{
  "protocol": {
    "device": "CPU",
    "test": "matrices",
    "dtype": "float32",
    "threads": 1,
    "warmup": 10,
    "samples_per_process": 30,
    "fresh_processes": 5,
    "synchronize_each_sample": true,
    "stable_realized_inputs": true,
    "torch_matrix_output_preallocated": true,
    "tiny_outputs_reused_by_tinyjit": true,
    "hcq2": false
  },
  "tinygrad_commit": "5b60a09ab0a7f20c3426e505251c48cac020d1fa",
  "validation": {
    "matrix_max_abs": 0.0015360576791749736,
    "passed": true
  },
  "summary": {
    "ratio": {
      "4x128x128": {
        "median": 55.86608300907912,
        "process_ratios": [
          71.39031705227079,
          72.71122536418167,
          53.31396083385976,
          55.86608300907912,
          55.75033333333334
        ]
      },
      "10000x128x128": {
        "median": 27.70036648553014,
        "process_ratios": [
          27.050078151605376,
          27.808214531203973,
          27.862924778761062,
          27.70036648553014,
          27.61590978319891
        ]
      },
      "128x10000x128": {
        "median": 43.70779287515898,
        "process_ratios": [
          43.5020887790147,
          45.37782709096029,
          43.70779287515898,
          44.41049226594301,
          43.62015233022637
        ]
      }
    },
    "torch": {
      "4x128x128": {
        "median_ms": 0.0015,
        "process_medians_ms": [
          0.001167,
          0.001167,
          0.001583,
          0.001542,
          0.0015
        ]
      },
      "10000x128x128": {
        "median_ms": 0.226,
        "process_medians_ms": [
          0.22904200000000002,
          0.224042,
          0.226,
          0.2255205,
          0.226521
        ]
      },
      "128x10000x128": {
        "median_ms": 0.23279149999999998,
        "process_medians_ms": [
          0.23279149999999998,
          0.2219375,
          0.233521,
          0.2303125,
          0.2346875
        ]
      }
    },
    "tiny": {
      "4x128x128": {
        "median_ms": 0.084396,
        "process_medians_ms": [
          0.08331250000000001,
          0.08485400000000001,
          0.084396,
          0.0861455,
          0.0836255
        ]
      },
      "10000x128x128": {
        "median_ms": 6.2470005,
        "process_medians_ms": [
          6.1956039999999994,
          6.230208,
          6.297021,
          6.2470005,
          6.2555835
        ]
      },
      "128x10000x128": {
        "median_ms": 10.206687500000001,
        "process_medians_ms": [
          10.1269165,
          10.0710415,
          10.206687500000001,
          10.228291500000001,
          10.237104500000001
        ]
      }
    }
  },
  "runs": [
    {
      "backend": "torch",
      "version": "2.8.0",
      "device": "cpu",
      "threads": 1,
      "process_index": 0,
      "results": {
        "4x128x128": {
          "median_ms": 0.001167,
          "p10_ms": 0.001166,
          "p90_ms": 0.00125,
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        },
        "10000x128x128": {
          "median_ms": 0.22904200000000002,
          "p10_ms": 0.222959,
          "p90_ms": 0.234708,
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        },
        "128x10000x128": {
          "median_ms": 0.23279149999999998,
          "p10_ms": 0.227417,
          "p90_ms": 0.246583,
          "error": {
            "max_abs": 0.0015360576791749736,
            "rms": 0.00017936516186356426
          }
        }
      }
    },
    {
      "backend": "tiny",
      "device": "CPU",
      "hcq2": false,
      "threads": 1,
      "process_index": 0,
      "results": {
        "4x128x128": {
          "median_ms": 0.08331250000000001,
          "p10_ms": 0.077166,
          "p90_ms": 0.087666,
          "direct_kernel_ms": 0.021666986867785454,
          "initial_ms": [
            33.525125,
            1.063917,
            0.097541
          ],
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        },
        "10000x128x128": {
          "median_ms": 6.1956039999999994,
          "p10_ms": 6.00975,
          "p90_ms": 6.281667,
          "direct_kernel_ms": 6.171084009110928,
          "initial_ms": [
            69.085667,
            7.743584,
            6.157292
          ],
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        },
        "128x10000x128": {
          "median_ms": 10.1269165,
          "p10_ms": 9.924542,
          "p90_ms": 10.265291,
          "direct_kernel_ms": 10.079708066768944,
          "initial_ms": [
            67.870375,
            13.065917,
            11.712333
          ],
          "error": {
            "max_abs": 0.0015360576773559842,
            "rms": 0.00017936516186414065
          }
        }
      }
    },
    {
      "backend": "tiny",
      "device": "CPU",
      "hcq2": false,
      "threads": 1,
      "process_index": 1,
      "results": {
        "10000x128x128": {
          "median_ms": 6.230208,
          "p10_ms": 6.073083,
          "p90_ms": 6.318833,
          "direct_kernel_ms": 6.197500042617321,
          "initial_ms": [
            91.662042,
            7.822625,
            6.222083
          ],
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        },
        "128x10000x128": {
          "median_ms": 10.0710415,
          "p10_ms": 9.922791,
          "p90_ms": 10.311834,
          "direct_kernel_ms": 10.043042013421655,
          "initial_ms": [
            61.62025,
            11.880042,
            10.387709
          ],
          "error": {
            "max_abs": 0.0015360576773559842,
            "rms": 0.00017936516186414065
          }
        },
        "4x128x128": {
          "median_ms": 0.08485400000000001,
          "p10_ms": 0.078958,
          "p90_ms": 0.11175,
          "direct_kernel_ms": 0.01962494570761919,
          "initial_ms": [
            11.955292,
            1.051584,
            0.096791
          ],
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        }
      }
    },
    {
      "backend": "torch",
      "version": "2.8.0",
      "device": "cpu",
      "threads": 1,
      "process_index": 1,
      "results": {
        "10000x128x128": {
          "median_ms": 0.224042,
          "p10_ms": 0.223125,
          "p90_ms": 0.22925,
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        },
        "128x10000x128": {
          "median_ms": 0.2219375,
          "p10_ms": 0.219208,
          "p90_ms": 0.231667,
          "error": {
            "max_abs": 0.0015360576791749736,
            "rms": 0.00017936516186356426
          }
        },
        "4x128x128": {
          "median_ms": 0.001167,
          "p10_ms": 0.001167,
          "p90_ms": 0.001209,
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        }
      }
    },
    {
      "backend": "torch",
      "version": "2.8.0",
      "device": "cpu",
      "threads": 1,
      "process_index": 2,
      "results": {
        "128x10000x128": {
          "median_ms": 0.233521,
          "p10_ms": 0.224291,
          "p90_ms": 0.250917,
          "error": {
            "max_abs": 0.0015360576791749736,
            "rms": 0.00017936516186356426
          }
        },
        "4x128x128": {
          "median_ms": 0.001583,
          "p10_ms": 0.001583,
          "p90_ms": 0.001625,
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        },
        "10000x128x128": {
          "median_ms": 0.226,
          "p10_ms": 0.22475,
          "p90_ms": 0.231833,
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        }
      }
    },
    {
      "backend": "tiny",
      "device": "CPU",
      "hcq2": false,
      "threads": 1,
      "process_index": 2,
      "results": {
        "128x10000x128": {
          "median_ms": 10.206687500000001,
          "p10_ms": 10.039834,
          "p90_ms": 10.349417,
          "direct_kernel_ms": 10.076249949634075,
          "initial_ms": [
            96.149291,
            11.618375,
            10.233292
          ],
          "error": {
            "max_abs": 0.0015360576773559842,
            "rms": 0.00017936516186414065
          }
        },
        "4x128x128": {
          "median_ms": 0.084396,
          "p10_ms": 0.082541,
          "p90_ms": 0.08825,
          "direct_kernel_ms": 0.021249987185001373,
          "initial_ms": [
            12.383292,
            1.091167,
            0.096375
          ],
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        },
        "10000x128x128": {
          "median_ms": 6.297021,
          "p10_ms": 6.205083,
          "p90_ms": 6.336792,
          "direct_kernel_ms": 6.283415947109461,
          "initial_ms": [
            56.048125,
            7.821333,
            6.222667
          ],
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        }
      }
    },
    {
      "backend": "tiny",
      "device": "CPU",
      "hcq2": false,
      "threads": 1,
      "process_index": 3,
      "results": {
        "4x128x128": {
          "median_ms": 0.0861455,
          "p10_ms": 0.083375,
          "p90_ms": 0.091166,
          "direct_kernel_ms": 0.021791085600852966,
          "initial_ms": [
            33.756667,
            1.107875,
            0.100917
          ],
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        },
        "10000x128x128": {
          "median_ms": 6.2470005,
          "p10_ms": 6.176542,
          "p90_ms": 6.336459,
          "direct_kernel_ms": 6.18537503760308,
          "initial_ms": [
            69.304041,
            7.894958,
            6.245459
          ],
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        },
        "128x10000x128": {
          "median_ms": 10.228291500000001,
          "p10_ms": 10.120875,
          "p90_ms": 10.372958,
          "direct_kernel_ms": 10.222207987681031,
          "initial_ms": [
            61.678417,
            11.825709,
            10.171459
          ],
          "error": {
            "max_abs": 0.0015360576773559842,
            "rms": 0.00017936516186414065
          }
        }
      }
    },
    {
      "backend": "torch",
      "version": "2.8.0",
      "device": "cpu",
      "threads": 1,
      "process_index": 3,
      "results": {
        "4x128x128": {
          "median_ms": 0.001542,
          "p10_ms": 0.0015,
          "p90_ms": 0.001625,
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        },
        "10000x128x128": {
          "median_ms": 0.2255205,
          "p10_ms": 0.222542,
          "p90_ms": 0.228584,
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        },
        "128x10000x128": {
          "median_ms": 0.2303125,
          "p10_ms": 0.220625,
          "p90_ms": 0.285125,
          "error": {
            "max_abs": 0.0015360576791749736,
            "rms": 0.00017936516186356426
          }
        }
      }
    },
    {
      "backend": "torch",
      "version": "2.8.0",
      "device": "cpu",
      "threads": 1,
      "process_index": 4,
      "results": {
        "10000x128x128": {
          "median_ms": 0.226521,
          "p10_ms": 0.224875,
          "p90_ms": 0.229792,
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        },
        "128x10000x128": {
          "median_ms": 0.2346875,
          "p10_ms": 0.224917,
          "p90_ms": 0.255375,
          "error": {
            "max_abs": 0.0015360576791749736,
            "rms": 0.00017936516186356426
          }
        },
        "4x128x128": {
          "median_ms": 0.0015,
          "p10_ms": 0.001458,
          "p90_ms": 0.001542,
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        }
      }
    },
    {
      "backend": "tiny",
      "device": "CPU",
      "hcq2": false,
      "threads": 1,
      "process_index": 4,
      "results": {
        "10000x128x128": {
          "median_ms": 6.2555835,
          "p10_ms": 6.175041,
          "p90_ms": 6.319083,
          "direct_kernel_ms": 6.213374901562929,
          "initial_ms": [
            90.938959,
            7.952542,
            6.182958
          ],
          "error": {
            "max_abs": 3.140326408868077e-05,
            "rms": 2.309186961552311e-06
          }
        },
        "128x10000x128": {
          "median_ms": 10.237104500000001,
          "p10_ms": 10.127541,
          "p90_ms": 10.333958,
          "direct_kernel_ms": 10.260291979648173,
          "initial_ms": [
            61.224375,
            11.736,
            10.209584
          ],
          "error": {
            "max_abs": 0.0015360576773559842,
            "rms": 0.00017936516186414065
          }
        },
        "4x128x128": {
          "median_ms": 0.0836255,
          "p10_ms": 0.081416,
          "p90_ms": 0.086458,
          "direct_kernel_ms": 0.020999927073717117,
          "initial_ms": [
            12.129458,
            1.0405,
            0.103084
          ],
          "error": {
            "max_abs": 9.594529398526674e-06,
            "rms": 2.2642906909800874e-06
          }
        }
      }
    }
  ]
}
