ChaiBapchya/mxnet_operator_benchmark_results_array_manipulation_cpu.md

## mxnet_operator_benchmark_results_array_manipulation_cpu.md

      
    Raw
  

              mxnet_operator_benchmark_results_array_manipulation_cpu.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
2.4547
1.3703
25165.8359
{'data': (32, 3, 256, 256), 'gamma': (3,), 'beta': (3,), 'moving_mean': (3,), 'moving_var': (3,)}


BatchNorm
6.0028
1.5685
38400.0117
{'data': (32, 3, 10000, 10), 'gamma': (3,), 'beta': (3,), 'moving_mean': (3,), 'moving_var': (3,)}


BlockGrad
0.5237
---
2097.1521
{'data': (1024, 1024)}


BlockGrad
0.0145
---
20.0
{'data': (10000, 1)}


BlockGrad
0.5857
---
2000.0
{'data': (10000, 100)}


CTCLoss
---
---
3492195.5
{'data': (1024, 100, 100), 'label': (100, 100)}


Convolution
0.5066
0.9342
98815.9375
{'data': (32, 3, 256), 'weight': (64, 3, 3), 'bias': (64,), 'kernel': (3,), 'stride': (1,), 'dilate': (1,), 'pad': (0,), 'num_filter': 64, 'layout': 'NCW'}


Deconvolution
5.2545
8.364
2835571.75
{'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
2.8362
0.6007
37748.7344
{'data': (32, 3, 256, 256), 'p': 0.5, 'mode': 'always'}


Dropout
0.2912
0.0546
400.0
{'data': (10000, 10), 'p': 0.5, 'mode': 'always'}


FullyConnected
2.465
86.5237
8.192
{'data': (32, 3, 256, 256), 'num_hidden': 64, 'weight': (64, 196608), 'bias': (64,), 'flatten': True}


FullyConnected
1.1209
32.8333
6291.4561
{'data': (32, 3, 256, 256), 'num_hidden': 64, 'weight': (64, 256), 'bias': (64,), 'flatten': False}


LeakyReLU
0.0696
0.0817
2097.1521
{'data': (1024, 1024), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.0486
0.0466
40.0
{'data': (10000, 1), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.0854
0.0839
2000.0
{'data': (10000, 100), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.1382
0.7319
2097.1521
{'data': (1024, 1024), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.0844
0.0497
20.0
{'data': (10000, 1), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.1359
0.6828
2000.0
{'data': (10000, 100), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.7239
0.2762
2097.1521
{'data': (1024, 1024), 'act_type': 'selu'}


LeakyReLU
0.0382
0.0347
20.0
{'data': (10000, 1), 'act_type': 'selu'}


LeakyReLU
0.7685
0.3158
4000.0
{'data': (10000, 100), 'act_type': 'selu'}


LeakyReLU
0.4101
8.3719
4194.3042
{'data': (1024, 1024), 'act_type': 'prelu', 'gamma': (1, 1024)}


LeakyReLU
0.0425
0.7114
20.0
{'data': (10000, 1), 'act_type': 'prelu', 'gamma': (1, 1)}


LeakyReLU
0.3988
7.9939
4000.0
{'data': (10000, 100), 'act_type': 'prelu', 'gamma': (1, 100)}


MakeLoss
0.2846
0.0365
3478006.75
{'data': (1024, 1024)}


MakeLoss
0.0324
0.0239
3475949.75
{'data': (10000, 1)}


MakeLoss
0.086
0.036
3479929.75
{'data': (10000, 100)}


Pooling
0.2163
0.2732
49.152
{'data': (32, 3, 256), 'kernel': 3, 'pool_type': 'avg', 'global_pool': 0, 'stride': 1, 'pad': 1}


SwapAxis
3.0576
3.1598
4194.3042
{'data': (1024, 1024), 'dim1': 0, 'dim2': 1}


SwapAxis
0.0548
0.0503
20.0
{'data': (10000, 1), 'dim1': 0, 'dim2': 1}


SwapAxis
2.5586
2.5601
2000.0
{'data': (10000, 100), 'dim1': 0, 'dim2': 1}


abs
0.0555
0.2705
2097.1521
{'data': (1024, 1024)}


abs
0.0163
0.0775
20.0
{'data': (10000, 1)}


abs
0.0544
0.3541
4000.0
{'data': (10000, 100)}


adam_update
0.3242
---
2770335.75
{'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
0.0376
---
2768278.5
{'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
0.6383
---
2772258.5
{'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
0.8253
0.6489
2097.1521
{'data': (1024, 1024)}


arccos
0.0488
0.0268
20.0
{'data': (10000, 1)}


arccos
0.8131
0.7397
2000.0
{'data': (10000, 100)}


arccosh
0.7674
0.8093
2097.1521
{'data': (1024, 1024)}


arccosh
0.0329
0.0293
20.0
{'data': (10000, 1)}


arccosh
0.7672
0.7945
2000.0
{'data': (10000, 100)}


arcsin
0.752
0.6436
2097.1521
{'data': (1024, 1024)}


arcsin
0.0312
0.0273
20.0
{'data': (10000, 1)}


arcsin
0.774
0.7634
2000.0
{'data': (10000, 100)}


arcsinh
1.4861
0.5057
4194.3042
{'data': (1024, 1024)}


arcsinh
0.0447
0.0333
40.0
{'data': (10000, 1)}


arcsinh
1.7063
0.6787
2000.0
{'data': (10000, 100)}


arctan
0.635
0.0452
2097.1521
{'data': (1024, 1024)}


arctan
0.0384
0.0246
20.0
{'data': (10000, 1)}


arctan
0.883
0.0811
2000.0
{'data': (10000, 100)}


arctanh
1.307
0.0802
2097.1521
{'data': (1024, 1024)}


arctanh
0.0547
0.0343
20.0
{'data': (10000, 1)}


arctanh
1.2393
0.0824
4000.0
{'data': (10000, 100)}


argmax
18.4832
---
2.048
{'data': (1024, 1024), 'axis': 0}


argmax
0.1647
---
0.002
{'data': (10000, 1), 'axis': 0}


argmax
14.4155
---
0.2
{'data': (10000, 100), 'axis': 0}


argmax_channel
0.6482
---
2.048
{'data': (1024, 1024)}


argmax_channel
0.0435
---
20.0
{'data': (10000, 1)}


argmax_channel
0.4859
---
20.0
{'data': (10000, 100)}


argmin
17.7454
---
2.048
{'data': (1024, 1024), 'axis': 0}


argmin
0.1595
---
0.002
{'data': (10000, 1), 'axis': 0}


argmin
14.3292
---
0.2
{'data': (10000, 100), 'axis': 0}


argsort
64.5372
---
2097.1521
{'data': (1024, 1024), 'axis': 0}


argsort
1.3778
---
20.0
{'data': (10000, 1), 'axis': 0}


argsort
59.7691
---
2000.0
{'data': (10000, 100), 'axis': 0}


batch_dot
535.6865
---
134217.7344
{'lhs': (32, 1024, 1024), 'rhs': (32, 1024, 1024)}


batch_dot
380.7793
---
128000.0
{'lhs': (32, 1000, 10), 'rhs': (32, 1000, 10), 'transpose_b': True}


batch_dot
4.269
---
12.8
{'lhs': (32, 1000, 1), 'rhs': (32, 100, 1000), 'transpose_a': True, 'transpose_b': True}


broadcast_add
0.0805
0.104
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_add
0.0389
0.0496
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_add
0.0286
0.0267
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_axes
0.0367
---
4.096
{'data': (1, 1024), 'axis': 0, 'size': 2}


broadcast_axes
0.0364
---
0.004
{'data': (1, 1), 'axis': 0, 'size': 2}


broadcast_axes
0.0357
---
0.4
{'data': (1, 100), 'axis': 0, 'size': 2}


broadcast_axis
0.0362
---
4.096
{'data': (1, 1024), 'axis': 0, 'size': 2}


broadcast_axis
0.0357
---
0.004
{'data': (1, 1), 'axis': 0, 'size': 2}


broadcast_axis
0.0362
---
0.4
{'data': (1, 100), 'axis': 0, 'size': 2}


broadcast_div
0.0696
0.1254
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_div
0.0392
0.0562
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_div
0.025
0.029
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_equal
0.0679
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_equal
0.0364
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_equal
0.0218
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_greater
0.065
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_greater
0.1909
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_greater
0.0237
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_greater_equal
0.0641
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_greater_equal
0.0362
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_greater_equal
0.0237
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_hypot
0.4676
1.0331
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_hypot
0.0744
0.1468
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_hypot
0.0369
0.0512
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_lesser
0.0637
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_lesser
0.0384
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_lesser
0.0235
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_lesser_equal
0.0665
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_lesser_equal
0.0366
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_lesser_equal
0.0236
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_like
0.9105
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_like
0.1282
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_like
0.052
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_logical_and
0.0664
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_logical_and
0.0367
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_and
0.0243
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_logical_or
0.0673
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_logical_or
0.0351
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_or
0.0248
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


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


broadcast_logical_xor
0.043
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_xor
0.0342
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_maximum
0.0975
0.1265
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_maximum
0.0435
0.0544
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_maximum
0.041
0.0382
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_minimum
0.0929
0.1232
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_minimum
0.0391
0.0526
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_minimum
0.0348
0.0348
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_minus
0.0547
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_minus
0.0365
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_minus
0.0248
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_mod
1.452
0.1408
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_mod
0.2005
0.0536
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_mod
0.0432
0.0611
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_mul
0.0719
0.1064
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_mul
0.0398
0.054
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_mul
0.0259
0.0287
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_not_equal
0.0593
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_not_equal
0.0362
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_not_equal
0.0223
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_plus
0.0623
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_plus
0.0379
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_plus
0.0235
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_power
2.4128
5.668
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_power
0.2574
0.5698
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_power
0.1101
0.0918
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_sub
0.0693
0.0848
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_sub
0.0387
0.0507
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_sub
0.0263
0.0257
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_to
0.9184
---
4194.3042
{'data': (1, 1024), 'shape': (1024, 1024)}


broadcast_to
0.0463
---
20.0
{'data': (1, 1), 'shape': (10000, 1)}


broadcast_to
1.4525
---
4000.0
{'data': (1, 100), 'shape': (10000, 100)}


cbrt
1.3946
0.0813
4194.3042
{'data': (1024, 1024)}


cbrt
0.0418
0.024
20.0
{'data': (10000, 1)}


cbrt
1.4034
0.0819
4000.0
{'data': (10000, 100)}


ceil
0.1926
---
2097.1521
{'data': (1024, 1024)}


ceil
0.054
---
20.0
{'data': (10000, 1)}


ceil
0.1758
---
4000.0
{'data': (10000, 100)}


cos
0.54
0.5433
4194.3042
{'data': (1024, 1024)}


cos
0.0329
0.0309
20.0
{'data': (10000, 1)}


cos
0.532
0.5122
2000.0
{'data': (10000, 100)}


cosh
0.6319
1.0282
4194.3042
{'data': (1024, 1024)}


cosh
0.0358
0.0352
20.0
{'data': (10000, 1)}


cosh
0.8056
1.4169
2000.0
{'data': (10000, 100)}


ctc_loss
---
2.9964
3559850.25
{'data': (1024, 100, 100), 'label': (100, 100)}


degrees
0.0807
0.0472
4194.3042
{'data': (1024, 1024)}


degrees
0.0264
0.021
40.0
{'data': (10000, 1)}


degrees
0.1904
0.0483
4000.0
{'data': (10000, 100)}


depth_to_space
0.0502
---
0.064
{'data': (1, 4, 2, 4), 'block_size': 2}


depth_to_space
0.9073
---
1000.0
{'data': (10, 25, 10, 100), 'block_size': 5}


diag
0.035
0.3348
2.046
{'data': (1024, 1024), 'k': 1}


diag
0.0298
0.046
---
{'data': (10000, 1), 'k': 1}


diag
0.05
0.3138
0.198
{'data': (10000, 100), 'k': 1}


dot
3.8458
5.8184
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


dot
0.4278
4.5966
4000.0
{'lhs': (1000, 10), 'rhs': (1000, 10), 'transpose_b': True}


dot
0.1662
0.24
0.2
{'lhs': (1000, 1), 'rhs': (100, 1000), 'transpose_a': True, 'transpose_b': True}


elemwise_add
0.0886
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_add
0.0635
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_add
0.0541
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_div
0.0763
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_div
0.0389
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_div
0.0224
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_mul
0.0768
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_mul
0.0383
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_mul
0.0217
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_sub
0.0767
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_sub
0.0375
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_sub
0.0221
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


erf
0.8126
0.6297
4194.3042
{'data': (1024, 1024)}


erf
0.0429
0.0347
40.0
{'data': (10000, 1)}


erf
0.8759
0.7495
4000.0
{'data': (10000, 100)}


erfinv
2.9241
0.6029
4194.3042
{'data': (1024, 1024)}


erfinv
0.0591
0.0358
20.0
{'data': (10000, 1)}


erfinv
3.3703
0.748
4000.0
{'data': (10000, 100)}


exp
0.6265
---
2097.1521
{'data': (1024, 1024)}


exp
0.032
---
20.0
{'data': (10000, 1)}


exp
0.604
---
2000.0
{'data': (10000, 100)}


expand_dims
0.0704
---
2097.1521
{'data': (1024, 1024), 'axis': 0}


expand_dims
0.0521
---
20.0
{'data': (10000, 1), 'axis': 0}


expand_dims
0.0698
---
2000.0
{'data': (10000, 100), 'axis': 0}


expm1
1.025
0.6943
4194.3042
{'data': (1024, 1024)}


expm1
0.0402
0.0322
40.0
{'data': (10000, 1)}


expm1
0.9514
0.6478
2000.0
{'data': (10000, 100)}


fix
0.3445
---
2097.1521
{'data': (1024, 1024)}


fix
0.0292
---
20.0
{'data': (10000, 1)}


fix
0.3399
---
4000.0
{'data': (10000, 100)}


flatten
0.0562
---
4194.3042
{'data': (1024, 1024)}


flatten
0.0369
---
20.0
{'data': (10000, 1)}


flatten
0.0532
---
2000.0
{'data': (10000, 100)}


flip
1.2305
---
2097.1521
{'data': (1024, 1024), 'axis': 0}


flip
0.0402
---
40.0
{'data': (10000, 1), 'axis': 0}


flip
1.1872
---
4000.0
{'data': (10000, 100), 'axis': 0}


floor
0.1904
---
4194.3042
{'data': (1024, 1024)}


floor
0.0619
---
40.0
{'data': (10000, 1)}


floor
0.1711
---
4000.0
{'data': (10000, 100)}


ftml_update
1.5383
---
2776550.0
{'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
0.0437
---
2774492.75
{'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
1.6264
---
2778472.75
{'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
1.1971
---
2782764.25
{'weight': (1024, 1024), 'grad': (1024, 1024), 'z': (1024, 1024), 'n': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


ftrl_update
0.0393
---
2780707.25
{'weight': (10000, 1), 'grad': (10000, 1), 'z': (10000, 1), 'n': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


ftrl_update
3.3391
---
2784687.25
{'weight': (10000, 100), 'grad': (10000, 100), 'z': (10000, 100), 'n': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


gamma
3.812
6.915
2097.1521
{'data': (1024, 1024)}


gamma
0.0609
0.0861
20.0
{'data': (10000, 1)}


gamma
3.6748
6.6537
2000.0
{'data': (10000, 100)}


gammaln
27.0604
2.3544
2097.1521
{'data': (1024, 1024)}


gammaln
0.2901
0.0429
20.0
{'data': (10000, 1)}


gammaln
25.7833
2.0524
2000.0
{'data': (10000, 100)}


hard_sigmoid
0.1153
0.1333
2097.1521
{'data': (1024, 1024), 'alpha': 0.25, 'beta': 0.5}


hard_sigmoid
0.0367
0.032
40.0
{'data': (10000, 1), 'alpha': 0.25, 'beta': 0.5}


hard_sigmoid
0.1108
0.1274
4000.0
{'data': (10000, 100), 'alpha': 0.25, 'beta': 0.5}


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


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


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


log
0.6242
0.0443
2097.1521
{'data': (1024, 1024)}


log
0.0389
0.025
20.0
{'data': (10000, 1)}


log
0.7693
0.0647
2000.0
{'data': (10000, 100)}


log10
1.0121
0.0825
4194.3042
{'data': (1024, 1024)}


log10
0.0394
0.0222
20.0
{'data': (10000, 1)}


log10
0.9536
0.0782
4000.0
{'data': (10000, 100)}


log1p
1.0159
0.0818
2097.1521
{'data': (1024, 1024)}


log1p
0.0405
0.0229
20.0
{'data': (10000, 1)}


log1p
0.9541
0.0797
2000.0
{'data': (10000, 100)}


log2
0.8358
0.0774
4194.3042
{'data': (1024, 1024)}


log2
0.0412
0.0225
40.0
{'data': (10000, 1)}


log2
0.7859
0.0756
2000.0
{'data': (10000, 100)}


log_softmax
1.6662
0.7953
2097.1521
{'data': (1024, 1024), 'axis': -1, 'temperature': 0.5}


log_softmax
0.0543
0.0452
20.0
{'data': (10000, 1), 'axis': -1, 'temperature': 0.5}


log_softmax
1.5958
0.7584
4000.0
{'data': (10000, 100), 'axis': -1, 'temperature': 0.5}


logical_not
0.0484
---
4194.3042
{'data': (1024, 1024)}


logical_not
0.0203
---
20.0
{'data': (10000, 1)}


logical_not
0.0487
---
4000.0
{'data': (10000, 100)}


make_loss
0.3954
---
2097.1521
{'data': (1024, 1024)}


make_loss
0.0193
---
20.0
{'data': (10000, 1)}


make_loss
0.3792
---
4000.0
{'data': (10000, 100)}


max
1.7964
1.5455
4.096
{'data': (1024, 1024), 'axis': 0}


max
0.3288
0.0406
0.002
{'data': (10000, 1), 'axis': 0}


max
1.0818
0.9315
0.2
{'data': (10000, 100), 'axis': 0}


max_axis
1.7976
---
4.096
{'data': (1024, 1024), 'axis': 0}


max_axis
0.3277
---
0.002
{'data': (10000, 1), 'axis': 0}


max_axis
1.5417
---
0.4
{'data': (10000, 100), 'axis': 0}


mean
1.9617
2.1988
4.096
{'data': (1024, 1024), 'axis': 0}


mean
0.5004
0.0677
0.004
{'data': (10000, 1), 'axis': 0}


mean
1.4045
1.8053
0.2
{'data': (10000, 100), 'axis': 0}


min
1.0851
0.978
2.048
{'data': (1024, 1024), 'axis': 0}


min
0.3317
0.0362
0.002
{'data': (10000, 1), 'axis': 0}


min
1.2383
1.4078
0.4
{'data': (10000, 100), 'axis': 0}


min_axis
1.7617
---
4.096
{'data': (1024, 1024), 'axis': 0}


min_axis
0.265
---
0.002
{'data': (10000, 1), 'axis': 0}


min_axis
1.0856
---
0.2
{'data': (10000, 100), 'axis': 0}


mp_sgd_mom_update
0.2292
---
2788978.5
{'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
0.045
---
2786921.5
{'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
0.2199
---
2790901.5
{'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
0.1828
---
2795193.0
{'weight': (1024, 1024), 'grad': (1024, 1024), 'weight32': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_update
0.0876
---
2793135.75
{'weight': (10000, 1), 'grad': (10000, 1), 'weight32': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_update
0.1723
---
2797115.75
{'weight': (10000, 100), 'grad': (10000, 100), 'weight32': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


nanprod
1.3615
1.1304
2.048
{'data': (1024, 1024), 'axis': 0}


nanprod
0.3334
0.0379
0.002
{'data': (10000, 1), 'axis': 0}


nanprod
1.3297
1.4296
0.4
{'data': (10000, 100), 'axis': 0}


nansum
1.8905
1.4722
4.096
{'data': (1024, 1024), 'axis': 0}


nansum
0.3359
0.0407
0.002
{'data': (10000, 1), 'axis': 0}


nansum
1.6127
1.411
0.4
{'data': (10000, 100), 'axis': 0}


negative
0.0472
---
4194.3042
{'data': (1024, 1024)}


negative
0.0205
---
40.0
{'data': (10000, 1)}


negative
0.1123
---
4000.0
{'data': (10000, 100)}


ones_like
0.0431
---
2097.1521
{'data': (1024, 1024)}


ones_like
0.0187
---
40.0
{'data': (10000, 1)}


ones_like
0.0431
---
2000.0
{'data': (10000, 100)}


pad
---
---
0.192
{'data': (1, 4, 2, 4), 'mode': 'constant', 'pad_width': (0, 0, 0, 0, 1, 1, 1, 1)}


pad
---
---
612.0
{'data': (10, 25, 10, 100), 'mode': 'constant', 'pad_width': (0, 0, 0, 0, 1, 1, 1, 1)}


prod
1.849
1.4723
2.048
{'data': (1024, 1024), 'axis': 0}


prod
0.3253
0.0415
0.002
{'data': (10000, 1), 'axis': 0}


prod
1.5403
1.4172
0.4
{'data': (10000, 100), 'axis': 0}


radians
0.0587
0.0526
2097.1521
{'data': (1024, 1024)}


radians
0.0251
0.0219
40.0
{'data': (10000, 1)}


radians
0.0551
0.0502
2000.0
{'data': (10000, 100)}


random_exponential
2.5054
---
4194.3042
{'shape': (1024, 1024)}


random_exponential
0.0538
---
20.0
{'shape': (10000, 1)}


random_exponential
2.4009
---
4000.0
{'shape': (10000, 100)}


random_gamma
7.2308
---
4194.3042
{'shape': (1024, 1024)}


random_gamma
0.0958
---
40.0
{'shape': (10000, 1)}


random_gamma
6.9175
---
4000.0
{'shape': (10000, 100)}


random_generalized_negative_binomial
9.6304
---
4194.3042
{'shape': (1024, 1024)}


random_generalized_negative_binomial
0.1185
---
20.0
{'shape': (10000, 1)}


random_generalized_negative_binomial
8.3206
---
4000.0
{'shape': (10000, 100)}


random_negative_binomial
8.1951
---
4194.3042
{'k': 1, 'p': 1, 'shape': (1024, 1024)}


random_negative_binomial
0.1167
---
20.0
{'k': 1, 'p': 1, 'shape': (10000, 1)}


random_negative_binomial
7.0832
---
4000.0
{'k': 1, 'p': 1, 'shape': (10000, 100)}


random_normal
2.7703
---
4194.3042
{'shape': (1024, 1024)}


random_normal
0.052
---
20.0
{'shape': (10000, 1)}


random_normal
2.8538
---
4000.0
{'shape': (10000, 100)}


random_pdf_dirichlet
0.036
---
0.004
{'sample': (2,), 'alpha': [0.0, 2.5]}


random_pdf_exponential
0.0362
---
0.004
{'sample': (2,), 'lam': [1.0, 8.5]}


random_pdf_gamma
0.0359
---
0.004
{'sample': (2,), 'alpha': [0.0, 2.5], 'beta': [1.0, 0.7]}


random_pdf_generalized_negative_binomial
0.0367
---
0.004
{'sample': (2,), 'mu': [2.0, 2.5], 'alpha': [0.0, 2.5]}


random_pdf_negative_binomial
0.0379
---
0.004
{'sample': (2,), 'k': [20, 49], 'p': [0.4, 0.77]}


random_pdf_normal
0.0382
---
0.004
{'sample': (2,), 'mu': [2.0, 2.5], 'sigma': [1.0, 3.7]}


random_pdf_poisson
0.0366
---
0.004
{'sample': (2,), 'lam': [1.0, 8.5]}


random_pdf_uniform
0.038
---
0.004
{'sample': (2,), 'low': [0.0, 2.5], 'high': [1.0, 3.7]}


random_poisson
1.7797
---
2097.1521
{'shape': (1024, 1024)}


random_poisson
0.0538
---
20.0
{'shape': (10000, 1)}


random_poisson
2.2778
---
4000.0
{'shape': (10000, 100)}


random_randint
1.6953
---
4194.3042
{'low': 0, 'high': 5, 'shape': (1024, 1024)}


random_randint
0.0494
---
20.0
{'low': 0, 'high': 5, 'shape': (10000, 1)}


random_randint
1.6299
---
4000.0
{'low': 0, 'high': 5, 'shape': (10000, 100)}


random_uniform
0.9135
---
4194.3042
{'low': 0, 'high': 5, 'shape': (1024, 1024)}


random_uniform
0.0407
---
20.0
{'low': 0, 'high': 5, 'shape': (10000, 1)}


random_uniform
0.8816
---
2000.0
{'low': 0, 'high': 5, 'shape': (10000, 100)}


rcbrt
1.3697
1.3793
2097.1521
{'data': (1024, 1024)}


rcbrt
0.0435
0.0393
20.0
{'data': (10000, 1)}


rcbrt
0.8965
0.8649
2000.0
{'data': (10000, 100)}


reciprocal
0.0572
0.0503
2097.1521
{'data': (1024, 1024)}


reciprocal
0.0265
0.0258
20.0
{'data': (10000, 1)}


reciprocal
0.0541
0.049
2000.0
{'data': (10000, 100)}


relu
0.0679
0.0725
2097.1521
{'data': (1024, 1024)}


relu
0.0307
0.0347
20.0
{'data': (10000, 1)}


relu
0.0631
0.0668
2000.0
{'data': (10000, 100)}


repeat
6.7251
9.2498
8388.6084
{'data': (1024, 1024), 'repeats': 2, 'axis': 0}


repeat
0.0821
0.0882
40.0
{'data': (10000, 1), 'repeats': 2, 'axis': 0}


repeat
6.6146
8.9965
8000.0
{'data': (10000, 100), 'repeats': 2, 'axis': 0}


reshape
0.1159
---
4194.3042
{'data': (1024, 1024), 'shape': (1024, 1024)}


reshape
0.0393
---
20.0
{'data': (10000, 1), 'shape': (10000, 1)}


reshape
0.1467
---
2000.0
{'data': (10000, 100), 'shape': (10000, 100)}


reshape_like
0.4189
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


reshape_like
0.0525
---
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


reshape_like
0.0183
---
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


reverse
1.348
1.2252
4194.3042
{'data': (1024, 1024), 'axis': 0}


reverse
0.0448
0.0382
20.0
{'data': (10000, 1), 'axis': 0}


reverse
1.2309
1.1925
2000.0
{'data': (10000, 100), 'axis': 0}


rint
0.2959
---
4194.3042
{'data': (1024, 1024)}


rint
0.0339
---
20.0
{'data': (10000, 1)}


rint
0.2864
---
2000.0
{'data': (10000, 100)}


rmsprop_update
0.3979
---
2801407.25
{'weight': (1024, 1024), 'grad': (1024, 1024), 'n': (1024, 1024), 'lr': 0.1, 'gamma1': 0.1, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmsprop_update
0.0359
---
2799350.0
{'weight': (10000, 1), 'grad': (10000, 1), 'n': (10000, 1), 'lr': 0.5, 'gamma1': 0.5, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmsprop_update
0.5375
---
2803330.0
{'weight': (10000, 100), 'grad': (10000, 100), 'n': (10000, 100), 'lr': 0.9, 'gamma1': 0.9, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmspropalex_update
0.5847
---
2807621.5
{'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
0.0501
---
2805564.25
{'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
1.0995
---
2809544.25
{'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
0.334
---
2097.1521
{'data': (1024, 1024)}


round
0.0812
---
40.0
{'data': (10000, 1)}


round
0.3124
---
2000.0
{'data': (10000, 100)}


rsqrt
0.8001
0.8138
4194.3042
{'data': (1024, 1024)}


rsqrt
0.04
0.0341
20.0
{'data': (10000, 1)}


rsqrt
0.9483
0.8989
4000.0
{'data': (10000, 100)}


sample_exponential
4.659
---
8388.6084
{'lam': [1.0, 8.5], 'shape': (1024, 1024)}


sample_exponential
0.0723
---
40.0
{'lam': [1.0, 8.5], 'shape': (10000, 1)}


sample_exponential
3.9018
---
4000.0
{'lam': [1.0, 8.5], 'shape': (10000, 100)}


sample_gamma
12.5227
---
8388.6084
{'alpha': [0.0, 2.5], 'shape': (1024, 1024), 'beta': [1.0, 0.7]}


sample_gamma
0.2556
---
40.0
{'alpha': [0.0, 2.5], 'shape': (10000, 1), 'beta': [1.0, 0.7]}


sample_gamma
16.9475
---
8000.0
{'alpha': [0.0, 2.5], 'shape': (10000, 100), 'beta': [1.0, 0.7]}


sample_generalized_negative_binomial
27.9329
---
4194.3042
{'mu': [2.0, 2.5], 'shape': (1024, 1024), 'alpha': [0.0, 2.5]}


sample_generalized_negative_binomial
0.304
---
40.0
{'mu': [2.0, 2.5], 'shape': (10000, 1), 'alpha': [0.0, 2.5]}


sample_generalized_negative_binomial
26.4842
---
8000.0
{'mu': [2.0, 2.5], 'shape': (10000, 100), 'alpha': [0.0, 2.5]}


sample_negative_binomial
348.2019
---
8388.6084
{'k': [20, 49], 'shape': (1024, 1024), 'p': [0.4, 0.77]}


sample_negative_binomial
3.392
---
80.0
{'k': [20, 49], 'shape': (10000, 1), 'p': [0.4, 0.77]}


sample_negative_binomial
333.3187
---
8000.0
{'k': [20, 49], 'shape': (10000, 100), 'p': [0.4, 0.77]}


sample_normal
5.4808
---
8388.6084
{'mu': [2.0, 2.5], 'shape': (1024, 1024), 'sigma': [1.0, 3.7]}


sample_normal
0.0865
---
40.0
{'mu': [2.0, 2.5], 'shape': (10000, 1), 'sigma': [1.0, 3.7]}


sample_normal
5.4253
---
8000.0
{'mu': [2.0, 2.5], 'shape': (10000, 100), 'sigma': [1.0, 3.7]}


sample_poisson
9.9675
---
8388.6084
{'lam': [1.0, 8.5], 'shape': (1024, 1024)}


sample_poisson
0.1336
---
80.0
{'lam': [1.0, 8.5], 'shape': (10000, 1)}


sample_poisson
8.4468
---
8000.0
{'lam': [1.0, 8.5], 'shape': (10000, 100)}


sample_uniform
1.8085
---
8388.6084
{'low': [0.0, 2.5], 'shape': (1024, 1024), 'high': [1.0, 3.7]}


sample_uniform
0.0503
---
40.0
{'low': [0.0, 2.5], 'shape': (10000, 1), 'high': [1.0, 3.7]}


sample_uniform
1.7414
---
4000.0
{'low': [0.0, 2.5], 'shape': (10000, 100), 'high': [1.0, 3.7]}


sgd_mom_update
0.2096
---
2813835.75
{'weight': (1024, 1024), 'grad': (1024, 1024), 'mom': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_mom_update
0.0435
---
2811778.75
{'weight': (10000, 1), 'grad': (10000, 1), 'mom': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_mom_update
0.2025
---
2815758.75
{'weight': (10000, 100), 'grad': (10000, 100), 'mom': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.184
---
2820050.0
{'weight': (1024, 1024), 'grad': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.0352
---
2817993.0
{'weight': (10000, 1), 'grad': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.1758
---
2821973.0
{'weight': (10000, 100), 'grad': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


shape_array
0.0154
---
0.016
{'data': (1024, 1024)}


shape_array
0.0167
---
0.016
{'data': (10000, 1)}


shape_array
0.0169
---
0.008
{'data': (10000, 100)}


shuffle
0.9945
---
4194.3042
{'data': (1024, 1024)}


shuffle
0.3834
---
20.0
{'data': (10000, 1)}


shuffle
1.2289
---
2000.0
{'data': (10000, 100)}


sigmoid
0.9567
0.0783
2097.1521
{'data': (1024, 1024)}


sigmoid
0.0308
0.0189
20.0
{'data': (10000, 1)}


sigmoid
0.7113
0.0588
2000.0
{'data': (10000, 100)}


sign
0.2668
0.0479
2097.1521
{'data': (1024, 1024)}


sign
0.0665
0.0161
20.0
{'data': (10000, 1)}


sign
0.2572
0.0461
2000.0
{'data': (10000, 100)}


signsgd_update
0.2092
---
2826264.5
{'weight': (1024, 1024), 'grad': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


signsgd_update
0.0347
---
2824207.25
{'weight': (10000, 1), 'grad': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


signsgd_update
0.1997
---
2828187.25
{'weight': (10000, 100), 'grad': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


signum_update
0.3595
---
2832478.75
{'weight': (1024, 1024), 'grad': (1024, 1024), 'mom': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


signum_update
0.038
---
2830421.5
{'weight': (10000, 1), 'grad': (10000, 1), 'mom': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


signum_update
0.3437
---
2834401.5
{'weight': (10000, 100), 'grad': (10000, 100), 'mom': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


sin
0.5153
0.5249
2097.1521
{'data': (1024, 1024)}


sin
0.0286
0.0261
20.0
{'data': (10000, 1)}


sin
0.4978
0.4983
2000.0
{'data': (10000, 100)}


sinh
1.6528
1.0047
4194.3042
{'data': (1024, 1024)}


sinh
0.0413
0.0326
20.0
{'data': (10000, 1)}


sinh
1.4752
0.9075
2000.0
{'data': (10000, 100)}


size_array
0.0118
---
0.008
{'data': (1024, 1024)}


size_array
0.012
---
0.004
{'data': (10000, 1)}


size_array
0.0119
---
0.008
{'data': (10000, 100)}


smooth_l1
0.3027
0.2908
3525181.5
{'data': (1024, 1024)}


smooth_l1
0.0653
0.059
3523124.5
{'data': (10000, 1)}


smooth_l1
0.4443
0.2827
3527104.5
{'data': (10000, 100)}


softmax
1.7664
0.1404
2097.1521
{'data': (1024, 1024), 'axis': -1, 'temperature': 0.5}


softmax
0.0518
0.0401
20.0
{'data': (10000, 1), 'axis': -1, 'temperature': 0.5}


softmax
1.7034
0.1508
2000.0
{'data': (10000, 100), 'axis': -1, 'temperature': 0.5}


softmax_cross_entropy
0.8424
---
3525104.5
{'data': (1024, 1024), 'label': (1024,)}


softsign
0.0648
0.0614
2097.1521
{'data': (1024, 1024)}


softsign
0.0205
0.0183
40.0
{'data': (10000, 1)}


softsign
0.1246
0.2147
2000.0
{'data': (10000, 100)}


sort
65.2196
---
6291.4561
{'data': (1024, 1024), 'axis': 0}


sort
1.4063
---
60.0
{'data': (10000, 1), 'axis': 0}


sort
60.064
---
6000.0
{'data': (10000, 100), 'axis': 0}


space_to_depth
0.0603
---
0.064
{'data': (1, 4, 2, 4), 'block_size': 2}


space_to_depth
1.1028
---
500.0
{'data': (10, 25, 10, 100), 'block_size': 5}


split
---
---
2097.1521
{'data': (1024, 1024), 'num_outputs': 1, 'axis': 0}


split
---
---
20.0
{'data': (10000, 1), 'num_outputs': 1, 'axis': 0}


split
---
---
2000.0
{'data': (10000, 100), 'num_outputs': 1, 'axis': 0}


sqrt
0.6781
0.0613
2097.1521
{'data': (1024, 1024)}


sqrt
0.0376
0.0253
20.0
{'data': (10000, 1)}


sqrt
0.7613
0.0631
4000.0
{'data': (10000, 100)}


square
0.1501
0.0604
4194.3042
{'data': (1024, 1024)}


square
0.0274
0.0226
40.0
{'data': (10000, 1)}


square
0.1878
0.0584
4000.0
{'data': (10000, 100)}


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


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


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


sum
1.9129
1.449
2.048
{'data': (1024, 1024), 'axis': 0}


sum
0.3474
0.0384
0.002
{'data': (10000, 1), 'axis': 0}


sum
1.6377
1.391
0.4
{'data': (10000, 100), 'axis': 0}


sum_axis
1.9243
---
4.096
{'data': (1024, 1024), 'axis': 0}


sum_axis
0.3433
---
0.004
{'data': (10000, 1), 'axis': 0}


sum_axis
1.6669
---
0.2
{'data': (10000, 100), 'axis': 0}


swapaxes
3.0577
---
4194.3042
{'data': (1024, 1024), 'dim1': 0, 'dim2': 1}


swapaxes
0.0575
---
20.0
{'data': (10000, 1), 'dim1': 0, 'dim2': 1}


swapaxes
2.7522
---
4000.0
{'data': (10000, 100), 'dim1': 0, 'dim2': 1}


tan
1.178
0.0876
2097.1521
{'data': (1024, 1024)}


tan
0.0437
0.0247
20.0
{'data': (10000, 1)}


tan
1.1238
0.0835
2000.0
{'data': (10000, 100)}


tanh
1.4582
0.0829
2097.1521
{'data': (1024, 1024)}


tanh
0.0471
0.0262
20.0
{'data': (10000, 1)}


tanh
1.3859
0.0836
4000.0
{'data': (10000, 100)}


tile
6.824
8.9998
8388.6084
{'data': (1024, 1024), 'reps': 2}


tile
0.169
0.1915
40.0
{'data': (10000, 1), 'reps': 2}


tile
6.7106
9.2244
8000.0
{'data': (10000, 100), 'reps': 2}


topk
58.6909
---
2.048
{'data': (1024, 1024), 'axis': 0, 'k': 1}


topk
0.625
---
0.004
{'data': (10000, 1), 'axis': 0, 'k': 1}


topk
55.8119
---
0.4
{'data': (10000, 100), 'axis': 0, 'k': 1}


transpose
0.0938
---
4194.3042
{'data': (1024, 1024)}


transpose
0.0436
---
40.0
{'data': (10000, 1)}


transpose
0.214
---
4000.0
{'data': (10000, 100)}


trunc
0.3037
---
4194.3042
{'data': (1024, 1024)}


trunc
0.0872
---
20.0
{'data': (10000, 1)}


trunc
0.3006
---
4000.0
{'data': (10000, 100)}


zeros_like
0.0465
---
4194.3042
{'data': (1024, 1024)}


zeros_like
0.0185
---
40.0
{'data': (10000, 1)}


zeros_like
0.0443
---
2000.0
{'data': (10000, 100)}


## mxnet_operator_benchmark_results_array_manipulation_gpu.md

      
    Raw
  

              mxnet_operator_benchmark_results_array_manipulation_gpu.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
1.187
2.3393
25165.8359
{'data': (32, 3, 256, 256), 'gamma': (3,), 'beta': (3,), 'moving_mean': (3,), 'moving_var': (3,)}


BatchNorm
1.8038
3.561
38400.0117
{'data': (32, 3, 10000, 10), 'gamma': (3,), 'beta': (3,), 'moving_mean': (3,), 'moving_var': (3,)}


BlockGrad
0.0461
---
2097.1521
{'data': (1024, 1024)}


BlockGrad
0.0697
---
40.0
{'data': (10000, 1)}


BlockGrad
0.043
---
4000.0
{'data': (10000, 100)}


CTCLoss
---
---
1259642.125
{'data': (1024, 100, 100), 'label': (100, 100)}


Convolution
0.1133
0.1788
17924.0508
{'data': (32, 3, 256), 'weight': (64, 3, 3), 'bias': (64,), 'kernel': (3,), 'stride': (1,), 'dilate': (1,), 'pad': (0,), 'num_filter': 64, 'layout': 'NCW'}


Deconvolution
0.1669
0.1472
648924.3125
{'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
0.1527
0.1032
37748.7344
{'data': (32, 3, 256, 256), 'p': 0.5, 'mode': 'always'}


Dropout
0.0665
0.0395
600.0
{'data': (10000, 10), 'p': 0.5, 'mode': 'always'}


FullyConnected
0.335
0.3779
8.192
{'data': (32, 3, 256, 256), 'num_hidden': 64, 'weight': (64, 196608), 'bias': (64,), 'flatten': True}


FullyConnected
0.1999
0.2843
6291.4561
{'data': (32, 3, 256, 256), 'num_hidden': 64, 'weight': (64, 256), 'bias': (64,), 'flatten': False}


LeakyReLU
0.0451
0.0484
4194.3042
{'data': (1024, 1024), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.0495
0.0348
40.0
{'data': (10000, 1), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.0431
0.0461
2000.0
{'data': (10000, 100), 'act_type': 'leaky', 'slope': 0.1}


LeakyReLU
0.0467
0.0501
2097.1521
{'data': (1024, 1024), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.032
0.0325
20.0
{'data': (10000, 1), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.0522
0.0519
4000.0
{'data': (10000, 100), 'act_type': 'elu', 'slope': 0.1}


LeakyReLU
0.0501
0.0504
4194.3042
{'data': (1024, 1024), 'act_type': 'selu'}


LeakyReLU
0.0608
0.0443
40.0
{'data': (10000, 1), 'act_type': 'selu'}


LeakyReLU
0.049
0.0504
4000.0
{'data': (10000, 100), 'act_type': 'selu'}


LeakyReLU
0.0734
0.1204
4194.3042
{'data': (1024, 1024), 'act_type': 'prelu', 'gamma': (1, 1024)}


LeakyReLU
0.0492
0.0748
40.0
{'data': (10000, 1), 'act_type': 'prelu', 'gamma': (1, 1)}


LeakyReLU
0.0695
0.1284
4000.0
{'data': (10000, 100), 'act_type': 'prelu', 'gamma': (1, 100)}


MakeLoss
0.0435
0.0333
1245453.375
{'data': (1024, 1024)}


MakeLoss
0.0292
0.0255
1243396.25
{'data': (10000, 1)}


MakeLoss
0.0389
0.0311
1247376.25
{'data': (10000, 100)}


Pooling
0.0242
0.0291
49.152
{'data': (32, 3, 256), 'kernel': 3, 'pool_type': 'avg', 'global_pool': 0, 'stride': 1, 'pad': 1}


SwapAxis
0.0563
0.0481
4194.3042
{'data': (1024, 1024), 'dim1': 0, 'dim2': 1}


SwapAxis
0.0315
0.0274
20.0
{'data': (10000, 1), 'dim1': 0, 'dim2': 1}


SwapAxis
0.0503
0.0433
2000.0
{'data': (10000, 100), 'dim1': 0, 'dim2': 1}


abs
0.0414
0.0412
4194.3042
{'data': (1024, 1024)}


abs
0.0278
0.025
20.0
{'data': (10000, 1)}


abs
0.0353
0.0369
4000.0
{'data': (10000, 100)}


adam_update
0.0801
---
583688.125
{'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
0.0481
---
581631.0
{'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
0.0792
---
585611.0
{'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
0.0373
0.0416
4194.3042
{'data': (1024, 1024)}


arccos
0.0255
0.024
20.0
{'data': (10000, 1)}


arccos
0.0375
0.0417
2000.0
{'data': (10000, 100)}


arccosh
0.047
0.0455
4194.3042
{'data': (1024, 1024)}


arccosh
0.0289
0.0271
40.0
{'data': (10000, 1)}


arccosh
0.0355
0.0417
2000.0
{'data': (10000, 100)}


arcsin
0.037
0.0442
2097.1521
{'data': (1024, 1024)}


arcsin
0.0241
0.0248
20.0
{'data': (10000, 1)}


arcsin
0.0329
0.041
2000.0
{'data': (10000, 100)}


arcsinh
0.0345
0.0414
2097.1521
{'data': (1024, 1024)}


arcsinh
0.0235
0.0255
20.0
{'data': (10000, 1)}


arcsinh
0.0335
0.0397
2000.0
{'data': (10000, 100)}


arctan
0.0496
0.0494
4194.3042
{'data': (1024, 1024)}


arctan
0.0294
0.0269
20.0
{'data': (10000, 1)}


arctan
0.0367
0.0411
2000.0
{'data': (10000, 100)}


arctanh
0.0473
0.0485
4194.3042
{'data': (1024, 1024)}


arctanh
0.0353
0.0312
20.0
{'data': (10000, 1)}


arctanh
0.0414
0.0414
2000.0
{'data': (10000, 100)}


argmax
0.3082
---
4.096
{'data': (1024, 1024), 'axis': 0}


argmax
1.3093
---
0.004
{'data': (10000, 1), 'axis': 0}


argmax
2.412
---
0.4
{'data': (10000, 100), 'axis': 0}


argmax_channel
0.3038
---
4.096
{'data': (1024, 1024)}


argmax_channel
0.0429
---
40.0
{'data': (10000, 1)}


argmax_channel
0.0662
---
40.0
{'data': (10000, 100)}


argmin
0.2968
---
4.096
{'data': (1024, 1024), 'axis': 0}


argmin
1.3119
---
0.004
{'data': (10000, 1), 'axis': 0}


argmin
2.4093
---
0.4
{'data': (10000, 100), 'axis': 0}


argsort
1.263
---
4194.3042
{'data': (1024, 1024), 'axis': 0}


argsort
0.2307
---
40.0
{'data': (10000, 1), 'axis': 0}


argsort
1.1869
---
4000.0
{'data': (10000, 100), 'axis': 0}


batch_dot
4.908
---
67108.8672
{'lhs': (32, 1024, 1024), 'rhs': (32, 1024, 1024)}


batch_dot
0.5305
---
64000.0
{'lhs': (32, 1000, 10), 'rhs': (32, 1000, 10), 'transpose_b': True}


batch_dot
0.0544
---
12.8
{'lhs': (32, 1000, 1), 'rhs': (32, 100, 1000), 'transpose_a': True, 'transpose_b': True}


broadcast_add
0.0667
0.0588
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_add
0.054
0.0408
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_add
0.052
0.0388
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_axes
0.038
---
8.192
{'data': (1, 1024), 'axis': 0, 'size': 2}


broadcast_axes
0.0383
---
0.008
{'data': (1, 1), 'axis': 0, 'size': 2}


broadcast_axes
0.0386
---
0.8
{'data': (1, 100), 'axis': 0, 'size': 2}


broadcast_axis
0.0386
---
8.192
{'data': (1, 1024), 'axis': 0, 'size': 2}


broadcast_axis
0.0384
---
0.008
{'data': (1, 1), 'axis': 0, 'size': 2}


broadcast_axis
0.0394
---
0.8
{'data': (1, 100), 'axis': 0, 'size': 2}


broadcast_div
0.0527
0.0756
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_div
0.0379
0.0427
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_div
0.0382
0.0419
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_equal
0.053
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_equal
0.0372
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_equal
0.0357
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_greater
0.0534
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_greater
0.0374
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_greater
0.035
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_greater_equal
0.0532
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_greater_equal
0.0383
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_greater_equal
0.0347
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_hypot
0.0512
0.0806
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_hypot
0.0378
0.0437
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_hypot
0.038
0.0429
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_lesser
0.0529
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_lesser
0.036
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_lesser
0.0343
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_lesser_equal
0.0533
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_lesser_equal
0.0361
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_lesser_equal
0.0364
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_like
0.1243
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_like
0.0439
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_like
0.0392
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_logical_and
0.0531
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_logical_and
0.0387
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_and
0.0334
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_logical_or
0.0485
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_logical_or
0.0373
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_or
0.0367
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_logical_xor
0.0535
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_logical_xor
0.0351
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_logical_xor
0.0357
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_maximum
0.0522
0.0771
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_maximum
0.0375
0.0413
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_maximum
0.0385
0.0419
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_minimum
0.0528
0.0799
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_minimum
0.0377
0.0425
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_minimum
0.0379
0.0422
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_minus
0.0529
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_minus
0.0382
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_minus
0.0357
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_mod
0.1119
0.062
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_mod
0.047
0.0443
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_mod
0.042
0.0424
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_mul
0.0533
0.0696
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_mul
0.038
0.0434
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_mul
0.0386
0.0433
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_not_equal
0.0542
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_not_equal
0.0365
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_not_equal
0.035
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_plus
0.0532
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_plus
0.0336
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_plus
0.0331
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_power
0.0513
0.0828
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_power
0.0387
0.0423
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_power
0.0383
0.0419
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_sub
0.0525
0.054
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


broadcast_sub
0.0366
0.0407
200.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


broadcast_sub
0.0376
0.0413
20.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


broadcast_to
0.1055
---
4194.3042
{'data': (1, 1024), 'shape': (1024, 1024)}


broadcast_to
0.0387
---
40.0
{'data': (1, 1), 'shape': (10000, 1)}


broadcast_to
0.0966
---
4000.0
{'data': (1, 100), 'shape': (10000, 100)}


cbrt
0.0459
0.048
2097.1521
{'data': (1024, 1024)}


cbrt
0.0358
0.0326
40.0
{'data': (10000, 1)}


cbrt
0.0422
0.0433
2000.0
{'data': (10000, 100)}


ceil
0.045
---
4194.3042
{'data': (1024, 1024)}


ceil
0.034
---
40.0
{'data': (10000, 1)}


ceil
0.0448
---
4000.0
{'data': (10000, 100)}


cos
0.0463
0.0475
2097.1521
{'data': (1024, 1024)}


cos
0.036
0.0334
40.0
{'data': (10000, 1)}


cos
0.0444
0.0433
2000.0
{'data': (10000, 100)}


cosh
0.0463
0.0469
4194.3042
{'data': (1024, 1024)}


cosh
0.0323
0.0308
20.0
{'data': (10000, 1)}


cosh
0.0412
0.0402
2000.0
{'data': (10000, 100)}


ctc_loss
---
0.1446
1286336.875
{'data': (1024, 100, 100), 'label': (100, 100)}


degrees
0.0458
0.0415
4194.3042
{'data': (1024, 1024)}


degrees
0.0335
0.0325
20.0
{'data': (10000, 1)}


degrees
0.0417
0.0347
2000.0
{'data': (10000, 100)}


depth_to_space
0.0373
---
0.128
{'data': (1, 4, 2, 4), 'block_size': 2}


depth_to_space
0.071
---
1000.0
{'data': (10, 25, 10, 100), 'block_size': 5}


diag
0.0369
0.0417
4.092
{'data': (1024, 1024), 'k': 1}


diag
0.0185
0.0335
---
{'data': (10000, 1), 'k': 1}


diag
0.0374
0.0407
0.198
{'data': (10000, 100), 'k': 1}


dot
0.2585
0.4344
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


dot
0.0725
0.096
4000.0
{'lhs': (1000, 10), 'rhs': (1000, 10), 'transpose_b': True}


dot
0.0806
0.0694
0.4
{'lhs': (1000, 1), 'rhs': (100, 1000), 'transpose_a': True, 'transpose_b': True}


elemwise_add
0.053
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_add
0.0349
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_add
0.0346
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_div
0.052
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_div
0.0352
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_div
0.0369
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_mul
0.0533
---
2097.1521
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_mul
0.0359
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_mul
0.0348
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


elemwise_sub
0.0538
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


elemwise_sub
0.0367
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


elemwise_sub
0.0369
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


erf
0.0466
0.0467
2097.1521
{'data': (1024, 1024)}


erf
0.0344
0.032
40.0
{'data': (10000, 1)}


erf
0.0457
0.0457
4000.0
{'data': (10000, 100)}


erfinv
0.122
0.0491
4194.3042
{'data': (1024, 1024)}


erfinv
0.0377
0.0318
20.0
{'data': (10000, 1)}


erfinv
0.1151
0.0459
2000.0
{'data': (10000, 100)}


exp
0.0457
---
4194.3042
{'data': (1024, 1024)}


exp
0.0312
---
40.0
{'data': (10000, 1)}


exp
0.0448
---
4000.0
{'data': (10000, 100)}


expand_dims
0.0533
---
4194.3042
{'data': (1024, 1024), 'axis': 0}


expand_dims
0.0438
---
40.0
{'data': (10000, 1), 'axis': 0}


expand_dims
0.0545
---
4000.0
{'data': (10000, 100), 'axis': 0}


expm1
0.047
0.0452
4194.3042
{'data': (1024, 1024)}


expm1
0.0305
0.0281
20.0
{'data': (10000, 1)}


expm1
0.0355
0.0401
2000.0
{'data': (10000, 100)}


fix
0.0471
---
4194.3042
{'data': (1024, 1024)}


fix
0.0361
---
40.0
{'data': (10000, 1)}


fix
0.0461
---
4000.0
{'data': (10000, 100)}


flatten
0.0522
---
4194.3042
{'data': (1024, 1024)}


flatten
0.0459
---
40.0
{'data': (10000, 1)}


flatten
0.0586
---
4000.0
{'data': (10000, 100)}


flip
0.0536
---
4194.3042
{'data': (1024, 1024), 'axis': 0}


flip
0.043
---
40.0
{'data': (10000, 1), 'axis': 0}


flip
0.0541
---
4000.0
{'data': (10000, 100), 'axis': 0}


floor
0.0452
---
4194.3042
{'data': (1024, 1024)}


floor
0.0359
---
40.0
{'data': (10000, 1)}


floor
0.0448
---
4000.0
{'data': (10000, 100)}


ftml_update
0.099
---
589902.4375
{'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
0.0409
---
587845.3125
{'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
0.1
---
591825.3125
{'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
0.0794
---
596116.75
{'weight': (1024, 1024), 'grad': (1024, 1024), 'z': (1024, 1024), 'n': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


ftrl_update
0.04
---
594059.625
{'weight': (10000, 1), 'grad': (10000, 1), 'z': (10000, 1), 'n': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


ftrl_update
0.0772
---
598039.625
{'weight': (10000, 100), 'grad': (10000, 100), 'z': (10000, 100), 'n': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


gamma
0.0675
0.0893
4194.3042
{'data': (1024, 1024)}


gamma
0.0337
0.0332
20.0
{'data': (10000, 1)}


gamma
0.0439
0.0752
2000.0
{'data': (10000, 100)}


gammaln
0.0564
0.0747
4194.3042
{'data': (1024, 1024)}


gammaln
0.0248
0.0241
20.0
{'data': (10000, 1)}


gammaln
0.0423
0.0624
2000.0
{'data': (10000, 100)}


hard_sigmoid
0.0341
0.036
2097.1521
{'data': (1024, 1024), 'alpha': 0.25, 'beta': 0.5}


hard_sigmoid
0.0237
0.0221
20.0
{'data': (10000, 1), 'alpha': 0.25, 'beta': 0.5}


hard_sigmoid
0.0328
0.0345
2000.0
{'data': (10000, 100), 'alpha': 0.25, 'beta': 0.5}


identity
---
---
4194.3042
{'data': (1024, 1024)}


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


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


log
0.0475
0.049
4194.3042
{'data': (1024, 1024)}


log
0.0327
0.032
20.0
{'data': (10000, 1)}


log
0.0402
0.0416
4000.0
{'data': (10000, 100)}


log10
0.0457
0.0479
4194.3042
{'data': (1024, 1024)}


log10
0.0294
0.0273
20.0
{'data': (10000, 1)}


log10
0.0371
0.0402
2000.0
{'data': (10000, 100)}


log1p
0.0463
0.0483
4194.3042
{'data': (1024, 1024)}


log1p
0.0349
0.0311
20.0
{'data': (10000, 1)}


log1p
0.0439
0.0431
2000.0
{'data': (10000, 100)}


log2
0.0431
0.0455
2097.1521
{'data': (1024, 1024)}


log2
0.0302
0.0279
20.0
{'data': (10000, 1)}


log2
0.0373
0.0398
2000.0
{'data': (10000, 100)}


log_softmax
0.0586
0.0544
2097.1521
{'data': (1024, 1024), 'axis': -1, 'temperature': 0.5}


log_softmax
0.0723
0.0556
40.0
{'data': (10000, 1), 'axis': -1, 'temperature': 0.5}


log_softmax
0.0587
0.0546
2000.0
{'data': (10000, 100), 'axis': -1, 'temperature': 0.5}


logical_not
0.0358
---
4194.3042
{'data': (1024, 1024)}


logical_not
0.0235
---
40.0
{'data': (10000, 1)}


logical_not
0.0354
---
4000.0
{'data': (10000, 100)}


make_loss
0.0404
---
4194.3042
{'data': (1024, 1024)}


make_loss
0.0289
---
40.0
{'data': (10000, 1)}


make_loss
0.0421
---
4000.0
{'data': (10000, 100)}


max
0.0807
0.1123
4.096
{'data': (1024, 1024), 'axis': 0}


max
0.0502
0.0368
0.002
{'data': (10000, 1), 'axis': 0}


max
0.0795
0.1086
0.4
{'data': (10000, 100), 'axis': 0}


max_axis
0.0574
---
4.096
{'data': (1024, 1024), 'axis': 0}


max_axis
0.0502
---
0.004
{'data': (10000, 1), 'axis': 0}


max_axis
0.063
---
0.4
{'data': (10000, 100), 'axis': 0}


mean
0.0722
0.1033
4.096
{'data': (1024, 1024), 'axis': 0}


mean
0.0535
0.0454
0.002
{'data': (10000, 1), 'axis': 0}


mean
0.0766
0.0994
0.4
{'data': (10000, 100), 'axis': 0}


min
0.0712
0.1112
4.096
{'data': (1024, 1024), 'axis': 0}


min
0.0511
0.0403
0.002
{'data': (10000, 1), 'axis': 0}


min
0.063
0.0995
0.2
{'data': (10000, 100), 'axis': 0}


min_axis
0.0478
---
4.096
{'data': (1024, 1024), 'axis': 0}


min_axis
0.0431
---
0.002
{'data': (10000, 1), 'axis': 0}


min_axis
0.0524
---
0.2
{'data': (10000, 100), 'axis': 0}


mp_sgd_mom_update
0.0696
---
602331.0625
{'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
0.0388
---
600273.875
{'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
0.0682
---
604253.875
{'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
0.0586
---
608545.375
{'weight': (1024, 1024), 'grad': (1024, 1024), 'weight32': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_update
0.0391
---
606488.1875
{'weight': (10000, 1), 'grad': (10000, 1), 'weight32': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


mp_sgd_update
0.0584
---
610468.1875
{'weight': (10000, 100), 'grad': (10000, 100), 'weight32': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


nanprod
0.0611
0.0996
4.096
{'data': (1024, 1024), 'axis': 0}


nanprod
0.0355
0.0231
0.002
{'data': (10000, 1), 'axis': 0}


nanprod
0.0559
0.0921
0.4
{'data': (10000, 100), 'axis': 0}


nansum
0.0531
0.0956
4.096
{'data': (1024, 1024), 'axis': 0}


nansum
0.0366
0.0215
0.002
{'data': (10000, 1), 'axis': 0}


nansum
0.0579
0.0916
0.4
{'data': (10000, 100), 'axis': 0}


negative
0.045
---
4194.3042
{'data': (1024, 1024)}


negative
0.0343
---
40.0
{'data': (10000, 1)}


negative
0.0458
---
4000.0
{'data': (10000, 100)}


ones_like
0.0339
---
4194.3042
{'data': (1024, 1024)}


ones_like
0.0244
---
40.0
{'data': (10000, 1)}


ones_like
0.0321
---
4000.0
{'data': (10000, 100)}


pad
---
---
0.384
{'data': (1, 4, 2, 4), 'mode': 'constant', 'pad_width': (0, 0, 0, 0, 1, 1, 1, 1)}


pad
---
---
1224.0
{'data': (10, 25, 10, 100), 'mode': 'constant', 'pad_width': (0, 0, 0, 0, 1, 1, 1, 1)}


prod
0.0643
0.0995
4.096
{'data': (1024, 1024), 'axis': 0}


prod
0.034
0.0228
0.002
{'data': (10000, 1), 'axis': 0}


prod
0.0526
0.092
0.4
{'data': (10000, 100), 'axis': 0}


radians
0.0361
0.0328
2097.1521
{'data': (1024, 1024)}


radians
0.0262
0.0248
20.0
{'data': (10000, 1)}


radians
0.0362
0.0321
2000.0
{'data': (10000, 100)}


random_exponential
2.916
---
2097.1521
{'shape': (1024, 1024)}


random_exponential
0.0508
---
20.0
{'shape': (10000, 1)}


random_exponential
2.5791
---
4000.0
{'shape': (10000, 100)}


random_gamma
7.8382
---
4194.3042
{'shape': (1024, 1024)}


random_gamma
0.1
---
20.0
{'shape': (10000, 1)}


random_gamma
7.7596
---
2000.0
{'shape': (10000, 100)}


random_generalized_negative_binomial
8.6073
---
2097.1521
{'shape': (1024, 1024)}


random_generalized_negative_binomial
0.0938
---
20.0
{'shape': (10000, 1)}


random_generalized_negative_binomial
6.5623
---
4000.0
{'shape': (10000, 100)}


random_negative_binomial
5.775
---
4194.3042
{'k': 1, 'p': 1, 'shape': (1024, 1024)}


random_negative_binomial
0.1201
---
40.0
{'k': 1, 'p': 1, 'shape': (10000, 1)}


random_negative_binomial
7.9775
---
4000.0
{'k': 1, 'p': 1, 'shape': (10000, 100)}


random_normal
2.493
---
4194.3042
{'shape': (1024, 1024)}


random_normal
0.0571
---
20.0
{'shape': (10000, 1)}


random_normal
2.363
---
4000.0
{'shape': (10000, 100)}


random_pdf_dirichlet
0.0456
---
0.004
{'sample': (2,), 'alpha': [0.0, 2.5]}


random_pdf_exponential
0.043
---
0.008
{'sample': (2,), 'lam': [1.0, 8.5]}


random_pdf_gamma
0.047
---
0.008
{'sample': (2,), 'alpha': [0.0, 2.5], 'beta': [1.0, 0.7]}


random_pdf_generalized_negative_binomial
0.0461
---
0.008
{'sample': (2,), 'mu': [2.0, 2.5], 'alpha': [0.0, 2.5]}


random_pdf_negative_binomial
0.0382
---
0.008
{'sample': (2,), 'k': [20, 49], 'p': [0.4, 0.77]}


random_pdf_normal
0.0378
---
0.008
{'sample': (2,), 'mu': [2.0, 2.5], 'sigma': [1.0, 3.7]}


random_pdf_poisson
0.0177
---
0.004
{'sample': (2,), 'lam': [1.0, 8.5]}


random_pdf_uniform
0.0236
---
0.008
{'sample': (2,), 'low': [0.0, 2.5], 'high': [1.0, 3.7]}


random_poisson
2.3348
---
2097.1521
{'shape': (1024, 1024)}


random_poisson
0.0527
---
20.0
{'shape': (10000, 1)}


random_poisson
1.6591
---
2000.0
{'shape': (10000, 100)}


random_randint
0.9378
---
2097.1521
{'low': 0, 'high': 5, 'shape': (1024, 1024)}


random_randint
0.0423
---
20.0
{'low': 0, 'high': 5, 'shape': (10000, 1)}


random_randint
0.8959
---
2000.0
{'low': 0, 'high': 5, 'shape': (10000, 100)}


random_uniform
0.3582
---
2097.1521
{'low': 0, 'high': 5, 'shape': (1024, 1024)}


random_uniform
0.0387
---
20.0
{'low': 0, 'high': 5, 'shape': (10000, 1)}


random_uniform
0.3482
---
2000.0
{'low': 0, 'high': 5, 'shape': (10000, 100)}


rcbrt
0.0482
0.0491
4194.3042
{'data': (1024, 1024)}


rcbrt
0.0354
0.0309
20.0
{'data': (10000, 1)}


rcbrt
0.0435
0.0415
2000.0
{'data': (10000, 100)}


reciprocal
0.0441
0.0453
4194.3042
{'data': (1024, 1024)}


reciprocal
0.0307
0.0266
40.0
{'data': (10000, 1)}


reciprocal
0.046
0.0459
4000.0
{'data': (10000, 100)}


relu
0.0514
0.0496
4194.3042
{'data': (1024, 1024)}


relu
0.0249
0.0241
40.0
{'data': (10000, 1)}


relu
0.0491
0.0472
4000.0
{'data': (10000, 100)}


repeat
0.1931
0.0951
8388.6084
{'data': (1024, 1024), 'repeats': 2, 'axis': 0}


repeat
0.0415
0.041
80.0
{'data': (10000, 1), 'repeats': 2, 'axis': 0}


repeat
0.1779
0.0945
8000.0
{'data': (10000, 100), 'repeats': 2, 'axis': 0}


reshape
0.058
---
4194.3042
{'data': (1024, 1024), 'shape': (1024, 1024)}


reshape
0.0417
---
20.0
{'data': (10000, 1), 'shape': (10000, 1)}


reshape
0.0528
---
2000.0
{'data': (10000, 100), 'shape': (10000, 100)}


reshape_like
0.0534
---
4194.3042
{'lhs': (1024, 1024), 'rhs': (1024, 1024)}


reshape_like
0.0458
---
400.0
{'lhs': (10000, 10), 'rhs': (10000, 10)}


reshape_like
0.0431
---
40.0
{'lhs': (10000, 1), 'rhs': (10000, 1)}


reverse
0.0442
0.0391
2097.1521
{'data': (1024, 1024), 'axis': 0}


reverse
0.0291
0.0279
20.0
{'data': (10000, 1), 'axis': 0}


reverse
0.0419
0.0399
2000.0
{'data': (10000, 100), 'axis': 0}


rint
0.0436
---
4194.3042
{'data': (1024, 1024)}


rint
0.0318
---
40.0
{'data': (10000, 1)}


rint
0.0426
---
4000.0
{'data': (10000, 100)}


rmsprop_update
0.0649
---
614759.6875
{'weight': (1024, 1024), 'grad': (1024, 1024), 'n': (1024, 1024), 'lr': 0.1, 'gamma1': 0.1, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmsprop_update
0.0386
---
612702.5
{'weight': (10000, 1), 'grad': (10000, 1), 'n': (10000, 1), 'lr': 0.5, 'gamma1': 0.5, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmsprop_update
0.064
---
616682.5
{'weight': (10000, 100), 'grad': (10000, 100), 'n': (10000, 100), 'lr': 0.9, 'gamma1': 0.9, 'epsilon': 1e-08, 'rescale_grad': 0.4}


rmspropalex_update
0.0926
---
620974.0
{'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
0.031
---
618916.8125
{'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
0.0881
---
622896.8125
{'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
0.069
---
4194.3042
{'data': (1024, 1024)}


round
0.0532
---
40.0
{'data': (10000, 1)}


round
0.0641
---
4000.0
{'data': (10000, 100)}


rsqrt
0.0478
0.0513
4194.3042
{'data': (1024, 1024)}


rsqrt
0.0391
0.0349
40.0
{'data': (10000, 1)}


rsqrt
0.0497
0.0515
4000.0
{'data': (10000, 100)}


sample_exponential
0.1656
---
8388.6084
{'lam': [1.0, 8.5], 'shape': (1024, 1024)}


sample_exponential
0.1043
---
80.0
{'lam': [1.0, 8.5], 'shape': (10000, 1)}


sample_exponential
0.1557
---
8000.0
{'lam': [1.0, 8.5], 'shape': (10000, 100)}


sample_gamma
0.3236
---
8388.6084
{'alpha': [0.0, 2.5], 'shape': (1024, 1024), 'beta': [1.0, 0.7]}


sample_gamma
0.2764
---
80.0
{'alpha': [0.0, 2.5], 'shape': (10000, 1), 'beta': [1.0, 0.7]}


sample_gamma
0.3106
---
8000.0
{'alpha': [0.0, 2.5], 'shape': (10000, 100), 'beta': [1.0, 0.7]}


sample_generalized_negative_binomial
0.9559
---
8388.6084
{'mu': [2.0, 2.5], 'shape': (1024, 1024), 'alpha': [0.0, 2.5]}


sample_generalized_negative_binomial
0.5525
---
80.0
{'mu': [2.0, 2.5], 'shape': (10000, 1), 'alpha': [0.0, 2.5]}


sample_generalized_negative_binomial
0.9381
---
8000.0
{'mu': [2.0, 2.5], 'shape': (10000, 100), 'alpha': [0.0, 2.5]}


sample_negative_binomial
1.4427
---
8388.6084
{'k': [20, 49], 'shape': (1024, 1024), 'p': [0.4, 0.77]}


sample_negative_binomial
0.8663
---
80.0
{'k': [20, 49], 'shape': (10000, 1), 'p': [0.4, 0.77]}


sample_negative_binomial
1.4467
---
8000.0
{'k': [20, 49], 'shape': (10000, 100), 'p': [0.4, 0.77]}


sample_normal
0.1471
---
8388.6084
{'mu': [2.0, 2.5], 'shape': (1024, 1024), 'sigma': [1.0, 3.7]}


sample_normal
0.0622
---
80.0
{'mu': [2.0, 2.5], 'shape': (10000, 1), 'sigma': [1.0, 3.7]}


sample_normal
0.1462
---
8000.0
{'mu': [2.0, 2.5], 'shape': (10000, 100), 'sigma': [1.0, 3.7]}


sample_poisson
0.441
---
8388.6084
{'lam': [1.0, 8.5], 'shape': (1024, 1024)}


sample_poisson
0.3674
---
80.0
{'lam': [1.0, 8.5], 'shape': (10000, 1)}


sample_poisson
0.4476
---
8000.0
{'lam': [1.0, 8.5], 'shape': (10000, 100)}


sample_uniform
0.165
---
8388.6084
{'low': [0.0, 2.5], 'shape': (1024, 1024), 'high': [1.0, 3.7]}


sample_uniform
0.0614
---
80.0
{'low': [0.0, 2.5], 'shape': (10000, 1), 'high': [1.0, 3.7]}


sample_uniform
0.1557
---
8000.0
{'low': [0.0, 2.5], 'shape': (10000, 100), 'high': [1.0, 3.7]}


sgd_mom_update
0.0586
---
627188.25
{'weight': (1024, 1024), 'grad': (1024, 1024), 'mom': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_mom_update
0.033
---
625131.125
{'weight': (10000, 1), 'grad': (10000, 1), 'mom': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_mom_update
0.049
---
629111.125
{'weight': (10000, 100), 'grad': (10000, 100), 'mom': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.0355
---
633402.5625
{'weight': (1024, 1024), 'grad': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.0197
---
631345.4375
{'weight': (10000, 1), 'grad': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4, 'lazy_update': 0}


sgd_update
0.0334
---
635325.4375
{'weight': (10000, 100), 'grad': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4, 'lazy_update': 0}


shape_array
0.0307
---
0.016
{'data': (1024, 1024)}


shape_array
0.0579
---
0.016
{'data': (10000, 1)}


shape_array
0.0313
---
0.016
{'data': (10000, 100)}


shuffle
0.1734
---
4194.3042
{'data': (1024, 1024)}


shuffle
0.3943
---
40.0
{'data': (10000, 1)}


shuffle
0.4167
---
4000.0
{'data': (10000, 100)}


sigmoid
0.0488
0.048
4194.3042
{'data': (1024, 1024)}


sigmoid
0.0315
0.0294
20.0
{'data': (10000, 1)}


sigmoid
0.0359
0.0398
2000.0
{'data': (10000, 100)}


sign
0.0454
0.041
2097.1521
{'data': (1024, 1024)}


sign
0.0319
0.0316
20.0
{'data': (10000, 1)}


sign
0.0436
0.0365
4000.0
{'data': (10000, 100)}


signsgd_update
0.0436
---
639616.875
{'weight': (1024, 1024), 'grad': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


signsgd_update
0.0272
---
637559.75
{'weight': (10000, 1), 'grad': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


signsgd_update
0.0426
---
641539.75
{'weight': (10000, 100), 'grad': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


signum_update
0.0553
---
645831.1875
{'weight': (1024, 1024), 'grad': (1024, 1024), 'mom': (1024, 1024), 'lr': 0.1, 'rescale_grad': 0.4}


signum_update
0.0299
---
643774.0625
{'weight': (10000, 1), 'grad': (10000, 1), 'mom': (10000, 1), 'lr': 0.5, 'rescale_grad': 0.4}


signum_update
0.0548
---
647754.0625
{'weight': (10000, 100), 'grad': (10000, 100), 'mom': (10000, 100), 'lr': 0.9, 'rescale_grad': 0.4}


sin
0.0435
0.0434
4194.3042
{'data': (1024, 1024)}


sin
0.0418
0.0339
40.0
{'data': (10000, 1)}


sin
0.0531
0.0488
4000.0
{'data': (10000, 100)}


sinh
0.0488
0.0488
2097.1521
{'data': (1024, 1024)}


sinh
0.0357
0.0322
40.0
{'data': (10000, 1)}


sinh
0.0385
0.0397
2000.0
{'data': (10000, 100)}


size_array
0.0273
---
0.008
{'data': (1024, 1024)}


size_array
0.0257
---
0.008
{'data': (10000, 1)}


size_array
0.0247
---
0.008
{'data': (10000, 100)}


smooth_l1
0.0327
0.0338
1251668.125
{'data': (1024, 1024)}


smooth_l1
0.0213
0.0196
1249611.0
{'data': (10000, 1)}


smooth_l1
0.0313
0.0331
1253591.0
{'data': (10000, 100)}


softmax
0.0773
0.0605
4194.3042
{'data': (1024, 1024), 'axis': -1, 'temperature': 0.5}


softmax
0.0751
0.0536
40.0
{'data': (10000, 1), 'axis': -1, 'temperature': 0.5}


softmax
0.0625
0.0563
4000.0
{'data': (10000, 100), 'axis': -1, 'temperature': 0.5}


softmax_cross_entropy
0.0495
---
1251591.0
{'data': (1024, 1024), 'label': (1024,)}


softsign
0.0432
0.0434
2097.1521
{'data': (1024, 1024)}


softsign
0.0322
0.0282
40.0
{'data': (10000, 1)}


softsign
0.0425
0.0413
2000.0
{'data': (10000, 100)}


sort
1.2868
---
6291.4561
{'data': (1024, 1024), 'axis': 0}


sort
0.2408
---
60.0
{'data': (10000, 1), 'axis': 0}


sort
1.2224
---
6000.0
{'data': (10000, 100), 'axis': 0}


space_to_depth
0.047
---
0.128
{'data': (1, 4, 2, 4), 'block_size': 2}


space_to_depth
0.0524
---
1000.0
{'data': (10, 25, 10, 100), 'block_size': 5}


split
---
---
4194.3042
{'data': (1024, 1024), 'num_outputs': 1, 'axis': 0}


split
---
---
40.0
{'data': (10000, 1), 'num_outputs': 1, 'axis': 0}


split
---
---
2000.0
{'data': (10000, 100), 'num_outputs': 1, 'axis': 0}


sqrt
0.0716
0.045
4194.3042
{'data': (1024, 1024)}


sqrt
0.0336
0.0309
20.0
{'data': (10000, 1)}


sqrt
0.0387
0.0421
2000.0
{'data': (10000, 100)}


square
0.0351
0.0405
2097.1521
{'data': (1024, 1024)}


square
0.0301
0.0282
20.0
{'data': (10000, 1)}


square
0.0366
0.0397
2000.0
{'data': (10000, 100)}


stop_gradient
---
---
4194.3042
{'data': (1024, 1024)}


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


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


sum
0.0616
0.0855
4.096
{'data': (1024, 1024), 'axis': 0}


sum
0.0483
0.0298
0.004
{'data': (10000, 1), 'axis': 0}


sum
0.0708
0.0843
0.4
{'data': (10000, 100), 'axis': 0}


sum_axis
0.0534
---
4.096
{'data': (1024, 1024), 'axis': 0}


sum_axis
0.0499
---
0.004
{'data': (10000, 1), 'axis': 0}


sum_axis
0.0677
---
0.4
{'data': (10000, 100), 'axis': 0}


swapaxes
0.0496
---
4194.3042
{'data': (1024, 1024), 'dim1': 0, 'dim2': 1}


swapaxes
0.036
---
40.0
{'data': (10000, 1), 'dim1': 0, 'dim2': 1}


swapaxes
0.0508
---
4000.0
{'data': (10000, 100), 'dim1': 0, 'dim2': 1}


tan
0.0432
0.0422
4194.3042
{'data': (1024, 1024)}


tan
0.0285
0.0252
20.0
{'data': (10000, 1)}


tan
0.0356
0.0384
2000.0
{'data': (10000, 100)}


tanh
0.0355
0.0384
2097.1521
{'data': (1024, 1024)}


tanh
0.027
0.0263
20.0
{'data': (10000, 1)}


tanh
0.035
0.0387
2000.0
{'data': (10000, 100)}


tile
0.2002
0.0998
8388.6084
{'data': (1024, 1024), 'reps': 2}


tile
0.0369
0.036
40.0
{'data': (10000, 1), 'reps': 2}


tile
0.1754
0.0927
4000.0
{'data': (10000, 100), 'reps': 2}


topk
0.1062
---
2.048
{'data': (1024, 1024), 'axis': 0, 'k': 1}


topk
0.0693
---
0.002
{'data': (10000, 1), 'axis': 0, 'k': 1}


topk
0.1154
---
0.2
{'data': (10000, 100), 'axis': 0, 'k': 1}


transpose
0.054
---
4194.3042
{'data': (1024, 1024)}


transpose
0.0319
---
40.0
{'data': (10000, 1)}


transpose
0.0526
---
4000.0
{'data': (10000, 100)}


trunc
0.0395
---
4194.3042
{'data': (1024, 1024)}


trunc
0.0308
---
40.0
{'data': (10000, 1)}


trunc
0.0399
---
4000.0
{'data': (10000, 100)}


zeros_like
0.0326
---
4194.3042
{'data': (1024, 1024)}


zeros_like
0.0283
---
40.0
{'data': (10000, 1)}


zeros_like
0.0321
---
4000.0
{'data': (10000, 100)}