 1200,
        "powerDrawW": 108.95
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 71,
        "smClockMhz": 1275,
        "powerDrawW": 114.5
      }
    },
    {
      "runIndex": 5,
      "startedAt": "2026-06-01T21:58:04.228Z",
      "finishedAt": "2026-06-01T21:58:08.573Z",
      "rawPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run5.ollama.json",
      "logPath": "public/benchmarks/evidence/local-ollama-rtx-3080-16gb-mobile-2026-06-01/codegemma-7b/run5.log",
      "decodeTokPerSec": 80.36630459344288,
      "prefillTokPerSec": 1964.6365422396857,
      "totalMs": 4269.2142,
      "loadMs": 365.2895,
      "promptEvalMs": 16.288,
      "evalMs": 3185.4146,
      "promptTokens": 32,
      "responseTokens": 256,
      "responseHashSha256": "2e69dca24becc4ec965c528af1582f8662a1cece3c70e4aefadc8ce52207d13b",
      "beforeGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 71,
        "smClockMhz": 1275,
        "powerDrawW": 109.87
      },
      "afterGpu": {
        "vendor": "nvidia",
        "name": "NVIDIA GeForce RTX 3080 Laptop GPU",
        "driver": "571.96",
        "memoryTotalMb": 16384,
        "memoryUsedMb": 9192,
        "tempC": 72,
        "smClockMhz": 1155,
        "powerDrawW": 111.42
      }
    }
  ],
  "errors": [],
  "uploadEligible": true
}