{
 "machine": {
  "safec": "safec 1.0.0",
  "clang": "Homebrew clang version 22.1.8",
  "rustc": "rustc 1.98.1 (48a229cea 2026-09-01)",
  "go": "go version go1.26.5 darwin/arm64",
  "zig": "0.16.0",
  "python": "Python 3.14.6",
  "torch": "2.14.1",
  "tensorflow": "2.21.0",
  "numpy": "2.5.4",
  "mlx": "0.32.4",
  "cpu": "Apple M1 Pro",
  "os": "macOS 26.6.2",
  "cores": 10
 },
 "attention": {
  "safec_mps": {
   "flash fp16 d64 s1024 c0": {
    "ms_per_call": 4.9261,
    "tflops": 3.4
   },
   "flash fp16 d64 s1024 c1": {
    "ms_per_call": 2.808,
    "tflops": 2.99
   },
   "flash fp16 d64 s4096 c0": {
    "ms_per_call": 77.9922,
    "tflops": 3.42
   },
   "flash fp16 d64 s4096 c1": {
    "ms_per_call": 39.9959,
    "tflops": 3.22
   },
   "flash fp16 d128 s1024 c0": {
    "ms_per_call": 10.958,
    "tflops": 3.07
   },
   "flash fp16 d128 s1024 c1": {
    "ms_per_call": 6.1592,
    "tflops": 2.72
   },
   "flash fp16 d128 s4096 c0": {
    "ms_per_call": 171.4948,
    "tflops": 3.17
   },
   "flash fp16 d128 s4096 c1": {
    "ms_per_call": 89.7568,
    "tflops": 2.97
   },
   "flash bf16 d64 s1024 c0": {
    "ms_per_call": 6.0435,
    "tflops": 2.77
   },
   "flash bf16 d64 s1024 c1": {
    "ms_per_call": 3.2488,
    "tflops": 2.55
   },
   "flash bf16 d64 s4096 c0": {
    "ms_per_call": 95.7106,
    "tflops": 2.8
   },
   "flash bf16 d64 s4096 c1": {
    "ms_per_call": 49.0912,
    "tflops": 2.76
   },
   "flash bf16 d128 s1024 c0": {
    "ms_per_call": 12.1084,
    "tflops": 2.82
   },
   "flash bf16 d128 s1024 c1": {
    "ms_per_call": 6.7619,
    "tflops": 2.52
   },
   "flash bf16 d128 s4096 c0": {
    "ms_per_call": 189.1073,
    "tflops": 2.9
   },
   "flash bf16 d128 s4096 c1": {
    "ms_per_call": 98.2186,
    "tflops": 2.8
   },
   "flash fp32 d64 s1024 c0": {
    "ms_per_call": 6.052,
    "tflops": 2.82
   },
   "flash fp32 d64 s1024 c1": {
    "ms_per_call": 3.2054,
    "tflops": 2.68
   },
   "flash fp32 d64 s4096 c0": {
    "ms_per_call": 93.7544,
    "tflops": 2.93
   },
   "flash fp32 d64 s4096 c1": {
    "ms_per_call": 47.8302,
    "tflops": 2.87
   },
   "flash fp32 d128 s1024 c0": {
    "ms_per_call": 12.1548,
    "tflops": 2.83
   },
   "flash fp32 d128 s1024 c1": {
    "ms_per_call": 6.7408,
    "tflops": 2.55
   },
   "flash fp32 d128 s4096 c0": {
    "ms_per_call": 187.0942,
    "tflops": 2.94
   },
   "flash fp32 d128 s4096 c1": {
    "ms_per_call": 97.8252,
    "tflops": 2.81
   },
   "flash bf16 gqa32/8 d128 s4096 c1": {
    "ms_per_call": 194.9229,
    "tflops": 2.82
   },
   "linear fp32 d64 s4096 c0": {
    "ms_per_call": 3.2679
   },
   "linear fp32 d64 s4096 c1": {
    "ms_per_call": 6.9853
   },
   "linear fp32 d64 s16384 c0": {
    "ms_per_call": 12.2808
   },
   "linear fp32 d64 s16384 c1": {
    "ms_per_call": 28.0978
   },
   "linear fp32 d128 s4096 c0": {
    "ms_per_call": 10.2875
   },
   "linear fp32 d128 s4096 c1": {
    "ms_per_call": 21.974
   },
   "linear bf16 d64 s4096 c0": {
    "ms_per_call": 3.1461
   },
   "linear bf16 d64 s4096 c1": {
    "ms_per_call": 6.6124
   },
   "linear bf16 d64 s16384 c0": {
    "ms_per_call": 11.8624
   },
   "linear bf16 d64 s16384 c1": {
    "ms_per_call": 26.5692
   },
   "linear bf16 d128 s4096 c0": {
    "ms_per_call": 9.92
   },
   "linear bf16 d128 s4096 c1": {
    "ms_per_call": 21.1353
   }
  },
  "pytorch_mps": {
   "flash fp16 d64 s1024 c0": {
    "ms_per_call": 4.8762,
    "tflops": 3.42
   },
   "flash fp16 d64 s1024 c1": {
    "ms_per_call": 2.7214,
    "tflops": 3.09
   },
   "flash fp16 d64 s4096 c0": {
    "ms_per_call": 76.2543,
    "tflops": 3.51
   },
   "flash fp16 d64 s4096 c1": {
    "ms_per_call": 40.0888,
    "tflops": 3.38
   },
   "flash fp16 d128 s1024 c0": {
    "ms_per_call": 9.66,
    "tflops": 3.54
   },
   "flash fp16 d128 s1024 c1": {
    "ms_per_call": 5.7901,
    "tflops": 2.96
   },
   "flash fp16 d128 s4096 c0": {
    "ms_per_call": 150.8943,
    "tflops": 3.64
   },
   "flash fp16 d128 s4096 c1": {
    "ms_per_call": 84.3793,
    "tflops": 3.23
   },
   "flash bf16 d64 s1024 c0": {
    "ms_per_call": 5.6872,
    "tflops": 2.96
   },
   "flash bf16 d64 s1024 c1": {
    "ms_per_call": 3.1099,
    "tflops": 2.64
   },
   "flash bf16 d64 s4096 c0": {
    "ms_per_call": 88.974,
    "tflops": 3.05
   },
   "flash bf16 d64 s4096 c1": {
    "ms_per_call": 45.83,
    "tflops": 2.96
   },
   "flash bf16 d128 s1024 c0": {
    "ms_per_call": 13.1964,
    "tflops": 2.55
   },
   "flash bf16 d128 s1024 c1": {
    "ms_per_call": 7.219,
    "tflops": 2.35
   },
   "flash bf16 d128 s4096 c0": {
    "ms_per_call": 200.924,
    "tflops": 2.73
   },
   "flash bf16 d128 s4096 c1": {
    "ms_per_call": 106.3428,
    "tflops": 2.53
   },
   "flash fp32 d64 s1024 c0": {
    "ms_per_call": 6.8666,
    "tflops": 2.43
   },
   "flash fp32 d64 s1024 c1": {
    "ms_per_call": 3.8327,
    "tflops": 2.22
   },
   "flash fp32 d64 s4096 c0": {
    "ms_per_call": 107.438,
    "tflops": 2.52
   },
   "flash fp32 d64 s4096 c1": {
    "ms_per_call": 56.146,
    "tflops": 2.42
   },
   "flash fp32 d128 s1024 c0": {
    "ms_per_call": 19.97,
    "tflops": 1.68
   },
   "flash fp32 d128 s1024 c1": {
    "ms_per_call": 10.9616,
    "tflops": 1.53
   },
   "flash fp32 d128 s4096 c0": {
    "ms_per_call": 317.2731,
    "tflops": 1.71
   },
   "flash fp32 d128 s4096 c1": {
    "ms_per_call": 163.358,
    "tflops": 1.65
   },
   "flash bf16 gqa32/8 d128 s4096 c1": {
    "ms_per_call": 213.7793,
    "tflops": 2.57
   },
   "linear fp32 d64 s4096 c0": {
    "ms_per_call": 10.6052
   },
   "linear fp32 d64 s4096 c1": {
    "ms_per_call": 28.3837
   },
   "linear fp32 d64 s16384 c0": {
    "ms_per_call": 41.8208
   },
   "linear fp32 d64 s16384 c1": {
    "ms_per_call": 113.0029
   },
   "linear fp32 d128 s4096 c0": {
    "ms_per_call": 21.8711
   },
   "linear fp32 d128 s4096 c1": {
    "ms_per_call": 57.9566
   },
   "linear bf16 d64 s4096 c0": {
    "ms_per_call": 12.1132
   },
   "linear bf16 d64 s4096 c1": {
    "ms_per_call": 29.2995
   },
   "linear bf16 d64 s16384 c0": {
    "ms_per_call": 47.5055
   },
   "linear bf16 d64 s16384 c1": {
    "ms_per_call": 117.0589
   },
   "linear bf16 d128 s4096 c0": {
    "ms_per_call": 25.1942
   },
   "linear bf16 d128 s4096 c1": {
    "ms_per_call": 60.3833
   }
  },
  "mlx": {
   "flash fp16 d64 s1024 c0": {
    "ms_per_call": 4.5917,
    "tflops": 3.74
   },
   "flash fp16 d64 s1024 c1": {
    "ms_per_call": 2.6474,
    "tflops": 3.24
   },
   "flash fp16 d64 s4096 c0": {
    "ms_per_call": 72.5699,
    "tflops": 3.79
   },
   "flash fp16 d64 s4096 c1": {
    "ms_per_call": 40.7321,
    "tflops": 3.31
   },
   "flash fp16 d128 s1024 c0": {
    "ms_per_call": 9.4037,
    "tflops": 3.63
   },
   "flash fp16 d128 s1024 c1": {
    "ms_per_call": 4.9514,
    "tflops": 3.47
   },
   "flash fp16 d128 s4096 c0": {
    "ms_per_call": 147.9744,
    "tflops": 3.72
   },
   "flash fp16 d128 s4096 c1": {
    "ms_per_call": 75.0561,
    "tflops": 3.66
   },
   "flash bf16 d64 s1024 c0": {
    "ms_per_call": 5.245,
    "tflops": 3.23
   },
   "flash bf16 d64 s1024 c1": {
    "ms_per_call": 2.7955,
    "tflops": 3.07
   },
   "flash bf16 d64 s4096 c0": {
    "ms_per_call": 82.4814,
    "tflops": 3.32
   },
   "flash bf16 d64 s4096 c1": {
    "ms_per_call": 42.5296,
    "tflops": 3.17
   },
   "flash bf16 d128 s1024 c0": {
    "ms_per_call": 12.0523,
    "tflops": 2.85
   },
   "flash bf16 d128 s1024 c1": {
    "ms_per_call": 6.5397,
    "tflops": 2.63
   },
   "flash bf16 d128 s4096 c0": {
    "ms_per_call": 189.5153,
    "tflops": 2.9
   },
   "flash bf16 d128 s4096 c1": {
    "ms_per_call": 96.6074,
    "tflops": 2.85
   },
   "flash fp32 d64 s1024 c0": {
    "ms_per_call": 6.1152,
    "tflops": 2.81
   },
   "flash fp32 d64 s1024 c1": {
    "ms_per_call": 3.4777,
    "tflops": 2.47
   },
   "flash fp32 d64 s4096 c0": {
    "ms_per_call": 97.4932,
    "tflops": 2.81
   },
   "flash fp32 d64 s4096 c1": {
    "ms_per_call": 51.9654,
    "tflops": 2.64
   },
   "flash fp32 d128 s1024 c0": {
    "ms_per_call": 16.3367,
    "tflops": 2.1
   },
   "flash fp32 d128 s1024 c1": {
    "ms_per_call": 9.2241,
    "tflops": 1.85
   },
   "flash fp32 d128 s4096 c0": {
    "ms_per_call": 264.8074,
    "tflops": 2.07
   },
   "flash fp32 d128 s4096 c1": {
    "ms_per_call": 124.825,
    "tflops": 1.93
   },
   "flash bf16 gqa32/8 d128 s4096 c1": {
    "ms_per_call": 196.9595,
    "tflops": 2.76
   },
   "linear fp32 d64 s4096 c0": {
    "ms_per_call": 10.9503
   },
   "linear fp32 d64 s4096 c1": {
    "ms_per_call": 21.1258
   },
   "linear fp32 d64 s16384 c0": {
    "ms_per_call": 40.7904
   },
   "linear fp32 d64 s16384 c1": {
    "ms_per_call": 81.0334
   },
   "linear fp32 d128 s4096 c0": {
    "ms_per_call": 22.9496
   },
   "linear fp32 d128 s4096 c1": {
    "ms_per_call": 43.4835
   },
   "linear bf16 d64 s4096 c0": {
    "ms_per_call": 12.5251
   },
   "linear bf16 d64 s4096 c1": {
    "ms_per_call": 22.6182
   },
   "linear bf16 d64 s16384 c0": {
    "ms_per_call": 48.9564
   },
   "linear bf16 d64 s16384 c1": {
    "ms_per_call": 87.5516
   },
   "linear bf16 d128 s4096 c0": {
    "ms_per_call": 26.3101
   },
   "linear bf16 d128 s4096 c1": {
    "ms_per_call": 45.6706
   }
  }
 },
 "fib": {
  "safec": {
   "debug": 0.1387,
   "release": 0.0825,
   "rss_mb": 5.6,
   "compile_s": 0.16
  },
  "c": {
   "debug": 0.1379,
   "release": 0.082,
   "rss_mb": 1.3,
   "compile_s": 0.131
  },
  "cpp": {
   "debug": 0.1395,
   "release": 0.0876,
   "rss_mb": 1.3,
   "compile_s": 0.13
  },
  "rust": {
   "debug": 0.1937,
   "release": 0.0805,
   "rss_mb": 1.5,
   "compile_s": 0.141
  },
  "zig": {
   "debug": 0.1638,
   "release": 0.0808,
   "rss_mb": 1.5,
   "compile_s": 5.396
  },
  "go": {
   "debug": 0.1143,
   "release": 0.0894,
   "rss_mb": 3.9,
   "compile_s": 2.287
  },
  "python": {
   "debug": 2.6994,
   "release": 2.7146,
   "rss_mb": 14.5,
   "compile_s": null
  }
 },
 "nbody": {
  "safec": {
   "debug": 0.6486,
   "release": 0.1191,
   "rss_mb": 5.6,
   "compile_s": 0.173
  },
  "c": {
   "debug": 0.4258,
   "release": 0.112,
   "rss_mb": 1.3,
   "compile_s": 0.173
  },
  "cpp": {
   "debug": 0.4255,
   "release": 0.1122,
   "rss_mb": 1.3,
   "compile_s": 0.165
  },
  "rust": {
   "debug": 0.6455,
   "release": 0.1183,
   "rss_mb": 1.5,
   "compile_s": 0.146
  },
  "zig": {
   "debug": 0.563,
   "release": 0.1165,
   "rss_mb": 1.5,
   "compile_s": 5.622
  },
  "go": {
   "debug": 0.2743,
   "release": 0.103,
   "rss_mb": 3.9,
   "compile_s": 0.142
  },
  "python": {
   "debug": 9.8941,
   "release": 9.8822,
   "rss_mb": 15.2,
   "compile_s": null
  }
 },
 "binarytrees": {
  "safec": {
   "debug": 0.3594,
   "release": 0.2219,
   "rss_mb": 29.6,
   "compile_s": 0.173
  },
  "c": {
   "debug": 1.6562,
   "release": 1.5619,
   "rss_mb": 17.5,
   "compile_s": 0.142
  },
  "cpp": {
   "debug": 1.9176,
   "release": 1.7201,
   "rss_mb": 17.5,
   "compile_s": 0.141
  },
  "rust": {
   "debug": 3.1155,
   "release": 1.7442,
   "rss_mb": 17.6,
   "compile_s": 0.124
  },
  "zig": {
   "debug": 4.9238,
   "release": 1.5491,
   "rss_mb": 17.6,
   "compile_s": 5.415
  },
  "go": {
   "debug": 2.5452,
   "release": 1.2738,
   "rss_mb": 39.1,
   "compile_s": 0.138
  },
  "python": {
   "debug": 22.5971,
   "release": 23.2538,
   "rss_mb": 87.1,
   "compile_s": null
  },
  "safec_heap": {
   "debug": 1.2336,
   "release": 0.8297,
   "rss_mb": 50.3,
   "compile_s": 0.166
  }
 },
 "binarytrees_mt": {
  "safec": {
   "release": 0.0638,
   "rss_mb": 93.9,
   "compile_s": 0.194
  },
  "c": {
   "release": 0.592,
   "rss_mb": 69.7,
   "compile_s": 0.146
  },
  "cpp": {
   "release": 0.6127,
   "rss_mb": 71.6,
   "compile_s": 0.605
  },
  "rust": {
   "release": 0.6834,
   "rss_mb": 78.3,
   "compile_s": 0.29
  },
  "zig": {
   "release": 0.5845,
   "rss_mb": 68.3,
   "compile_s": 5.675
  },
  "go": {
   "release": 0.4956,
   "rss_mb": 141.8,
   "compile_s": 0.149
  },
  "python": {
   "release": 25.3591,
   "rss_mb": 355.7,
   "compile_s": null
  },
  "safec_heap": {
   "release": 0.1602,
   "rss_mb": 203.6,
   "compile_s": 0.178
  }
 },
 "simd": {
  "safec": {
   "scalar": 0.0451,
   "vec": 0.0308
  },
  "c": {
   "scalar": 0.0438,
   "vec": 0.031
  },
  "cpp": {
   "scalar": 0.0435,
   "vec": 0.0315
  },
  "rust": {
   "scalar": 0.046,
   "vec": 0.0408
  },
  "zig": {
   "scalar": 0.049,
   "vec": 0.0328
  },
  "go": {
   "scalar": 0.0617
  },
  "python": {
   "scalar": 3.6263,
   "vec": 0.1518
  }
 },
 "collections": {
  "safec": {
   "vec push": 0.966,
   "vec sort": 56.383,
   "map insert": 31.145,
   "map get": 16.612,
   "list push+pop": 4.866,
   "queue enq+deq": 1.661,
   "bst insert": 352.01,
   "bst get": 278.485
  },
  "c": {
   "vec push": 0.79,
   "vec sort": 88.764,
   "map insert": 28.802,
   "map get": 12.874,
   "list push+pop": 26.092,
   "queue enq+deq": 1.937,
   "bst insert": 326.815,
   "bst get": 396.735
  }
 },
 "memalloc": {
  "std_ns_per_pair": 12.22,
  "malloc_ns_per_pair": 27.6,
  "std_ms": 114.66,
  "malloc_ms": 232.9
 },
 "web": {
  "safec": {
   "rps": 32925.12,
   "p50_ms": 1,
   "p99_ms": 3,
   "failed": 0,
   "rss_mb": 6.6
  },
  "c": {
   "rps": 26111.06,
   "p50_ms": 2,
   "p99_ms": 3,
   "failed": 0,
   "rss_mb": 2.1
  },
  "cpp": {
   "rps": 26180.61,
   "p50_ms": 2,
   "p99_ms": 3,
   "failed": 0,
   "rss_mb": 2.2
  },
  "zig": {
   "rps": 24750.55,
   "p50_ms": 2,
   "p99_ms": 3,
   "failed": 0,
   "rss_mb": 2.3
  },
  "go": {
   "rps": 26423.6,
   "p50_ms": 2,
   "p99_ms": 3,
   "failed": 0,
   "rss_mb": 19.8
  },
  "rust": {
   "rps": 29106.36,
   "p50_ms": 2,
   "p99_ms": 2,
   "failed": 0,
   "rss_mb": 3.5
  },
  "python": {
   "rps": 4812.8,
   "p50_ms": 10,
   "p99_ms": 18,
   "failed": 0,
   "rss_mb": 56.8
  }
 },
 "dsp": {
  "safec_cpu": {
   "fir stream n=1048576 taps=127": {
    "ms_per_call": 5.819,
    "checksum": 965.503541
   },
   "fir block n=1048576 taps=127": {
    "ms_per_call": 0.926,
    "checksum": 965.503541
   },
   "fftconv n=1048576 taps=4096": {
    "ms_per_call": 3.065,
    "checksum": 11438.877046
   },
   "biquad n=1048576": {
    "ms_per_call": 0.666,
    "checksum": -646.722217
   },
   "stft n=1048576 frame=1024 hop=256": {
    "ms_per_call": 3.794,
    "frames": 4093.0
   },
   "resample n=1048576 44100->48000": {
    "ms_per_call": 4.132,
    "checksum": -704.289165
   },
   "conv2d 1024x1024 k=5x5": {
    "ms_per_call": 0.589,
    "checksum": 205.652439
   },
   "fir stream n=1048576 taps=127 f32": {
    "ms_per_call": 3.45,
    "checksum": 965.503683
   },
   "fir block n=1048576 taps=127 f32": {
    "ms_per_call": 0.523,
    "checksum": 965.503384
   },
   "fftconv n=1048576 taps=4096 f32": {
    "ms_per_call": 1.973,
    "checksum": 11438.881763
   },
   "biquad n=1048576 f32": {
    "ms_per_call": 0.661,
    "checksum": -646.722416
   },
   "stft n=1048576 frame=1024 hop=256 f32": {
    "ms_per_call": 2.202,
    "frames": 4093.0
   },
   "resample n=1048576 44100->48000 f32": {
    "ms_per_call": 4.774,
    "checksum": -704.289136
   }
  },
  "scipy": {
   "fir stream n=1048576 taps=127": {
    "ms_per_call": 22.378
   },
   "fir block n=1048576 taps=127": {
    "ms_per_call": 21.385
   },
   "fftconv n=1048576 taps=4096": {
    "ms_per_call": 22.666
   },
   "biquad n=1048576": {
    "ms_per_call": 7.967
   },
   "stft n=1048576 frame=1024 hop=256": {
    "ms_per_call": 57.741
   },
   "resample n=1048576 44100->48000": {
    "ms_per_call": 14.901
   },
   "conv2d 1024x1024 k=5x5": {
    "ms_per_call": 36.134
   },
   "fir stream n=1048576 taps=127 f32": {
    "ms_per_call": 15.928
   },
   "fir block n=1048576 taps=127 f32": {
    "ms_per_call": 15.08
   },
   "fftconv n=1048576 taps=4096 f32": {
    "ms_per_call": 16.83
   },
   "biquad n=1048576 f32": {
    "ms_per_call": 7.845
   },
   "stft n=1048576 frame=1024 hop=256 f32": {
    "ms_per_call": 56.016
   },
   "resample n=1048576 44100->48000 f32": {
    "ms_per_call": 16.837
   }
  },
  "kfr": {
   "fft1d n=1048576 c128 cpu": {
    "ms_per_call": 6.278
   },
   "fft2d 1024x1024 c128 cpu": {
    "ms_per_call": 8.755
   },
   "batch 1024x4096 c2c cpu": {
    "ms_per_call": 12.254
   },
   "rfft 1024x4096 r2c cpu": {
    "ms_per_call": 7.076
   },
   "fir stream n=1048576 taps=127": {
    "ms_per_call": 44.767
   },
   "fftconv n=1048576 taps=4096": {
    "ms_per_call": 8.065
   },
   "biquad n=1048576": {
    "ms_per_call": 5.078
   },
   "resample n=1048576 44100->48000": {
    "ms_per_call": 162.172
   },
   "resample draft n=1048576 44100->48000": {
    "ms_per_call": 32.873
   },
   "fft1d n=1048576 c64 cpu": {
    "ms_per_call": 3.198
   },
   "fft2d 1024x1024 c64 cpu": {
    "ms_per_call": 3.112
   },
   "batch 1024x4096 c2c f32 cpu": {
    "ms_per_call": 6.05
   },
   "rfft 1024x4096 r2c f32 cpu": {
    "ms_per_call": 3.815
   },
   "fir stream n=1048576 taps=127 f32": {
    "ms_per_call": 63.467
   },
   "fftconv n=1048576 taps=4096 f32": {
    "ms_per_call": 4.054
   },
   "biquad n=1048576 f32": {
    "ms_per_call": 4.189
   },
   "resample n=1048576 44100->48000 f32": {
    "ms_per_call": 90.002
   },
   "resample draft n=1048576 44100->48000 f32": {
    "ms_per_call": 34.01
   }
  },
  "q_madronalib": {
   "biquad q n=1048576 f32": {
    "ms_per_call": 2.619,
    "checksum": 996.364585
   },
   "biquad madronalib n=1048576 f32": {
    "ms_per_call": 4.379,
    "checksum": 996.365938
   }
  },
  "rust": {
   "fft1d n=1048576 c128 cpu": {
    "ms_per_call": 8.593
   },
   "batch 1024x4096 c2c cpu": {
    "ms_per_call": 16.165
   },
   "fft1d n=1048576 c64 cpu": {
    "ms_per_call": 4.231
   },
   "batch 1024x4096 c2c f32 cpu": {
    "ms_per_call": 8.999
   },
   "resample n=1048576 44100->48000": {
    "ms_per_call": 291.417
   }
  }
 },
 "ml": {
  "small/safec_cpu": {
   ":train_ms": 10.509,
   ":last_loss": 100.333267,
   ":inference_ms_per_1000": 9.105
  },
  "small/safec_blas": {
   ":train_ms": 10.908,
   ":last_loss": 100.333267,
   ":inference_ms_per_1000": 9.122
  },
  "big/safec_cpu": {
   ":train_ms": 65.052,
   ":last_loss": 95.428734,
   ":inference_ms_per_200": 59.238
  },
  "big/safec_blas": {
   ":train_ms": 66.798,
   ":last_loss": 95.428734,
   ":inference_ms_per_200": 58.878
  },
  "small/safec_mps": {
   ":train_ms": 40.887,
   ":last_loss": 100.33329,
   ":inference_ms_per_1000": 65.31
  },
  "big/safec_mps": {
   ":train_ms": 47.703,
   ":last_loss": 95.428223,
   ":inference_ms_per_200": 48.784
  },
  "small/pytorch_cpu": {
   "cpu:train_ms": 8.669,
   ":last_loss": 100.018272,
   "cpu:inference_ms_per_1000": 13.068
  },
  "small/pytorch_mps": {
   "mps:train_ms": 61.172,
   ":last_loss": 103.423889,
   "mps:inference_ms_per_1000": 347.57
  },
  "big/pytorch_cpu": {
   "cpu:train_ms": 53.655,
   ":last_loss": 106.548233,
   "cpu:inference_ms_per_200": 48.623
  },
  "big/pytorch_mps": {
   "mps:train_ms": 41.233,
   ":last_loss": 108.710899,
   "mps:inference_ms_per_200": 89.045
  },
  "small/mlx": {
   "mlx:train_ms": 70.609,
   ":last_loss": 96.632431,
   "mlx:inference_ms_per_1000": 298.859
  },
  "big/mlx": {
   "mlx:train_ms": 58.765,
   ":last_loss": 88.7621,
   "mlx:inference_ms_per_200": 78.613
  },
  "small/tensorflow_cpu": {
   "cpu:train_ms": 75.79,
   ":last_loss": 106.182999,
   "cpu:inference_ms_per_1000": 229.698
  },
  "big/tensorflow_cpu": {
   "cpu:train_ms": 140.927,
   ":last_loss": 316.538452,
   "cpu:inference_ms_per_200": 184.939
  }
 },
 "conv": {
  "safec_cpu": {
   "conv64 b=1 cpu": {
    "ms_per_image": 0.6,
    "checksum": 528.229
   },
   "unet128 b=1 cpu": {
    "ms_per_image": 2.433,
    "checksum": 994785.1006
   }
  },
  "pytorch_cpu": {
   "conv64 b=1 cpu": {
    "ms_per_batch": 0.3428,
    "ms_per_image": 0.34278
   },
   "unet128 b=1 cpu": {
    "ms_per_batch": 9.2867,
    "ms_per_image": 9.2867
   }
  },
  "pytorch_mps": {
   "conv64 b=1 mps": {
    "ms_per_batch": 0.0728,
    "ms_per_image": 0.07283
   },
   "conv64 b=32 mps": {
    "ms_per_batch": 1.4321,
    "ms_per_image": 0.04475
   },
   "unet128 b=1 mps": {
    "ms_per_batch": 0.6642,
    "ms_per_image": 0.66417
   },
   "unet128 b=16 mps": {
    "ms_per_batch": 9.3093,
    "ms_per_image": 0.58183
   }
  }
 },
 "fft": {
  "safec_cpu": {
   "fft1d n=1048576 c128 cpu": {
    "ms_per_call": 2.003
   },
   "fft2d 1024x1024 c128 cpu": {
    "ms_per_call": 1.591
   },
   "batch 1024x4096 c2c cpu": {
    "ms_per_call": 2.657
   },
   "rfft 1024x4096 r2c cpu": {
    "ms_per_call": 1.452
   },
   "fft1d n=1048576 c64 cpu": {
    "ms_per_call": 1.575
   },
   "fft2d 1024x1024 c64 cpu": {
    "ms_per_call": 0.865
   },
   "batch 1024x4096 c2c f32 cpu": {
    "ms_per_call": 1.172
   },
   "rfft 1024x4096 r2c f32 cpu": {
    "ms_per_call": 0.805
   }
  },
  "pytorch_cpu": {
   "fft1d n=1048576 c128 cpu": {
    "ms_per_call": 16.4915
   },
   "fft2d 1024x1024 c128 cpu": {
    "ms_per_call": 12.4478
   },
   "batch 1024x4096 c2c cpu": {
    "ms_per_call": 9.0187
   },
   "rfft 1024x4096 r2c cpu": {
    "ms_per_call": 4.1237
   }
  },
  "pytorch_mps": {
   "batch 1024x4096 c2c mps": {
    "ms_per_call": 0.3969
   },
   "rfft 1024x4096 r2c mps": {
    "ms_per_call": 0.2194
   }
  },
  "fftw_1t": {
   "fft1d n=1048576 c128 st": {
    "ms_per_call": 16.339
   },
   "fft2d 1024x1024 c128 st": {
    "ms_per_call": 37.192
   },
   "batch 1024x4096 c2c st": {
    "ms_per_call": 16.64
   },
   "rfft 1024x4096 r2c st": {
    "ms_per_call": 8.457
   },
   "fft1d n=1048576 c64 st": {
    "ms_per_call": 7.909
   },
   "fft2d 1024x1024 c64 st": {
    "ms_per_call": 24.683
   },
   "batch 1024x4096 c2c f32 st": {
    "ms_per_call": 10.481
   },
   "rfft 1024x4096 r2c f32 st": {
    "ms_per_call": 5.305
   }
  },
  "fftw_16t": {
   "fft1d n=1048576 c128 mt": {
    "ms_per_call": 3.649
   },
   "fft2d 1024x1024 c128 mt": {
    "ms_per_call": 7.071
   },
   "batch 1024x4096 c2c mt": {
    "ms_per_call": 3.201
   },
   "rfft 1024x4096 r2c mt": {
    "ms_per_call": 1.809
   },
   "fft1d n=1048576 c64 mt": {
    "ms_per_call": 2.932
   },
   "fft2d 1024x1024 c64 mt": {
    "ms_per_call": 4.838
   },
   "batch 1024x4096 c2c f32 mt": {
    "ms_per_call": 2.021
   },
   "rfft 1024x4096 r2c f32 mt": {
    "ms_per_call": 1.063
   }
  },
  "kfr": {
   "fft1d n=1048576 c128 cpu": {
    "ms_per_call": 6.85
   },
   "fft2d 1024x1024 c128 cpu": {
    "ms_per_call": 9.226
   },
   "batch 1024x4096 c2c cpu": {
    "ms_per_call": 12.705
   },
   "rfft 1024x4096 r2c cpu": {
    "ms_per_call": 7.383
   },
   "fir stream n=1048576 taps=127": {
    "ms_per_call": 45.88
   },
   "fftconv n=1048576 taps=4096": {
    "ms_per_call": 8.213
   },
   "biquad n=1048576": {
    "ms_per_call": 5.141
   },
   "resample n=1048576 44100->48000": {
    "ms_per_call": 165.11
   },
   "resample draft n=1048576 44100->48000": {
    "ms_per_call": 33.134
   },
   "fft1d n=1048576 c64 cpu": {
    "ms_per_call": 3.293
   },
   "fft2d 1024x1024 c64 cpu": {
    "ms_per_call": 3.314
   },
   "batch 1024x4096 c2c f32 cpu": {
    "ms_per_call": 6.211
   },
   "rfft 1024x4096 r2c f32 cpu": {
    "ms_per_call": 3.922
   },
   "fir stream n=1048576 taps=127 f32": {
    "ms_per_call": 64.14
   },
   "fftconv n=1048576 taps=4096 f32": {
    "ms_per_call": 4.198
   },
   "biquad n=1048576 f32": {
    "ms_per_call": 4.204
   },
   "resample n=1048576 44100->48000 f32": {
    "ms_per_call": 91.701
   },
   "resample draft n=1048576 44100->48000 f32": {
    "ms_per_call": 34.123
   }
  },
  "rust": {
   "fft1d n=1048576 c128 cpu": {
    "ms_per_call": 9.436
   },
   "batch 1024x4096 c2c cpu": {
    "ms_per_call": 17.618
   },
   "fft1d n=1048576 c64 cpu": {
    "ms_per_call": 4.625
   },
   "batch 1024x4096 c2c f32 cpu": {
    "ms_per_call": 10.261
   },
   "resample n=1048576 44100->48000": {
    "ms_per_call": 301.362
   }
  },
  "safec_cpu_1t": {
   "fft1d n=1048576 c128 cpu": {
    "ms_per_call": 10.368
   },
   "fft2d 1024x1024 c128 cpu": {
    "ms_per_call": 8.538
   },
   "batch 1024x4096 c2c cpu": {
    "ms_per_call": 15.447
   },
   "rfft 1024x4096 r2c cpu": {
    "ms_per_call": 8.928
   },
   "fft1d n=1048576 c64 cpu": {
    "ms_per_call": 7.551
   },
   "fft2d 1024x1024 c64 cpu": {
    "ms_per_call": 4.432
   },
   "batch 1024x4096 c2c f32 cpu": {
    "ms_per_call": 8.703
   },
   "rfft 1024x4096 r2c f32 cpu": {
    "ms_per_call": 5.241
   },
   "note": "SAFEC_NUM_THREADS=1 ./build/fft/safec_cpu, best of 3"
  }
 },
 "ml_nn": {
  "safec_cpu": {
   "mlp_small": {
    "train_ms": 5.792,
    "last_loss": 0.0297226
   },
   "mlp_big": {
    "train_ms": 29.273,
    "last_loss": 0.0
   },
   "vae": {
    "train_ms": 18.786,
    "last_loss": 31.5691
   },
   "gan": {
    "train_ms": 12.547,
    "last_loss": 1.40552
   },
   "dqn": {
    "train_ms": 72.131,
    "last_loss": 0.138698
   },
   "ppo": {
    "train_ms": 33.232,
    "last_loss": 48.7648
   },
   "lora": {
    "train_ms": 18.385,
    "last_loss": 0.0219908
   },
   "conv_ae": {
    "train_ms": 51.64,
    "last_loss": 0.255294
   }
  },
  "safec_mps": {
   "mlp_small": {
    "train_ms": 92.018,
    "last_loss": 0.0297224
   },
   "mlp_big": {
    "train_ms": 22.207,
    "last_loss": 0.0
   },
   "vae": {
    "train_ms": 157.846,
    "last_loss": 31.5691
   },
   "gan": {
    "train_ms": 318.498,
    "last_loss": 1.40552
   },
   "dqn": {
    "train_ms": 1642.72,
    "last_loss": 0.138697
   },
   "ppo": {
    "train_ms": 373.301,
    "last_loss": 48.7648
   },
   "lora": {
    "train_ms": 199.333,
    "last_loss": 0.0219904
   },
   "conv_ae": {
    "train_ms": 41.116,
    "last_loss": 0.255294
   }
  },
  "pytorch_cpu": {
   "mlp_small": {
    "train_ms": 161.118,
    "last_loss": 0.0331768
   },
   "mlp_big": {
    "train_ms": 78.022,
    "last_loss": 0.0
   },
   "vae": {
    "train_ms": 108.603,
    "last_loss": 50.6271
   },
   "gan": {
    "train_ms": 121.186,
    "last_loss": 1.42086
   },
   "dqn": {
    "train_ms": 518.67,
    "last_loss": 0.126866
   },
   "ppo": {
    "train_ms": 326.975,
    "last_loss": 41.1927
   },
   "lora": {
    "train_ms": 203.802,
    "last_loss": 0.0405097
   },
   "conv_ae": {
    "train_ms": 912.88,
    "last_loss": 0.484414
   }
  },
  "pytorch_mps": {
   "mlp_small": {
    "train_ms": 965.951,
    "last_loss": 0.0256877
   },
   "mlp_big": {
    "train_ms": 99.555,
    "last_loss": 0.0
   },
   "vae": {
    "train_ms": 430.751,
    "last_loss": 50.467
   },
   "gan": {
    "train_ms": 671.545,
    "last_loss": 1.48902
   },
   "dqn": {
    "train_ms": 4000.093,
    "last_loss": 0.223788
   },
   "ppo": {
    "train_ms": 1261.356,
    "last_loss": 34.5712
   },
   "lora": {
    "train_ms": 1260.209,
    "last_loss": 0.0458703
   },
   "conv_ae": {
    "train_ms": 265.456,
    "last_loss": 0.498442
   }
  }
 },
 "ml_nn_gpu": {
  "safec_cpu": {
   "mlp": {
    "train_ms": 301.553,
    "last_loss": 0.0
   },
   "cnn": {
    "train_ms": 1098.19,
    "last_loss": 0.080832
   },
   "resnet": {
    "train_ms": 3492.26,
    "last_loss": 0.131147
   },
   "unet": {
    "train_ms": 2190.6,
    "last_loss": 0.137156
   },
   "dcgan": {
    "train_ms": 869.467,
    "last_loss": 0.444124
   },
   "convvae": {
    "train_ms": 166.667,
    "last_loss": 393.525
   }
  },
  "safec_mps": {
   "mlp_fp32": {
    "train_ms": 126.221,
    "last_loss": 0.0
   },
   "cnn_fp32": {
    "train_ms": 172.628,
    "last_loss": 0.080832
   },
   "resnet_fp32": {
    "train_ms": 560.589,
    "last_loss": 0.131156
   },
   "unet_fp32": {
    "train_ms": 373.972,
    "last_loss": 0.137155
   },
   "dcgan_fp32": {
    "train_ms": 309.814,
    "last_loss": 0.44448
   },
   "convvae_fp32": {
    "train_ms": 153.779,
    "last_loss": 393.525
   },
   "mlp_fp16": {
    "train_ms": 131.399,
    "last_loss": 0.0
   },
   "cnn_fp16": {
    "train_ms": 160.302,
    "last_loss": 0.0808863
   },
   "resnet_fp16": {
    "train_ms": 513.286,
    "last_loss": 0.131143
   },
   "unet_fp16": {
    "train_ms": 394.306,
    "last_loss": 0.137228
   },
   "dcgan_fp16": {
    "train_ms": 316.386,
    "last_loss": 0.445209
   },
   "convvae_fp16": {
    "train_ms": 159.35,
    "last_loss": 393.548
   },
   "mlp_bf16": {
    "train_ms": 185.306,
    "last_loss": 0.0
   },
   "cnn_bf16": {
    "train_ms": 185.741,
    "last_loss": 0.0813884
   },
   "resnet_bf16": {
    "train_ms": 580.5,
    "last_loss": 0.131229
   },
   "unet_bf16": {
    "train_ms": 423.401,
    "last_loss": 0.137551
   },
   "dcgan_bf16": {
    "train_ms": 339.133,
    "last_loss": 0.444145
   },
   "convvae_bf16": {
    "train_ms": 165.325,
    "last_loss": 393.526
   }
  },
  "pytorch_mps": {
   "mlp_fp32": {
    "train_ms": 277.351,
    "last_loss": 0.0
   },
   "cnn_fp32": {
    "train_ms": 266.562,
    "last_loss": 1.38584
   },
   "resnet_fp32": {
    "train_ms": 538.153,
    "last_loss": 0.148273
   },
   "unet_fp32": {
    "train_ms": 462.699,
    "last_loss": 0.396342
   },
   "dcgan_fp32": {
    "train_ms": 342.127,
    "last_loss": 0.70854
   },
   "convvae_fp32": {
    "train_ms": 165.243,
    "last_loss": 444.771
   },
   "mlp_fp16": {
    "train_ms": 472.725,
    "last_loss": 0.0
   },
   "cnn_fp16": {
    "train_ms": 349.092,
    "last_loss": 1.3946
   },
   "resnet_fp16": {
    "train_ms": 651.651,
    "last_loss": 0.179025
   },
   "unet_fp16": {
    "train_ms": 536.779,
    "last_loss": 0.414297
   },
   "mlp_bf16": {
    "train_ms": 324.207,
    "last_loss": 0.0
   },
   "cnn_bf16": {
    "train_ms": 310.259,
    "last_loss": 1.38806
   },
   "resnet_bf16": {
    "train_ms": 671.916,
    "last_loss": 0.166652
   },
   "unet_bf16": {
    "train_ms": 483.376,
    "last_loss": 0.40853
   }
  }
 }
}