NB: The benchmarking results are released with approval from General Motors.
This Jupyter Notebook compares the performance (execution time, memory consumption):
$ uname -a
Linux velociti 4.4.0-45-generic #66-Ubuntu SMP Wed Oct 19 14:12:37 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
$ cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=16.04
DISTRIB_CODENAME=xenial
DISTRIB_DESCRIPTION="Ubuntu 16.04.1 LTS"
NVIDIA GeForce GTX 1080 "Founders Edition":
using 14 Caffe libraries:
tag
] Branch (revision hash, date): math libraries.cpu
] Master (4ba654f, 5/Oct/2016): with OpenBLAS 0.2.19;cuda
] Master (4ba654f, 5/Oct/2016): with cuBLAS (part of CUDA Toolkit 8.0.44);cudnn
] Master (4ba654f, 5/Oct/2016): with cuDNN 5.1;nvidia-cuda
] NVIDIA v0.15 (1024d34, 17/Nov/2016): with cuBLAS (part of CUDA Toolkit 8.0.44);nvidia-cudnn
] NVIDIA v0.15 (1024d34, 17/Nov/2016): with cuDNN 5.1;nvidia-fp16-cuda
] NVIDIA experimental/fp16 (fca1cf4, 11/Jul/2016): with cuBLAS (part of CUDA Toolkit 8.0.44);nvidia-fp16-cudnn
] NVIDIA experimental/fp16 (fca1cf4, 11/Jul/2016): with cuDNN 5.1;clblas
] OpenCL (9abafdc, 7/Oct/2016): with ViennaCL 1.7.1 and clBLAS 2.10;clblast
] OpenCL (9abafdc, 7/Oct/2016): with ViennaCL 1.7.1 and CLBlast 0.9.0;viennacl
] OpenCL (9abafdc, 7/Oct/2016): with ViennaCL 1.7.1 only;libdnn-cuda
] OpenCL (cfaaae1, 25/Oct/2016): with libDNN and cuBLAS;libdnn-clblas
] OpenCL (cfaaae1, 25/Oct/2016): with libDNN, ViennaCL 1.7.1 and clBLAS 2.10;libdnn-clblast
] OpenCL (cfaaae1, 25/Oct/2016): with libDNN, ViennaCL 1.7.1 and CLBlast 0.9.0;libdnn-viennacl
] OpenCL (cfaaae1, 25/Oct/2016): with libDNN and ViennaCL 1.7.1.using 4 CNN models:
with the batch size varying from 2 to 16 with step 2.
fw = [ 'forward' ]
fwbw = [ 'forward', 'backward' ]
# Set to fw for inference; to fwbw for training.
direction = fw
direction
['forward']
if direction==fw:
time_ms = 'time_fw_ms'
else: # direction==fwbw
time_ms = 'time_fwbw_ms'
time_ms
'time_fw_ms'
def images_per_second(time_in_milliseconds):
return 1000.0 / time_in_milliseconds
NB: Please ignore this section if you are not interested in re-running or modifying this notebook.
import os
import sys
import json
import re
If some of the scientific packages are missing, please install them using:
# pip install jupyter pandas numpy matplotlib
import IPython as ip
import pandas as pd
import numpy as np
import matplotlib as mp
print ('IPython version: %s' % ip.__version__)
print ('Pandas version: %s' % pd.__version__)
print ('NumPy version: %s' % np.__version__)
print ('Matplotlib version: %s' % mp.__version__)
IPython version: 5.3.0 Pandas version: 0.19.2 NumPy version: 1.11.0 Matplotlib version: 2.0.0
from IPython.display import Image
from IPython.core.display import HTML
import matplotlib.pyplot as plt
from matplotlib import cm
%matplotlib inline
default_title = 'NVIDIA GTX 1080'
default_ylabel = 'Execution time (ms)'
default_colormap = cm.autumn
default_figsize = [16, 8]
default_dpi = 200
default_fontsize = 16
if mp.__version__[0]=='2': mp.style.use('classic')
mp.rcParams['figure.figsize'] = default_figsize
mp.rcParams['figure.dpi'] = default_dpi
mp.rcParams['font.size'] = default_fontsize
mp.rcParams['legend.fontsize'] = 'medium'
If CK is not installed, please install it using:
# pip install ck
import ck.kernel as ck
print ('CK version: %s' % ck.__version__)
CK version: 1.9.1.1
repo_uoa = 'ck-caffe-nvidia-gtx1080'
def get_experimental_results(repo_uoa, tags):
module_uoa = 'experiment'
r = ck.access({'action':'search', 'repo_uoa':repo_uoa, 'module_uoa':module_uoa, 'tags':tags})
if r['return']>0:
print ("Error: %s" % r['error'])
exit(1)
experiments = r['lst']
dfs = []
for experiment in experiments:
data_uoa = experiment['data_uoa']
r = ck.access({'action':'list_points', 'repo_uoa':repo_uoa, 'module_uoa':module_uoa, 'data_uoa':data_uoa})
if r['return']>0:
print ("Error: %s" % r['error'])
exit(1)
# Get (lib_tag, model_tag) from a list of tags that should be available in r['dict']['tags'].
# Tags include 2 of the 3 irrelevant tags, a model tag and a lib tag.
# NB: Since it's easier to list all model tags than all lib tags, the latter list is not expicitly specified.
tags = r['dict']['tags']
irrelevant_tags = [ 'explore-batch-size-libs-models','time_gpu','time_cpu','time_gpu_fp16' ]
model_tags = [ 'bvlc-alexnet','bvlc-googlenet','deepscale-squeezenet-1.0','deepscale-squeezenet-1.1' ]
lib_model_tags = [ tag for tag in tags if tag not in irrelevant_tags ]
model_tags = [ tag for tag in lib_model_tags if tag in model_tags ]
lib_tags = [ tag for tag in lib_model_tags if tag not in model_tags ]
if len(lib_tags)==1 and len(model_tags)==1:
(lib, model) = (lib_tags[0], model_tags[0])
else:
continue
for point in r['points']:
with open(os.path.join(r['path'], 'ckp-%s.0001.json' % point)) as point_file:
point_data_raw = json.load(point_file)
# Obtain column data.
characteristics = [
{
'time (ms)' : characteristics['run'].get(time_ms,+1e9), # "positive infinity"
'memory (MB)' : characteristics['run'].get('memory_mbytes',-1),
'success?' : characteristics['run'].get('run_success','n/a'),
'per layer info' : characteristics['run'].get('per_layer_info',[])
}
for characteristics in point_data_raw['characteristics_list']
]
# Deal with missing column data (resulting from failed runs).
if len(characteristics)==1:
repetitions = point_data_raw['features'].get('statistical_repetitions',1)
characteristics = characteristics * repetitions
# Construct a DataFrame.
df = pd.DataFrame(characteristics)
# Set columns and index names.
df.columns.name = 'run characteristic'
df.index.name = 'repetition'
# Set indices.
df['lib'] = lib
df['model'] = model
df['batch size'] = point_data_raw['choices']['env']['CK_CAFFE_BATCH_SIZE']
df = df.set_index(['lib', 'model', 'batch size'], append=True)
df = df.reorder_levels(('model', 'lib', 'batch size', 'repetition'))
# Append to the list of similarly constructed DataFrames.
dfs.append(df)
# Concatenate all constructed DataFrames (i.e. stack on top of each other).
result = pd.concat(dfs)
return result.sortlevel(result.index.names)
def plot(mean, std, title=default_title, ylabel=default_ylabel, rot=0, ymax=0):
ymax = mean.max().max() if ymax==0 else ymax
ax = mean.plot(kind='bar', yerr=std, grid=True, legend=True, rot=rot, ylim=[0,ymax*1.05],
fontsize=default_fontsize, figsize=default_figsize, colormap=default_colormap)
ax.set_title(title, fontsize=default_fontsize)
ax.set_xlabel(mean.index.name, fontsize=default_fontsize)
ax.set_ylabel(ylabel, fontsize=default_fontsize)
return ax
pretty_print_libs = {
'cpu': 'OpenBLAS (CPU)',
'libdnn-cuda':'libDNN-fp32',
'nvidia-cuda':'cuBLAS-fp32',
'nvidia-fp16-cuda':'cuBLAS-fp16',
'nvidia-cudnn':'cuDNN-fp32',
'nvidia-fp16-cudnn':'cuDNN-fp16'
}
pretty_print_models = {
'bvlc-alexnet':'AlexNet',
'bvlc-googlenet':'GoogleNet',
'deepscale-squeezenet-1.0':'SqueezeNet 1.0',
'deepscale-squeezenet-1.1':'SqueezeNet 1.1'
}
speedup_sort_models = [
'OpenBLAS (CPU)',
'libDNN-fp32',
'cuBLAS-fp32',
'cuBLAS-fp16',
'cuDNN-fp32',
'cuDNN-fp16'
]
# ['cuda', 'cudnn'] are roughly equivalent to ['nvidia-cuda', 'nvidia-cudnn'], so can be dropped.
def plot_max_num_images_per_second(df_mean_time_per_image, libs_to_drop=['cuda', 'cudnn'], rot=0, fontsize=None):
min_time_per_image = df_mean_time_per_image.min(axis=1).unstack('lib')
max_num_images_per_second = images_per_second(min_time_per_image) \
.drop(libs_to_drop, axis=1) \
.rename(columns=pretty_print_libs, index=pretty_print_models) \
.reindex(columns=speedup_sort_models)
ax = max_num_images_per_second \
.plot(kind='bar', rot=rot, width=0.95, grid=True, legend=True,
fontsize=default_fontsize, figsize=default_figsize, colormap=default_colormap)
ax.set_title(default_title, fontsize=default_fontsize)
ax.set_xlabel(max_num_images_per_second.index.name, fontsize=default_fontsize)
ax.set_ylabel('Images/s (with the best even batch size between 2 and 16)', fontsize=default_fontsize)
ax.legend(loc='upper center');
for patch in ax.patches:
ax.annotate(str(int(patch.get_height()+0.5)), (patch.get_x()*1.00, patch.get_height()*1.01), fontsize=fontsize)
# ['cuda', 'cudnn'] are roughly equivalent to ['nvidia-cuda', 'nvidia-cudnn'], so can be dropped.
def plot_speedup_over_baseline(df_mean_time_per_image, baseline='cpu', libs_to_drop=['cuda', 'cudnn'], rot=0, fontsize=None):
speedup_over_baseline = df_mean_time_per_image.min(axis=1).unstack('model').ix[baseline] / \
df_mean_time_per_image.min(axis=1).unstack('model')
speedup_over_baseline = speedup_over_baseline.T \
.drop(libs_to_drop, axis=1) \
.rename(index=pretty_print_models, columns=pretty_print_libs) \
.reindex(columns=speedup_sort_models)
ax = speedup_over_baseline \
.plot(kind='bar', rot=rot, width=0.95, grid=True, legend=True,
fontsize=default_fontsize, figsize=default_figsize, colormap=default_colormap)
ax.set_title(default_title, fontsize=default_fontsize)
ax.set_xlabel(speedup_over_baseline.index.name, fontsize=default_fontsize)
ax.set_ylabel('Speedup over the given baseline (%s)' % pretty_print_libs[baseline], fontsize=default_fontsize)
for patch in ax.patches:
ax.annotate('{0:.2f}'.format(patch.get_height())[0:4], (patch.get_x()*1.00, patch.get_height()*1.01),
fontsize=fontsize)
# This transformation is time consuming, hence only call it once for multiple plots.
def get_per_layer_info(df_all):
df_per_layer_info = df_all['per layer info']
row_dfs = []
for (row_info, row_id) in zip(df_per_layer_info, range(len(df_per_layer_info))):
# Skip constructing a DataFrame when no layer info is available.
if not row_info: continue
# Augment each layer info with the row index: (model, lib, batch size, repetition).
for layer_info in row_info:
layer_info.update({ k : v for k, v in zip(df_per_layer_info.index.names, df_per_layer_info.index[row_id]) })
# Construct a DataFrame and move the row index to where it belongs.
row_df = pd.DataFrame(data=row_info).set_index(df_per_layer_info.index.names)
row_dfs.append(row_df)
return pd.concat(row_dfs)
def plot_time_per_image_per_layer(df_per_layer_info, model, libs, batch_sizes,
direction=['forward'], lower=0.0, upper=1.0, ymax=0, rot=90):
df_time_per_batch = df_per_layer_info.loc[model, libs, batch_sizes] \
.set_index(['direction', 'label'], append=True) \
.reorder_levels(['direction', 'label', 'model', 'lib', 'batch size', 'repetition' ]) \
.ix[direction] \
.reorder_levels(['label', 'model', 'lib', 'batch size', 'repetition', 'direction' ]) \
.groupby(level=['label', 'model', 'lib', 'batch size', 'repetition']).sum() \
['time_ms']
df_time_per_image = df_time_per_batch.unstack('batch size') / batch_sizes
df = df_time_per_image.unstack(['lib', 'model'])
df = df.reorder_levels(['model', 'lib', 'batch size'], axis=1)
mean = df.groupby(level='label').mean()
std = df.groupby(level='label').std()
select = (lower*mean.sum() <= mean).any(axis=1) & (mean <= upper*mean.sum()).any(axis=1)
ymax = mean[select].max().max() if ymax==0 else ymax
ax = plot(mean=mean[select], std=std[select], ylabel='Execution per image time per layer (ms)', ymax=ymax, rot=rot)
ax.set_xlabel('Layer', fontsize=default_fontsize)
# The ideal adaptive solution for each layer selects the best performing library from the 'libs_for_adaptation' list.
# FIXME: add batch_sizes as explicit parameter.
def get_ideal_adaptive_solution(df_per_layer_info, libs_for_adaptation, direction):
df_for_adaptation = df_per_layer_info \
.set_index(['direction', 'label'], append=True) \
.reorder_levels(['direction', 'lib', 'model', 'label', 'batch size', 'repetition']) \
.ix[direction] \
.reorder_levels(['lib', 'model', 'label', 'batch size', 'repetition', 'direction']) \
.ix[libs_for_adaptation] \
.reorder_levels(['model', 'label', 'lib', 'batch size', 'repetition', 'direction']) \
['time_ms']
# With every step, reduce the rightmost dimension until the min time per model is reached.
df_cum_time_per_repetition = df_for_adaptation.groupby(level=df_for_adaptation.index.names[:-1]).sum()
df_min_time_per_repetition = df_cum_time_per_repetition.groupby(level=df_cum_time_per_repetition.index.names[:-1]).min()
df_min_time_per_batch = df_min_time_per_repetition.unstack('batch size') / batch_sizes
df_min_time_per_image = df_min_time_per_batch.min(axis=1)
df_min_time_per_layer = df_min_time_per_image.groupby(level=df_min_time_per_image.index.names[:-1]).min()
#df_min_time_per_model = df_min_time_per_layer.groupby(level=df_min_time_per_layer.index.names[:-1]).sum()
# Transform to get the models in the index and the libs in the columns.
df_min_time_per_layer_idx = df_min_time_per_image.groupby(level=df_min_time_per_image.index.names[:-1]).idxmin()
df_ideal = df_min_time_per_image[df_min_time_per_layer_idx] \
.reorder_levels(['model', 'lib', 'label']) \
.groupby(level=['model', 'lib']).sum() \
.unstack('lib')
# Sort in the order of increasing time per model.
df_ideal_sorted = df_ideal.ix[df_ideal.sum(axis=1).sort_values(ascending=True).index]
return df_ideal_sorted
def plot_ideal_adaptive_solution(df_ideal, df_real, tag=""):
figsize=[15, 3]
if not tag=="": figsize=[10, 2] # good for dumping png (e.g. 3 graphs fit well onto a slide).
for model in df_ideal.index:
df_data = {}; df_data['adaptive'] = df_ideal.ix[model]
for lib in df_ideal.columns:
df_data[lib] = pd.Series(index=df_ideal.columns)
df_data[lib][lib] = df_real.ix[model, lib]
df = pd.DataFrame(df_data).T \
.rename(index={'cpu': 'OpenBLAS only', 'cuda':'cuBLAS only', 'cudnn':'cuDNN only', 'libdnn-cuda': 'libDNN only'},
columns={'cpu': 'OpenBLAS', 'cuda':'cuBLAS', 'cudnn':'cuDNN', 'libdnn-cuda': 'libDNN'})
ax = df.ix[df.sum(axis=1).sort_values(ascending=True).index] \
.plot(kind='barh', stacked=True, width=0.9, grid=True, legend=True,
fontsize=default_fontsize, figsize=figsize, colormap=cm.summer_r)
#.legend(loc='lower right')
ax.set_title('%s - execution time per image (ms)' % model, fontsize=default_fontsize)
if not tag=="": ax.get_figure().savefig('%s.%s.png' % (tag, model))
def plot_time_per_image_and_memory_consumption(df_all, model, lib):
df = df_all[['time (ms)', 'memory (MB)']] \
.groupby(level=df_all.index.names[:-1]).mean() \
.loc[model, lib]
df['time per image (ms)'] = df['time (ms)'].divide(df.index, axis=0)
df['memory per image (MB)'] = df['memory (MB)'].divide(df.index, axis=0)
df = df.drop('time (ms)', axis=1).sortlevel(axis=1)
ax = df.plot(secondary_y=['memory (MB)', 'memory per image (MB)'], mark_right=False, grid=True,
figsize=[12, 8], fontsize=default_fontsize, colormap=cm.winter)
ax.set_title('%s w/ %s' % (model, lib), fontsize=default_fontsize)
ax.set_xlabel(df.index.name, fontsize=default_fontsize)
ax.set_ylabel('execution time (ms)', fontsize=default_fontsize); ax.legend(loc='center left'); ax.set_ylim(0)
ax.right_ax.set_ylabel('memory consumption (MB)', fontsize=default_fontsize); ax.right_ax.legend(loc='center right')
NB: Please ignore this section if you are not interested in re-running or modifying this notebook.
The Caffe experimental data was collected on the experimental platform (after installing all Caffe libraries and models of interest) as follows:
$ cd `ck find ck-caffe:script:explore-batch-size-libs-models`
$ python explore-batch-size-libs-models-benchmark.py
It can be downloaded from GitHub via CK as follows:
$ ck pull repo:ck-caffe-nvidia-gtx1080 --url=https://github.com/dividiti/ck-caffe-nvidia-gtx1080
df_all = get_experimental_results(repo_uoa=repo_uoa, tags='explore-batch-size-libs-models')
pd.options.display.max_columns = len(df_all.columns)
pd.options.display.max_rows = len(df_all.index)
df_all
run characteristic | memory (MB) | per layer info | success? | time (ms) | |||
---|---|---|---|---|---|---|---|
model | lib | batch size | repetition | ||||
bvlc-alexnet | clblas | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.94669 |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.20973 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 9.24682 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.23080 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 13.54340 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.47150 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 17.72540 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 21.72930 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 21.72930 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 21.72930 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 25.90130 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 25.90820 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 25.90310 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.57250 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.18980 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.56030 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 34.77500 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 34.40120 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 38.66000 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 38.66000 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 38.66000 | |||
clblast | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 24.84130 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 24.84130 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 24.84130 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 31.49390 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 31.49390 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 31.49390 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 43.37870 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 43.37870 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 43.37870 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 60.31870 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 60.31870 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 60.31870 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 77.92740 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 77.92740 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 77.92740 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 78.77240 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 78.77240 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 78.77240 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 86.10000 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 86.10000 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 86.10000 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 105.13100 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 105.13100 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 105.13100 | |||
cpu | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 234.21200 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 229.52200 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 226.04700 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 332.20800 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 332.01600 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 331.48400 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 439.48100 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 442.57000 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 437.23500 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 541.65000 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 538.10100 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 537.45800 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 626.30600 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 658.54800 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 633.62000 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 725.67100 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 746.46200 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 735.05400 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 818.67600 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 822.47200 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 822.19200 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 896.59600 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 895.84600 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 894.85800 | |||
cuda | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.31162 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.40726 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.41606 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.41325 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.44371 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.57187 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.81779 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.83734 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.83315 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.73280 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.65923 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.63539 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.59990 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.63220 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.61400 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.06930 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.06130 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.04820 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.40870 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.40380 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.42740 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.68920 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.77720 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.68200 | |||
cudnn | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.29418 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.29888 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.29994 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.17645 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.46275 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.18464 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.40416 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.31066 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.31469 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.85306 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.86992 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.87277 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.54246 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.66202 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.54394 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.09728 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.76614 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.78051 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.90950 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.93123 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.93114 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.95302 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.97142 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.96934 | |||
libdnn-clblas | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.16272 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.15498 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.45392 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.51354 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.20838 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.23578 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.15500 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.39460 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.39250 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.42790 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.80570 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.48430 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.24850 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.58030 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.46780 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.65670 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.99430 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.55820 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.68710 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.08740 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.13460 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.18730 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.84320 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.82580 | |||
libdnn-clblast | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.95990 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.91520 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.80160 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.59760 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.59010 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.69120 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.49730 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.37090 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.43320 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.87470 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.88590 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.83510 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.70780 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.65430 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.63800 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.04290 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.03540 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.03100 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.17500 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.13800 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.12210 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.21090 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.19490 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.11800 | |||
libdnn-cuda | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.05811 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.03398 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.04106 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.38010 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.38112 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.36480 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.97958 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.98595 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.97456 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.35891 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.35891 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.26493 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.80582 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.74746 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.73843 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.35980 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.39650 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.33920 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.52090 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.54650 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.35500 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.01820 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.99850 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.98620 | |||
libdnn-viennacl | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.56240 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.56140 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.60620 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.69090 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.93480 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.71340 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.98080 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.92630 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.66130 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.53180 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.62750 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.48700 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.10010 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.04890 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.09080 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.99140 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.88370 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.64510 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.34250 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.32480 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.34760 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.26780 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.05660 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.07420 | |||
nvidia-cuda | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.33517 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.37514 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.32813 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.44870 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.43642 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.42618 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.89562 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.90381 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.89955 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.60710 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.65680 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.60672 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.61580 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.62600 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.72980 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.05910 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.53690 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.49000 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.70210 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.68780 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.71030 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.98890 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.90890 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.91710 | |||
nvidia-cudnn | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.53763 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.53747 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.63475 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.50557 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.50733 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.52179 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.84102 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.85741 | |||
2 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.85600 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.49024 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.50294 | |||
2 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.53939 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.28282 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.38931 | |||
2 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.50605 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.46656 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.53376 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.47616 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.50342 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.18394 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.51366 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.82234 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.93043 | |||
2 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.61776 | |||
nvidia-fp16-cuda | 2 | 0 | 7.704896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.55315 | |
1 | 7.704896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.46611 | |||
2 | 7.704896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.54477 | |||
4 | 0 | 15.409792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.68966 | ||
1 | 15.409792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.61594 | |||
2 | 15.409792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.59136 | |||
6 | 0 | 23.114688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.63370 | ||
1 | 23.114688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.62850 | |||
2 | 23.114688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.64590 | |||
8 | 0 | 30.819584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.84900 | ||
1 | 30.819584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.84290 | |||
2 | 30.819584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.82240 | |||
10 | 0 | 38.524480 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.22520 | ||
1 | 38.524480 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.24420 | |||
2 | 38.524480 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.28350 | |||
12 | 0 | 46.229376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 22.87720 | ||
1 | 46.229376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 22.88230 | |||
2 | 46.229376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 22.80760 | |||
14 | 0 | 53.934272 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.28420 | ||
1 | 53.934272 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.30570 | |||
2 | 53.934272 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.31390 | |||
16 | 0 | 61.639168 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.19220 | ||
1 | 61.639168 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.34780 | |||
2 | 61.639168 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.31400 | |||
nvidia-fp16-cudnn | 2 | 0 | 7.704896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 64.45600 | |
1 | 7.704896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 62.91660 | |||
2 | 7.704896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 62.26330 | |||
4 | 0 | 15.409792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 77.78200 | ||
1 | 15.409792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 77.79840 | |||
2 | 15.409792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 77.75440 | |||
6 | 0 | 23.114688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 94.27120 | ||
1 | 23.114688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 92.14330 | |||
2 | 23.114688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 94.28270 | |||
8 | 0 | 30.819584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 131.13100 | ||
1 | 30.819584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 130.30200 | |||
2 | 30.819584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 131.53200 | |||
10 | 0 | 38.524480 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 144.08100 | ||
1 | 38.524480 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 144.09500 | |||
2 | 38.524480 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 144.23700 | |||
12 | 0 | 46.229376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 161.19600 | ||
1 | 46.229376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 151.15500 | |||
2 | 46.229376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 151.10500 | |||
14 | 0 | 53.934272 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 163.90300 | ||
1 | 53.934272 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 163.97700 | |||
2 | 53.934272 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 163.96800 | |||
16 | 0 | 61.639168 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 207.02700 | ||
1 | 61.639168 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 207.11100 | |||
2 | 61.639168 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 207.08700 | |||
viennacl | 2 | 0 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 33.41230 | |
1 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 33.49610 | |||
2 | 16.646488 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 33.65170 | |||
4 | 0 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 51.38020 | ||
1 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 51.38020 | |||
2 | 33.292976 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 51.38020 | |||
6 | 0 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 79.09160 | ||
1 | 49.939464 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 77.74000 | |||
8 | 0 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 95.74090 | ||
1 | 66.585952 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 98.87530 | |||
10 | 0 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 120.53700 | ||
1 | 83.232440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 117.68100 | |||
12 | 0 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 138.57700 | ||
1 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 138.57700 | |||
2 | 99.878928 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 138.57700 | |||
14 | 0 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 157.27800 | ||
1 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 157.27800 | |||
2 | 116.525416 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 157.27800 | |||
16 | 0 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 178.99800 | ||
1 | 133.171904 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 182.17500 | |||
bvlc-googlenet | clblas | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.22710 |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.78840 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.28130 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 31.97640 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 31.97640 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 31.97640 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 42.58410 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 42.60470 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 52.47200 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 53.04000 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 67.66080 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 63.24330 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 75.85490 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 75.85490 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 75.85490 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 85.13860 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 91.58330 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 94.05310 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 95.11530 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 95.11530 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 95.11530 | |||
clblast | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 84.74520 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 84.74520 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 84.74520 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 287.85700 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 287.85700 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 287.85700 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 220.17700 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 220.17700 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 220.17700 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 296.28900 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 296.28900 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 296.28900 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 426.49800 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 426.49800 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 426.49800 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 436.97200 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 436.97200 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 436.97200 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 472.28300 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 472.28300 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 472.28300 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 541.09600 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 541.09600 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 541.09600 | |||
cpu | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 402.63700 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 406.11800 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 403.49200 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 730.00200 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 730.53000 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 731.39300 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1041.55000 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1038.72000 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1037.39000 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1353.39000 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1353.79000 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1356.01000 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1656.32000 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1654.63000 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1660.07000 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1971.41000 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1958.12000 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 1955.44000 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2293.01000 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2264.95000 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2264.29000 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3201.15000 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2565.12000 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2562.73000 | |||
cuda | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.08450 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.56200 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.35040 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.85640 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.84110 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.38580 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.20130 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.11290 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.14200 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 36.50560 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 36.58980 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 37.16610 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 44.96490 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 44.81130 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 44.73140 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 53.20840 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 53.06280 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 52.64590 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 61.60620 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 61.54550 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 61.45020 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 69.85760 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 70.75560 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 69.80650 | |||
cudnn | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.62266 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.64112 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.68614 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.13626 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.10643 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.81008 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.00420 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.05000 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.03260 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.30750 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.34120 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.28190 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.97170 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.97350 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.98200 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.44810 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.48620 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.56580 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.67530 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.71670 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.69440 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.09960 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.07530 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.08330 | |||
libdnn-clblas | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.80500 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.06040 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.74980 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.00500 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.20200 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.92320 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.65030 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 22.28020 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 22.43360 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.03050 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 25.74750 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 25.73000 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.68090 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.26180 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.51780 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 33.78010 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 33.63600 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 33.87060 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 38.18060 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 39.51210 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 38.20610 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 42.26880 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 42.34240 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 43.58680 | |||
libdnn-clblast | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.02710 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.46900 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.64980 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.03000 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.37170 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.33110 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.06460 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.05950 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.71290 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.77370 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.54520 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.94700 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.24520 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.42380 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.30100 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 34.65320 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 36.10830 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 34.51800 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 40.66950 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 39.05420 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 40.66610 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 43.06650 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 44.71500 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 43.03550 | |||
libdnn-cuda | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.75860 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.76270 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.76290 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.23850 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.16590 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.14800 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.12220 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.08140 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.03440 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.39760 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.37920 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.54280 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.08870 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.09870 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.19330 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.41600 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.65960 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.47530 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 34.87990 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 34.96240 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 34.97390 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 39.19170 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 39.28490 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 41.95650 | |||
libdnn-viennacl | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.94190 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.96850 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.25600 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.14050 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.24870 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.49710 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.77420 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 22.61190 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 22.55870 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.32840 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.30720 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 25.86210 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.56990 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.54140 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.85410 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 35.16420 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 35.36280 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 35.21640 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 39.78750 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 38.50850 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 38.51140 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 42.69880 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 44.08730 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 42.82770 | |||
nvidia-cuda | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.06870 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.43810 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.39960 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.31980 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.54360 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.02120 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.32700 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.72140 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.36290 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 34.85120 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.90830 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 33.72460 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 41.16090 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 40.32510 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 40.27910 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 47.75550 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 47.81980 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 47.96310 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 55.13340 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 55.09040 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 55.17950 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 63.46880 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 62.82440 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 62.55530 | |||
nvidia-cudnn | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.83389 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.78554 | |||
2 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.78150 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.07950 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.70973 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.53210 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.14110 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.08710 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.16900 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.63660 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.99700 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.13590 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.82750 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.86850 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.84390 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.64720 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.68510 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.92150 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.97000 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.74600 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.38590 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.92090 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.76730 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.36020 | |||
nvidia-fp16-cuda | 2 | 0 | 54.551232 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.59890 | |
1 | 54.551232 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.64320 | |||
2 | 54.551232 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.57250 | |||
4 | 0 | 109.102464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.07710 | ||
1 | 109.102464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.98180 | |||
2 | 109.102464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.21090 | |||
6 | 0 | 163.653696 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 41.46280 | ||
1 | 163.653696 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 41.80170 | |||
2 | 163.653696 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 38.66210 | |||
8 | 0 | 218.204928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 48.30310 | ||
1 | 218.204928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 51.27480 | |||
2 | 218.204928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 54.92430 | |||
10 | 0 | 272.756160 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 62.68370 | ||
1 | 272.756160 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 68.42060 | |||
2 | 272.756160 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 62.85370 | |||
12 | 0 | 327.307392 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 75.83850 | ||
1 | 327.307392 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 75.73090 | |||
2 | 327.307392 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 75.59990 | |||
14 | 0 | 381.858624 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 87.32030 | ||
1 | 381.858624 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 88.64360 | |||
2 | 381.858624 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 83.24590 | |||
16 | 0 | 436.409856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 97.39670 | ||
1 | 436.409856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 96.81410 | |||
2 | 436.409856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 98.84780 | |||
nvidia-fp16-cudnn | 2 | 0 | 54.551232 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 189.63100 | |
1 | 54.551232 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 189.92100 | |||
2 | 54.551232 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 189.88300 | |||
4 | 0 | 109.102464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 242.86500 | ||
1 | 109.102464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 247.95500 | |||
2 | 109.102464 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 247.90700 | |||
6 | 0 | 163.653696 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 269.28500 | ||
1 | 163.653696 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 269.24600 | |||
2 | 163.653696 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 269.19600 | |||
8 | 0 | 218.204928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 359.08100 | ||
1 | 218.204928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 358.73900 | |||
2 | 218.204928 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 359.10200 | |||
10 | 0 | 272.756160 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 409.44700 | ||
1 | 272.756160 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 397.51500 | |||
2 | 272.756160 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 409.49300 | |||
12 | 0 | 327.307392 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 435.99300 | ||
1 | 327.307392 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 434.65900 | |||
2 | 327.307392 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 434.67200 | |||
14 | 0 | 381.858624 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 568.72800 | ||
1 | 381.858624 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 562.51900 | |||
2 | 381.858624 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 563.15700 | |||
16 | 0 | 436.409856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 540.86300 | ||
1 | 436.409856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 547.85400 | |||
2 | 436.409856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 535.67400 | |||
viennacl | 2 | 0 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 69.31140 | |
1 | 110.306688 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 66.63680 | |||
4 | 0 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 129.35200 | ||
1 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 129.35200 | |||
2 | 220.613376 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 129.35200 | |||
6 | 0 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 177.67200 | ||
1 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 177.67200 | |||
2 | 330.920064 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 177.67200 | |||
8 | 0 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 223.95300 | ||
1 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 223.95300 | |||
2 | 441.226752 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 223.95300 | |||
10 | 0 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 277.49800 | ||
1 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 277.49800 | |||
2 | 551.533440 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 277.49800 | |||
12 | 0 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 327.71300 | ||
1 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 327.71300 | |||
2 | 661.840128 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 327.71300 | |||
14 | 0 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 380.06300 | ||
1 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 380.06300 | |||
2 | 772.146816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 380.06300 | |||
16 | 0 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 437.90700 | ||
1 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 437.90700 | |||
2 | 882.453504 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 437.90700 | |||
deepscale-squeezenet-1.0 | clblas | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.37062 |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 9.38189 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.78730 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.48280 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.69590 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 23.27930 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 23.27930 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 23.27930 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 27.98590 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 27.98590 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 27.98590 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 34.14440 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 34.14440 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 34.14440 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 42.60320 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 40.27190 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 42.73460 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 49.04860 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 49.04860 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 49.04860 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 52.76880 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 52.76880 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 52.76880 | |||
clblast | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 51.72840 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 51.72840 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 51.72840 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 80.18130 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 80.18130 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 80.18130 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 91.16290 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 91.16290 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 91.16290 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 173.55000 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 173.55000 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 173.55000 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 198.66100 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 198.66100 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 198.66100 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 241.68700 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 241.68700 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 241.68700 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 207.42100 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 207.42100 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 207.42100 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 298.16600 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 298.16600 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 298.16600 | |||
cpu | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 143.62800 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 129.41900 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 121.00900 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 273.39200 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 256.95700 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 277.89300 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 388.01400 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 375.02500 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 368.99000 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 488.15500 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 488.58100 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 487.88700 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 598.73100 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 599.37400 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 600.14100 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 709.80100 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 710.50800 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 710.05400 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 816.99500 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 816.69800 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 834.05400 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 920.77400 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 920.03500 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 917.35100 | |||
cuda | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.52448 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.77594 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.49741 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.76704 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.04790 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.07117 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.96410 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.23020 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.89430 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.35240 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.35340 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.35760 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.21960 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.26160 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.23910 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.45380 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.49740 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.97710 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.00400 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.66870 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.65270 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.83320 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.84940 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.12170 | |||
cudnn | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.43040 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.43981 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.46362 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.50810 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.01792 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.45894 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.13869 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.10045 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.48154 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.72858 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.06240 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.76237 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.37310 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.34570 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.34470 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.34050 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.34430 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.31590 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.02680 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.10800 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.02680 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.95290 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.23860 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.94370 | |||
libdnn-clblas | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.76381 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.96323 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.99824 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.15300 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.21350 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.21460 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.77370 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.82580 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.40610 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.13500 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.15240 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.71480 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.77170 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.73350 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.75000 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.04500 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 25.75740 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.05430 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.42120 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.33960 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.51960 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.47410 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.05130 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.03690 | |||
libdnn-clblast | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.80982 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.97024 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.99699 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.40080 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.21130 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.08040 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.58180 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.46860 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.59670 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.15240 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.72600 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.11330 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.80210 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.81540 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.43440 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.86170 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.91270 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.00520 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.51020 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.46650 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.50330 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.00310 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.01740 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 29.77280 | |||
libdnn-cuda | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.33152 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.10640 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.34790 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.80755 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.80960 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.50490 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.21533 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.20474 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.19354 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.54050 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.15080 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.20300 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.26430 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.25310 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.88690 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.68760 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.71780 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.34960 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.09870 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.67630 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.16720 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.46930 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.48040 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.51430 | |||
libdnn-viennacl | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.00944 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.62618 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.02976 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.18170 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.03200 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.15100 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.32020 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.55020 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.24520 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.69160 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.09310 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.74340 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.42110 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.70070 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.77590 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.05950 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.07280 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.51790 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.38110 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 28.54170 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.38850 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 31.99270 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.62240 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.92160 | |||
nvidia-cuda | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.50662 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.52403 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.27910 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.79059 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.57453 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.77318 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.93730 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.93330 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.41100 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.04830 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.37450 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.02370 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.88850 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.54820 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.47630 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.98500 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.86820 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.88070 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.78760 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 22.52400 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.88350 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.75620 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.71730 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.78590 | |||
nvidia-cudnn | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.28704 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.28115 | |||
2 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.30240 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.01677 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.97382 | |||
2 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.96550 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.71277 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.71264 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.71539 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.16253 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.34294 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.01821 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.05510 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.03030 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.19900 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.95930 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.76880 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.73640 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.09200 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.44410 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.46300 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.04560 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.23950 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.30160 | |||
nvidia-fp16-cuda | 2 | 0 | 52.441856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.38557 | |
1 | 52.441856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.38989 | |||
2 | 52.441856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.46054 | |||
4 | 0 | 104.883712 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.23296 | ||
1 | 104.883712 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.29338 | |||
2 | 104.883712 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.29843 | |||
6 | 0 | 157.325568 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.10470 | ||
1 | 157.325568 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.14540 | |||
2 | 157.325568 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.17510 | |||
8 | 0 | 209.767424 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.97240 | ||
1 | 209.767424 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.03690 | |||
2 | 209.767424 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.95390 | |||
10 | 0 | 262.209280 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.77450 | ||
1 | 262.209280 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.83490 | |||
2 | 262.209280 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.80830 | |||
12 | 0 | 314.651136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.84990 | ||
1 | 314.651136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.78960 | |||
2 | 314.651136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.78810 | |||
14 | 0 | 367.092992 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.57120 | ||
1 | 367.092992 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.57290 | |||
2 | 367.092992 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 27.58530 | |||
16 | 0 | 419.534848 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 31.59830 | ||
1 | 419.534848 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 30.78350 | |||
2 | 419.534848 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 31.55150 | |||
nvidia-fp16-cudnn | 2 | 0 | 52.441856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 56.25750 | |
1 | 52.441856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 62.38920 | |||
2 | 52.441856 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 61.96840 | |||
4 | 0 | 104.883712 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 95.31680 | ||
1 | 104.883712 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 95.44400 | |||
2 | 104.883712 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 95.28910 | |||
6 | 0 | 157.325568 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 119.73200 | ||
1 | 157.325568 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 125.52200 | |||
2 | 157.325568 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 117.37600 | |||
8 | 0 | 209.767424 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 144.42900 | ||
1 | 209.767424 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 143.40400 | |||
2 | 209.767424 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 141.64000 | |||
10 | 0 | 262.209280 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 188.53600 | ||
1 | 262.209280 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 188.61300 | |||
2 | 262.209280 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 188.58700 | |||
12 | 0 | 314.651136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 228.28700 | ||
1 | 314.651136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 228.74800 | |||
2 | 314.651136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 228.65900 | |||
14 | 0 | 367.092992 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 246.20300 | ||
1 | 367.092992 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 246.24600 | |||
2 | 367.092992 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 246.12500 | |||
16 | 0 | 419.534848 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 259.10800 | ||
1 | 419.534848 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 258.15100 | |||
2 | 419.534848 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 258.40400 | |||
viennacl | 2 | 0 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.21420 | |
1 | 106.120408 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 31.97130 | |||
4 | 0 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 58.89050 | ||
1 | 212.240816 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 59.99030 | |||
6 | 0 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 88.22270 | ||
1 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 87.39960 | |||
2 | 318.361224 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 87.24580 | |||
8 | 0 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 114.61900 | ||
1 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 114.61900 | |||
2 | 424.481632 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 114.61900 | |||
10 | 0 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 137.82800 | ||
1 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 136.13200 | |||
2 | 530.602040 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 134.60200 | |||
12 | 0 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 158.09600 | ||
1 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 158.75300 | |||
2 | 636.722448 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 160.88200 | |||
14 | 0 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 183.79100 | ||
1 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 183.79100 | |||
2 | 742.842856 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 183.79100 | |||
16 | 0 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 207.45100 | ||
1 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 207.45100 | |||
2 | 848.963264 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 207.45100 | |||
deepscale-squeezenet-1.1 | clblas | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.47315 |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.78710 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 8.49402 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.92820 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 15.93110 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.66620 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.69050 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.68740 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.78660 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.92900 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 21.86990 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.85200 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 26.32680 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.98170 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 28.09350 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 28.09350 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 28.09350 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 32.01740 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 35.37310 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 32.08160 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 36.09910 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 36.13080 | |||
clblast | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 40.40430 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 40.40430 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 40.40430 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 57.97760 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 57.97760 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 57.97760 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 83.33630 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 83.33630 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 83.33630 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 139.91600 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 139.91600 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 139.91600 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 196.89800 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 196.89800 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 196.89800 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 196.91800 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 196.91800 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 196.91800 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 211.28200 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 211.28200 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 211.28200 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 218.17300 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 218.17300 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 218.17300 | |||
cpu | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 63.30800 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 74.72900 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 97.53600 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 179.51000 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 171.27300 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 181.27700 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 247.68800 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 259.48600 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 258.57300 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 299.98500 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 328.25100 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 295.46100 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 382.67800 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 359.50200 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 358.31700 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 418.99800 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 413.87400 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 412.18800 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 473.47100 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 469.10100 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 469.32800 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 528.70300 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 524.17600 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 529.63500 | |||
cuda | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.06016 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.75808 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.80854 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.18298 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.19043 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.48806 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.90272 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.61917 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.95078 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.35210 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.05660 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.54390 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.46910 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.53240 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.77800 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.87510 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.97540 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.92840 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.94100 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.31430 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.30920 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.03600 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.79570 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.82240 | |||
cudnn | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.97171 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.73840 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.73715 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.73043 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.75091 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.85744 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.78208 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.72678 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.74957 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.70413 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.92998 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.67808 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.76797 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.84070 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.83418 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.80598 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.79590 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.80835 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.04602 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.06547 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.03514 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.49277 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.75603 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.52035 | |||
libdnn-clblas | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.77350 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.91450 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.88083 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.98416 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.96528 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.07856 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.21430 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.09660 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.49350 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.28600 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.80240 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.89980 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.40560 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.07280 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.06350 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.69040 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.48410 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.50480 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.70130 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.75080 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.98890 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.99000 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.05550 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.96450 | |||
libdnn-clblast | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.10918 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.94944 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.78502 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.12384 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.98310 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.01776 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.55070 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.18160 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.13960 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.88740 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.49640 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.26970 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.12010 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.13290 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 14.27350 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.49660 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.55420 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.60810 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.65340 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.68800 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.01760 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.08130 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.04830 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.36150 | |||
libdnn-cuda | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.26362 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.29114 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.27885 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.47693 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.52813 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.47014 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.60378 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.62483 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.59110 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.22534 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.22035 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.20282 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.44928 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.43411 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.42445 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.61843 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.67098 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.64515 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.00020 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.99790 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.00630 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.30990 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.31060 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.27570 | |||
libdnn-viennacl | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.90643 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.89926 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.80960 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.06240 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.53260 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.97667 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.21760 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.10890 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.17320 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.29540 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 13.94280 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.21420 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.21090 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.17000 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.08120 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.47510 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.52330 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.52160 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.95720 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.65920 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.02370 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 23.72790 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.93060 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 20.97250 | |||
nvidia-cuda | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.04173 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.74989 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.17203 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.18394 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.20454 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.50518 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.07526 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.01203 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.43264 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 11.10960 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.26690 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 10.47380 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.79590 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.79920 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.83990 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.09790 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.48690 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.23100 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 16.97520 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.73760 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 17.49400 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.73450 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.66930 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 19.17950 | |||
nvidia-cudnn | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.65626 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.88867 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 2.62349 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.58400 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.59629 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.61517 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.45850 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.46400 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 4.43187 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.72202 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.47738 | |||
2 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 5.45894 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.43686 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.40922 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.46861 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.48256 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.27654 | |||
2 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 7.03898 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.51866 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.49862 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 8.50240 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.27341 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.59594 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.28666 | |||
nvidia-fp16-cuda | 2 | 0 | 31.696448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.48262 | |
1 | 31.696448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.48979 | |||
2 | 31.696448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 3.49798 | |||
4 | 0 | 63.392896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.40397 | ||
1 | 63.392896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.43789 | |||
2 | 63.392896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 6.49421 | |||
6 | 0 | 95.089344 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.43104 | ||
1 | 95.089344 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.40851 | |||
2 | 95.089344 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 9.48291 | |||
8 | 0 | 126.785792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.41980 | ||
1 | 126.785792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.50680 | |||
2 | 126.785792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 12.40780 | |||
10 | 0 | 158.482240 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.42760 | ||
1 | 158.482240 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.42350 | |||
2 | 158.482240 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 15.42550 | |||
12 | 0 | 190.178688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.57080 | ||
1 | 190.178688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.46780 | |||
2 | 190.178688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 18.58760 | |||
14 | 0 | 221.875136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.54700 | ||
1 | 221.875136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.57140 | |||
2 | 221.875136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 21.54800 | |||
16 | 0 | 253.571584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.52480 | ||
1 | 253.571584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.61290 | |||
2 | 253.571584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 24.45410 | |||
nvidia-fp16-cudnn | 2 | 0 | 31.696448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 41.20930 | |
1 | 31.696448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 41.54160 | |||
2 | 31.696448 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 41.89110 | |||
4 | 0 | 63.392896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 54.68850 | ||
1 | 63.392896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 58.69200 | |||
2 | 63.392896 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 56.28020 | |||
6 | 0 | 95.089344 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 59.68280 | ||
1 | 95.089344 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 69.86240 | |||
2 | 95.089344 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 69.88790 | |||
8 | 0 | 126.785792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 83.08120 | ||
1 | 126.785792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 87.91240 | |||
2 | 126.785792 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 81.46330 | |||
10 | 0 | 158.482240 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 105.15400 | ||
1 | 158.482240 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 110.48000 | |||
2 | 158.482240 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 110.49200 | |||
12 | 0 | 190.178688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 117.14100 | ||
1 | 190.178688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 110.69500 | |||
2 | 190.178688 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 113.73100 | |||
14 | 0 | 221.875136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 147.53200 | ||
1 | 221.875136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 150.28900 | |||
2 | 221.875136 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 148.49500 | |||
16 | 0 | 253.571584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 133.65500 | ||
1 | 253.571584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 130.36300 | |||
2 | 253.571584 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 133.14300 | |||
viennacl | 2 | 0 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 22.35900 | |
1 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 22.35900 | |||
2 | 64.629592 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 22.35900 | |||
4 | 0 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 36.00480 | ||
1 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 36.00480 | |||
2 | 129.259184 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 36.00480 | |||
6 | 0 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 52.40320 | ||
1 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 53.95870 | |||
2 | 193.888776 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 52.37770 | |||
8 | 0 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 69.15310 | ||
1 | 258.518368 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 70.35600 | |||
10 | 0 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 85.80200 | ||
1 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 85.80200 | |||
2 | 323.147960 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 85.80200 | |||
12 | 0 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 102.10300 | ||
1 | 387.777552 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 102.51200 | |||
14 | 0 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 116.32100 | ||
1 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 116.32100 | |||
2 | 452.407144 | [{u'index': 0, u'direction': u'forward', u'tim... | no | 116.32100 | |||
16 | 0 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 131.20700 | ||
1 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 129.80700 | |||
2 | 517.036736 | [{u'index': 0, u'direction': u'forward', u'tim... | yes | 130.60700 |
df_time = df_all['time (ms)'].unstack(df_all.index.names[:-1])
pd.options.display.max_columns = len(df_time.columns)
pd.options.display.max_rows = len(df_time.index)
df_time
model | bvlc-alexnet | bvlc-googlenet | deepscale-squeezenet-1.0 | deepscale-squeezenet-1.1 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
lib | clblas | clblast | cpu | cuda | cudnn | libdnn-clblas | libdnn-clblast | libdnn-cuda | libdnn-viennacl | nvidia-cuda | nvidia-cudnn | nvidia-fp16-cuda | nvidia-fp16-cudnn | viennacl | clblas | clblast | cpu | cuda | cudnn | libdnn-clblas | libdnn-clblast | libdnn-cuda | libdnn-viennacl | nvidia-cuda | nvidia-cudnn | nvidia-fp16-cuda | nvidia-fp16-cudnn | viennacl | clblas | clblast | cpu | cuda | cudnn | libdnn-clblas | libdnn-clblast | libdnn-cuda | libdnn-viennacl | nvidia-cuda | nvidia-cudnn | nvidia-fp16-cuda | nvidia-fp16-cudnn | viennacl | clblas | clblast | cpu | cuda | cudnn | libdnn-clblas | libdnn-clblast | libdnn-cuda | libdnn-viennacl | nvidia-cuda | nvidia-cudnn | nvidia-fp16-cuda | nvidia-fp16-cudnn | viennacl | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
batch size | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 |
repetition | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
0 | 8.94669 | 13.2308 | 17.4715 | 21.7293 | 25.9013 | 30.5725 | 34.7750 | 38.66 | 24.8413 | 31.4939 | 43.3787 | 60.3187 | 77.9274 | 78.7724 | 86.1 | 105.131 | 234.212 | 332.208 | 439.481 | 541.650 | 626.306 | 725.671 | 818.676 | 896.596 | 3.31162 | 5.41325 | 6.81779 | 8.73280 | 10.5999 | 13.0693 | 14.4087 | 16.6892 | 2.29418 | 3.17645 | 3.40416 | 3.85306 | 4.54246 | 6.09728 | 5.90950 | 6.95302 | 7.16272 | 8.51354 | 10.1550 | 11.4279 | 13.2485 | 14.6567 | 16.6871 | 18.1873 | 10.9599 | 12.5976 | 13.4973 | 14.8747 | 17.7078 | 18.0429 | 21.1750 | 21.2109 | 4.05811 | 5.38010 | 6.97958 | 8.35891 | 9.80582 | 12.3598 | 13.5209 | 15.0182 | 12.5624 | 13.6909 | 15.9808 | 17.5318 | 20.1001 | 21.9914 | 24.3425 | 27.2678 | 3.33517 | 5.44870 | 6.89562 | 8.60710 | 10.6158 | 13.0591 | 13.7021 | 15.9889 | 2.53763 | 3.50557 | 3.84102 | 4.49024 | 5.28282 | 6.46656 | 6.50342 | 7.82234 | 5.55315 | 8.68966 | 11.6337 | 14.8490 | 18.2252 | 22.8772 | 24.2842 | 29.1922 | 64.4560 | 77.7820 | 94.2712 | 131.131 | 144.081 | 161.196 | 163.903 | 207.027 | 33.4123 | 51.3802 | 79.0916 | 95.7409 | 120.537 | 138.577 | 157.278 | 178.998 | 20.2271 | 31.9764 | 42.5841 | 52.472 | 67.6608 | 75.8549 | 85.1386 | 95.1153 | 84.7452 | 287.857 | 220.177 | 296.289 | 426.498 | 436.972 | 472.283 | 541.096 | 402.637 | 730.002 | 1041.55 | 1353.39 | 1656.32 | 1971.41 | 2293.01 | 3201.15 | 12.0845 | 19.8564 | 28.2013 | 36.5056 | 44.9649 | 53.2084 | 61.6062 | 69.8576 | 7.62266 | 9.13626 | 11.0042 | 12.3075 | 13.9717 | 15.4481 | 17.6753 | 19.0996 | 14.8050 | 18.0050 | 23.6503 | 27.0305 | 30.6809 | 33.7801 | 38.1806 | 42.2688 | 17.0271 | 21.0300 | 23.0646 | 26.7737 | 30.2452 | 34.6532 | 40.6695 | 43.0665 | 11.7586 | 15.2385 | 19.1222 | 23.3976 | 26.0887 | 30.4160 | 34.8799 | 39.1917 | 14.9419 | 18.1405 | 23.7742 | 27.3284 | 29.5699 | 35.1642 | 39.7875 | 42.6988 | 12.0687 | 19.3198 | 26.3270 | 34.8512 | 41.1609 | 47.7555 | 55.1334 | 63.4688 | 7.83389 | 10.07950 | 11.1411 | 13.6366 | 14.8275 | 16.6472 | 18.9700 | 19.9209 | 14.5989 | 28.0771 | 41.4628 | 48.3031 | 62.6837 | 75.8385 | 87.3203 | 97.3967 | 189.631 | 242.865 | 269.285 | 359.081 | 409.447 | 435.993 | 568.728 | 540.863 | 69.3114 | 129.352 | 177.672 | 223.953 | 277.498 | 327.713 | 380.063 | 437.907 | 9.37062 | 15.7873 | 23.2793 | 27.9859 | 34.1444 | 42.6032 | 49.0486 | 52.7688 | 51.7284 | 80.1813 | 91.1629 | 173.55 | 198.661 | 241.687 | 207.421 | 298.166 | 143.628 | 273.392 | 388.014 | 488.155 | 598.731 | 709.801 | 816.995 | 920.774 | 4.52448 | 7.76704 | 10.9641 | 14.3524 | 17.2196 | 20.4538 | 24.0040 | 26.8332 | 3.43040 | 5.50810 | 7.13869 | 8.72858 | 10.3731 | 12.3405 | 14.0268 | 15.9529 | 7.76381 | 10.1530 | 15.7737 | 18.1350 | 19.7717 | 23.0450 | 26.4212 | 32.4741 | 7.80982 | 10.4008 | 14.5818 | 18.1524 | 19.8021 | 24.8617 | 26.5102 | 32.0031 | 4.33152 | 6.80755 | 9.21533 | 11.5405 | 14.2643 | 16.6876 | 19.0987 | 21.4693 | 7.00944 | 10.1817 | 13.3202 | 16.6916 | 21.4211 | 23.0595 | 26.3811 | 31.9927 | 4.50662 | 7.79059 | 10.9373 | 14.0483 | 15.8885 | 18.9850 | 21.7876 | 24.7562 | 3.28704 | 5.01677 | 6.71277 | 8.16253 | 10.0551 | 11.9593 | 13.0920 | 15.0456 | 4.38557 | 8.23296 | 12.1047 | 15.9724 | 19.7745 | 23.8499 | 27.5712 | 31.5983 | 56.2575 | 95.3168 | 119.732 | 144.429 | 188.536 | 228.287 | 246.203 | 259.108 | 32.2142 | 58.8905 | 88.2227 | 114.619 | 137.828 | 158.096 | 183.791 | 207.451 | 7.47315 | 12.9282 | 15.6662 | 21.7866 | 23.8520 | 28.0935 | 32.0174 | 36.0991 | 40.4043 | 57.9776 | 83.3363 | 139.916 | 196.898 | 196.918 | 211.282 | 218.173 | 63.308 | 179.510 | 247.688 | 299.985 | 382.678 | 418.998 | 473.471 | 528.703 | 4.06016 | 6.18298 | 8.90272 | 11.3521 | 13.4691 | 15.8751 | 18.9410 | 21.0360 | 2.97171 | 3.73043 | 4.78208 | 5.70413 | 6.76797 | 7.80598 | 9.04602 | 9.49277 | 6.77350 | 7.98416 | 10.2143 | 12.2860 | 14.4056 | 16.6904 | 18.7013 | 20.9900 | 6.10918 | 9.12384 | 11.5507 | 13.8874 | 16.1201 | 16.4966 | 18.6534 | 21.0813 | 3.26362 | 4.47693 | 5.60378 | 7.22534 | 8.44928 | 9.61843 | 11.0002 | 12.3099 | 5.90643 | 9.06240 | 10.2176 | 12.2954 | 16.2109 | 16.4751 | 20.9572 | 23.7279 | 4.04173 | 6.18394 | 8.07526 | 11.1096 | 12.7959 | 15.0979 | 16.9752 | 19.7345 | 2.65626 | 3.58400 | 4.45850 | 5.72202 | 6.43686 | 7.48256 | 8.51866 | 9.27341 | 3.48262 | 6.40397 | 9.43104 | 12.4198 | 15.4276 | 18.5708 | 21.5470 | 24.5248 | 41.2093 | 54.6885 | 59.6828 | 83.0812 | 105.154 | 117.141 | 147.532 | 133.655 | 22.359 | 36.0048 | 52.4032 | 69.1531 | 85.802 | 102.103 | 116.321 | 131.207 |
1 | 9.20973 | 13.5434 | 17.7254 | 21.7293 | 25.9082 | 30.1898 | 34.4012 | 38.66 | 24.8413 | 31.4939 | 43.3787 | 60.3187 | 77.9274 | 78.7724 | 86.1 | 105.131 | 229.522 | 332.016 | 442.570 | 538.101 | 658.548 | 746.462 | 822.472 | 895.846 | 3.40726 | 5.44371 | 6.83734 | 8.65923 | 10.6322 | 13.0613 | 14.4038 | 16.7772 | 2.29888 | 3.46275 | 3.31066 | 3.86992 | 4.66202 | 5.76614 | 5.93123 | 6.97142 | 7.15498 | 8.20838 | 10.3946 | 11.8057 | 13.5803 | 14.9943 | 17.0874 | 17.8432 | 11.9152 | 12.5901 | 14.3709 | 14.8859 | 16.6543 | 19.0354 | 20.1380 | 21.1949 | 4.03398 | 5.38112 | 6.98595 | 8.35891 | 9.74746 | 12.3965 | 13.5465 | 14.9985 | 12.5614 | 13.9348 | 15.9263 | 17.6275 | 20.0489 | 21.8837 | 24.3248 | 28.0566 | 3.37514 | 5.43642 | 6.90381 | 8.65680 | 10.6260 | 12.5369 | 13.6878 | 15.9089 | 2.53747 | 3.50733 | 3.85741 | 4.50294 | 5.38931 | 6.53376 | 6.18394 | 7.93043 | 5.46611 | 8.61594 | 11.6285 | 14.8429 | 18.2442 | 22.8823 | 24.3057 | 29.3478 | 62.9166 | 77.7984 | 92.1433 | 130.302 | 144.095 | 151.155 | 163.977 | 207.111 | 33.4961 | 51.3802 | 77.7400 | 98.8753 | 117.681 | 138.577 | 157.278 | 182.175 | 18.7884 | 31.9764 | 42.6047 | 53.040 | 63.2433 | 75.8549 | 91.5833 | 95.1153 | 84.7452 | 287.857 | 220.177 | 296.289 | 426.498 | 436.972 | 472.283 | 541.096 | 406.118 | 730.530 | 1038.72 | 1353.79 | 1654.63 | 1958.12 | 2264.95 | 2565.12 | 11.5620 | 19.8411 | 28.1129 | 36.5898 | 44.8113 | 53.0628 | 61.5455 | 70.7556 | 7.64112 | 9.10643 | 11.0500 | 12.3412 | 13.9735 | 15.4862 | 17.7167 | 19.0753 | 16.0604 | 19.2020 | 22.2802 | 25.7475 | 29.2618 | 33.6360 | 39.5121 | 42.3424 | 16.4690 | 19.3717 | 23.0595 | 26.5452 | 30.4238 | 36.1083 | 39.0542 | 44.7150 | 11.7627 | 15.1659 | 19.0814 | 23.3792 | 26.0987 | 30.6596 | 34.9624 | 39.2849 | 14.9685 | 18.2487 | 22.6119 | 27.3072 | 29.5414 | 35.3628 | 38.5085 | 44.0873 | 11.4381 | 18.5436 | 26.7214 | 32.9083 | 40.3251 | 47.8198 | 55.0904 | 62.8244 | 7.78554 | 9.70973 | 11.0871 | 12.9970 | 14.8685 | 16.6851 | 19.7460 | 19.7673 | 14.6432 | 27.9818 | 41.8017 | 51.2748 | 68.4206 | 75.7309 | 88.6436 | 96.8141 | 189.921 | 247.955 | 269.246 | 358.739 | 397.515 | 434.659 | 562.519 | 547.854 | 66.6368 | 129.352 | 177.672 | 223.953 | 277.498 | 327.713 | 380.063 | 437.907 | 9.38189 | 17.4828 | 23.2793 | 27.9859 | 34.1444 | 40.2719 | 49.0486 | 52.7688 | 51.7284 | 80.1813 | 91.1629 | 173.55 | 198.661 | 241.687 | 207.421 | 298.166 | 129.419 | 256.957 | 375.025 | 488.581 | 599.374 | 710.508 | 816.698 | 920.035 | 4.77594 | 8.04790 | 11.2302 | 14.3534 | 17.2616 | 20.4974 | 23.6687 | 26.8494 | 3.43981 | 6.01792 | 7.10045 | 9.06240 | 10.3457 | 12.3443 | 14.1080 | 16.2386 | 6.96323 | 10.2135 | 15.8258 | 18.1524 | 19.7335 | 25.7574 | 29.3396 | 32.0513 | 6.97024 | 10.2113 | 13.4686 | 16.7260 | 19.8154 | 24.9127 | 29.4665 | 32.0174 | 5.10640 | 6.80960 | 9.20474 | 12.1508 | 14.2531 | 16.7178 | 19.6763 | 21.4804 | 8.62618 | 12.0320 | 15.5502 | 18.0931 | 19.7007 | 23.0728 | 28.5417 | 32.6224 | 4.52403 | 7.57453 | 10.9333 | 13.3745 | 16.5482 | 18.8682 | 22.5240 | 24.7173 | 3.28115 | 4.97382 | 6.71264 | 8.34294 | 10.0303 | 11.7688 | 13.4441 | 15.2395 | 4.38989 | 8.29338 | 12.1454 | 16.0369 | 19.8349 | 23.7896 | 27.5729 | 30.7835 | 62.3892 | 95.4440 | 125.522 | 143.404 | 188.613 | 228.748 | 246.246 | 258.151 | 31.9713 | 59.9903 | 87.3996 | 114.619 | 136.132 | 158.753 | 183.791 | 207.451 | 10.78710 | 15.9311 | 15.6905 | 21.9290 | 26.3268 | 28.0935 | 35.3731 | 36.1308 | 40.4043 | 57.9776 | 83.3363 | 139.916 | 196.898 | 196.918 | 211.282 | 218.173 | 74.729 | 171.273 | 259.486 | 328.251 | 359.502 | 413.874 | 469.101 | 524.176 | 3.75808 | 6.19043 | 8.61917 | 11.0566 | 13.5324 | 15.9754 | 18.3143 | 20.7957 | 2.73840 | 3.75091 | 4.72678 | 5.92998 | 6.84070 | 7.79590 | 9.06547 | 9.75603 | 7.91450 | 8.96528 | 10.0966 | 15.8024 | 16.0728 | 20.4841 | 18.7508 | 21.0555 | 5.94944 | 7.98310 | 10.1816 | 13.4964 | 16.1329 | 16.5542 | 18.6880 | 21.0483 | 3.29114 | 4.52813 | 5.62483 | 7.22035 | 8.43411 | 9.67098 | 10.9979 | 12.3106 | 5.89926 | 10.53260 | 10.1089 | 13.9428 | 16.1700 | 16.5233 | 18.6592 | 20.9306 | 3.74989 | 6.20454 | 8.01203 | 10.2669 | 12.7992 | 15.4869 | 17.7376 | 19.6693 | 2.88867 | 3.59629 | 4.46400 | 5.47738 | 6.40922 | 7.27654 | 8.49862 | 9.59594 | 3.48979 | 6.43789 | 9.40851 | 12.5068 | 15.4235 | 18.4678 | 21.5714 | 24.6129 | 41.5416 | 58.6920 | 69.8624 | 87.9124 | 110.480 | 110.695 | 150.289 | 130.363 | 22.359 | 36.0048 | 53.9587 | 70.3560 | 85.802 | 102.512 | 116.321 | 129.807 |
2 | 9.24682 | NaN | NaN | 21.7293 | 25.9031 | 30.5603 | NaN | 38.66 | 24.8413 | 31.4939 | 43.3787 | 60.3187 | 77.9274 | 78.7724 | 86.1 | 105.131 | 226.047 | 331.484 | 437.235 | 537.458 | 633.620 | 735.054 | 822.192 | 894.858 | 3.41606 | 5.57187 | 6.83315 | 8.63539 | 10.6140 | 13.0482 | 14.4274 | 16.6820 | 2.29994 | 3.18464 | 3.31469 | 3.87277 | 4.54394 | 5.78051 | 5.93114 | 6.96934 | 7.45392 | 8.23578 | 10.3925 | 11.4843 | 13.4678 | 14.5582 | 17.1346 | 17.8258 | 11.8016 | 11.6912 | 14.4332 | 15.8351 | 16.6380 | 19.0310 | 21.1221 | 21.1180 | 4.04106 | 5.36480 | 6.97456 | 8.26493 | 9.73843 | 12.3392 | 13.3550 | 14.9862 | 12.6062 | 13.7134 | 15.6613 | 17.4870 | 20.0908 | 21.6451 | 24.3476 | 28.0742 | 3.32813 | 5.42618 | 6.89955 | 8.60672 | 10.7298 | 12.4900 | 13.7103 | 15.9171 | 2.63475 | 3.52179 | 3.85600 | 4.53939 | 5.50605 | 6.47616 | 6.51366 | 7.61776 | 5.54477 | 8.59136 | 11.6459 | 14.8224 | 18.2835 | 22.8076 | 24.3139 | 29.3140 | 62.2633 | 77.7544 | 94.2827 | 131.532 | 144.237 | 151.105 | 163.968 | 207.087 | 33.6517 | 51.3802 | NaN | NaN | NaN | 138.577 | 157.278 | NaN | 20.2813 | 31.9764 | NaN | NaN | NaN | 75.8549 | 94.0531 | 95.1153 | 84.7452 | 287.857 | 220.177 | 296.289 | 426.498 | 436.972 | 472.283 | 541.096 | 403.492 | 731.393 | 1037.39 | 1356.01 | 1660.07 | 1955.44 | 2264.29 | 2562.73 | 11.3504 | 20.3858 | 28.1420 | 37.1661 | 44.7314 | 52.6459 | 61.4502 | 69.8065 | 7.68614 | 9.81008 | 11.0326 | 12.2819 | 13.9820 | 15.5658 | 17.6944 | 19.0833 | 14.7498 | 17.9232 | 22.4336 | 25.7300 | 29.5178 | 33.8706 | 38.2061 | 43.5868 | 15.6498 | 19.3311 | 24.7129 | 27.9470 | 30.3010 | 34.5180 | 40.6661 | 43.0355 | 11.7629 | 15.1480 | 19.0344 | 23.5428 | 26.1933 | 30.4753 | 34.9739 | 41.9565 | 16.2560 | 19.4971 | 22.5587 | 25.8621 | 30.8541 | 35.2164 | 38.5114 | 42.8277 | 11.3996 | 20.0212 | 26.3629 | 33.7246 | 40.2791 | 47.9631 | 55.1795 | 62.5553 | 7.78150 | 10.53210 | 11.1690 | 13.1359 | 14.8439 | 15.9215 | 18.3859 | 20.3602 | 14.5725 | 28.2109 | 38.6621 | 54.9243 | 62.8537 | 75.5999 | 83.2459 | 98.8478 | 189.883 | 247.907 | 269.196 | 359.102 | 409.493 | 434.672 | 563.157 | 535.674 | NaN | 129.352 | 177.672 | 223.953 | 277.498 | 327.713 | 380.063 | 437.907 | NaN | 15.6959 | 23.2793 | 27.9859 | 34.1444 | 42.7346 | 49.0486 | 52.7688 | 51.7284 | 80.1813 | 91.1629 | 173.55 | 198.661 | 241.687 | 207.421 | 298.166 | 121.009 | 277.893 | 368.990 | 487.887 | 600.141 | 710.054 | 834.054 | 917.351 | 4.49741 | 8.07117 | 10.8943 | 14.3576 | 17.2391 | 20.9771 | 23.6527 | 27.1217 | 3.46362 | 5.45894 | 7.48154 | 8.76237 | 10.3447 | 12.3159 | 14.0268 | 15.9437 | 6.99824 | 10.2146 | 15.4061 | 16.7148 | 19.7500 | 23.0543 | 28.5196 | 32.0369 | 6.99699 | 11.0804 | 15.5967 | 18.1133 | 21.4344 | 23.0052 | 26.5033 | 29.7728 | 4.34790 | 7.50490 | 9.19354 | 12.2030 | 14.8869 | 17.3496 | 19.1672 | 21.5143 | 7.02976 | 10.1510 | 13.2452 | 16.7434 | 19.7759 | 23.5179 | 26.3885 | 32.9216 | 5.27910 | 7.77318 | 10.4110 | 13.0237 | 16.4763 | 18.8807 | 21.8835 | 24.7859 | 3.30240 | 4.96550 | 6.71539 | 8.01821 | 10.1990 | 11.7364 | 13.4630 | 15.3016 | 4.46054 | 8.29843 | 12.1751 | 15.9539 | 19.8083 | 23.7881 | 27.5853 | 31.5515 | 61.9684 | 95.2891 | 117.376 | 141.640 | 188.587 | 228.659 | 246.125 | 258.404 | NaN | NaN | 87.2458 | 114.619 | 134.602 | 160.882 | 183.791 | 207.451 | 8.49402 | NaN | 15.6874 | 21.8699 | 23.9817 | 28.0935 | 32.0816 | NaN | 40.4043 | 57.9776 | 83.3363 | 139.916 | 196.898 | 196.918 | 211.282 | 218.173 | 97.536 | 181.277 | 258.573 | 295.461 | 358.317 | 412.188 | 469.328 | 529.635 | 3.80854 | 6.48806 | 8.95078 | 11.5439 | 13.7780 | 15.9284 | 18.3092 | 20.8224 | 2.73715 | 3.85744 | 4.74957 | 5.67808 | 6.83418 | 7.80835 | 9.03514 | 9.52035 | 5.88083 | 9.07856 | 11.4935 | 13.8998 | 16.0635 | 16.5048 | 20.9889 | 20.9645 | 6.78502 | 8.01776 | 10.1396 | 12.2697 | 14.2735 | 18.6081 | 21.0176 | 21.3615 | 3.27885 | 4.47014 | 5.59110 | 7.20282 | 8.42445 | 9.64515 | 11.0063 | 12.2757 | 6.80960 | 7.97667 | 10.1732 | 12.2142 | 16.0812 | 16.5216 | 21.0237 | 20.9725 | 4.17203 | 6.50518 | 8.43264 | 10.4738 | 12.8399 | 15.2310 | 17.4940 | 19.1795 | 2.62349 | 3.61517 | 4.43187 | 5.45894 | 6.46861 | 7.03898 | 8.50240 | 9.28666 | 3.49798 | 6.49421 | 9.48291 | 12.4078 | 15.4255 | 18.5876 | 21.5480 | 24.4541 | 41.8911 | 56.2802 | 69.8879 | 81.4633 | 110.492 | 113.731 | 148.495 | 133.143 | 22.359 | 36.0048 | 52.3777 | NaN | 85.802 | NaN | 116.321 | 130.607 |
df_mean_time_per_batch = df_time.describe().ix['mean'].unstack(level='batch size')
pd.options.display.max_columns = len(df_mean_time_per_batch.columns)
pd.options.display.max_rows = len(df_mean_time_per_batch.index)
df_mean_time_per_batch
batch size | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | |
---|---|---|---|---|---|---|---|---|---|
model | lib | ||||||||
bvlc-alexnet | clblas | 9.134413 | 13.387100 | 17.598450 | 21.729300 | 25.904200 | 30.440867 | 34.588100 | 38.660000 |
clblast | 24.841300 | 31.493900 | 43.378700 | 60.318700 | 77.927400 | 78.772400 | 86.100000 | 105.131000 | |
cpu | 229.927000 | 331.902667 | 439.762000 | 539.069667 | 639.491333 | 735.729000 | 821.113333 | 895.766667 | |
cuda | 3.378313 | 5.476277 | 6.829427 | 8.675807 | 10.615367 | 13.059600 | 14.413300 | 16.716133 | |
cudnn | 2.297667 | 3.274613 | 3.343170 | 3.865250 | 4.582807 | 5.881310 | 5.923957 | 6.964593 | |
libdnn-clblas | 7.257207 | 8.319233 | 10.314033 | 11.572633 | 13.432200 | 14.736400 | 16.969700 | 17.952100 | |
libdnn-clblast | 11.558900 | 12.292967 | 14.100467 | 15.198567 | 17.000033 | 18.703100 | 20.811700 | 21.174600 | |
libdnn-cuda | 4.044383 | 5.375340 | 6.980030 | 8.327583 | 9.763903 | 12.365167 | 13.474133 | 15.000967 | |
libdnn-viennacl | 12.576667 | 13.779700 | 15.856133 | 17.548767 | 20.079933 | 21.840067 | 24.338300 | 27.799533 | |
nvidia-cuda | 3.346147 | 5.437100 | 6.899660 | 8.623540 | 10.657200 | 12.695333 | 13.700067 | 15.938300 | |
nvidia-cudnn | 2.569950 | 3.511563 | 3.851477 | 4.510857 | 5.392727 | 6.492160 | 6.400340 | 7.790177 | |
nvidia-fp16-cuda | 5.521343 | 8.632320 | 11.636033 | 14.838100 | 18.250967 | 22.855700 | 24.301267 | 29.284667 | |
nvidia-fp16-cudnn | 63.211967 | 77.778267 | 93.565733 | 130.988333 | 144.137667 | 154.485333 | 163.949333 | 207.075000 | |
viennacl | 33.520033 | 51.380200 | 78.415800 | 97.308100 | 119.109000 | 138.577000 | 157.278000 | 180.586500 | |
bvlc-googlenet | clblas | 19.765600 | 31.976400 | 42.594400 | 52.756000 | 65.452050 | 75.854900 | 90.258333 | 95.115300 |
clblast | 84.745200 | 287.857000 | 220.177000 | 296.289000 | 426.498000 | 436.972000 | 472.283000 | 541.096000 | |
cpu | 404.082333 | 730.641667 | 1039.220000 | 1354.396667 | 1657.006667 | 1961.656667 | 2274.083333 | 2776.333333 | |
cuda | 11.665633 | 20.027767 | 28.152067 | 36.753833 | 44.835867 | 52.972367 | 61.533967 | 70.139900 | |
cudnn | 7.649973 | 9.350923 | 11.028933 | 12.310200 | 13.975733 | 15.500033 | 17.695467 | 19.086067 | |
libdnn-clblas | 15.205067 | 18.376733 | 22.788033 | 26.169333 | 29.820167 | 33.762233 | 38.632933 | 42.732667 | |
libdnn-clblast | 16.381967 | 19.910933 | 23.612333 | 27.088633 | 30.323333 | 35.093167 | 40.129933 | 43.605667 | |
libdnn-cuda | 11.761400 | 15.184133 | 19.079333 | 23.439867 | 26.126900 | 30.516967 | 34.938733 | 40.144367 | |
libdnn-viennacl | 15.388800 | 18.628767 | 22.981600 | 26.832567 | 29.988467 | 35.247800 | 38.935800 | 43.204600 | |
nvidia-cuda | 11.635467 | 19.294867 | 26.470433 | 33.828033 | 40.588367 | 47.846133 | 55.134433 | 62.949500 | |
nvidia-cudnn | 7.800310 | 10.107110 | 11.132400 | 13.256500 | 14.846633 | 16.417933 | 19.033967 | 20.016133 | |
nvidia-fp16-cuda | 14.604867 | 28.089933 | 40.642200 | 51.500733 | 64.652667 | 75.723100 | 86.403267 | 97.686200 | |
nvidia-fp16-cudnn | 189.811667 | 246.242333 | 269.242333 | 358.974000 | 405.485000 | 435.108000 | 564.801333 | 541.463667 | |
viennacl | 67.974100 | 129.352000 | 177.672000 | 223.953000 | 277.498000 | 327.713000 | 380.063000 | 437.907000 | |
deepscale-squeezenet-1.0 | clblas | 9.376255 | 16.322000 | 23.279300 | 27.985900 | 34.144400 | 41.869900 | 49.048600 | 52.768800 |
clblast | 51.728400 | 80.181300 | 91.162900 | 173.550000 | 198.661000 | 241.687000 | 207.421000 | 298.166000 | |
cpu | 131.352000 | 269.414000 | 377.343000 | 488.207667 | 599.415333 | 710.121000 | 822.582333 | 919.386667 | |
cuda | 4.599277 | 7.962037 | 11.029533 | 14.354467 | 17.240100 | 20.642767 | 23.775133 | 26.934767 | |
cudnn | 3.444610 | 5.661653 | 7.240227 | 8.851117 | 10.354500 | 12.333567 | 14.053867 | 16.045067 | |
libdnn-clblas | 7.241760 | 10.193700 | 15.668533 | 17.667400 | 19.751733 | 23.952233 | 28.093467 | 32.187433 | |
libdnn-clblast | 7.259017 | 10.564167 | 14.549033 | 17.663900 | 20.350633 | 24.259867 | 27.493333 | 31.264433 | |
libdnn-cuda | 4.595273 | 7.040683 | 9.204537 | 11.964767 | 14.468100 | 16.918333 | 19.314067 | 21.488000 | |
libdnn-viennacl | 7.555127 | 10.788233 | 14.038533 | 17.176033 | 20.299233 | 23.216733 | 27.103767 | 32.512233 | |
nvidia-cuda | 4.769917 | 7.712767 | 10.760533 | 13.482167 | 16.304333 | 18.911300 | 22.065033 | 24.753133 | |
nvidia-cudnn | 3.290197 | 4.985363 | 6.713600 | 8.174560 | 10.094800 | 11.821500 | 13.333033 | 15.195567 | |
nvidia-fp16-cuda | 4.412000 | 8.274923 | 12.141733 | 15.987733 | 19.805900 | 23.809200 | 27.576467 | 31.311100 | |
nvidia-fp16-cudnn | 60.205033 | 95.349967 | 120.876667 | 143.157667 | 188.578667 | 228.564667 | 246.191333 | 258.554333 | |
viennacl | 32.092750 | 59.440400 | 87.622700 | 114.619000 | 136.187333 | 159.243667 | 183.791000 | 207.451000 | |
deepscale-squeezenet-1.1 | clblas | 8.918090 | 14.429650 | 15.681367 | 21.861833 | 24.720167 | 28.093500 | 33.157367 | 36.114950 |
clblast | 40.404300 | 57.977600 | 83.336300 | 139.916000 | 196.898000 | 196.918000 | 211.282000 | 218.173000 | |
cpu | 78.524333 | 177.353333 | 255.249000 | 307.899000 | 366.832333 | 415.020000 | 470.633333 | 527.504667 | |
cuda | 3.875593 | 6.287157 | 8.824223 | 11.317533 | 13.593167 | 15.926300 | 18.521500 | 20.884700 | |
cudnn | 2.815753 | 3.779593 | 4.752810 | 5.770730 | 6.814283 | 7.803410 | 9.048877 | 9.589717 | |
libdnn-clblas | 6.856277 | 8.676000 | 10.601467 | 13.996067 | 15.513967 | 17.893100 | 19.480333 | 21.003333 | |
libdnn-clblast | 6.281213 | 8.374900 | 10.623967 | 13.217833 | 15.508833 | 17.219633 | 19.453000 | 21.163700 | |
libdnn-cuda | 3.277870 | 4.491733 | 5.606570 | 7.216170 | 8.435947 | 9.644853 | 11.001467 | 12.298733 | |
libdnn-viennacl | 6.205097 | 9.190557 | 10.166567 | 12.817467 | 16.154033 | 16.506667 | 20.213367 | 21.877000 | |
nvidia-cuda | 3.987883 | 6.297887 | 8.173310 | 10.616767 | 12.811667 | 15.271933 | 17.402267 | 19.527767 | |
nvidia-cudnn | 2.722807 | 3.598487 | 4.451457 | 5.552780 | 6.438230 | 7.266027 | 8.506560 | 9.385337 | |
nvidia-fp16-cuda | 3.490130 | 6.445357 | 9.440820 | 12.444800 | 15.425533 | 18.542067 | 21.555467 | 24.530600 | |
nvidia-fp16-cudnn | 41.547333 | 56.553567 | 66.477700 | 84.152300 | 108.708667 | 113.855667 | 148.772000 | 132.387000 | |
viennacl | 22.359000 | 36.004800 | 52.913200 | 69.754550 | 85.802000 | 102.307500 | 116.321000 | 130.540333 |
batch_sizes = df_mean_time_per_batch.columns.tolist()
# batch_sizes
df_mean_time_per_image = df_mean_time_per_batch / batch_sizes
pd.options.display.max_columns = len(df_mean_time_per_image.columns)
pd.options.display.max_rows = len(df_mean_time_per_image.index)
df_mean_time_per_image
batch size | 2 | 4 | 6 | 8 | 10 | 12 | 14 | 16 | |
---|---|---|---|---|---|---|---|---|---|
model | lib | ||||||||
bvlc-alexnet | clblas | 4.567207 | 3.346775 | 2.933075 | 2.716162 | 2.590420 | 2.536739 | 2.470579 | 2.416250 |
clblast | 12.420650 | 7.873475 | 7.229783 | 7.539837 | 7.792740 | 6.564367 | 6.150000 | 6.570688 | |
cpu | 114.963500 | 82.975667 | 73.293667 | 67.383708 | 63.949133 | 61.310750 | 58.650952 | 55.985417 | |
cuda | 1.689157 | 1.369069 | 1.138238 | 1.084476 | 1.061537 | 1.088300 | 1.029521 | 1.044758 | |
cudnn | 1.148833 | 0.818653 | 0.557195 | 0.483156 | 0.458281 | 0.490109 | 0.423140 | 0.435287 | |
libdnn-clblas | 3.628603 | 2.079808 | 1.719006 | 1.446579 | 1.343220 | 1.228033 | 1.212121 | 1.122006 | |
libdnn-clblast | 5.779450 | 3.073242 | 2.350078 | 1.899821 | 1.700003 | 1.558592 | 1.486550 | 1.323412 | |
libdnn-cuda | 2.022192 | 1.343835 | 1.163338 | 1.040948 | 0.976390 | 1.030431 | 0.962438 | 0.937560 | |
libdnn-viennacl | 6.288333 | 3.444925 | 2.642689 | 2.193596 | 2.007993 | 1.820006 | 1.738450 | 1.737471 | |
nvidia-cuda | 1.673073 | 1.359275 | 1.149943 | 1.077943 | 1.065720 | 1.057944 | 0.978576 | 0.996144 | |
nvidia-cudnn | 1.284975 | 0.877891 | 0.641913 | 0.563857 | 0.539273 | 0.541013 | 0.457167 | 0.486886 | |
nvidia-fp16-cuda | 2.760672 | 2.158080 | 1.939339 | 1.854762 | 1.825097 | 1.904642 | 1.735805 | 1.830292 | |
nvidia-fp16-cudnn | 31.605983 | 19.444567 | 15.594289 | 16.373542 | 14.413767 | 12.873778 | 11.710667 | 12.942187 | |
viennacl | 16.760017 | 12.845050 | 13.069300 | 12.163512 | 11.910900 | 11.548083 | 11.234143 | 11.286656 | |
bvlc-googlenet | clblas | 9.882800 | 7.994100 | 7.099067 | 6.594500 | 6.545205 | 6.321242 | 6.447024 | 5.944706 |
clblast | 42.372600 | 71.964250 | 36.696167 | 37.036125 | 42.649800 | 36.414333 | 33.734500 | 33.818500 | |
cpu | 202.041167 | 182.660417 | 173.203333 | 169.299583 | 165.700667 | 163.471389 | 162.434524 | 173.520833 | |
cuda | 5.832817 | 5.006942 | 4.692011 | 4.594229 | 4.483587 | 4.414364 | 4.395283 | 4.383744 | |
cudnn | 3.824987 | 2.337731 | 1.838156 | 1.538775 | 1.397573 | 1.291669 | 1.263962 | 1.192879 | |
libdnn-clblas | 7.602533 | 4.594183 | 3.798006 | 3.271167 | 2.982017 | 2.813519 | 2.759495 | 2.670792 | |
libdnn-clblast | 8.190983 | 4.977733 | 3.935389 | 3.386079 | 3.032333 | 2.924431 | 2.866424 | 2.725354 | |
libdnn-cuda | 5.880700 | 3.796033 | 3.179889 | 2.929983 | 2.612690 | 2.543081 | 2.495624 | 2.509023 | |
libdnn-viennacl | 7.694400 | 4.657192 | 3.830267 | 3.354071 | 2.998847 | 2.937317 | 2.781129 | 2.700287 | |
nvidia-cuda | 5.817733 | 4.823717 | 4.411739 | 4.228504 | 4.058837 | 3.987178 | 3.938174 | 3.934344 | |
nvidia-cudnn | 3.900155 | 2.526778 | 1.855400 | 1.657063 | 1.484663 | 1.368161 | 1.359569 | 1.251008 | |
nvidia-fp16-cuda | 7.302433 | 7.022483 | 6.773700 | 6.437592 | 6.465267 | 6.310258 | 6.171662 | 6.105387 | |
nvidia-fp16-cudnn | 94.905833 | 61.560583 | 44.873722 | 44.871750 | 40.548500 | 36.259000 | 40.342952 | 33.841479 | |
viennacl | 33.987050 | 32.338000 | 29.612000 | 27.994125 | 27.749800 | 27.309417 | 27.147357 | 27.369187 | |
deepscale-squeezenet-1.0 | clblas | 4.688128 | 4.080500 | 3.879883 | 3.498238 | 3.414440 | 3.489158 | 3.503471 | 3.298050 |
clblast | 25.864200 | 20.045325 | 15.193817 | 21.693750 | 19.866100 | 20.140583 | 14.815786 | 18.635375 | |
cpu | 65.676000 | 67.353500 | 62.890500 | 61.025958 | 59.941533 | 59.176750 | 58.755881 | 57.461667 | |
cuda | 2.299638 | 1.990509 | 1.838256 | 1.794308 | 1.724010 | 1.720231 | 1.698224 | 1.683423 | |
cudnn | 1.722305 | 1.415413 | 1.206704 | 1.106390 | 1.035450 | 1.027797 | 1.003848 | 1.002817 | |
libdnn-clblas | 3.620880 | 2.548425 | 2.611422 | 2.208425 | 1.975173 | 1.996019 | 2.006676 | 2.011715 | |
libdnn-clblast | 3.629508 | 2.641042 | 2.424839 | 2.207987 | 2.035063 | 2.021656 | 1.963810 | 1.954027 | |
libdnn-cuda | 2.297637 | 1.760171 | 1.534089 | 1.495596 | 1.446810 | 1.409861 | 1.379576 | 1.343000 | |
libdnn-viennacl | 3.777563 | 2.697058 | 2.339756 | 2.147004 | 2.029923 | 1.934728 | 1.935983 | 2.032015 | |
nvidia-cuda | 2.384958 | 1.928192 | 1.793422 | 1.685271 | 1.630433 | 1.575942 | 1.576074 | 1.547071 | |
nvidia-cudnn | 1.645098 | 1.246341 | 1.118933 | 1.021820 | 1.009480 | 0.985125 | 0.952360 | 0.949723 | |
nvidia-fp16-cuda | 2.206000 | 2.068731 | 2.023622 | 1.998467 | 1.980590 | 1.984100 | 1.969748 | 1.956944 | |
nvidia-fp16-cudnn | 30.102517 | 23.837492 | 20.146111 | 17.894708 | 18.857867 | 19.047056 | 17.585095 | 16.159646 | |
viennacl | 16.046375 | 14.860100 | 14.603783 | 14.327375 | 13.618733 | 13.270306 | 13.127929 | 12.965687 | |
deepscale-squeezenet-1.1 | clblas | 4.459045 | 3.607413 | 2.613561 | 2.732729 | 2.472017 | 2.341125 | 2.368383 | 2.257184 |
clblast | 20.202150 | 14.494400 | 13.889383 | 17.489500 | 19.689800 | 16.409833 | 15.091571 | 13.635813 | |
cpu | 39.262167 | 44.338333 | 42.541500 | 38.487375 | 36.683233 | 34.585000 | 33.616667 | 32.969042 | |
cuda | 1.937797 | 1.571789 | 1.470704 | 1.414692 | 1.359317 | 1.327192 | 1.322964 | 1.305294 | |
cudnn | 1.407877 | 0.944898 | 0.792135 | 0.721341 | 0.681428 | 0.650284 | 0.646348 | 0.599357 | |
libdnn-clblas | 3.428138 | 2.169000 | 1.766911 | 1.749508 | 1.551397 | 1.491092 | 1.391452 | 1.312708 | |
libdnn-clblast | 3.140607 | 2.093725 | 1.770661 | 1.652229 | 1.550883 | 1.434969 | 1.389500 | 1.322731 | |
libdnn-cuda | 1.638935 | 1.122933 | 0.934428 | 0.902021 | 0.843595 | 0.803738 | 0.785819 | 0.768671 | |
libdnn-viennacl | 3.102548 | 2.297639 | 1.694428 | 1.602183 | 1.615403 | 1.375556 | 1.443812 | 1.367312 | |
nvidia-cuda | 1.993942 | 1.574472 | 1.362218 | 1.327096 | 1.281167 | 1.272661 | 1.243019 | 1.220485 | |
nvidia-cudnn | 1.361403 | 0.899622 | 0.741909 | 0.694098 | 0.643823 | 0.605502 | 0.607611 | 0.586584 | |
nvidia-fp16-cuda | 1.745065 | 1.611339 | 1.573470 | 1.555600 | 1.542553 | 1.545172 | 1.539676 | 1.533162 | |
nvidia-fp16-cudnn | 20.773667 | 14.138392 | 11.079617 | 10.519038 | 10.870867 | 9.487972 | 10.626571 | 8.274188 | |
viennacl | 11.179500 | 9.001200 | 8.818867 | 8.719319 | 8.580200 | 8.525625 | 8.308643 | 8.158771 |
df_mean_time_per_image.min(axis=1)
model lib bvlc-alexnet clblas 2.416250 clblast 6.150000 cpu 55.985417 cuda 1.029521 cudnn 0.423140 libdnn-clblas 1.122006 libdnn-clblast 1.323412 libdnn-cuda 0.937560 libdnn-viennacl 1.737471 nvidia-cuda 0.978576 nvidia-cudnn 0.457167 nvidia-fp16-cuda 1.735805 nvidia-fp16-cudnn 11.710667 viennacl 11.234143 bvlc-googlenet clblas 5.944706 clblast 33.734500 cpu 162.434524 cuda 4.383744 cudnn 1.192879 libdnn-clblas 2.670792 libdnn-clblast 2.725354 libdnn-cuda 2.495624 libdnn-viennacl 2.700287 nvidia-cuda 3.934344 nvidia-cudnn 1.251008 nvidia-fp16-cuda 6.105387 nvidia-fp16-cudnn 33.841479 viennacl 27.147357 deepscale-squeezenet-1.0 clblas 3.298050 clblast 14.815786 cpu 57.461667 cuda 1.683423 cudnn 1.002817 libdnn-clblas 1.975173 libdnn-clblast 1.954027 libdnn-cuda 1.343000 libdnn-viennacl 1.934728 nvidia-cuda 1.547071 nvidia-cudnn 0.949723 nvidia-fp16-cuda 1.956944 nvidia-fp16-cudnn 16.159646 viennacl 12.965687 deepscale-squeezenet-1.1 clblas 2.257184 clblast 13.635813 cpu 32.969042 cuda 1.305294 cudnn 0.599357 libdnn-clblas 1.312708 libdnn-clblast 1.322731 libdnn-cuda 0.768671 libdnn-viennacl 1.367312 nvidia-cuda 1.220485 nvidia-cudnn 0.586584 nvidia-fp16-cuda 1.533162 nvidia-fp16-cudnn 8.274188 viennacl 8.158771 dtype: float64
plot_max_num_images_per_second(df_mean_time_per_image, libs_to_drop=[], fontsize=14)
# What is the batch size that gives the minimum time per image (or the maximum number of images per second)?
df_mean_time_per_image.idxmin(axis=1)
model lib bvlc-alexnet clblas 16 clblast 14 cpu 16 cuda 14 cudnn 14 libdnn-clblas 16 libdnn-clblast 16 libdnn-cuda 16 libdnn-viennacl 16 nvidia-cuda 14 nvidia-cudnn 14 nvidia-fp16-cuda 14 nvidia-fp16-cudnn 14 viennacl 14 bvlc-googlenet clblas 16 clblast 14 cpu 14 cuda 16 cudnn 16 libdnn-clblas 16 libdnn-clblast 16 libdnn-cuda 14 libdnn-viennacl 16 nvidia-cuda 16 nvidia-cudnn 16 nvidia-fp16-cuda 16 nvidia-fp16-cudnn 16 viennacl 14 deepscale-squeezenet-1.0 clblas 16 clblast 14 cpu 16 cuda 16 cudnn 16 libdnn-clblas 10 libdnn-clblast 16 libdnn-cuda 16 libdnn-viennacl 12 nvidia-cuda 16 nvidia-cudnn 16 nvidia-fp16-cuda 16 nvidia-fp16-cudnn 16 viennacl 16 deepscale-squeezenet-1.1 clblas 16 clblast 16 cpu 16 cuda 16 cudnn 16 libdnn-clblas 16 libdnn-clblast 16 libdnn-cuda 16 libdnn-viennacl 16 nvidia-cuda 16 nvidia-cudnn 16 nvidia-fp16-cuda 16 nvidia-fp16-cudnn 16 viennacl 16 dtype: int64
# Focus on e.g. nvidia-fp16-cuda, for which the batch size of 16 is not always the best.
df_mean_time_per_image.idxmin(axis=1).reorder_levels(['lib', 'model']).loc['nvidia-fp16-cuda']
model bvlc-alexnet 14 bvlc-googlenet 16 deepscale-squeezenet-1.0 16 deepscale-squeezenet-1.1 16 dtype: int64
# # Is the same answer as via .min(axis=1).values?
# df_mean_time_per_image.lookup(df_mean_time_per_image.index, df_mean_time_per_image.idxmin(axis=1)) \
# == df_mean_time_per_image.min(axis=1).values
df_time_per_image = df_time / (batch_sizes*(len(df_time.columns)/len(batch_sizes)))
df_min_time_per_image_index = pd.DataFrame(df_mean_time_per_image.idxmin(axis=1)).set_index(0, append=True).index.values
df_model_lib = df_time_per_image[df_min_time_per_image_index] \
.stack(['model', 'lib']).reorder_levels(['model','lib','repetition']).sum(axis=1)
df_model_lib_mean = df_model_lib.groupby(level=['model', 'lib']).mean()
df_model_lib_std = df_model_lib.groupby(level=['model', 'lib']).std()
zero_positive_infinity = df_model_lib_mean > 1e5
df_model_lib_mean[zero_positive_infinity] = 0
# exclude_positive_infinity = df_model_lib_mean < 1e6
# df_model_lib_mean = df_model_lib_mean[exclude_positive_infinity]
# df_model_lib_std = df_model_lib_std[exclude_positive_infinity]
mean = df_model_lib_mean.unstack('lib')
std = df_model_lib_std.unstack('lib')
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f7d321290>
mean = df_model_lib_mean.unstack('lib').drop('cpu', axis=1)
std = df_model_lib_std.unstack('lib').drop('cpu', axis=1)
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f8626c690>
cuda_level_performance = ['nvidia-cuda', 'nvidia-cudnn', 'libdnn-cuda']
mean = df_model_lib_mean.reorder_levels(['lib', 'model'])[cuda_level_performance].unstack('lib')
std = df_model_lib_std.reorder_levels(['lib', 'model'])[cuda_level_performance].unstack('lib')
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f84d79090>
cublas_libs = ['cuda', 'nvidia-cuda', 'nvidia-fp16-cuda']
mean = df_model_lib_mean.reorder_levels(['lib', 'model'])[cublas_libs].unstack('lib')
std = df_model_lib_std.reorder_levels(['lib', 'model'])[cublas_libs].unstack('lib')
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f7887f1d0>
# With cuBLAS, BVLC's master is up to 11% slower than NVIDIA's mainline.
cuda_vs_nvidia_cuda = mean['cuda'] / mean['nvidia-cuda']
cuda_vs_nvidia_cuda
model bvlc-alexnet 1.052061 bvlc-googlenet 1.114225 deepscale-squeezenet-1.0 1.088136 deepscale-squeezenet-1.1 1.069487 dtype: float64
# With cuBLAS, NVIDIA's fp16 branch is up to 77% slower than NVIDIA's fp32 mainline. Some of this difference
# may be explained by the fp16 branch not being maintained, hence not including the latest improvements.
# NB: But see below a more drastic performance difference and info on GTX 1080 support of fp16.
nvidia_fp16_cuda_vs_nvidia_fp32_cuda = mean['nvidia-fp16-cuda'] / mean['nvidia-cuda']
nvidia_fp16_cuda_vs_nvidia_fp32_cuda
model bvlc-alexnet 1.773806 bvlc-googlenet 1.551819 deepscale-squeezenet-1.0 1.264935 deepscale-squeezenet-1.1 1.256191 dtype: float64
cudnn_libs = ['cudnn', 'nvidia-cudnn', 'nvidia-fp16-cudnn']
mean = df_model_lib_mean.reorder_levels(['lib', 'model'])[cudnn_libs].unstack('lib')
std = df_model_lib_std.reorder_levels(['lib', 'model'])[cudnn_libs].unstack('lib')
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f786a2350>
# With cuDNN, BVLC's master is between 8% faster and 5% slower than NVIDIA's mainline (i.e. they are roughly equivalent).
cuda_vs_nvidia_cuda = mean['cudnn'] / mean['nvidia-cudnn']
cuda_vs_nvidia_cuda
model bvlc-alexnet 0.925569 bvlc-googlenet 0.953534 deepscale-squeezenet-1.0 1.055904 deepscale-squeezenet-1.1 1.021777 dtype: float64
# With cuDNN, NVIDIA's fp16 branch is up to 27 times slower than NVIDIA's fp32 mainline.
nvidia_fp16_cudnn_vs_nvidia_fp32_cudnn = mean['nvidia-fp16-cudnn'] / mean['nvidia-cudnn']
nvidia_fp16_cudnn_vs_nvidia_fp32_cudnn
model bvlc-alexnet 25.615722 bvlc-googlenet 27.051362 deepscale-squeezenet-1.0 17.015116 deepscale-squeezenet-1.1 14.105727 dtype: float64
http://www.anandtech.com/show/10325/the-nvidia-geforce-gtx-1080-and-1070-founders-edition-review/5
Low precision operations are in turn seen by NVIDIA as one of the keys into further growing their increasingly important datacenter market, as deep learning and certain other tasks are themselves rapidly growing fields. Pascal isn’t just faster than Maxwell overall, but when it comes to FP16 operations on the FP16x2 core, Pascal is a lot faster, with theoretical throughput over similar Maxwell GPUs increasing by over three-fold thanks to the combination of overall speed improvements and double speed FP16 execution.
GeForce GTX 1080, on the other hand, is not faster at FP16. In fact it’s downright slow. For their consumer cards, NVIDIA has severely limited FP16 CUDA performance. GTX 1080’s FP16 instruction rate is 1/128th its FP32 instruction rate, or after you factor in vec2 packing, the resulting theoretical performance (in FLOPs) is 1/64th the FP32 rate, or about 138 GFLOPs.
mean = df_model_lib_mean.unstack('model')
std = df_model_lib_std.unstack('model')
plot(mean, std, rot=30)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f785b6910>
mean = df_model_lib_mean.unstack('model').drop('cpu', axis=0)
std = df_model_lib_std.unstack('model').drop('cpu', axis=0)
plot(mean, std, rot=30)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f783e9550>
alexnet_level_accuracy = ['bvlc-alexnet','deepscale-squeezenet-1.0','deepscale-squeezenet-1.1']
# On this platform with all the libraries, SqueezeNet 1.0 is always slower than AlexNet
# despite a 50x reduction in weights (5 MB vs. 250 MB).
mean = df_model_lib_mean[alexnet_level_accuracy].unstack('model')
std = df_model_lib_std[alexnet_level_accuracy].unstack('model')
plot(mean, std, rot=30)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f78226dd0>
# SqueezeNet 1.1 is 41% faster than AlexNet with OpenBLAS (on the CPU).
mean = df_model_lib_mean[alexnet_level_accuracy].unstack('model').ix[['cpu']]
std = df_model_lib_std[alexnet_level_accuracy].unstack('model').ix[['cpu']]
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f73f58f50>
mean['deepscale-squeezenet-1.1'] / mean['bvlc-alexnet']
lib cpu 0.588886 dtype: float64
# SqueezeNet 1.0 is slower than AlexNet. SqueezeNet 1.1 is 18% faster than AlexNet with libDNN-CUDA.
mean = df_model_lib_mean[alexnet_level_accuracy].unstack('model').ix[cuda_level_performance]
std = df_model_lib_std[alexnet_level_accuracy].unstack('model').ix[cuda_level_performance]
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f73f08a50>
mean['deepscale-squeezenet-1.0'] / mean['bvlc-alexnet']
lib nvidia-cuda 1.580941 nvidia-cudnn 2.077409 libdnn-cuda 1.432441 dtype: float64
mean['deepscale-squeezenet-1.1'] / mean['bvlc-alexnet']
lib nvidia-cuda 1.247205 nvidia-cudnn 1.283083 libdnn-cuda 0.819863 dtype: float64
libdnn_libs = [ 'libdnn-cuda', 'libdnn-clblas', 'libdnn-clblast', 'libdnn-viennacl' ]
# With the libDNN libs, SqueezeNet 1.1 is roughly equivalent to AlexNet.
mean = df_model_lib_mean[alexnet_level_accuracy].unstack('model').ix[libdnn_libs]
std = df_model_lib_std[alexnet_level_accuracy].unstack('model').ix[libdnn_libs]
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f73d2e2d0>
mean['deepscale-squeezenet-1.1'] / mean['bvlc-alexnet']
lib libdnn-cuda 0.819863 libdnn-clblas 1.169965 libdnn-clblast 0.999485 libdnn-viennacl 0.786956 dtype: float64
opencl_libs = [ 'clblas', 'clblast', 'viennacl' ]
# SqueezeNet 1.0 is slower than AlexNet with all the OpenCL BLAS libs.
# SqueezeNet 1.1 is 28% faster than AlexNet with ViennaCL and 6.5% faster with clBLAS,
# but over 2 times slower with CLBlast.
mean = df_model_lib_mean[alexnet_level_accuracy].unstack('model').ix[opencl_libs]
std = df_model_lib_std[alexnet_level_accuracy].unstack('model').ix[opencl_libs]
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f73ac4f10>
mean['deepscale-squeezenet-1.1'] / mean['bvlc-alexnet']
lib clblas 0.934168 clblast 2.217205 viennacl 0.726248 dtype: float64
df_per_layer_info = get_per_layer_info(df_all)
# pd.options.display.max_columns = len(df_per_layer_info.columns)
# pd.options.display.max_rows = len(df_per_layer_info.index)
# df_per_layer_info
# Plot for a list of batch sizes.
# NB: This suggests that the fully connected layers benefit the most from increasing the batch size.
plot_time_per_image_per_layer(df_per_layer_info, model='bvlc-alexnet', libs='nvidia-cuda',
batch_sizes=[2, 8, 16], direction=direction)
# Plot for a list of batch sizes. Only plot layers that consume at least 10% of the total execution time.
plot_time_per_image_per_layer(df_per_layer_info, model='bvlc-alexnet', libs='nvidia-cudnn',
batch_sizes=[8, 16], direction=direction, lower=0.10, rot=0)
# Plot for a list of libs.
# NB: cuDNN and cuBLAS perform about the same on the fully connected layers (which suggests that
# cuDNN falls back to cuBLAS for these).
# Unsurprisingly, cuDNN performs better than cuBLAS on the convolution layers.
# Surprisingly, cuBLAS performs a bit better than cuDNN on the relu layers.
plot_time_per_image_per_layer(df_per_layer_info, model='bvlc-alexnet', libs=['nvidia-cuda','nvidia-cudnn'],
batch_sizes=16, direction=direction)
# Plot for a list of libs.
# NB: This suggests that libDNN is faster than cuDNN on the expand1x1 layers, but slower on the squeeze1x1,
# expand3x3, conv/pool10 layers.
plot_time_per_image_per_layer(df_per_layer_info, model='deepscale-squeezenet-1.1', libs=['nvidia-cudnn', 'libdnn-cuda'],
batch_sizes=16, direction=direction, ymax=0.05)
# Plot for a list of libs. Only plot layers that consume between 5% and 10% of the total execution time.
# NB: libDNN is slower than cuDNN on the expand3x3 layers and conv10 layers, but a bit faster on the conv1 layer.
plot_time_per_image_per_layer(df_per_layer_info, model='deepscale-squeezenet-1.1', libs=['nvidia-cudnn', 'libdnn-cuda'],
batch_sizes=16, direction=direction, lower=0.05, upper=0.10, rot=10)
# Plot for a list of libs and a list of batch sizes. (This works but might not be terribly legible).
plot_time_per_image_per_layer(df_per_layer_info, model='bvlc-alexnet', libs=['nvidia-cudnn', 'nvidia-cuda'],
batch_sizes=[4,6], direction=direction)
Overall, using cuDNN typically results in the minimum execution time. For some layers, however, other libraries may outperform cuDNN (e.g. libDNN from the OpenCL branch of Caffe). As we show below, using the best performing library per layer results in up to 20% execution time reduction over using cuDNN alone. For other models and on other platforms such adaptation can potentially results even in higher savings.
NB: Currently, the savings are hypothetical. However, Caffe allows for manual adaptation, i.e. the user can specify the engine to use for each layer in the model file (*.prototxt
). We are working on generating the optimized model file automatically from the obtained ideal adaptive solution. Please contact us if you are interested.
We only include libs built from the master and OpenCL branches because per layer adaptation implies building from the same source. The OpenCL branch is kept in sync with the master, while the NVIDIA branches are not.
all_libs = df_per_layer_info.index.get_level_values('lib').drop_duplicates() \
.drop(['nvidia-cuda', 'nvidia-cudnn', 'nvidia-fp16-cuda', 'nvidia-fp16-cudnn'])
all_libs
Index([u'clblas', u'clblast', u'cpu', u'cuda', u'cudnn', u'libdnn-clblas', u'libdnn-clblast', u'libdnn-cuda', u'libdnn-viennacl', u'viennacl'], dtype='object', name=u'lib')
Each row specifies an ideal adaptive solution for a model. Each column specifies the execution time (in ms per image) that the ideal adaptive solution would cumulatively spend using a particular library.
df_ideal_all = get_ideal_adaptive_solution(df_per_layer_info, all_libs, direction)
df_ideal_all
lib | cpu | cuda | cudnn | libdnn-clblast | libdnn-cuda | libdnn-viennacl |
---|---|---|---|---|---|---|
model | ||||||
bvlc-alexnet | NaN | 0.142984 | 0.253353 | NaN | NaN | NaN |
deepscale-squeezenet-1.1 | 0.000509 | 0.112327 | 0.293295 | NaN | 0.062514 | NaN |
deepscale-squeezenet-1.0 | 0.000446 | 0.151060 | 0.616261 | 0.014336 | 0.074738 | NaN |
bvlc-googlenet | 0.000250 | 0.134322 | 0.941497 | NaN | NaN | 0.020264 |
plot_ideal_adaptive_solution(df_ideal_all, df_model_lib_mean)
# Up to 20% execution time reduction compared to the best non-adaptive solution (i.e. cuDNN).
df_best_lib = df_model_lib_mean.reorder_levels(['lib', 'model'])[cuda_level_performance].unstack('lib')
df_ideal_all.sum(axis=1) / df_best_lib.min(axis=1)
model bvlc-alexnet 0.866942 bvlc-googlenet 0.876359 deepscale-squeezenet-1.0 0.902201 deepscale-squeezenet-1.1 0.798941 dtype: float64
df_ideal_cuda = get_ideal_adaptive_solution(df_per_layer_info, ['cuda', 'cudnn', 'libdnn-cuda'], direction)
df_ideal_cuda
lib | cuda | cudnn | libdnn-cuda |
---|---|---|---|
model | |||
bvlc-alexnet | 0.142984 | 0.253353 | NaN |
deepscale-squeezenet-1.1 | 0.112673 | 0.293636 | 0.062514 |
deepscale-squeezenet-1.0 | 0.151420 | 0.616526 | 0.089374 |
bvlc-googlenet | 0.134546 | 0.950613 | 0.012608 |
plot_ideal_adaptive_solution(df_ideal_cuda, df_model_lib_mean)
# Hypothetical execution time reduction compared to the best non-adaptive solution (i.e. cuDNN).
df_best_lib = df_model_lib_mean.reorder_levels(['lib', 'model'])[cuda_level_performance].unstack('lib')
df_ideal_cuda.sum(axis=1) / df_best_lib.min(axis=1)
model bvlc-alexnet 0.866942 bvlc-googlenet 0.877506 deepscale-squeezenet-1.0 0.902704 deepscale-squeezenet-1.1 0.799243 dtype: float64
# Up to 0.1% worse performance when using the CUDA-level performance libs only.
df_ideal_cuda.sum(axis=1) / df_ideal_all.sum(axis=1)
model bvlc-alexnet 1.000000 deepscale-squeezenet-1.1 1.000378 deepscale-squeezenet-1.0 1.000558 bvlc-googlenet 1.001308 dtype: float64
df_ideal_cudnn_cublas = get_ideal_adaptive_solution(df_per_layer_info, ['cudnn', 'cuda'], direction)
df_ideal_cudnn_cublas
lib | cudnn | cuda |
---|---|---|
model | ||
bvlc-alexnet | 0.253353 | 0.142984 |
deepscale-squeezenet-1.1 | 0.427118 | 0.105697 |
deepscale-squeezenet-1.0 | 0.757035 | 0.144636 |
bvlc-googlenet | 0.999683 | 0.102204 |
plot_ideal_adaptive_solution(df_ideal_cudnn_cublas, df_model_lib_mean)
# Hypothetical execution time reduction compared to the best non-adaptive solution (i.e. cuDNN).
df_best_lib = df_model_lib_mean.reorder_levels(['lib', 'model'])[cuda_level_performance].unstack('lib')
df_ideal_cudnn_cublas.sum(axis=1) / df_best_lib.min(axis=1)
model bvlc-alexnet 0.866942 bvlc-googlenet 0.880799 deepscale-squeezenet-1.0 0.949404 deepscale-squeezenet-1.1 0.908336 dtype: float64
# Up to 14% worse performance when using cuDNN+cuBLAS only.
df_ideal_cudnn_cublas.sum(axis=1) / df_ideal_all.sum(axis=1)
model bvlc-alexnet 1.000000 deepscale-squeezenet-1.1 1.136925 deepscale-squeezenet-1.0 1.052320 bvlc-googlenet 1.005066 dtype: float64
df_ideal_cudnn_libdnn = get_ideal_adaptive_solution(df_per_layer_info, ['cudnn', 'libdnn-cuda'], direction)
df_ideal_cudnn_libdnn
lib | cudnn | libdnn-cuda |
---|---|---|
model | ||
bvlc-alexnet | 0.309263 | 0.089684 |
deepscale-squeezenet-1.1 | 0.382638 | 0.090092 |
deepscale-squeezenet-1.0 | 0.706316 | 0.155610 |
bvlc-googlenet | 1.040084 | 0.060236 |
plot_ideal_adaptive_solution(df_ideal_cudnn_libdnn, df_model_lib_mean)
# Hypothetical execution time reduction compared to the best non-adaptive solution (i.e. cuDNN).
df_best_lib = df_model_lib_mean.reorder_levels(['lib', 'model'])[cuda_level_performance].unstack('lib')
df_ideal_cudnn_libdnn.sum(axis=1) / df_best_lib.min(axis=1)
model bvlc-alexnet 0.872649 bvlc-googlenet 0.879546 deepscale-squeezenet-1.0 0.907556 deepscale-squeezenet-1.1 0.805904 dtype: float64
# Less than 1% worse performance when using cuDNN+libDNN only.
df_ideal_cudnn_libdnn.sum(axis=1) / df_ideal_all.sum(axis=1)
model bvlc-alexnet 1.006583 deepscale-squeezenet-1.1 1.008715 deepscale-squeezenet-1.0 1.005935 bvlc-googlenet 1.003636 dtype: float64
df_memory = df_all['memory (MB)']
# Batch size of 4; repetition 0 (should always be available).
df_memory = df_memory.unstack(['model','lib']).loc[4].loc[0].unstack('lib')
plot(mean=df_memory, std=pd.DataFrame(), ylabel='Memory size (MB)')
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f72036510>
The above, however, does not tell the full story. The memory consumption, as reported by Caffe, increases linearly with the batch size. In other words, the memory consumption per image is constant. (Note that extra memory may be required e.g. for GPU buffers in host memory.)
The execution time per image, however, decreases asymptotically. Since minimizing the execution time almost always should be balanced with minimizing the memory consumption, we should select the batch size that results in "good enough" performance.
We give several examples below. Note that the execution time per batch is omitted to make the execution time per image more pronounced.
# Is the batch size of 8 "good enough"?
plot_time_per_image_and_memory_consumption(df_all, 'bvlc-alexnet', 'nvidia-cudnn')
# Is the batch size of 2 "good enough"?
plot_time_per_image_and_memory_consumption(df_all, 'deepscale-squeezenet-1.1', 'cpu')
# SqueezeNet consumes about 4 times more memory than AlexNet.
df_memory.ix[['bvlc-alexnet', 'deepscale-squeezenet-1.1']].iloc[1] / \
df_memory.ix[['bvlc-alexnet', 'deepscale-squeezenet-1.1']].iloc[0]
lib clblas 3.882476 clblast 3.882476 cpu 3.882476 cuda 3.882476 cudnn 3.882476 libdnn-clblas 3.882476 libdnn-clblast 3.882476 libdnn-cuda 3.882476 libdnn-viennacl 3.882476 nvidia-cuda 3.882476 nvidia-cudnn 3.882476 nvidia-fp16-cuda 4.113806 nvidia-fp16-cudnn 4.113806 viennacl 3.882476 dtype: float64
mean = df_model_lib_mean[['bvlc-alexnet', 'deepscale-squeezenet-1.1']].unstack('lib')
std = df_model_lib_std[['bvlc-alexnet', 'deepscale-squeezenet-1.1']].unstack('lib')
plot(mean, std)
<matplotlib.axes._subplots.AxesSubplot at 0x7f4f73500b90>
df_model_lib_mean[['bvlc-alexnet', 'deepscale-squeezenet-1.1']].unstack('lib').iloc[1] / \
df_model_lib_mean[['bvlc-alexnet', 'deepscale-squeezenet-1.1']].unstack('lib').iloc[0]
lib clblas 0.934168 clblast 2.217205 cpu 0.588886 cuda 1.267865 cudnn 1.416452 libdnn-clblas 1.169965 libdnn-clblast 0.999485 libdnn-cuda 0.819863 libdnn-viennacl 0.786956 nvidia-cuda 1.247205 nvidia-cudnn 1.283083 nvidia-fp16-cuda 0.883257 nvidia-fp16-cudnn 0.706551 viennacl 0.726248 dtype: float64
# cuDNN-fp32 is up to 129x faster than the CPU.
plot_speedup_over_baseline(df_mean_time_per_image, baseline='cpu', libs_to_drop=[], fontsize=14)
# cuDNN-fp32 is up to 3.1x faster than cuBLAS-fp32.
plot_speedup_over_baseline(df_mean_time_per_image, baseline='nvidia-cuda', libs_to_drop=[], fontsize=14)
# AlexNet and SqueezeNet 1.1 have very similar performance with cuBLAS-fp32 and cuDNN-fp32, as well as very similar accuracy!
# At the same time, SqueezeNet requires about 4 times more activation memory and 50 times less memory for
# the weights than AlexNet.
plot_speedup_over_baseline(df_mean_time_per_image.ix[['bvlc-alexnet', 'deepscale-squeezenet-1.1']],
baseline='nvidia-cuda', libs_to_drop=[], fontsize=20)
A suite of open-source tools for collecting knowledge on optimising AI: