connorgoggins/random_sampling_all_ops_gpu_5_warmup_5_runs.md Secret

## random_sampling_all_ops_gpu_5_warmup_5_runs.md

      
    Raw
  

              random_sampling_all_ops_gpu_5_warmup_5_runs.md
            
          
    Runtime Features


BLAS_APPLE : ✖ BLAS_APPLE
BLAS_ATLAS : ✖ BLAS_ATLAS
BLAS_MKL : ✖ BLAS_MKL
BLAS_OPEN : ✔ BLAS_OPEN
CAFFE : ✖ CAFFE
CPU_AVX : ✔ CPU_AVX
CPU_AVX2 : ✖ CPU_AVX2
CPU_SSE : ✔ CPU_SSE
CPU_SSE2 : ✔ CPU_SSE2
CPU_SSE3 : ✔ CPU_SSE3
CPU_SSE4A : ✖ CPU_SSE4A
CPU_SSE4_1 : ✔ CPU_SSE4_1
CPU_SSE4_2 : ✔ CPU_SSE4_2
CUDA : ✔ CUDA
CUDA_RTC : ✖ CUDA_RTC
CUDNN : ✔ CUDNN
CXX14 : ✖ CXX14
DEBUG : ✖ DEBUG
DIST_KVSTORE : ✖ DIST_KVSTORE
F16C : ✔ F16C
INT64_TENSOR_SIZE : ✖ INT64_TENSOR_SIZE
JEMALLOC : ✖ JEMALLOC
LAPACK : ✔ LAPACK
MKLDNN : ✔ MKLDNN
NCCL : ✖ NCCL
OPENCV : ✖ OPENCV
OPENMP : ✔ OPENMP
PROFILER : ✖ PROFILER
SIGNAL_HANDLER : ✖ SIGNAL_HANDLER
SSE : ✖ SSE
TENSORRT : ✖ TENSORRT
TVM_OP : ✖ TVM_OP

Benchmark Results


Operator
Avg Forward Time (ms)
Avg. Backward Time (ms)
Max Mem Usage (Storage) (Bytes)
Inputs


BatchNorm
1.8868
3.0016
54781.8867
{'data': (32, 3, 256, 256), 'gamma': (3,), 'beta': (3,), 'moving_mean': (3,), 'moving_var': (3,)}


BatchNorm
3.1232
4.5588
80598.9766
{'data': (32, 3, 10000, 10), 'gamma': (3,), 'beta': (3,), 'moving_mean': (3,), 'moving_var': (3,)}


BilinearSampler
3.1522
4.2186
16777.2168
{'data': (32, 2, 256, 256), 'grid': (32, 2, 256, 256)}


BlockGrad
0.1528
---
2097.1521
{'data': (1024, 1024)}


BlockGrad
0.168
---
20.0
{'data': (10000, 1)}


BlockGrad
0.1518
---
2000.0
{'data': (10000, 100)}


Convolution
0.1784
0.2836
182852.9531
{'data': (32, 3, 256), 'weight': (64, 3, 3), 'bias': (64,), 'kernel': (3,), 'stride': (1,), 'dilate': (1,), 'pad': (0,), 'num_filter': 64, 'layout': 'NCW'}


Deconvolution
0.2878
0.3518
814107.1875
{'data': (32, 3, 256), 'weight': (3, 64, 3), 'bias': (64,), 'kernel': (3,), 'stride': (1,), 'dilate': (1,), 'pad': (0,), 'adj': (0,), 'num_filter': 64, 'no_bias': False, 'layout': 'NCW'}


Dropout
0.48
0.3756
41798.9648
{'data': (32, 3, 256, 256), 'p': 0.5, 'mode': 'always'}


Dropout
0.069
0.0448
17233.1406
{'data': (10000, 10), 'p': 0.5, 'mode': 'always'}


FullyConnected
1.1966
1.9694
8.192
{'data': (32, 3, 256, 256), 'num_hidden': 64, 'weight': (64, 196608), 'bias': (64,), 'flatten': True}


FullyConnected
1.1624
1.4946
6291.4561
{'data': (32, 3, 256, 256), 'num_hidden': 64, 'weight': (64, 256), 'bias': (64,), 'flatten': False}


GridGenerator
1.6232
1.1788
8650.752
{'data': (32, 2, 256, 256), 'transform_type': 'warp'}


LeakyReLU
0.1088
0.1276
67690.4219
{'data': (1024, 1024), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.1216
0.0426
65633.2656
{'data': (10000, 1), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.1026
0.1208
69613.2656
{'data': (10000, 100), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.1116
0.1206
71807.5703
{'data': (1024, 1024), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.0478
0.0488
69750.4219
{'data': (10000, 1), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.1234
0.1256
73730.4219
{'data': (10000, 100), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.1266
0.135
75924.7266
{'data': (1024, 1024), 'act_type': 'selu'}


LeakyReLU
0.0668
0.0476
73867.5703
{'data': (10000, 1), 'act_type': 'selu'}


LeakyReLU
0.1158
0.1232
77847.5703
{'data': (10000, 100), 'act_type': 'selu'}


LeakyReLU
0.1426
0.3824
80041.875
{'data': (1024, 1024), 'act_type': 'prelu', 'gamma': (1, 1024)}


LeakyReLU
0.0646
0.0908
77984.7266
{'data': (10000, 1), 'act_type': 'prelu', 'gamma': (1, 1)}


LeakyReLU
0.1404
0.3404
81964.7266
{'data': (10000, 100), 'act_type': 'prelu', 'gamma': (1, 100)}


Pooling
0.0764
0.057
100755.0938
{'data': (32, 3, 256), 'kernel': 3, 'pool_type': 'avg', 'global_pool': 0, 'stride': 1, 'pad': 1}


abs
0.145
0.1388
4194.3042
{'data': (1024, 1024)}


abs
0.131
0.0464
20.0
{'data': (10000, 1)}


abs
0.128
0.1326
4000.0
{'data': (10000, 100)}


adam_update
0.2538
---
748871.0
{'weight': (1024, 1024), 'grad': (1024, 1024), 'mean': (1024, 1024), 'var': (1024, 1024), 'lr': 0.1, 'beta1': 0.1, 'beta2': 0.1, 'epsilon': 1e-08, 'rescale_grad': 0.4, 'lazy_update': 0}


adam_update
0.0546
---
746793.875
{'weight': (10000, 1), 'grad': (10000, 1), 'mean': (10000, 1), 'var': (10000, 1), 'lr': 0.5, 'beta1': 0.5, 'beta2': 0.5, 'epsilon': 1e-08, 'rescale_grad': 0.4, 'lazy_update': 0}


adam_update
0.2554
---
750793.875
{'weight': (10000, 100), 'grad': (10000, 100), 'mean': (10000, 100), 'var': (10000, 100), 'lr': 0.9, 'beta1': 0.9, 'beta2': 0.9, 'epsilon': 1e-08, 'rescale_grad': 0.4, 'lazy_update': 0}


arccos
0.1692
0.1828
4194.3042
{'data': (1024, 1024)}


arccos
0.0544
0.0526
20.0
{'data': (10000, 1)}


arccos
0.166
0.1734
4000.0
{'data': (10000, 100)}


arccosh
0.1826
0.1768
4194.3042
{'data': (1024, 1024)}


arccosh
0.0568
0.053
40.0
{'data': (10000, 1)}


arccosh
0.1748
0.1674
4000.0
{'data': (10000, 100)}


arcsin
0.1658
0.1754
4194.3042
{'data': (1024, 1024)}


arcsin
0.0582
0.046
20.0
{'data': (10000, 1)}


arcsin
0.1564
0.1706
4000.0
{'data': (10000, 100)}


arcsinh
0.1588
0.1512
4194.3042
{'data': (1024, 1024)}


arcsinh
0.051
0.042
20.0
{'data': (10000, 1)}


arcsinh
0.1428
0.1446
2000.0
{'data': (10000, 100)}


arctan
0.1444
0.1396
2097.1521
{'data': (1024, 1024)}


arctan
0.0532
0.042
20.0
{'data': (10000, 1)}


arctan
0.143
0.1354
2000.0
{'data': (10000, 100)}


arctanh
0.1348
0.1398
2097.1521
{'data': (1024, 1024)}


arctanh
0.049
0.04
20.0
{'data': (10000, 1)}


arctanh
0.1306
0.1418
2000.0
{'data': (10000, 100)}


argmax
0.7928
---
2.048
{'data': (1024, 1024), 'axis': 0}


argmax
5.2594
---
0.002
{'data': (10000, 1), 'axis': 0}


argmax
7.1912
---
0.2
{'data': (10000, 100), 'axis': 0}


argmax_channel
1.8202
---
4.096
{'data': (1024, 1024)}


argmax_channel
0.0474
---
40.0
{'data': (10000, 1)}


argmax_channel
0.4128
---
20.0
{'data': (10000, 100)}


argmin
0.7884
---
2.048
{'data': (1024, 1024), 'axis': 0}


argmin
5.2424
---
0.002
{'data': (10000, 1), 'axis': 0}


argmin
7.1766
---
0.2
{'data': (10000, 100), 'axis': 0}


argsort
5.6142
---
2097.1521
{'data': (1024, 1024), 'axis': 0}


argsort
0.5848
---
20.0
{'data': (10000, 1), 'axis': 0}


argsort
5.4684
---
4000.0
{'data': (10000, 100), 'axis': 0}


batch_dot
31.1065
---
67108.8672
{'lhs': (32, 1024, 1024), 'rhs': (32, 1024, 1024)}


batch_dot
2.9587
---
64000.0
{'lhs': (32, 1000, 10), 'rhs': (32, 1000, 10), 'transpose_b': True}


batch_dot
0.3451
---
6.4
{'lhs': (32, 1000, 1), 'rhs': (32, 100, 1000), 'transpose_a': True, 'transpose_b': True}


broadcast_add
0.1498
0.198
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_add
0.283
0.0536
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_add
0.1458
0.048
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_div
0.1632
0.301
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_div
0.0726
0.0626
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_div
0.0528
0.0502
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_equal
0.1498
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_equal
0.0472
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_equal
0.043
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_greater
0.1614
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_greater
0.061
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_greater
0.055
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_greater_equal
0.1698
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_greater_equal
0.0612
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_greater_equal
0.0562
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_hypot
0.1816
0.334
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_hypot
0.0582
0.0718
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_hypot
0.0452
0.0606
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_lesser
0.1706
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_lesser
0.064
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_lesser
0.0606
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_lesser_equal
0.1544
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_lesser_equal
0.0574
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_lesser_equal
0.051
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_logical_and
0.1706
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_logical_and
0.0604
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_and
0.0552
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_logical_or
0.1486
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_logical_or
0.0632
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_or
0.0534
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_logical_xor
0.168
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_logical_xor
0.0634
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_xor
0.0598
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_maximum
0.1694
0.2976
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_maximum
0.0536
0.0634
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_maximum
0.0456
0.0556
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_minimum
0.1724
0.3008
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_minimum
0.06
0.0662
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_minimum
0.0536
0.054
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_minus
0.1604
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_minus
0.0674
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_minus
0.0594
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_mod
0.5136
0.2524
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_mod
0.0978
0.0648
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_mod
0.0606
0.0506
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_mul
0.163
0.2388
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_mul
0.0584
0.0606
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_mul
0.0466
0.0576
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_not_equal
0.1744
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_not_equal
0.066
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_not_equal
0.059
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_plus
0.1682
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_plus
0.065
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_plus
0.0638
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_power
0.2116
0.3432
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_power
0.0674
0.078
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_power
0.0552
0.0546
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_sub
0.1622
0.1834
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_sub
0.0602
0.058
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_sub
0.0516
0.0488
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


cbrt
0.1292
0.1408
4194.3042
{'data': (1024, 1024)}


cbrt
0.047
0.0352
20.0
{'data': (10000, 1)}


cbrt
0.1192
0.1296
2000.0
{'data': (10000, 100)}


ceil
0.1404
---
4194.3042
{'data': (1024, 1024)}


ceil
0.064
---
40.0
{'data': (10000, 1)}


ceil
0.1348
---
2000.0
{'data': (10000, 100)}


cos
0.145
0.1404
4194.3042
{'data': (1024, 1024)}


cos
0.0568
0.0506
20.0
{'data': (10000, 1)}


cos
0.2172
0.1316
2000.0
{'data': (10000, 100)}


cosh
0.1358
0.1406
4194.3042
{'data': (1024, 1024)}


cosh
0.0536
0.0416
20.0
{'data': (10000, 1)}


cosh
0.1184
0.136
2000.0
{'data': (10000, 100)}


degrees
0.1328
0.107
2097.1521
{'data': (1024, 1024)}


degrees
0.0604
0.0454
20.0
{'data': (10000, 1)}


degrees
0.1178
0.1004
2000.0
{'data': (10000, 100)}


depth_to_space
0.06
---
0.128
{'data': (1, 4, 2, 4), 'block_size': 2}


depth_to_space
0.407
---
500.0
{'data': (10, 25, 10, 100), 'block_size': 5}


dot
1.2968
2.8156
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


dot
0.1986
0.4518
2000.0
{'lhs': (1000, 10), 'rhs': (1000, 10), 'transpose_b': True}


dot
0.3662
0.0856
0.2
{'lhs': (1000, 1), 'rhs': (100, 1000), 'transpose_a': True, 'transpose_b': True}


elemwise_add
0.164
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_add
0.0646
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_add
0.061
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_div
0.1612
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_div
0.0608
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_div
0.0572
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_mul
0.168
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_mul
0.0648
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_mul
0.0618
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_sub
0.1638
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_sub
0.0778
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_sub
0.063
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


erf
0.1394
0.1382
2097.1521
{'data': (1024, 1024)}


erf
0.049
0.043
20.0
{'data': (10000, 1)}


erf
0.1326
0.1296
2000.0
{'data': (10000, 100)}


erfinv
0.7784
0.1408
4194.3042
{'data': (1024, 1024)}


erfinv
0.0608
0.048
40.0
{'data': (10000, 1)}


erfinv
0.7246
0.1318
2000.0
{'data': (10000, 100)}


exp
0.1398
---
2097.1521
{'data': (1024, 1024)}


exp
0.0576
---
40.0
{'data': (10000, 1)}


exp
0.131
---
2000.0
{'data': (10000, 100)}


expm1
0.1268
0.1374
2097.1521
{'data': (1024, 1024)}


expm1
0.0444
0.042
20.0
{'data': (10000, 1)}


expm1
0.127
0.1266
2000.0
{'data': (10000, 100)}


fix
0.1434
---
2097.1521
{'data': (1024, 1024)}


fix
0.0618
---
40.0
{'data': (10000, 1)}


fix
0.1404
---
2000.0
{'data': (10000, 100)}


flatten
0.1428
---
2097.1521
{'data': (1024, 1024)}


flatten
0.0818
---
20.0
{'data': (10000, 1)}


flatten
0.15
---
2000.0
{'data': (10000, 100)}


flip
0.1494
---
2097.1521
{'data': (1024, 1024), 'axis': 0}


flip
0.0572
---
40.0
{'data': (10000, 1), 'axis': 0}


flip
0.1438
---
2000.0
{'data': (10000, 100), 'axis': 0}


floor
0.1394
---
4194.3042
{'data': (1024, 1024)}


floor
0.0738
---
20.0
{'data': (10000, 1)}


floor
0.1388
---
2000.0
{'data': (10000, 100)}


ftml_update
0.3468
---
755085.3125
{'weight': (1024, 1024), 'grad': (1024, 1024), 'd': (1024, 1024), 'v': (1024, 1024), 'z': (1024, 1024), 'lr': 0.1, 'beta1': 0.1, 'beta2': 0.1, 'epsilon': 1e-08, 't': 1, 'rescale_grad': 0.4, 'clip_grad': -1.0}


ftml_update
0.0628
---
753008.125
{'weight': (10000, 1), 'grad': (10000, 1), 'd': (10000, 1), 'v': (10000, 1), 'z': (10000, 1), 'lr': 0.5, 'beta1': 0.5, 'beta2': 0.5, 'epsilon': 1e-08, 't': 1, 'rescale_grad': 0.4, 'clip_grad': -1.0}


ftml_update
0.3464
---
757008.125
{'weight': (10000, 100), 'grad': (10000, 100), 'd': (10000, 100), 'v': (10000, 100), 'z': (10000, 100), 'lr': 0.9, 'beta1': 0.9, 'beta2': 0.9, 'epsilon': 1e-08, 't': 1, 'rescale_grad': 0.4, 'clip_grad': -1.0}


ftrl_update
0.3416
---
761299.625
{'weight': (1024, 1024), 'grad': (1024, 1024), 'z': (1024, 1024), 'n': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


ftrl_update
0.0608
---
759222.4375
{'weight': (10000, 1), 'grad': (10000, 1), 'z': (10000, 1), 'n': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


ftrl_update
0.3272
---
761222.4375
{'weight': (10000, 100), 'grad': (10000, 100), 'z': (10000, 100), 'n': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


gamma
0.2116
0.637
2097.1521
{'data': (1024, 1024)}


gamma
0.0752
0.0664
40.0
{'data': (10000, 1)}


gamma
0.2062
0.6068
2000.0
{'data': (10000, 100)}


gammaln
0.246
0.5472
2097.1521
{'data': (1024, 1024)}


gammaln
0.0648
0.0522
40.0
{'data': (10000, 1)}


gammaln
0.2328
0.5188
4000.0
{'data': (10000, 100)}


hard_sigmoid
0.1212
0.133
86256.1797
{'data': (1024, 1024), 'alpha': 0.25, 'beta': 0.5}


hard_sigmoid
0.072
0.0482
84199.0312
{'data': (10000, 1), 'alpha': 0.25, 'beta': 0.5}


hard_sigmoid
0.1182
0.134
88179.0312
{'data': (10000, 100), 'alpha': 0.25, 'beta': 0.5}


identity
---
---
2097.1521
{'data': (1024, 1024)}


identity
---
---
20.0
{'data': (10000, 1)}


identity
---
---
4000.0
{'data': (10000, 100)}


log
0.137
0.1398
4194.3042
{'data': (1024, 1024)}


log
0.0526
0.0454
20.0
{'data': (10000, 1)}


log
0.1288
0.1356
2000.0
{'data': (10000, 100)}


log10
0.1316
0.141
2097.1521
{'data': (1024, 1024)}


log10
0.0464
0.0426
20.0
{'data': (10000, 1)}


log10
0.1196
0.135
2000.0
{'data': (10000, 100)}


log1p
0.1334
0.1426
2097.1521
{'data': (1024, 1024)}


log1p
0.0454
0.043
20.0
{'data': (10000, 1)}


log1p
0.137
0.143
4000.0
{'data': (10000, 100)}


log2
0.1336
0.143
2097.1521
{'data': (1024, 1024)}


log2
0.054
0.0456
20.0
{'data': (10000, 1)}


log2
0.1288
0.14
2000.0
{'data': (10000, 100)}


log_softmax
0.19
0.1562
100684.7891
{'data': (1024, 1024), 'axis': -1, 'temperature': 0.5}


log_softmax
0.3862
0.2212
98627.6328
{'data': (10000, 1), 'axis': -1, 'temperature': 0.5}


log_softmax
0.191
0.1744
102607.6328
{'data': (10000, 100), 'axis': -1, 'temperature': 0.5}


logical_not
0.1416
---
2097.1521
{'data': (1024, 1024)}


logical_not
0.066
---
40.0
{'data': (10000, 1)}


logical_not
0.1384
---
2000.0
{'data': (10000, 100)}


make_loss
0.145
---
2097.1521
{'data': (1024, 1024)}


make_loss
0.0772
---
20.0
{'data': (10000, 1)}


make_loss
0.1506
---
2000.0
{'data': (10000, 100)}


max
0.1472
0.4112
2.048
{'data': (1024, 1024), 'axis': 0}


max
0.0962
0.0612
0.002
{'data': (10000, 1), 'axis': 0}


max
0.1434
0.4114
0.2
{'data': (10000, 100), 'axis': 0}


max_axis
0.148
---
2.048
{'data': (1024, 1024), 'axis': 0}


max_axis
0.0728
---
0.004
{'data': (10000, 1), 'axis': 0}


max_axis
0.135
---
0.2
{'data': (10000, 100), 'axis': 0}


mean
0.1624
0.4206
2.048
{'data': (1024, 1024), 'axis': 0}


mean
0.0764
0.053
0.002
{'data': (10000, 1), 'axis': 0}


mean
0.1562
0.3948
0.4
{'data': (10000, 100), 'axis': 0}


min
0.1682
0.4236
4.096
{'data': (1024, 1024), 'axis': 0}


min
0.0892
0.0566
0.002
{'data': (10000, 1), 'axis': 0}


min
0.1442
0.4044
0.4
{'data': (10000, 100), 'axis': 0}


min_axis
0.1622
---
2.048
{'data': (1024, 1024), 'axis': 0}


min_axis
0.0908
---
0.002
{'data': (10000, 1), 'axis': 0}


min_axis
0.1466
---
0.2
{'data': (10000, 100), 'axis': 0}


mp_sgd_mom_update
0.2172
---
765416.75
{'weight': (1024, 1024), 'grad': (1024, 1024), 'mom': (1024, 1024), 'weight32': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_mom_update
0.0654
---
765436.75
{'weight': (10000, 1), 'grad': (10000, 1), 'mom': (10000, 1), 'weight32': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_mom_update
0.2078
---
767436.75
{'weight': (10000, 100), 'grad': (10000, 100), 'mom': (10000, 100), 'weight32': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_update
0.1586
---
771631.0625
{'weight': (1024, 1024), 'grad': (1024, 1024), 'weight32': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_update
0.0616
---
771651.0625
{'weight': (10000, 1), 'grad': (10000, 1), 'weight32': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_update
0.1566
---
773651.0625
{'weight': (10000, 100), 'grad': (10000, 100), 'weight32': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


nanprod
0.1514
0.4164
4.096
{'data': (1024, 1024), 'axis': 0}


nanprod
0.0638
0.0584
0.002
{'data': (10000, 1), 'axis': 0}


nanprod
0.1468
0.4008
0.4
{'data': (10000, 100), 'axis': 0}


nansum
0.1736
0.4152
4.096
{'data': (1024, 1024), 'axis': 0}


nansum
0.0684
0.05
0.002
{'data': (10000, 1), 'axis': 0}


nansum
0.1436
0.3926
0.4
{'data': (10000, 100), 'axis': 0}


negative
0.1388
---
2097.1521
{'data': (1024, 1024)}


negative
0.0568
---
40.0
{'data': (10000, 1)}


negative
0.1358
---
2000.0
{'data': (10000, 100)}


ones_like
0.0832
---
4194.3042
{'data': (1024, 1024)}


ones_like
0.058
---
40.0
{'data': (10000, 1)}


ones_like
0.0824
---
4000.0
{'data': (10000, 100)}


prod
0.1356
0.414
4.096
{'data': (1024, 1024), 'axis': 0}


prod
0.0684
0.046
0.002
{'data': (10000, 1), 'axis': 0}


prod
0.1078
0.4052
0.2
{'data': (10000, 100), 'axis': 0}


radians
0.1132
0.0992
2097.1521
{'data': (1024, 1024)}


radians
0.043
0.0416
20.0
{'data': (10000, 1)}


radians
0.1188
0.0958
2000.0
{'data': (10000, 100)}


random_exponential
4.2034
---
2097.1521
{'shape': (1024, 1024)}


random_exponential
0.064
---
20.0
{'shape': (10000, 1)}


random_exponential
3.595
---
2000.0
{'shape': (10000, 100)}


random_gamma
7.982
---
4194.3042
{'shape': (1024, 1024)}


random_gamma
0.0988
---
20.0
{'shape': (10000, 1)}


random_gamma
7.4146
---
2000.0
{'shape': (10000, 100)}


random_generalized_negative_binomial
12.6448
---
4194.3042
{'shape': (1024, 1024)}


random_generalized_negative_binomial
0.1362
---
20.0
{'shape': (10000, 1)}


random_generalized_negative_binomial
10.8568
---
2000.0
{'shape': (10000, 100)}


random_negative_binomial
9.7516
---
2097.1521
{'k': 1, 'p': 1, 'shape': (1024, 1024)}


random_negative_binomial
0.1168
---
20.0
{'k': 1, 'p': 1, 'shape': (10000, 1)}


random_negative_binomial
8.6556
---
2000.0
{'k': 1, 'p': 1, 'shape': (10000, 100)}


random_normal
3.3208
---
4194.3042
{'shape': (1024, 1024)}


random_normal
0.0676
---
20.0
{'shape': (10000, 1)}


random_normal
3.596
---
2000.0
{'shape': (10000, 100)}


random_pdf_dirichlet
0.0692
---
0.002
{'sample': (2,), 'alpha': [0.0, 2.5]}


random_pdf_exponential
0.0622
---
0.004
{'sample': (2,), 'lam': [1.0, 8.5]}


random_pdf_gamma
0.08
---
0.004
{'sample': (2,), 'alpha': [0.0, 2.5], 'beta': [1.0, 0.7]}


random_pdf_generalized_negative_binomial
0.0826
---
0.004
{'sample': (2,), 'mu': [2.0, 2.5], 'alpha': [0.0, 2.5]}


random_pdf_negative_binomial
0.0804
---
0.004
{'sample': (2,), 'k': [20, 49], 'p': [0.4, 0.77]}


random_pdf_normal
0.0782
---
0.004
{'sample': (2,), 'mu': [2.0, 2.5], 'sigma': [1.0, 3.7]}


random_pdf_poisson
0.0774
---
0.004
{'sample': (2,), 'lam': [1.0, 8.5]}


random_pdf_uniform
0.0688
---
0.008
{'sample': (2,), 'low': [0.0, 2.5], 'high': [1.0, 3.7]}


random_poisson
2.7044
---
2097.1521
{'shape': (1024, 1024)}


random_poisson
0.0444
---
20.0
{'shape': (10000, 1)}


random_poisson
2.3736
---
2000.0
{'shape': (10000, 100)}


random_randint
2.0544
---
2097.1521
{'low': 0, 'high': 5, 'shape': (1024, 1024)}


random_randint
0.0404
---
20.0
{'low': 0, 'high': 5, 'shape': (10000, 1)}


random_randint
1.7864
---
2000.0
{'low': 0, 'high': 5, 'shape': (10000, 100)}


random_uniform
0.6446
---
2097.1521
{'low': 0, 'high': 5, 'shape': (1024, 1024)}


random_uniform
0.0242
---
20.0
{'low': 0, 'high': 5, 'shape': (10000, 1)}


random_uniform
0.6248
---
2000.0
{'low': 0, 'high': 5, 'shape': (10000, 100)}


rcbrt
0.1246
0.1422
2097.1521
{'data': (1024, 1024)}


rcbrt
0.0418
0.0346
20.0
{'data': (10000, 1)}


rcbrt
0.1186
0.1364
2000.0
{'data': (10000, 100)}


reciprocal
0.1104
0.141
2097.1521
{'data': (1024, 1024)}


reciprocal
0.0396
0.0328
20.0
{'data': (10000, 1)}


reciprocal
0.1086
0.1436
2000.0
{'data': (10000, 100)}


relu
0.1198
0.1326
4194.3042
{'data': (1024, 1024)}


relu
0.0482
0.0344
20.0
{'data': (10000, 1)}


relu
0.1102
0.1268
2000.0
{'data': (10000, 100)}


rint
0.1308
---
2097.1521
{'data': (1024, 1024)}


rint
0.0582
---
40.0
{'data': (10000, 1)}


rint
0.1384
---
2000.0
{'data': (10000, 100)}


rmsprop_update
0.2006
---
779942.5
{'weight': (1024, 1024), 'grad': (1024, 1024), 'n': (1024, 1024), 'lr': 0.1, 'gamma1': 0.1, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmsprop_update
0.063
---
777865.375
{'weight': (10000, 1), 'grad': (10000, 1), 'n': (10000, 1), 'lr': 0.5, 'gamma1': 0.5, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmsprop_update
0.2098
---
779865.375
{'weight': (10000, 100), 'grad': (10000, 100), 'n': (10000, 100), 'lr': 0.9, 'gamma1': 0.9, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmspropalex_update
0.3118
---
786156.8125
{'weight': (1024, 1024), 'grad': (1024, 1024), 'n': (1024, 1024), 'g': (1024, 1024), 'delta': (1024, 1024), 'lr': 0.1, 'gamma1': 0.1, 'gamma2': 0.1, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmspropalex_update
0.0684
---
784079.6875
{'weight': (10000, 1), 'grad': (10000, 1), 'n': (10000, 1), 'g': (10000, 1), 'delta': (10000, 1), 'lr': 0.5, 'gamma1': 0.5, 'gamma2': 0.5, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmspropalex_update
0.2998
---
788079.6875
{'weight': (10000, 100), 'grad': (10000, 100), 'n': (10000, 100), 'g': (10000, 100), 'delta': (10000, 100), 'lr': 0.9, 'gamma1': 0.9, 'gamma2': 0.9, 'epsilon': 1e-08, 'rescale_grad': 0.4}


round
0.1496
---
2097.1521
{'data': (1024, 1024)}


round
0.0588
---
20.0
{'data': (10000, 1)}


round
0.135
---
2000.0
{'data': (10000, 100)}


rsqrt
0.1802
0.1782
4194.3042
{'data': (1024, 1024)}


rsqrt
0.0444
0.034
20.0
{'data': (10000, 1)}


rsqrt
0.15
0.1704
2000.0
{'data': (10000, 100)}


sample_exponential
2.216
---
4194.3042
{'lam': [1.0, 8.5], 'shape': (1024, 1024)}


sample_exponential
0.3616
---
40.0
{'lam': [1.0, 8.5], 'shape': (10000, 1)}


sample_exponential
2.3746
---
4000.0
{'lam': [1.0, 8.5], 'shape': (10000, 100)}


sample_gamma
2.4246
---
4194.3042
{'alpha': [0.0, 2.5], 'shape': (1024, 1024), 'beta': [1.0, 0.7]}


sample_gamma
0.7358
---
80.0
{'alpha': [0.0, 2.5], 'shape': (10000, 1), 'beta': [1.0, 0.7]}


sample_gamma
2.2882
---
8000.0
{'alpha': [0.0, 2.5], 'shape': (10000, 100), 'beta': [1.0, 0.7]}


sample_generalized_negative_binomial
5.2148
---
8388.6084
{'mu': [2.0, 2.5], 'shape': (1024, 1024), 'alpha': [0.0, 2.5]}


sample_generalized_negative_binomial
1.4182
---
80.0
{'mu': [2.0, 2.5], 'shape': (10000, 1), 'alpha': [0.0, 2.5]}


sample_generalized_negative_binomial
4.9334
---
8000.0
{'mu': [2.0, 2.5], 'shape': (10000, 100), 'alpha': [0.0, 2.5]}


sample_multinomial
6315.751
---
67108.8672
{'data': (32, 32), 'shape': (1024, 1024)}


sample_multinomial
60.8282
---
1280.0
{'data': (32, 32), 'shape': (10000, 1)}


sample_multinomial
5990.1196
---
64000.0
{'data': (32, 32), 'shape': (10000, 100)}


sample_negative_binomial
7.7644
---
8388.6084
{'k': [20, 49], 'shape': (1024, 1024), 'p': [0.4, 0.77]}


sample_negative_binomial
2.0136
---
80.0
{'k': [20, 49], 'shape': (10000, 1), 'p': [0.4, 0.77]}


sample_negative_binomial
7.719
---
8000.0
{'k': [20, 49], 'shape': (10000, 100), 'p': [0.4, 0.77]}


sample_normal
2.0614
---
8388.6084
{'mu': [2.0, 2.5], 'shape': (1024, 1024), 'sigma': [1.0, 3.7]}


sample_normal
0.1656
---
40.0
{'mu': [2.0, 2.5], 'shape': (10000, 1), 'sigma': [1.0, 3.7]}


sample_normal
1.9742
---
8000.0
{'mu': [2.0, 2.5], 'shape': (10000, 100), 'sigma': [1.0, 3.7]}


sample_poisson
3.5398
---
8388.6084
{'lam': [1.0, 8.5], 'shape': (1024, 1024)}


sample_poisson
0.8752
---
80.0
{'lam': [1.0, 8.5], 'shape': (10000, 1)}


sample_poisson
3.0426
---
8000.0
{'lam': [1.0, 8.5], 'shape': (10000, 100)}


sample_uniform
2.014
---
8388.6084
{'low': [0.0, 2.5], 'shape': (1024, 1024), 'high': [1.0, 3.7]}


sample_uniform
0.1252
---
40.0
{'low': [0.0, 2.5], 'shape': (10000, 1), 'high': [1.0, 3.7]}


sample_uniform
1.8966
---
8000.0
{'low': [0.0, 2.5], 'shape': (10000, 100), 'high': [1.0, 3.7]}


sgd_mom_update
0.1904
---
790274.0
{'weight': (1024, 1024), 'grad': (1024, 1024), 'mom': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_mom_update
0.0554
---
790314.0
{'weight': (10000, 1), 'grad': (10000, 1), 'mom': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_mom_update
0.1804
---
792294.0
{'weight': (10000, 100), 'grad': (10000, 100), 'mom': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.1394
---
798585.4375
{'weight': (1024, 1024), 'grad': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.0526
---
796528.25
{'weight': (10000, 1), 'grad': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.1448
---
798508.25
{'weight': (10000, 100), 'grad': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


shape_array
0.0464
---
0.008
{'data': (1024, 1024)}


shape_array
0.0502
---
0.008
{'data': (10000, 1)}


shape_array
0.0504
---
0.008
{'data': (10000, 100)}


shuffle
0.2984
---
2097.1521
{'data': (1024, 1024)}


shuffle
1.1608
---
60.0
{'data': (10000, 1)}


shuffle
0.948
---
2000.0
{'data': (10000, 100)}


sigmoid
0.1548
0.1354
2097.1521
{'data': (1024, 1024)}


sigmoid
0.0564
0.0494
20.0
{'data': (10000, 1)}


sigmoid
0.1272
0.1354
4000.0
{'data': (10000, 100)}


sign
0.1244
0.1084
4194.3042
{'data': (1024, 1024)}


sign
0.0456
0.0414
20.0
{'data': (10000, 1)}


sign
0.1194
0.0962
2000.0
{'data': (10000, 100)}


signsgd_update
0.1386
---
804799.75
{'weight': (1024, 1024), 'grad': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


signsgd_update
0.0542
---
802722.5625
{'weight': (10000, 1), 'grad': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


signsgd_update
0.1348
---
806722.5625
{'weight': (10000, 100), 'grad': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


signum_update
0.1884
---
808916.875
{'weight': (1024, 1024), 'grad': (1024, 1024), 'mom': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


signum_update
0.0546
---
808956.875
{'weight': (10000, 1), 'grad': (10000, 1), 'mom': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


signum_update
0.1814
---
810936.875
{'weight': (10000, 100), 'grad': (10000, 100), 'mom': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


sin
0.1238
0.1336
2097.1521
{'data': (1024, 1024)}


sin
0.0514
0.0428
20.0
{'data': (10000, 1)}


sin
0.1212
0.1278
2000.0
{'data': (10000, 100)}


sinh
0.1404
0.1344
2097.1521
{'data': (1024, 1024)}


sinh
0.0494
0.0418
20.0
{'data': (10000, 1)}


sinh
0.129
0.1274
2000.0
{'data': (10000, 100)}


size_array
0.0562
---
0.008
{'data': (1024, 1024)}


size_array
0.127
---
0.008
{'data': (10000, 1)}


size_array
0.0442
---
0.008
{'data': (10000, 100)}


softmax
0.2136
0.1568
92470.4844
{'data': (1024, 1024), 'axis': -1, 'temperature': 0.5}


softmax
0.3888
0.224
90413.3281
{'data': (10000, 1), 'axis': -1, 'temperature': 0.5}


softmax
0.2164
0.1734
94393.3281
{'data': (10000, 100), 'axis': -1, 'temperature': 0.5}


softsign
0.1282
0.1414
2097.1521
{'data': (1024, 1024)}


softsign
0.0482
0.0446
20.0
{'data': (10000, 1)}


softsign
0.121
0.1356
2000.0
{'data': (10000, 100)}


sort
5.889
---
4194.3042
{'data': (1024, 1024), 'axis': 0}


sort
0.6066
---
40.0
{'data': (10000, 1), 'axis': 0}


sort
5.7126
---
6000.0
{'data': (10000, 100), 'axis': 0}


space_to_depth
0.062
---
0.128
{'data': (1, 4, 2, 4), 'block_size': 2}


space_to_depth
0.158
---
500.0
{'data': (10, 25, 10, 100), 'block_size': 5}


sqrt
0.1312
0.1522
2097.1521
{'data': (1024, 1024)}


sqrt
0.046
0.0486
20.0
{'data': (10000, 1)}


sqrt
0.1214
0.1452
2000.0
{'data': (10000, 100)}


square
0.1068
0.1288
2097.1521
{'data': (1024, 1024)}


square
0.045
0.0424
20.0
{'data': (10000, 1)}


square
0.1084
0.1232
2000.0
{'data': (10000, 100)}


stop_gradient
---
---
2097.1521
{'data': (1024, 1024)}


stop_gradient
---
---
40.0
{'data': (10000, 1)}


stop_gradient
---
---
2000.0
{'data': (10000, 100)}


sum
0.177
0.3356
4.096
{'data': (1024, 1024), 'axis': 0}


sum
0.0678
0.0424
0.002
{'data': (10000, 1), 'axis': 0}


sum
0.1254
0.3388
0.4
{'data': (10000, 100), 'axis': 0}


sum_axis
0.1798
---
2.048
{'data': (1024, 1024), 'axis': 0}


sum_axis
0.0728
---
0.004
{'data': (10000, 1), 'axis': 0}


sum_axis
0.145
---
0.2
{'data': (10000, 100), 'axis': 0}


swapaxes
0.3194
---
2097.1521
{'data': (1024, 1024), 'dim1': 0, 'dim2': 1}


swapaxes
0.0424
---
40.0
{'data': (10000, 1), 'dim1': 0, 'dim2': 1}


swapaxes
0.3238
---
2000.0
{'data': (10000, 100), 'dim1': 0, 'dim2': 1}


tan
0.1472
0.1312
2097.1521
{'data': (1024, 1024)}


tan
0.0488
0.0368
20.0
{'data': (10000, 1)}


tan
0.14
0.1252
2000.0
{'data': (10000, 100)}


tanh
0.1326
0.1308
2097.1521
{'data': (1024, 1024)}


tanh
0.0462
0.0432
20.0
{'data': (10000, 1)}


tanh
0.127
0.1316
2000.0
{'data': (10000, 100)}


topk
0.5652
---
4.096
{'data': (1024, 1024), 'axis': 0, 'k': 1}


topk
0.1332
---
0.002
{'data': (10000, 1), 'axis': 0, 'k': 1}


topk
0.4552
---
0.2
{'data': (10000, 100), 'axis': 0, 'k': 1}


transpose
0.1128
---
2097.1521
{'data': (1024, 1024)}


transpose
0.057
---
40.0
{'data': (10000, 1)}


transpose
0.1328
---
4000.0
{'data': (10000, 100)}


trunc
0.1412
---
2097.1521
{'data': (1024, 1024)}


trunc
0.0534
---
20.0
{'data': (10000, 1)}


trunc
0.116
---
2000.0
{'data': (10000, 100)}


zeros_like
0.0902
---
4194.3042
{'data': (1024, 1024)}


zeros_like
0.0578
---
20.0
{'data': (10000, 1)}


zeros_like
0.0834
---
4000.0
{'data': (10000, 100)}