{
  "fixed_work_sweep": [
    {
      "microbatches": 1,
      "overhead": 0.0,
      "duration_per_operation": 32.0,
      "elapsed": 256.0,
      "utilization": 0.25,
      "compute_utilization": 0.25
    },
    {
      "microbatches": 2,
      "overhead": 0.0,
      "duration_per_operation": 16.0,
      "elapsed": 160.0,
      "utilization": 0.4,
      "compute_utilization": 0.4
    },
    {
      "microbatches": 4,
      "overhead": 0.0,
      "duration_per_operation": 8.0,
      "elapsed": 112.0,
      "utilization": 0.5714285714285714,
      "compute_utilization": 0.5714285714285714
    },
    {
      "microbatches": 8,
      "overhead": 0.0,
      "duration_per_operation": 4.0,
      "elapsed": 88.0,
      "utilization": 0.7272727272727273,
      "compute_utilization": 0.7272727272727273
    },
    {
      "microbatches": 16,
      "overhead": 0.0,
      "duration_per_operation": 2.0,
      "elapsed": 76.0,
      "utilization": 0.8421052631578947,
      "compute_utilization": 0.8421052631578947
    },
    {
      "microbatches": 32,
      "overhead": 0.0,
      "duration_per_operation": 1.0,
      "elapsed": 70.0,
      "utilization": 0.9142857142857143,
      "compute_utilization": 0.9142857142857143
    },
    {
      "microbatches": 64,
      "overhead": 0.0,
      "duration_per_operation": 0.5,
      "elapsed": 67.0,
      "utilization": 0.9552238805970149,
      "compute_utilization": 0.9552238805970149
    },
    {
      "microbatches": 128,
      "overhead": 0.0,
      "duration_per_operation": 0.25,
      "elapsed": 65.5,
      "utilization": 0.9770992366412213,
      "compute_utilization": 0.9770992366412213
    },
    {
      "microbatches": 1,
      "overhead": 0.1,
      "duration_per_operation": 32.1,
      "elapsed": 256.8,
      "utilization": 0.25,
      "compute_utilization": 0.24922118380062305
    },
    {
      "microbatches": 2,
      "overhead": 0.1,
      "duration_per_operation": 16.1,
      "elapsed": 161.0,
      "utilization": 0.4,
      "compute_utilization": 0.39751552795031053
    },
    {
      "microbatches": 4,
      "overhead": 0.1,
      "duration_per_operation": 8.1,
      "elapsed": 113.4,
      "utilization": 0.5714285714285714,
      "compute_utilization": 0.564373897707231
    },
    {
      "microbatches": 8,
      "overhead": 0.1,
      "duration_per_operation": 4.1,
      "elapsed": 90.20000000000002,
      "utilization": 0.7272727272727271,
      "compute_utilization": 0.7095343680709533
    },
    {
      "microbatches": 16,
      "overhead": 0.1,
      "duration_per_operation": 2.1,
      "elapsed": 79.80000000000003,
      "utilization": 0.8421052631578945,
      "compute_utilization": 0.802005012531328
    },
    {
      "microbatches": 32,
      "overhead": 0.1,
      "duration_per_operation": 1.1,
      "elapsed": 77.00000000000004,
      "utilization": 0.9142857142857138,
      "compute_utilization": 0.8311688311688307
    },
    {
      "microbatches": 64,
      "overhead": 0.1,
      "duration_per_operation": 0.6,
      "elapsed": 80.4000000000001,
      "utilization": 0.9552238805970137,
      "compute_utilization": 0.7960199004975114
    },
    {
      "microbatches": 128,
      "overhead": 0.1,
      "duration_per_operation": 0.35,
      "elapsed": 91.70000000000022,
      "utilization": 0.977099236641219,
      "compute_utilization": 0.6979280261722993
    },
    {
      "microbatches": 1,
      "overhead": 0.5,
      "duration_per_operation": 32.5,
      "elapsed": 260.0,
      "utilization": 0.25,
      "compute_utilization": 0.24615384615384617
    },
    {
      "microbatches": 2,
      "overhead": 0.5,
      "duration_per_operation": 16.5,
      "elapsed": 165.0,
      "utilization": 0.4,
      "compute_utilization": 0.3878787878787879
    },
    {
      "microbatches": 4,
      "overhead": 0.5,
      "duration_per_operation": 8.5,
      "elapsed": 119.0,
      "utilization": 0.5714285714285714,
      "compute_utilization": 0.5378151260504201
    },
    {
      "microbatches": 8,
      "overhead": 0.5,
      "duration_per_operation": 4.5,
      "elapsed": 99.0,
      "utilization": 0.7272727272727273,
      "compute_utilization": 0.6464646464646465
    },
    {
      "microbatches": 16,
      "overhead": 0.5,
      "duration_per_operation": 2.5,
      "elapsed": 95.0,
      "utilization": 0.8421052631578947,
      "compute_utilization": 0.6736842105263158
    },
    {
      "microbatches": 32,
      "overhead": 0.5,
      "duration_per_operation": 1.5,
      "elapsed": 105.0,
      "utilization": 0.9142857142857143,
      "compute_utilization": 0.6095238095238096
    },
    {
      "microbatches": 64,
      "overhead": 0.5,
      "duration_per_operation": 1.0,
      "elapsed": 134.0,
      "utilization": 0.9552238805970149,
      "compute_utilization": 0.47761194029850745
    },
    {
      "microbatches": 128,
      "overhead": 0.5,
      "duration_per_operation": 0.75,
      "elapsed": 196.5,
      "utilization": 0.9770992366412213,
      "compute_utilization": 0.3256997455470738
    },
    {
      "microbatches": 1,
      "overhead": 1.0,
      "duration_per_operation": 33.0,
      "elapsed": 264.0,
      "utilization": 0.25,
      "compute_utilization": 0.24242424242424243
    },
    {
      "microbatches": 2,
      "overhead": 1.0,
      "duration_per_operation": 17.0,
      "elapsed": 170.0,
      "utilization": 0.4,
      "compute_utilization": 0.3764705882352941
    },
    {
      "microbatches": 4,
      "overhead": 1.0,
      "duration_per_operation": 9.0,
      "elapsed": 126.0,
      "utilization": 0.5714285714285714,
      "compute_utilization": 0.5079365079365079
    },
    {
      "microbatches": 8,
      "overhead": 1.0,
      "duration_per_operation": 5.0,
      "elapsed": 110.0,
      "utilization": 0.7272727272727273,
      "compute_utilization": 0.5818181818181818
    },
    {
      "microbatches": 16,
      "overhead": 1.0,
      "duration_per_operation": 3.0,
      "elapsed": 114.0,
      "utilization": 0.8421052631578947,
      "compute_utilization": 0.5614035087719298
    },
    {
      "microbatches": 32,
      "overhead": 1.0,
      "duration_per_operation": 2.0,
      "elapsed": 140.0,
      "utilization": 0.9142857142857143,
      "compute_utilization": 0.45714285714285713
    },
    {
      "microbatches": 64,
      "overhead": 1.0,
      "duration_per_operation": 1.5,
      "elapsed": 201.0,
      "utilization": 0.9552238805970149,
      "compute_utilization": 0.31840796019900497
    },
    {
      "microbatches": 128,
      "overhead": 1.0,
      "duration_per_operation": 1.25,
      "elapsed": 327.5,
      "utilization": 0.9770992366412213,
      "compute_utilization": 0.19541984732824427
    }
  ],
  "total_compute_per_stage_per_direction": 32,
  "balanced_forward": [
    {
      "stage": 0,
      "microbatch": 0,
      "start": 0.0,
      "finish": 1.0
    },
    {
      "stage": 1,
      "microbatch": 0,
      "start": 1.0,
      "finish": 2.0
    },
    {
      "stage": 2,
      "microbatch": 0,
      "start": 2.0,
      "finish": 3.0
    },
    {
      "stage": 3,
      "microbatch": 0,
      "start": 3.0,
      "finish": 4.0
    },
    {
      "stage": 0,
      "microbatch": 1,
      "start": 1.0,
      "finish": 2.0
    },
    {
      "stage": 1,
      "microbatch": 1,
      "start": 2.0,
      "finish": 3.0
    },
    {
      "stage": 2,
      "microbatch": 1,
      "start": 3.0,
      "finish": 4.0
    },
    {
      "stage": 3,
      "microbatch": 1,
      "start": 4.0,
      "finish": 5.0
    },
    {
      "stage": 0,
      "microbatch": 2,
      "start": 2.0,
      "finish": 3.0
    },
    {
      "stage": 1,
      "microbatch": 2,
      "start": 3.0,
      "finish": 4.0
    },
    {
      "stage": 2,
      "microbatch": 2,
      "start": 4.0,
      "finish": 5.0
    },
    {
      "stage": 3,
      "microbatch": 2,
      "start": 5.0,
      "finish": 6.0
    },
    {
      "stage": 0,
      "microbatch": 3,
      "start": 3.0,
      "finish": 4.0
    },
    {
      "stage": 1,
      "microbatch": 3,
      "start": 4.0,
      "finish": 5.0
    },
    {
      "stage": 2,
      "microbatch": 3,
      "start": 5.0,
      "finish": 6.0
    },
    {
      "stage": 3,
      "microbatch": 3,
      "start": 6.0,
      "finish": 7.0
    }
  ],
  "imbalanced_forward": [
    {
      "stage": 0,
      "microbatch": 0,
      "start": 0.0,
      "finish": 1.0
    },
    {
      "stage": 1,
      "microbatch": 0,
      "start": 1.0,
      "finish": 2.0
    },
    {
      "stage": 2,
      "microbatch": 0,
      "start": 2.0,
      "finish": 4.0
    },
    {
      "stage": 3,
      "microbatch": 0,
      "start": 4.0,
      "finish": 5.0
    },
    {
      "stage": 0,
      "microbatch": 1,
      "start": 1.0,
      "finish": 2.0
    },
    {
      "stage": 1,
      "microbatch": 1,
      "start": 2.0,
      "finish": 3.0
    },
    {
      "stage": 2,
      "microbatch": 1,
      "start": 4.0,
      "finish": 6.0
    },
    {
      "stage": 3,
      "microbatch": 1,
      "start": 6.0,
      "finish": 7.0
    },
    {
      "stage": 0,
      "microbatch": 2,
      "start": 2.0,
      "finish": 3.0
    },
    {
      "stage": 1,
      "microbatch": 2,
      "start": 3.0,
      "finish": 4.0
    },
    {
      "stage": 2,
      "microbatch": 2,
      "start": 6.0,
      "finish": 8.0
    },
    {
      "stage": 3,
      "microbatch": 2,
      "start": 8.0,
      "finish": 9.0
    },
    {
      "stage": 0,
      "microbatch": 3,
      "start": 3.0,
      "finish": 4.0
    },
    {
      "stage": 1,
      "microbatch": 3,
      "start": 4.0,
      "finish": 5.0
    },
    {
      "stage": 2,
      "microbatch": 3,
      "start": 8.0,
      "finish": 10.0
    },
    {
      "stage": 3,
      "microbatch": 3,
      "start": 10.0,
      "finish": 11.0
    }
  ]
}
