root/branches/livecd/RWTHMindstormsNXT/doc/programming/OptimizePerformance.html @ 465

Revision 465, 4.3 KB (checked in by behrens, 5 years ago)

NXTmotor: remove "silent toggling" and move adaptive default settings in constructor
SendToNXT: add warning of ignoring Speedregulation in the case of MultipleMotorPorts?, disabled brakeAtTachoLimit=0 warning, add missing %
Line 
1
2<!DOCTYPE html
3  PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
4<html>
5   <head>
6      <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7   
8      <!--
9This HTML is auto-generated from an M-file.
10To make changes, update the M-file and republish this document.
11      -->
12      <title>Optimizing performance</title>
13      <meta name="generator" content="MATLAB 7.7">
14      <meta name="date" content="2008-11-05">
15      <meta name="m-file" content="OptimizePerformance">
16<link type="text/css" rel="stylesheet" href="../style.css">
17  </head>
18   <body>
19<p class="header">RWTH - Mindstorms NXT Toolbox</p>
20      <div class="content">
21         <h1>Optimizing performance</h1>
22         <!--introduction-->
23         <p>When working with RWTH - Mindstorms NXT Toolbox and USB connections, high packet rates of more than 100 sensor or motor readings
24            per second are common. In order to achieve such short latency for MATLAB - NXT communication, high CPU load is generated,
25            being the speed limit on slow machines.
26         </p>
27         <!--/introduction-->
28         <p>To improve performance, a function called</p><pre class="codeinput">OptimizeToolboxPerformance
29</pre><p>is provided. It will replace the frequently called function <tt>typecast</tt> with a faster binary version (<tt>typecastc</tt>). The binary MEX-files are already included in your MATLAB installation directory. But since they reside in a private directory,
30            they cannot be accessed from the outside. That is why <tt>OptimizeToolboxPerformance</tt> will try to copy all binary files to the toolboxes own private directory. So, before calling the function, make sure you
31            got write access where you installed RWTH - Mindstorms NXT Toolbox.
32         </p>
33         <p>The function <tt>OptimizeToolboxPerformance</tt> is fail-safe and verifies each step. The user is asked before action is taken. Just run the command to improve your toolbox
34            performance by a factor of up to 3 (depending on your machine of course). If it succeeds, the lowlevel functions <tt>wordbytes2dec</tt> and <tt>dec2wordbytes</tt> will have been replaced by faster versions, leading to less CPU load needed (and hence faster toolbox speed).
35         </p>
36         <p>Please note that the function only copies files that are already present in you MATLAB installation. So the optimization will
37            most likely only work for the specific MATLAB version, and only for the operating system you ran the command from.
38         </p>
39         <p class="footer"><br>
40            Published with MATLAB&reg; 7.7<br></p>
41      </div>
42      <!--
43##### SOURCE BEGIN #####
44%% Optimizing performance
45% When working with RWTH - Mindstorms NXT Toolbox and USB connections, high
46% packet rates of more than 100 sensor or motor readings per second are
47% common. In order to achieve such short latency for MATLAB - NXT
48% communication, high CPU load is generated, being the speed limit on slow
49% machines.
50
51%%
52% To improve performance, a function called
53
54OptimizeToolboxPerformance
55
56%%
57% is provided. It will replace the frequently called function
58% |typecast| with a faster binary version (|typecastc|). The
59% binary MEX-files are already included in your MATLAB installation
60% directory. But since they reside in a private directory, they cannot be
61% accessed from the outside. That is why |OptimizeToolboxPerformance| will
62% try to copy all binary files to the toolboxes own private directory. So,
63% before calling the function, make sure you got write access where you
64% installed RWTH - Mindstorms NXT Toolbox.
65
66%%
67% The function |OptimizeToolboxPerformance| is fail-safe and verifies each step.
68% The user is asked before action is taken. Just run the command to improve
69% your toolbox performance by a factor of up to 3 (depending on your
70% machine of course). If it succeeds, the lowlevel functions
71% |wordbytes2dec| and |dec2wordbytes| will have been replaced by faster
72% versions, leading to less CPU load needed (and hence faster toolbox speed).
73%
74
75%%
76% Please note that the function only copies files that are already present
77% in you MATLAB installation. So the optimization will most likely only
78% work for the specific MATLAB version, and only for the operating system
79% you ran the command from.
80
81##### SOURCE END #####
82-->
83   </body>
84</html>
Note: See TracBrowser for help on using the browser.