Hpt
Home
GitHub
crate.io
Benchmarks
Home
GitHub
crate.io
Benchmarks
  • Benchmarks

    • unary
    • binary
    • reduce
    • conv(f32)
    • conv(f16)
    • pooling
    • normalization
    • matmul
    • fft
    • nn

      • resnet(f32)
      • resnet(f16)
      • lstm

Resnet(f16) Benchmark

::: chartjs Input (batch = 1, in channel = 3)

{
  "type": "line",
  "data": {
    "labels": [64, 96, 128, 160, 192, 224, 256, 288, 320, 352, 384, 416, 448, 480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800, 832, 864, 896, 928, 960, 992, 1024],
    "datasets": [
      {
        "label": "Hpt (v0.1.3)",
        "data": [14.7748, 16.19328, 17.925169, 19.528278, 23.089699, 26.73416, 31.39434, 35.393353, 37.929993, 41.459568, 46.637234, 52.75692, 56.36174, 60.739742, 67.25677, 73.78967, 82.85325, 92.596466, 97.668015, 103.78764, 114.536476, 124.007614, 132.66156, 143.69586, 154.80894, 164.15437, 176.43008, 188.24593, 200.0232, 213.79866, 228.82953],
        "backgroundColor": "rgb(116, 211, 28)",
        "borderColor": "rgb(116, 211, 28)",
        "borderWidth": 1
      },
      {
        "label": "Torch Script",
        "data": [299.4229793548584, 644.977331161499, 1134.8223447799683, 1732.719087600708, 2585.9309434890747, 3453.8393020629883, 7186.537027359009, 6336.400032043457, 7202.898263931274, 8352.205514907837, 11397.365093231201, 11668.107271194458, 13675.772428512573, 15396.769046783447, 39403.24521064758, 19892.587900161743, 22498.664617538452, 25368.01815032959, 30897.54343032837, 31315.685987472534, 36826.29299163818, 43631.5860748291, 76190.76585769653, 56240.63730239868, 63325.66165924072, 59643.19705963135, 67817.62623786926, 69325.19102096558, 74669.29650306702, 81244.25292015076, 179291.4478778839],
        "backgroundColor": "rgb(255, 99, 133)",
        "borderColor": "rgb(255, 99, 132)",
        "borderWidth": 1
      },
      {
        "label": "OnnxRuntime",
        "data": [15.118646621704102, 16.502833366394043, 18.494415283203125, 22.799158096313477, 25.002050399780273, 32.93938636779785, 31.113386154174805, 38.520193099975586, 42.95964241027832, 50.511789321899414, 55.59840202331543, 65.62092304229736, 72.0242977142334, 83.10179710388184, 90.3639554977417, 98.08757305145264, 108.04955959320068, 118.44785213470459, 131.69190883636475, 143.71705055236816, 161.15565299987793, 164.81671333312988, 174.99463558197021, 186.02945804595947, 203.90276908874512, 218.6570167541504, 234.1932773590088, 253.23929786682132, 266.5415287017822, 281.5079927444458, 308.5894823074341],
        "backgroundColor": "rgb(255, 206, 86)",
        "borderColor": "rgb(255, 206, 86)",
        "borderWidth": 1
      }
    ]
  },
  "options": {
    "animation": false,
    "responsive": true,
    "plugins": {
      "legend": {
        "position": "top"
      }
    },
    "scales": {
      "y": {
        "beginAtZero": true,
        "title": {
          "display": true,
          "text": "Time (ms)"
        }
      },
        "x": {
        "beginAtZero": true,
        "title": {
          "display": true,
          "text": "width and height"
        }
      }
    }
  }
}

:::

Compilation config

[profile.release]
opt-level = 3
incremental = true
debug = true
lto = "fat"
codegen-units = 1

Running Threads

10

Device specification

CPU: 12th Gen Intel(R) Core(TM) i5-12600K 3.69 GHz

RAM: G.SKILL Trident Z Royal Series (Intel XMP) DDR4 64GB

System: Windows 11 Pro 23H2

最近更新: 2025/6/24 21:23
Contributors: Jianqoq
Prev
resnet(f32)
Next
lstm