Skip to content

Commit a5b48e5

Browse files
zoybaizoybai
authored
Merge pull request #48 from zoybai/jupyter-notebook
Jupyter notebook
2 parents c017418 + e3124d0 commit a5b48e5

File tree

3 files changed

+261
-0
lines changed

3 files changed

+261
-0
lines changed

‎benchmarks/lockhammer/README.rst‎

Lines changed: 29 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -58,3 +58,32 @@ the format lh_[testname]. Each lockhammer binary accepts the following options:
5858
[-p parallel] Parallelizable section in loop iterations, default 0
5959
[-s] Run in safe mode with normal priority threads instead of RT_FIFO priority, default no
6060

61+
62+
Plotting
63+
========
64+
65+
The default plotting script utilizes jupyter-notebook, matplotlib, seaborn
66+
and pandas under python3 environment. For Ubuntu on x86_64 machine, the
67+
following packages have to be installed:
68+
apt install build-essential python3 python3-pip jupyter-notebook
69+
70+
For aarch64 machine, additional packages are also needed:
71+
apt install pkg-config libfreetype6-dev python3-scipy
72+
73+
Then pip3 can install all plotting related libraries with the following cmd:
74+
pip3 install matplotlib seaborn pandas numpy
75+
76+
Note, seaborn has to be installed without scipy as dependency on aarch64:
77+
pip3 install seaborn --no-dependencies
78+
79+
The jupyter-notebook can be started with:
80+
jupyter-notebook --ip 0.0.0.0 --port=8888
81+
82+
Now any browser should be able to access the jupyter notebook called:
83+
lockhammer-jupyter-notebook.ipynb
84+
85+
Start a browser, with IP address set to the jupyter server IP and port 8888:
86+
e.g. http://example.test.com:8888
87+
88+
Click the notebook named lockhammer-jupyter-notebook.ipynb, run each cell one
89+
by one and jupyter should be able to generate the png graph locally.

‎benchmarks/lockhammer/scripts/lockhammer-all.csv.xz‎

2.6 MB
Binary file not shown.

‎benchmarks/lockhammer/scripts/lockhammer-jupyter-notebook.ipynb‎

Lines changed: 232 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,232 @@
1+
{
2+
"cells": [
3+
{
4+
"cell_type": "markdown",
5+
"metadata": {},
6+
"source": [
7+
"# License"
8+
]
9+
},
10+
{
11+
"cell_type": "markdown",
12+
"metadata": {},
13+
"source": [
14+
"Copyright (c) 2018, ARM Limited. All rights reserved.\n",
15+
"\n",
16+
"SPDX-License-Identifier: BSD-3-Clause\n",
17+
"\n",
18+
"Redistribution and use in source and binary forms, with or without\n",
19+
"modification, are permitted provided that the following conditions are met:\n",
20+
"\n",
21+
"Redistributions of source code must retain the above copyright notice, this\n",
22+
"list of conditions and the following disclaimer.\n",
23+
"\n",
24+
"Redistributions in binary form must reproduce the above copyright notice, this\n",
25+
"list of conditions and the following disclaimer in the documentation and/or\n",
26+
"other materials provided with the distribution.\n",
27+
"\n",
28+
"Neither the name of ARM Limited nor the names of its contributors may be used\n",
29+
"to endorse or promote products derived from this software without specific\n",
30+
"prior written permission.\n",
31+
"\n",
32+
"THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS \"AS IS\"\n",
33+
"AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE\n",
34+
"IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE\n",
35+
"DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE\n",
36+
"FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL\n",
37+
"DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR\n",
38+
"SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER\n",
39+
"CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR\n",
40+
"TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE\n",
41+
"OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE."
42+
]
43+
},
44+
{
45+
"cell_type": "markdown",
46+
"metadata": {},
47+
"source": [
48+
"# Prerequisite Libraries"
49+
]
50+
},
51+
{
52+
"cell_type": "markdown",
53+
"metadata": {},
54+
"source": [
55+
"- ## python3\n",
56+
"`apt install python3 python3-pip`\n",
57+
"- ## jupyter-notebook\n",
58+
"`apt install jupyter-notebook`\n",
59+
"- ## matplotlib seaborn pandas numpy\n",
60+
"`pip3 install matplotlib seaborn pandas numpy`"
61+
]
62+
},
63+
{
64+
"cell_type": "markdown",
65+
"metadata": {},
66+
"source": [
67+
"# Matplotlib and Seaborn Settings"
68+
]
69+
},
70+
{
71+
"cell_type": "code",
72+
"execution_count": null,
73+
"metadata": {
74+
"collapsed": false
75+
},
76+
"outputs": [],
77+
"source": [
78+
"%matplotlib inline\n",
79+
"import matplotlib\n",
80+
"import matplotlib.pyplot as plt\n",
81+
"import seaborn as sns\n",
82+
"import pandas as pd\n",
83+
"import numpy as np\n",
84+
"import warnings\n",
85+
"warnings.filterwarnings('ignore')\n",
86+
"sns.set()\n",
87+
"\n",
88+
"# default 12 colors and markers\n",
89+
"default_palette = [\n",
90+
" '#765f97', #Purple\n",
91+
" '#1b9e77', #Dark Green\n",
92+
" '#8c5c20', #Brown\n",
93+
" '#0038bd', #Blue\n",
94+
" '#cf364a', #Red\n",
95+
" '#343434', #Jet Black\n",
96+
" '#878681', #Titanium Gray\n",
97+
" '#f561dd', #Magenta\n",
98+
" '#a6cee3', #Calico Blue\n",
99+
" '#dea0dd', #Plum\n",
100+
" '#7fc97f', #Grass Green\n",
101+
" '#fdc086', #Pale Yellow\n",
102+
" ]\n",
103+
"\n",
104+
"default_markers=['^', '*', 'd', 'x', 'D', 'o', 'v', 's', 'p', '>', '<', '.']\n",
105+
"default_marker_size = 100\n",
106+
"\n",
107+
"# seaborn settings\n",
108+
"sns.set(context=\"notebook\", style=\"darkgrid\", font_scale=2, rc={\"lines.linewidth\": 3, \"xtick.major.size\": 4, \"ytick.major.size\": 4})\n",
109+
"sns.set_palette(default_palette)\n",
110+
"sns.palplot(sns.color_palette())"
111+
]
112+
},
113+
{
114+
"cell_type": "markdown",
115+
"metadata": {},
116+
"source": [
117+
"# Lockhammer Common Settings"
118+
]
119+
},
120+
{
121+
"cell_type": "code",
122+
"execution_count": null,
123+
"metadata": {
124+
"collapsed": true
125+
},
126+
"outputs": [],
127+
"source": [
128+
"# common variables\n",
129+
"lock_workloads = [\"cas_event_mutex\", \"cas_lockref\", \"cas_rw_lock\", \"incdec_refcount\", \"osq_lock\", \"queued_spinlock\", \"ticket_spinlock\", \"jvm_objectmonitor\", \"swap_mutex\", \"tbb_spin_rw_mutex\", \"event_mutex\", \"empty\"]\n",
130+
"lock_hosts = [\"x86-1\", \"x86-2\", \"x86-3\", \"x86-4\"]\n",
131+
"lock_parameters = [[\"-c\", \"200ns\", \"-p\", \"1000ns\"]]\n",
132+
"exectx_count_max = 88\n",
133+
"exectx_count_gap = 4\n",
134+
"lh_csv_result_header = [\"num_threads\", \"avg_exectx\", \"scheduled_time_per_access\", \"real_time_per_access\", \"access_rate\", \"avg_lock_depth\",\n",
135+
" \"date\", \"fqdn\", \"exec\", \"t\", \"tv\", \"a\", \"av\", \"c\", \"cv\", \"p\", \"pv\", \"o\", \"ov\"]\n",
136+
"\n",
137+
"param_name = \"contended_system_latency (ns)\"\n",
138+
"lock_yaxis_name = param_name\n",
139+
"\n",
140+
"# lockhammer-all.csv.xz is a xz-compressed aggregated file of different machines' raw csv result\n",
141+
"raw_csv_filename = \"lockhammer-all.csv.xz\"\n",
142+
"raw_df = pd.read_csv(raw_csv_filename, sep=', ', header=None, names=lh_csv_result_header, engine='python')\n",
143+
"\n",
144+
"# common functions\n",
145+
"def plot_lines_only(dataf):\n",
146+
" # each test repeat 9 times, but we only plot the median latency (access_rate)\n",
147+
" median_list = []\n",
148+
" for hst, grp0 in dataf.groupby(\"host\"):\n",
149+
" for nth, grp1 in grp0.groupby(\"num_threads\"):\n",
150+
" median_list.append({\"host\": hst, \"num_threads\": nth, param_name: grp1.median()[param_name]})\n",
151+
" median_df = pd.DataFrame(median_list)\n",
152+
" \n",
153+
" from matplotlib.colors import ListedColormap\n",
154+
" cmap = ListedColormap(sns.color_palette(default_palette).as_hex())\n",
155+
" for i, (hst, grp2) in enumerate(median_df.groupby(\"host\")):\n",
156+
" plt.plot(\"num_threads\", param_name, data=grp2, color=cmap(i))"
157+
]
158+
},
159+
{
160+
"cell_type": "markdown",
161+
"metadata": {},
162+
"source": [
163+
"# Lockhammer all workloads, raw contended system latency, 2018.11.06."
164+
]
165+
},
166+
{
167+
"cell_type": "code",
168+
"execution_count": null,
169+
"metadata": {
170+
"collapsed": false,
171+
"scrolled": false
172+
},
173+
"outputs": [],
174+
"source": [
175+
"# use lmplot (not catplot) to plot raw system latencies\n",
176+
"for param in lock_parameters:\n",
177+
" for workload in lock_workloads:\n",
178+
" tidy_df = pd.DataFrame()\n",
179+
" for sut in sorted(lock_hosts):\n",
180+
" host_df = raw_df.loc[(raw_df['fqdn'].str.startswith(sut) & raw_df['exec'].str.endswith(workload))]\n",
181+
" test_df = host_df.loc[(host_df['cv'] == param[1]) & (host_df['pv'] == param[3])]\n",
182+
" copy_df = test_df.copy()\n",
183+
" copy_df['host'] = sut\n",
184+
" all_df = pd.melt(copy_df, id_vars=['host', 'num_threads'], value_vars=['access_rate'], value_name=param_name)\n",
185+
" tidy_df = pd.concat([tidy_df, all_df])\n",
186+
" \n",
187+
" # because lmplot doesn't plot lines, we have to use plot_lines_only to plot them\n",
188+
" sns.lmplot(x=\"num_threads\", y=param_name, hue=\"host\", data=tidy_df, x_estimator=np.median, x_ci=50,\n",
189+
" height=10, aspect=2, fit_reg=False, markers=default_markers[:len(lock_hosts)], scatter_kws={\"s\": default_marker_size})\n",
190+
" \n",
191+
" # plot lines which connect lmplot dots\n",
192+
" plot_lines_only(tidy_df)\n",
193+
" \n",
194+
" # change title / axis and save the figure\n",
195+
" plt.title(\"lockhammer workload: {}, critical_time: {}, parallel_time: {}\".format(workload, param[1], param[3]))\n",
196+
" plt.xlim(0, exectx_count_max)\n",
197+
" plt.xticks(np.arange(0, exectx_count_max+1, exectx_count_gap))\n",
198+
" plt.savefig(\"github_lockhammer_all_common_20181106_{}_{}_{}.png\".format(workload, param[1], param[3]))"
199+
]
200+
},
201+
{
202+
"cell_type": "code",
203+
"execution_count": null,
204+
"metadata": {
205+
"collapsed": true
206+
},
207+
"outputs": [],
208+
"source": []
209+
}
210+
],
211+
"metadata": {
212+
"kernelspec": {
213+
"display_name": "Python 3",
214+
"language": "python",
215+
"name": "python3"
216+
},
217+
"language_info": {
218+
"codemirror_mode": {
219+
"name": "ipython",
220+
"version": 3
221+
},
222+
"file_extension": ".py",
223+
"mimetype": "text/x-python",
224+
"name": "python",
225+
"nbconvert_exporter": "python",
226+
"pygments_lexer": "ipython3",
227+
"version": "3.6.3"
228+
}
229+
},
230+
"nbformat": 4,
231+
"nbformat_minor": 2
232+
}

0 commit comments

Comments
 (0)