{
 "returncode": 0,
 "stdout": "ladder_d4.py:custom_kernel at 2.30% error: 15 timed calls, each in a fresh process\n  review flag: 18,670 bytes, close to the 20,480-byte limit\n  NVIDIA A100 80GB PCIe, torch 2.12.0+cu130, sandbox on\n  call  1/15: 3,480.668 ms, scorer's clock 3,481.448 ms\n  call  2/15: 3,479.799 ms, scorer's clock 3,482.080 ms\n  call  3/15: 3,483.643 ms, scorer's clock 3,485.880 ms\n  call  4/15: 3,482.619 ms, scorer's clock 3,484.709 ms\n  call  5/15: 3,482.448 ms, scorer's clock 3,484.626 ms\n  call  6/15: 3,480.071 ms, scorer's clock 3,482.300 ms\n  call  7/15: 3,480.738 ms, scorer's clock 3,483.045 ms\n  call  8/15: 3,479.596 ms, scorer's clock 3,481.799 ms\n  call  9/15: 3,482.873 ms, scorer's clock 3,485.141 ms\n  call 10/15: 3,482.725 ms, scorer's clock 3,484.819 ms\n  call 11/15: 3,479.916 ms, scorer's clock 3,482.093 ms\n  call 12/15: 3,480.177 ms, scorer's clock 3,482.513 ms\n  call 13/15: 3,479.610 ms, scorer's clock 3,481.797 ms\n  call 14/15: 3,483.005 ms, scorer's clock 3,485.258 ms\n  call 15/15: 3,479.999 ms, scorer's clock 3,482.272 ms\nMNIST 97.92% (107,714/110,000), 3481.336 ms/call; hold-out (fashion) 87.98%, 3480.796 ms/call; score 3481.336 ms\nenergy: one more fresh process runs the method back to back on fresh MNIST draws for 20 s, between idle windows in which it is frozen\n  NVIDIA A100 80GB PCIe: idle 81.8 W with the method frozen; telemetry reference 18.5 TFLOP/s at 7.58 J/TFLOP above idle (a healthy NVIDIA A100 80GB PCIe reads 6-11)\n  the round trip alone: 17.8 mJ per call over 3,636 empty calls, subtracted\n  the method: 6 calls in 20.8 s, MNIST 98.06%, 3,466.109 ms per call, 214.6 W on average\nenergy 460763.342 mJ per call above idle\n",
 "stderr": "",
 "record": {
  "version": "mnist-a100/1.2.0 (protections of sutro-mnist-medium/3.0.0, tightened; energy column)",
  "file": "ladder_d4.py",
  "function": "custom_kernel",
  "source_sha256": "425575f48e7daded863d1c002266a73a3467e2d6bf7010b2ca0afd3404256482",
  "band_bp": 230,
  "sandboxed": true,
  "holdout": "fashion",
  "device": "NVIDIA A100 80GB PCIe",
  "torch": "2.12.0+cu130",
  "calls": [
   {
    "dataset": "fashion",
    "ms": 3480.668212890625,
    "parent_ms": 3483.6139420000018,
    "correct": 8789,
    "total": 10000,
    "overhead_ms": 2.166421998910728
   },
   {
    "dataset": "mnist",
    "ms": 3479.798828125,
    "parent_ms": 3484.0292000000018,
    "correct": 9793,
    "total": 10000,
    "overhead_ms": 1.9490699990400913
   },
   {
    "dataset": "mnist",
    "ms": 3483.642822265625,
    "parent_ms": 3487.879333999999,
    "correct": 9765,
    "total": 10000,
    "overhead_ms": 1.9992744993393785
   },
   {
    "dataset": "mnist",
    "ms": 3482.618896484375,
    "parent_ms": 3486.7219280000013,
    "correct": 9806,
    "total": 10000,
    "overhead_ms": 2.0126925011169874
   },
   {
    "dataset": "fashion",
    "ms": 3482.44775390625,
    "parent_ms": 3486.567281999996,
    "correct": 8831,
    "total": 10000,
    "overhead_ms": 1.941228499734052
   },
   {
    "dataset": "fashion",
    "ms": 3480.0712890625,
    "parent_ms": 3484.171611000008,
    "correct": 8767,
    "total": 10000,
    "overhead_ms": 1.8720410012138018
   },
   {
    "dataset": "mnist",
    "ms": 3480.73779296875,
    "parent_ms": 3485.001771,
    "correct": 9799,
    "total": 10000,
    "overhead_ms": 1.9567219992941318
   },
   {
    "dataset": "mnist",
    "ms": 3479.595947265625,
    "parent_ms": 3483.700763999991,
    "correct": 9803,
    "total": 10000,
    "overhead_ms": 1.9021175006893145
   },
   {
    "dataset": "mnist",
    "ms": 3482.872802734375,
    "parent_ms": 3487.0534509999798,
    "correct": 9799,
    "total": 10000,
    "overhead_ms": 1.9124120006432577
   },
   {
    "dataset": "mnist",
    "ms": 3482.725341796875,
    "parent_ms": 3486.780842000002,
    "correct": 9793,
    "total": 10000,
    "overhead_ms": 1.961949998967384
   },
   {
    "dataset": "mnist",
    "ms": 3479.91552734375,
    "parent_ms": 3483.9369840000245,
    "correct": 9775,
    "total": 10000,
    "overhead_ms": 1.8442024999814066
   },
   {
    "dataset": "mnist",
    "ms": 3480.1767578125,
    "parent_ms": 3484.3713349999916,
    "correct": 9799,
    "total": 10000,
    "overhead_ms": 1.8584880010521374
   },
   {
    "dataset": "mnist",
    "ms": 3479.6103515625,
    "parent_ms": 3483.685997000009,
    "correct": 9804,
    "total": 10000,
    "overhead_ms": 1.889482498789448
   },
   {
    "dataset": "mnist",
    "ms": 3483.0048828125,
    "parent_ms": 3487.132308999975,
    "correct": 9778,
    "total": 10000,
    "overhead_ms": 1.8743280005061251
   },
   {
    "dataset": "fashion",
    "ms": 3479.99853515625,
    "parent_ms": 3484.1562900000154,
    "correct": 8807,
    "total": 10000,
    "overhead_ms": 1.883901499299668
   }
  ],
  "problems": [],
  "summary": "MNIST 97.92% (107,714/110,000), 3481.336 ms/call; hold-out (fashion) 87.98%, 3480.796 ms/call",
  "ranked_ms": 3481.336359197443,
  "energy": {
   "mj_per_call": 460763.34167068644,
   "problems": [],
   "net_mj_before_control": 460781.1534844383,
   "control_mj_per_call": 17.81181375188914,
   "control_calls": 3636,
   "gross_mj_per_call": 744585.0,
   "idle_w": 81.78835931780631,
   "active_w": 214.57914073518234,
   "calls": 6,
   "seconds": 20.819870863000006,
   "ms_median": 3466.109375,
   "timed_mnist_ms_median": 3480.73779296875,
   "correct": 58837,
   "total": 60000,
   "reference": {
    "tflops_per_s": 18.465493825967446,
    "j_per_tflop": 7.5817949123518,
    "exact": true,
    "band": [
     [
      6.0,
      11.0
     ],
     [
      15.0,
      23.0
     ]
    ]
   },
   "device": {
    "name": "NVIDIA A100 80GB PCIe",
    "uuid": "GPU-96bef2b2-13fa-4778-bf5f-b9b490dbe3da",
    "vbios": "92.00.A0.00.05",
    "driver": "580.95.05",
    "power_limit_w": 300.0
   },
   "windows": [
    {
     "name": "idle 1",
     "seconds": 5.004081366999998,
     "joules": 399.32,
     "utilization_max": 0,
     "contexts": 1
    },
    {
     "name": "reference",
     "seconds": 5.180338663999976,
     "joules": 1143.101,
     "matmuls": 696,
     "dim": 4096,
     "exact": true
    },
    {
     "name": "idle 2",
     "seconds": 5.003959261000006,
     "joules": 407.927,
     "utilization_max": 0,
     "contexts": 1
    },
    {
     "name": "control",
     "seconds": 5.003630724000004,
     "joules": 470.903,
     "calls": 3636
    },
    {
     "name": "idle 3",
     "seconds": 5.00341699400002,
     "joules": 404.361,
     "utilization_max": 0,
     "contexts": 1
    },
    {
     "name": "method",
     "seconds": 20.819870863000006,
     "joules": 4467.51,
     "calls": 6,
     "ms": [
      3466.205078125,
      3465.884765625,
      3466.013671875,
      3466.503173828125,
      3465.957275390625,
      3466.260498046875
     ],
     "correct": [
      9795,
      9794,
      9813,
      9809,
      9818,
      9808
     ]
    },
    {
     "name": "idle 4",
     "seconds": 5.0035917909999625,
     "joules": 414.096,
     "utilization_max": 0,
     "contexts": 1
    }
   ]
  }
 }
}
