• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1page.title=Analyzing Display and Performance with Systrace
2parent.title=Debugging
3parent.link=index.html
4@jd:body
5
6<div id="qv-wrapper">
7  <div id="qv">
8    <h2>In this document</h2>
9    <ol>
10      <li><a href="#overview">Overview</a>
11      </li>
12      <li><a href="#generate">Generating Traces</a>
13        <ol>
14          <li><a href="#limit-trace">Limiting trace data</a></li>
15          <li><a href="#config-categories">Configuring trace data categories</a></li>
16          <li><a href="#running">Running a trace</a></li>
17        </ol>
18      </li>
19      <li><a href="#analysis">Analyzing Traces</a>
20        <ol>
21          <li><a href="#long-processes">Long running processes</a></li>
22          <li><a href="#display-interupts">Interruptions in display execution</a></li>
23        </ol>
24      </li>
25    </ol>
26    <h2>See also</h2>
27    <ol>
28      <li><a href="{@docRoot}tools/help/systrace.html">Systrace</a>
29      </li>
30    </ol>
31  </div>
32</div>
33
34<p>After building features, eliminating bugs and cleaning up your code, you should spend some
35  time looking at the performance of your application. The speed and smoothness with which your
36  application draws pixels and performs operations has an significant impact on your users'
37  experience.</p>
38
39<p>Android applications operate within a shared resource environment, and the performance of
40  your application can be impacted by how efficiently it interacts with those resources in
41  the larger system. Applications also operate in a multithreaded environment, competing with other
42  threaded processes for resources, which can cause performance problems that are hard to diagnose.
43</p>
44
45<p>The {@code systrace} tool allows you to collect and review code execution data for your
46  application and the Android system. You can use this data to diagnose execution problems and
47  improve the performance of your application.</p>
48
49
50<h2 id="overview">Overview</h2>
51
52<p>{@code systrace} helps you analyze how the execution of your application fits into the larger
53  Android environment, letting you see system and applications process execution on a common
54  timeline. The tool allows you to generate highly detailed, interactive reports from devices
55  running Android 4.1 and higher, such as the report in figure 1.</p>
56
57<img src="{@docRoot}images/systrace/report.png" alt="Systrace example report" id="figure1" />
58<p class="img-caption">
59  <strong>Figure 1.</strong> An example {@code systrace} report on 5 seconds of process execution
60  for a running application and related Android system processes.
61</p>
62
63
64<h2 id="generate">Generating Traces</h2>
65
66<p>In order to create a trace of your application, you must perform a few setup steps. First, you
67  must have a device running Android 4.1 or higher. Setup the device for
68  <a href="{@docRoot}tools/device.html#setting-up">debugging</a>, connect it to your development
69  system and install your application. Some types of trace information, specifically disk activity
70  and kernel work queues, require root access to the device, but most {@code systrace} log data
71  only requires that the device be enabled for developer debugging.</p>
72
73
74<h3 id="limit-trace">Limiting trace data</h3>
75
76<p>The {@code systrace} tool can generate a potentially huge amount of data from applications
77  and system sources. To limit the amount of data the tool collects and make the data more relevant
78  to your analysis, use the following options:</p>
79
80<ul>
81  <li>Limit the amount of time covered by the trace with the {@code -t, --time} option. The default
82    length of a trace is 5 seconds.</li>
83  <li>Limit the size of the data collected by the trace with the {@code -b, --buf-size} option.</li>
84  <li>Specify what types of processes are traced using the {@code --set-tags} option and the
85  {@code --disk}, {@code --cpu-freq}, {@code --cpu-idle}, {@code --cpu-load} options.</li>
86</ul>
87
88
89<h3 id="config-categories">Configuring trace data categories</h3>
90
91<p>To use {@code systrace} effectively, you must specify the types of processes you want to trace.
92  The tool can gather the following types of process information:</p>
93
94<ul>
95  <li>General system processes such as graphics, audio and input processes (selected using trace
96    <a href="{@docRoot}tools/help/systrace.html#tags">Tags</a>).</li>
97  <li>Low level system information such as CPU, kernel and disk activity (selected using
98    <a href="{@docRoot}tools/help/systrace.html#options">Options</a>).</li>
99</ul>
100
101<p>To set trace tags for {@code systrace} using the command-line:</p>
102
103<ol>
104  <li>Use the {@code --set-tags} option:
105<pre>
106$> python systrace.py --set-tags=gfx,view,wm
107</pre>
108  </li>
109  <li>Stop and restart the {@code adb} shell to enable tracing of these processes.
110<pre>
111$> adb shell stop
112$> adb shell start
113</pre></li>
114</ol>
115
116<p>To set trace tags for {@code systrace} using the device user interface:</p>
117
118<ol>
119  <li>On the device connected for tracing, navigate to: <strong>Settings &gt;
120      Developer options &gt; Monitoring &gt; Enable traces</strong>.</li>
121  <li>Select the categories of processes to be traced and click <strong>OK</strong>.</li>
122</ol>
123
124<p class="note">
125  <strong>Note:</strong> The {@code adb} shell does not have to be stopped and restarted when
126  selecting trace tags using this method.
127</p>
128
129
130<h3 id="running">Running a trace</h3>
131
132<p>After you have configured the category tags for your trace, you can start collecting
133  information for analysis.</p>
134
135<p>To run a trace using the current trace tag settings:</p>
136
137<ol>
138  <li>Make sure the device is connected through a USB cable and is
139  <a href="{@docRoot}tools/device.html#setting-up">enabled for debugging</a>.</li>
140  <li>Run the trace with the low-level system trace options and limits you want, for example:
141<pre>
142$> python systrace.py --cpu-freq --cpu-load --time=10 -o mytracefile.html
143</pre>
144  </li>
145  <li>On the device, execute any user actions you want be included in the trace.</li>
146</ol>
147
148
149<h2 id="analysis">Analyzing Traces</h2>
150
151<p>After you have generated a trace using {@code systrace}, it lists the location of the output
152  file and you can open the report using a web browser.
153  How you use the trace data depends on the performance issues you are investigating. However,
154  this section provides some general instructions on how to analyze a trace.</p>
155
156<p>The reports generated by {@code systrace} are interactive, allowing you to zoom into and out of
157  the process execution details. Use the <em>W</em> key to zoom in, the <em>S</em>
158  key to zoom out, the <em>A</em> key to pan left and the <em>D</em> key to pan
159  right. Select a task in timeline using your mouse to get more information about the task.
160  For more information about the using the keyboard navigation shortcuts and navigation, see the
161  <a href="{@docRoot}tools/help/systrace.html#viewing-options">Systrace</a> reference
162  documentation.</p>
163
164<h3 id="long-processes">Long running processes</h3>
165
166<p>A well-behaved application executes many small operations quickly and with a regular rhythm,
167  with individual operations completing within few milliseconds, depending on the device
168  and the processes being performed, as shown in figure 2:</p>
169
170<img src="{@docRoot}images/systrace/process-rhythm.png" alt="Systrace exerpt of app processing"
171id="figure2" />
172<p class="img-caption">
173  <strong>Figure 2.</strong> Excerpt from a trace of a smoothly running application with a regular
174  execution rhythm.
175</p>
176
177<p>The trace excerpt in figure 2 shows a well-behaved application with
178  a regular process rhythm (1). The lower section of figure 2 shows a magnified section of
179  the trace indicated by the dotted outline, which reveals some irregularity in the process
180  execution. In particular, one of the wider task bars, indicated by (2), is taking slightly
181  longer (14 milliseconds) than other, similar tasks on this thread, which are averaging between
182  9 and 12 milliseconds to complete. This particular task execution length is likely not noticeable
183  to a user, unless it impacts another process with specific timing, such as a screen update.</p>
184
185<p>Long running processes show up as thicker than usual execution bars in a trace. These thicker
186  bars can indicate a problem in your application performance. When they show up in your
187  trace, zoom in on the process using the
188  <a href="{@docRoot}tools/help/systrace.html#viewing-options">keyboard navigation</a> shortcuts to
189  identify the task causing the problem, and click on the task to get more information. You should
190  also look at other processes running at the same time, looking for a thread in one process that is
191  being blocked by another process.</p>
192
193
194<h3 id="display-interupts">Interruptions in display execution</h3>
195
196<p>The {@code systrace} tool is particularly useful in analyzing application display slowness,
197  or pauses in animations, because it shows you the execution of your application across multiple
198  system processes. With display execution, drawing screen frames with a regular rhythm is essential
199  for good performance. Having a regular rhythm for display ensures that animations and motion are
200  smooth on screen. If an application drops out of this rhythm, the display can become jerky or slow
201  from the users perspective.</p>
202
203<p>If you are analyzing an application for this type of problem, examine the
204  <strong>SurfaceFlinger</strong> process in the {@code systrace} report where your application is
205  also executing to look for places where it drops out of its regular rhythm.</p>
206
207<img src="{@docRoot}images/systrace/display-rhythm.png" alt="Systrace exerpt of display processing"
208id="figure3" />
209<p class="img-caption">
210  <strong>Figure 3.</strong> Excerpt from a trace of an application showing interruptions in
211  display processing.
212</p>
213
214<p>The trace excerpt in figure 3 shows an section of a trace that indicates an interruption in the
215  device display. The section of the <strong>SurfaceFlinger</strong> process in top excerpt,
216  indicated by (1), shows that display frames are being missed. These
217  dropped frames are potentially causing the display to stutter or halt. Zooming into this problem
218  area in the lower trace, shows that a memory operation (image buffer dequeuing and allocation) in
219  the <strong>surfaceflinger</strong> secondary thread is taking a long time (2). This delay
220  causes the application to miss the display update window, indicated by the dotted
221  line. As the developer of this application, you should investigate other threads in your
222  application that may also be trying to allocate memory at the same time or otherwise blocking
223  memory allocation with another request or task.</p>
224
225<p>Regular, rhythmic execution of the <strong>SurfaceFlinger</strong> process is essential to smooth
226  display of screen content, particularly for animations and motion. Interruptions in the regular
227  execution pattern of this thread is not always an indication of a display problem with your
228  application. Further testing is required to determine if this is actually a performance problem
229  from a user perspective. Being able to identify display execution patterns like the example above
230  can help you detect display problems and build a smooth-running, high-performance application.
231</p>
232
233<p class="note">
234  <strong>Note:</strong> When using {@code systrace} to analyze display problems, make sure
235  you activate the tracing tags for <strong>Graphics</strong> and <strong>Views</strong>.
236</p>
237
238<p>For more information on the command line options and keyboard controls for {@code systrace},
239see the <a href="{@docRoot}tools/help/systrace.html">Systrace</a> reference page.</p>