In [ ]:
%matplotlib inline
import numpy as np
import scipy as sp
import pandas as pd
import scipy.stats as sps
import statsmodels.api as sm
import matplotlib as mpl
import matplotlib.pyplot as plt
import matplotlib.mlab as mlab
import matplotlib.ticker as ticker
import matplotlib.gridspec as gridspec
import radical.utils as ru
import radical.pilot as rp
import radical.analytics as ra
from IPython.display import display
from scipy.stats import linregress
pd.set_option('expand_frame_repr', False)
pd.set_option('display.max_rows', 50)
In [ ]:
# Global configurations
# ---------------------
# Matplotlib style
plt.style.use('seaborn-ticks')
# Use LaTeX and its body font for the diagrams' text.
mpl.rcParams['text.usetex'] = True
mpl.rcParams['font.family'] = 'serif'
mpl.rcParams['font.serif'] = ['Nimbus Roman Becker No9L']
# Font sizes
SIZE = 20
plt.rc('font', size=SIZE) # controls default text sizes
plt.rc('axes', titlesize=SIZE) # fontsize of the axes title
plt.rc('axes', labelsize=SIZE) # fontsize of the x any y labels
plt.rc('xtick', labelsize=SIZE) # fontsize of the tick labels
plt.rc('ytick', labelsize=SIZE) # fontsize of the tick labels
plt.rc('legend', fontsize=SIZE-2) # legend fontsize
plt.rc('figure', titlesize=SIZE) # # size of the figure title
# Use thinner lines for axes to avoid distractions.
mpl.rcParams['axes.linewidth'] = 0.75
mpl.rcParams['xtick.major.width'] = 0.75
mpl.rcParams['xtick.minor.width'] = 0.75
mpl.rcParams['ytick.major.width'] = 0.75
mpl.rcParams['ytick.minor.width'] = 0.75
mpl.rcParams['lines.linewidth'] = 2
# Do not use a box for the legend to avoid distractions.
mpl.rcParams['legend.frameon'] = False
# Restore part of matplotlib 1.5 behavior
mpl.rcParams['patch.force_edgecolor'] = True
mpl.rcParams['errorbar.capsize'] = 3
# Helpers
# -------
# Use coordinated colors. These are the "Tableau 20" colors as
# RGB. Each pair is strong/light. For a theory of color
tableau20 = [(31 , 119, 180), (174, 199, 232), # blue [ 0,1 ]
(255, 127, 14 ), (255, 187, 120), # orange [ 2,3 ]
(44 , 160, 44 ), (152, 223, 138), # green [ 4,5 ]
(214, 39 , 40 ), (255, 152, 150), # red [ 6,7 ]
(148, 103, 189), (197, 176, 213), # purple [ 8,9 ]
(140, 86 , 75 ), (196, 156, 148), # brown [10,11]
(227, 119, 194), (247, 182, 210), # pink [12,13]
(188, 189, 34 ), (219, 219, 141), # yellow [14,15]
(23 , 190, 207), (158, 218, 229), # cyan [16,17]
(65 , 68 , 81 ), (96 , 99 , 106), # gray [18,19]
(127, 127, 127), (143, 135, 130), # gray [20,21]
(165, 172, 175), (199, 199, 199), # gray [22,23]
(207, 207, 207)] # gray [24]
# Scale the RGB values to the [0, 1] range, which is the format
# matplotlib accepts.
for i in range(len(tableau20)):
r, g, b = tableau20[i]
tableau20[i] = (r / 255., g / 255., b / 255.)
# Return a single plot without right and top axes
def fig_setup():
fig = plt.figure(figsize=(13,7))
ax = fig.add_subplot(111)
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)
ax.get_xaxis().tick_bottom()
ax.get_yaxis().tick_left()
return fig, ax
In [ ]:
sxsede = pd.read_csv('data/sessions.csv', index_col=0)
pxsede = pd.read_csv('data/pilots.csv', index_col=0)
uxsede = pd.read_csv('data/units.csv', index_col=0)
In [ ]:
sxsede.loc[(sxsede['nunit'] >= 4 ) & (sxsede['nunit'] < 8 ), 'nunit_requested'] = 4
sxsede.loc[(sxsede['nunit'] >= 8 ) & (sxsede['nunit'] < 16 ), 'nunit_requested'] = 8
sxsede.loc[(sxsede['nunit'] >= 16 ) & (sxsede['nunit'] < 32 ), 'nunit_requested'] = 16
sxsede.loc[(sxsede['nunit'] >= 32 ) & (sxsede['nunit'] < 64 ), 'nunit_requested'] = 32
sxsede.loc[(sxsede['nunit'] >= 64 ) & (sxsede['nunit'] < 128 ), 'nunit_requested'] = 64
sxsede.loc[(sxsede['nunit'] >= 128 ) & (sxsede['nunit'] < 256 ), 'nunit_requested'] = 128
sxsede.loc[(sxsede['nunit'] >= 256 ) & (sxsede['nunit'] < 512 ), 'nunit_requested'] = 256
sxsede.loc[(sxsede['nunit'] >= 512 ) & (sxsede['nunit'] < 1024), 'nunit_requested'] = 512
sxsede.loc[(sxsede['nunit'] >= 1024) & (sxsede['nunit'] < 2048), 'nunit_requested'] = 1024
sxsede.loc[(sxsede['nunit'] >= 2048) & (sxsede['nunit'] < 4096), 'nunit_requested'] = 2048
sxsede.loc[(sxsede['nunit'] >= 4096) & (sxsede['nunit'] < 8192), 'nunit_requested'] = 4096
sxsede['nunit_rescheduled'] = sxsede['nunit'] - sxsede['nunit_requested']
In [ ]:
stampede = ['rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017261.0002', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017261.0003', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0001', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0007']
comet = ['rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017261.0000', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0004', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017267.0013', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017269.0001']
supermic_matteo = ['rp.session.one.mturilli.017239.0009', 'rp.session.one.mturilli.017238.0011', 'rp.session.one.mturilli.017239.0010', 'rp.session.one.mturilli.017238.0012', 'rp.session.one.mturilli.017239.0011', 'rp.session.one.mturilli.017238.0013', 'rp.session.one.mturilli.017239.0014', 'rp.session.one.mturilli.017238.0014', 'rp.session.one.mturilli.017239.0017', 'rp.session.one.mturilli.017238.0017', 'rp.session.one.mturilli.017239.0018', 'rp.session.one.mturilli.017238.0018', 'rp.session.one.mturilli.017239.0019', 'rp.session.one.mturilli.017239.0003', 'rp.session.one.mturilli.017239.0020', 'rp.session.one.mturilli.017239.0005', 'rp.session.one.mturilli.017239.0021', 'rp.session.one.mturilli.017239.0006', 'rp.session.one.mturilli.017239.0022', 'rp.session.one.mturilli.017239.0007']
supermic_ming = ['rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017261.0004', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0000', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0002', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0003', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0004', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0005', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0006', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0007', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0008', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0009', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017262.0010', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017263.0001', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0001', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0002', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0003', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0004', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0005', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0006', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0007', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0008', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0009', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0010', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0011', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0012', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0013', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0014', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0015', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0016', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0017', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0018', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0019', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0020', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0022', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0023', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0024', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0025', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0026', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0027', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0028', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0029', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0030', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0031', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0032', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0033', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0034', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0035', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0036', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0037', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0038', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0039', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0040', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0041', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0042', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0043', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0044', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017265.0045', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0005', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0008', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0009', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0010', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0011', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0012', 'rp.session.Ubuntu-1604-xenial-64-minimal.mingtha.017266.0013']
supermic = supermic_matteo+supermic_ming
ustampede = uxsede.loc[uxsede.sid.isin(stampede)]
ucomet = uxsede.loc[uxsede.sid.isin(comet)]
usupermic_matteo = uxsede.loc[uxsede.sid.isin(supermic_matteo)]
usupermic_ming = uxsede.loc[uxsede.sid.isin(supermic_ming)]
usupermic = uxsede.loc[uxsede.sid.isin(supermic)]
In [ ]:
print 'Number of units run on Stampede: %s' % ustampede.shape[0]
print 'Number of units run on Comet: %s' % ucomet.shape[0]
print 'Number of units run on SuperMIC (Matteo\'s experiments): %s' % usupermic_matteo.shape[0]
print 'Number of units run on SuperMIC (Ming\'s experiments): %s' % usupermic_ming.shape[0]
print 'Number of units run on SuperMIC: %s' % usupermic.shape[0]
In [ ]:
# Load wrangled data saved in .csv files.
sosg = pd.read_csv('../../OSG/analysis/data/sessions.csv', index_col=0)
posg = pd.read_csv('../../OSG/analysis/data/pilots.csv', index_col=0)
uosg = pd.read_csv('../../OSG/analysis/data/units.csv', index_col=0)
# Drop experiments with at least 8 tasks but less than 8 posg.
sosg = sosg[sosg.experiment != 'exp1']
posg = posg[posg.experiment != 'exp1']
uosg = uosg[uosg.experiment != 'exp1']
sosg = sosg[sosg.experiment != 'exp2']
posg = posg[posg.experiment != 'exp2']
uosg = uosg[uosg.experiment != 'exp2']
sosg = sosg[sosg.experiment != 'exp3']
posg = posg[posg.experiment != 'exp3']
uosg = uosg[uosg.experiment != 'exp3']
# Check for spurious data due to bugs and drop them
sosg = sosg.drop(['rp.session.radical.mturilli.017128.0000',
'rp.session.radical.mturilli.017139.0003',
# 'rp.session.radical.mturilli.017131.0005', # outlier Tq
'rp.session.radical.mturilli.017230.0003', # outlier TTC (Tr)
'rp.session.radical.mturilli.017212.0011', # inconsistency in concurrency
'rp.session.radical.mturilli.017223.0000', # inconsistency in concurrency
'rp.session.radical.mturilli.017228.0001']) # inconsistency in concurrency
# 'rp.session.radical.mturilli.017229.0003']) # inconsistency in concurrency
posg = posg[posg.sid != 'rp.session.radical.mturilli.017139.0003']
posg = posg[posg.sid != 'rp.session.radical.mturilli.017128.0000']
# posg = posg[posg.sid != 'rp.session.radical.mturilli.017131.0005']
posg = posg[posg.sid != 'rp.session.radical.mturilli.017230.0003']
posg = posg[posg.sid != 'rp.session.radical.mturilli.017212.0011']
posg = posg[posg.sid != 'rp.session.radical.mturilli.017223.0000']
posg = posg[posg.sid != 'rp.session.radical.mturilli.017228.0001']
posg = posg[posg.sid != 'rp.session.radical.mturilli.017229.0003']
uosg = uosg[uosg.sid != 'rp.session.radical.mturilli.017139.0003']
uosg = uosg[uosg.sid != 'rp.session.radical.mturilli.017128.0000']
# uosg = uosg[uosg.sid != 'rp.session.radical.mturilli.017131.0005']
uosg = uosg[uosg.sid != 'rp.session.radical.mturilli.017230.0003']
uosg = uosg[uosg.sid != 'rp.session.radical.mturilli.017212.0011']
uosg = uosg[uosg.sid != 'rp.session.radical.mturilli.017223.0000']
uosg = uosg[uosg.sid != 'rp.session.radical.mturilli.017228.0001']
uosg = uosg[uosg.sid != 'rp.session.radical.mturilli.017229.0003']
# Derive number of uosg requested and number of uosg
# rescheduled due to pilot halted/migrated by condor
sosg.loc[(sosg['nunit'] >= 8 ) & (sosg['nunit'] < 16 ), 'nunit_requested'] = 8
sosg.loc[(sosg['nunit'] >= 16 ) & (sosg['nunit'] < 32 ), 'nunit_requested'] = 16
sosg.loc[(sosg['nunit'] >= 32 ) & (sosg['nunit'] < 64 ), 'nunit_requested'] = 32
sosg.loc[(sosg['nunit'] >= 64 ) & (sosg['nunit'] < 128 ), 'nunit_requested'] = 64
sosg.loc[(sosg['nunit'] >= 128 ) & (sosg['nunit'] < 256 ), 'nunit_requested'] = 128
sosg.loc[(sosg['nunit'] >= 256 ) & (sosg['nunit'] < 512 ), 'nunit_requested'] = 256
sosg.loc[(sosg['nunit'] >= 512 ) & (sosg['nunit'] < 1024), 'nunit_requested'] = 512
sosg.loc[(sosg['nunit'] >= 1024) & (sosg['nunit'] < 2048), 'nunit_requested'] = 1024
sosg.loc[(sosg['nunit'] >= 2048) & (sosg['nunit'] < 4096), 'nunit_requested'] = 2048
sosg['nunit_rescheduled'] = sosg['nunit'] - sosg['nunit_requested']
# Drop runs with nunit_requested < npilot
# runs = sosg[sosg.nunit_requested < sosg.npilot]['sid'].tolist()
# sosg = sosg[sosg.nunit_requested >= sosg.npilot]
# for run in runs:
# posg = posg[posg.sid != run]
# uosg = uosg[uosg.sid != run]
In [ ]:
print 'Number of successful runs: %s' % sosg.shape[0]
print 'Number of pilots: %s' % posg.shape[0]
print 'Number of units: %s' % uosg.shape[0]
In [ ]:
fig, ax = fig_setup()
title='SuperMIC XSEDE\nFrequency of Synapse Emulation Execution Time ($T_x$) for the %s executions.\n.' % len(usupermic_matteo['U_AGENT_EXECUTING'].dropna().tolist())
# KDE execution time
# display(sps.normaltest(usupermic_matteo['U_AGENT_EXECUTING'].dropna().tolist()))
# usupermic_matteo[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax, color='black', title=title)
# Normed Histogram
usupermic_matteo[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color='gray',
edgecolor="k",
bins=30,
title=title,
alpha=0.20)
# normed=True)
plt.axvline(usupermic_matteo['U_AGENT_EXECUTING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
# ax.set_xlim(0,1400)
#plt.ylabel('Percentage')
plt.xlabel('Time (s)')
ax.legend(labels=['Median (%is)' % usupermic_matteo['U_AGENT_EXECUTING'].dropna().median(),
'Task Execution Time (Tx)'])
plt.savefig('figures/supermic_matteo_tx_percentage.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title='SuperMIC XSEDE\nFrequency of Synapse Emulation Execution Time ($T_x$) for the %s executions.\n.' % len(usupermic_ming['U_AGENT_EXECUTING'].dropna().tolist())
# KDE execution time
#display(sps.normaltest(usupermic_ming['U_AGENT_EXECUTING'].dropna().tolist()))
#usupermic_ming[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax, color='black', title=title)
# Normed Histogram
usupermic_ming[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color='gray',
edgecolor="k",
title=title,
bins=30,
alpha = 0.20)
# normed=True)
plt.axvline(usupermic_ming['U_AGENT_EXECUTING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
# ax.set_xlim(0,1400)
plt.xlabel('Time (s)')
ax.legend(labels=['Median', 'Task Execution Time (Tx)'])
plt.savefig('figures/supermic_ming_tx_frequency.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title='SuperMIC XSEDE\nFrequency of Synapse Emulation Execution Time ($T_x$) for the %s executions.\n.' % len(usupermic['U_AGENT_EXECUTING'].dropna().tolist())
# KDE execution time
#display(sps.normaltest(usupermic['U_AGENT_EXECUTING'].dropna().tolist()))
#usupermic[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax, color='black', title=title)
# Normed Histogram
print usupermic.shape[0]
usupermic[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color='gray',
edgecolor="k",
bins=30,
title=title,
alpha = 0.20) #,
# normed=True)
plt.axvline(usupermic['U_AGENT_EXECUTING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
#ax.set_xlim(0,1400)
plt.xlabel('Time (s)')
ax.legend(labels=['Median', 'Task Execution Time (Tx)'])
# plt.savefig('figures/supermic_tx_frequency.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title='Stampede XSEDE\nDensity of Synapse Emulation Execution Time'
# KDE execution time
display(sps.normaltest(ustampede['U_AGENT_EXECUTING'].dropna().tolist()))
# ustampede[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.kde(ax=ax, color='black', title=title)
# Normed Histogram
ustampede[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color='gray',
edgecolor="k",
bins=20,
alpha = 0.20)#,
# normed=True)
plt.axvline(ustampede['U_AGENT_EXECUTING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
# plt.axvline(ustampede['U_AGENT_EXECUTING'].dropna().mode(), color='b', linestyle='dashed', linewidth=1)
# plt.axvline(ustampede['U_AGENT_EXECUTING'].dropna().median(), color='g', linestyle='dashed', linewidth=1)
smin = min(ustampede['U_AGENT_EXECUTING'].dropna().tolist())
smax = max(ustampede['U_AGENT_EXECUTING'].dropna().tolist())
vettore = np.linspace(501, 897, num=20)
vdue = [0]*len(vettore)
sample = ustampede['U_AGENT_EXECUTING'].dropna().tolist()
sumsample = len(sample)
sumtot = 0
# for s in ustampede['U_AGENT_EXECUTING'].dropna().tolist():
# v = 0
# while vettore[v]<s:
# v = v+1
# vdue[v] += (1.0/sumsample)*((smax-smin)/sumsample)
# sumtot += 1.0/sumsample
# print smin
# print sumtot
# print vdue
#ax.set_xlim(0,1400)
plt.xlabel('Time (s)')
ax.legend(labels=['Median', 'Task Execution Time (Tx)'])
# plt.savefig('figures/stampede_tx_frequency.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title='Stampede XSEDE\nDensity of Synapse Emulation Execution Time'
# KDE execution time
# display(sps.normaltest(ucomet['U_AGENT_EXECUTING'].dropna().tolist()))
# ucomet[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax, color='black', title=title)
# Normed Histogram
ucomet[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color='gray',
edgecolor="k",
bins=40,
alpha = 0.20)
# normed=True)
plt.axvline(ucomet['U_AGENT_EXECUTING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
# ax.set_xlim(400,900)
plt.xlabel('Time (s)')
ax.legend(labels=['Median', 'Task Execution Time (Tx)'])
plt.savefig('figures/stampede_tx_frequency.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title='XSEDE OSG Virtual Cluster\nDensity of Units Execution Time (Tx)\nNote: Tail truncated at 2000s.'
uosg[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(ax=ax,
color=tableau20[7],
bins=100,
alpha = 0.25)#,
# normed=True)
# units[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax,
# color='black',
# title=title)
plt.axvline(uosg['U_AGENT_EXECUTING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
print uosg['U_AGENT_EXECUTING'].dropna().mode()
ax.set_xlim(0, 3000)
ax.set_xlabel('$T_x$ (s)')
ax.legend(labels=['Median (%is)' % uosg['U_AGENT_EXECUTING'].dropna().median(), 'Units Execution Time ($T_x$)'])
plt.savefig('figures/osg_dstrb_tx_all.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title='Stampede, Comet and SuperMIC XSEDE\nFrequency of Synapse Emulation Execution Time ($T_x$)'
# KDE execution time
# ustampede[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax, color=tableau20[0], title=title)
# ucomet[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax, color=tableau20[2], title=title)
# usupermic[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax, color=tableau20[4], title=title)
# Normed Histogram
uosg[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color=tableau20[7],
bins=100,
alpha = 0.25)#,
# normed=True)
ustampede[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color=tableau20[1],
edgecolor="k",
bins=20,
alpha = 0.25)
# normed=True)
ucomet[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color=tableau20[3],
edgecolor="k",
bins=40,
alpha = 0.25)
# normed=True)
usupermic[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color=tableau20[5],
edgecolor="k",
bins=30,
alpha = 0.25)
# normed=True)
plt.axvline(uosg['U_AGENT_EXECUTING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
plt.axvline(ustampede['U_AGENT_EXECUTING'].dropna().median(), color=tableau20[0], linestyle='dashed', linewidth=1)
plt.axvline(ucomet['U_AGENT_EXECUTING'].dropna().median(), color=tableau20[2], linestyle='dashed', linewidth=1)
plt.axvline(usupermic['U_AGENT_EXECUTING'].dropna().median(), color=tableau20[4], linestyle='dashed', linewidth=1)
ax.set_xlim(0, 2000)
plt.xlabel('Time (s)')
ax.legend(labels=['Median $T_x$ OSG VC (%is)' % uosg['U_AGENT_EXECUTING'].dropna().median(),
'Median $T_x$ Stampede (%is)' % ustampede['U_AGENT_EXECUTING'].dropna().median(),
'Median $T_x$ Comet (%is)' % ucomet['U_AGENT_EXECUTING'].dropna().median(),
'Median $T_x$ SuperMIC (%is)' % usupermic['U_AGENT_EXECUTING'].dropna().median(),
'$T_x$ OSG VC','$T_x$ Stampede', '$T_x$ Comet', '$T_x$ SuperMIC', ])
plt.savefig('figures/xsede_tx_frequency.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title_head = 'XSEDE Stampede, Comet, SuperMIC, OSG Virtual Cluster'
title_desc = 'Median $T_x$ for each resource with $\sigma$'
title = '%s\n%s' % (title_head, title_desc)
fig.suptitle(title)
index = ['OSG VC','Stampede','Comet','SuperMIC']
medians = [uosg['U_AGENT_EXECUTING'].dropna().median(),
ustampede['U_AGENT_EXECUTING'].dropna().median(),
ucomet['U_AGENT_EXECUTING'].dropna().median(),
usupermic['U_AGENT_EXECUTING'].dropna().median()]
stds = [uosg['U_AGENT_EXECUTING'].dropna().std(),
ustampede['U_AGENT_EXECUTING'].dropna().std(),
ucomet['U_AGENT_EXECUTING'].dropna().std(),
usupermic['U_AGENT_EXECUTING'].dropna().std()]
txs = pd.DataFrame({'medians': pd.Series(medians, index),
'stds' : pd.Series(stds , index)})
colors = [tableau20[6],tableau20[1],tableau20[3],tableau20[5]]
# Print sample sizes
print 'OSG sample size: %i' % len(uosg['U_AGENT_EXECUTING'].dropna())
print 'Stampede sample size: %i' % len(ustampede['U_AGENT_EXECUTING'].dropna())
print 'Comet sample size: %i' % len(ucomet['U_AGENT_EXECUTING'].dropna())
print 'SuperMIC sample size: %i' % len(usupermic['U_AGENT_EXECUTING'].dropna())
txs.plot(y='medians', yerr='stds',
color=colors,`
ax=ax,
kind='bar',
legend=False,
rot=0)
ax.set_ylabel('Time (s)')
ax.set_xlabel('Median $T_x$')
plt.savefig('figures/xsede_tx_frequency.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title='XSEDE OSG Virtual Cluster\nDensity of Units Execution Time (Tx)\nNote: Tail truncated at 2000s.'
posg[['P_LRMS_RUNNING']].dropna().reset_index(drop=True).plot.hist(ax=ax,
color='gray',#tableau20[7],
bins=80,
alpha = 0.25)#,
# normed=True)
# units[['U_AGENT_EXECUTING']].dropna().reset_index(drop=True).plot.density(ax=ax,
# color='black',
# title=title)
plt.axvline(posg['P_LRMS_RUNNING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
ax.set_xlim(0, 86400)
ax.set_xlabel('$T_r$ (s)')
ax.legend(labels=['Median (%is)' % posg['P_LRMS_RUNNING'].dropna().median(), 'Pilot running time ($T_r$)'])
#plt.savefig('figures/osg_dstrb_tx_all.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title= None#'XSEDE OSG Virtual Cluster\nDensity of Units Execution Time (Tx)\n.'
# Select pilots with:
# - units
# - FAILED final state
# - rescheduled units
# - without duplicated failed time value
# - with failed valued < 10% TTC (heuristic)
#
# 1. termination issues makes the FAILED state unreliable: it
# is used also for pilots that were killed by RP.
# 2. Pilot killed by RP are killed in bulk. Therefore, elimitate
# pilots with duplicated failed state.
# 3. Pilots available until the end of the session will have
# a failed value near the TTC of the session. Eliminate all
# the pilots with failed value within 1% (heuristic).
pfailed = posg[(posg.nunit > 0) &
(posg.FAILED > 0) &
(posg.sid.isin(sosg[sosg.nunit_rescheduled > 0].sid.tolist()))
].drop_duplicates(subset='FAILED')
for sid in pfailed.sid.tolist():
pfailed.loc[pfailed.sid == sid, 'TTC'] = sosg[sosg.sid == sid].TTC.values
pfailed = pfailed[pfailed.FAILED < pfailed.TTC-(pfailed.TTC/100.0*10)]
pfailed[['P_LRMS_RUNNING']].dropna().reset_index(drop=True).plot.hist(
ax=ax,
color='gray',#tableau20[7],
bins=30,
alpha = 0.25)#,
# normed=True)
# pfailed[['P_LRMS_RUNNING']].dropna().reset_index(drop=True).plot.density(ax=ax,
# color='black',
# title=title)
plt.axvline(pfailed['P_LRMS_RUNNING'].dropna().median(), color='r', linestyle='dashed', linewidth=1)
ax.set_xlim(0)
ax.set_xlabel('$T_r$ (s)')
ax.legend(labels=['Median (%is)' % pfailed['P_LRMS_RUNNING'].dropna().median(), 'Pilot running time ($T_r$)'])
# plt.savefig('figures/osg_dstrb_tx_all.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig, ax = fig_setup()
title= 'XSEDE OSG Virtual Cluster\nMedian duration ($T_r$) and $\sigma$ of preempted pilots for bags of 128, 256, 512, 1024, 2048 tasks\n.'
# Select pilots with:
# - units
# - FAILED final state
# - rescheduled units
# - without duplicated failed time value
# - with failed valued < 10% TTC (heuristic)
#
# 1. termination issues makes the FAILED state unreliable: it
# is used also for pilots that were killed by RP.
# 2. Pilot killed by RP are killed in bulk. Therefore, elimitate
# pilots with duplicated failed state.
# 3. Pilots available until the end of the session will have
# a failed value near the TTC of the session. Eliminate all
# the pilots with failed value within 1% (heuristic).
pfailed = posg[(posg.nunit > 0) &
(posg.FAILED > 0) &
(posg.sid.isin(sosg[sosg.nunit_rescheduled > 0].sid.tolist()))
].drop_duplicates(subset='FAILED')
for sid in pfailed.sid.tolist():
pfailed.loc[pfailed.sid == sid, 'TTC'] = sosg[sosg.sid == sid].TTC.values
pfailed = pfailed[pfailed.FAILED < pfailed.TTC-(pfailed.TTC/100.0*10)]
# Partition the number of pilots that failed before the end of the session
# based on the number of units to compute. Use only sessions where all the
# requested units were computed (either successfully or unsuccessfully).
# nunit_requested nunit_done
for sid in pfailed.sid.tolist():
pfailed.loc[pfailed.sid == sid, 'nunit_requested'] = sosg[sosg.sid == sid].nunit_requested.values
pfailed.loc[pfailed.sid == sid, 'nunit_done'] = sosg[sosg.sid == sid].nunit_done.values
pfailed = pfailed[pfailed.nunit_requested == pfailed.nunit_done]
medians = []
stds = []
index = pfailed.nunit_requested.unique().tolist()
for nunit in sorted(index):
print '%s: %s' % (nunit, pfailed[pfailed.nunit_requested == nunit].shape[0])
medians.append(pfailed[pfailed.nunit_requested == nunit]['P_LRMS_RUNNING'].median())
stds.append(pfailed[pfailed.nunit_requested == nunit]['P_LRMS_RUNNING'].std())
trs = pd.DataFrame({'medians': pd.Series(medians, sorted(index)),
'stds' : pd.Series(stds, sorted(index))})
colors = [tableau20[9],tableau20[11],tableau20[13],
tableau20[15],tableau20[17],tableau20[19]]
trs.plot(y='medians', yerr='stds',
color=colors,
ax=ax,
kind='bar',
legend=False,
rot=0,
title=title)
ax.set_ylabel('Time (s)')
ax.set_xlabel('Median $T_r$')
# plt.savefig('figures/xsede_tx_frequency.pdf', dpi=600, bbox_inches='tight')
In [ ]:
fig = plt.figure(figsize=(7,13))
ax = fig.add_subplot(111)
ax.spines["top"].set_visible(False)
ax.spines["right"].set_visible(False)
ax.get_xaxis().tick_bottom()
ax.get_yaxis().tick_left()
title_head = 'XSEDE Stampede, Comet, SuperMIC, OSG Virtual Cluster'
title_desc = 'Boxplot of $T_x$ for each resource'
title = '%s\n%s' % (title_head, title_desc)
fig.suptitle(title)
index = ['OSG VC','Stampede','Comet','SuperMIC']
medians = [uosg['U_AGENT_EXECUTING'].dropna(),
ustampede['U_AGENT_EXECUTING'].dropna(),
ucomet['U_AGENT_EXECUTING'].dropna(),
usupermic['U_AGENT_EXECUTING'].dropna()]
stds = [uosg['U_AGENT_EXECUTING'].dropna().std(),
ustampede['U_AGENT_EXECUTING'].dropna().std(),
ucomet['U_AGENT_EXECUTING'].dropna().std(),
usupermic['U_AGENT_EXECUTING'].dropna().std()]
txs = pd.concat([uosg['U_AGENT_EXECUTING'].dropna(),
ustampede['U_AGENT_EXECUTING'].dropna(),
ucomet['U_AGENT_EXECUTING'].dropna(),
usupermic['U_AGENT_EXECUTING'].dropna()],
axis=1,
keys=['OSG VC',
'Stampede',
'Comet',
'SuperMIC'])
colors = [tableau20[6],tableau20[1],tableau20[3],tableau20[5]]
# Print sample sizes
print 'OSG sample size: %i' % len(uosg['U_AGENT_EXECUTING'].dropna())
print 'Stampede sample size: %i' % len(ustampede['U_AGENT_EXECUTING'].dropna())
print 'Comet sample size: %i' % len(ucomet['U_AGENT_EXECUTING'].dropna())
print 'SuperMIC sample size: %i' % len(usupermic['U_AGENT_EXECUTING'].dropna())
bp = txs.boxplot(column=['OSG VC','Stampede','Comet','SuperMIC'], ax=ax, return_type='dict')
# Set color boxes
bp['boxes'][0].set_color(colors[0])
bp['boxes'][1].set_color(colors[1])
bp['boxes'][2].set_color(colors[2])
bp['boxes'][3].set_color(colors[3])
# Set color medians
bp['medians'][0].set_color(colors[0])
bp['medians'][1].set_color(colors[1])
bp['medians'][2].set_color(colors[2])
bp['medians'][3].set_color(colors[3])
# Set color whiskers
bp['whiskers'][0].set_color(colors[0])
bp['whiskers'][1].set_color(colors[0])
bp['whiskers'][2].set_color(colors[1])
bp['whiskers'][3].set_color(colors[1])
bp['whiskers'][4].set_color(colors[2])
bp['whiskers'][5].set_color(colors[2])
bp['whiskers'][6].set_color(colors[3])
bp['whiskers'][7].set_color(colors[3])
ax.set_ylim(0, 2000)
ax.set_ylabel('Time (s)')
ax.set_xlabel('$T_x$')
#plt.savefig('figures/xsede_tx_frequency.pdf', dpi=600, bbox_inches='tight')
In [ ]: