2015/gpututorial.html at master · iscs-sssihl/2015 · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
<!DOCTYPE html>
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<meta name="viewport" content="width=1200px, initial-scale=0.3, minimum-scale=0.3, maximum-scale=1.0, user-scalable=yes" />
<title>GPU tutorial </title>
<link href="/css/main.css" rel="stylesheet" type="text/css" />
<link href="/css/edetail.css" rel="stylesheet" type="text/css"/>
<script src="/js/jquery-1.9.1.min.js" type="text/javascript"></script>
<script src="/js/pretreatment.js" type="text/javascript"></script>
</head>

<body bgcolor="ffffcd">
<img align = "left" src="https://cloud.githubusercontent.com/assets/16099374/11537433/f58c0876-9942-11e5-96cb-0b90c44cdf49.png" height = "135" width = "135" alt="Swami"> <br>
<img align = "right" src="https://cloud.githubusercontent.com/assets/16099374/11537637/f413deb4-9943-11e5-8183-07b85a654c69.jpg" height = "120" width = "120" alt="SSSIHL">


	<div class="wrap" id="a">
		<div id="main">
		<div class="title"><h1 style="text-align:center;color:#333399;"> International Symposium on Computational Science(ISCS) 2015  <h2 style="text-align:center; color:#333399;">Department of Mathematics and Computer Science <br/>Sri Sathya Sai Institute of Higher Learning, Prashanthi Nilayam Campus</h2></h1> </div>
		<br>
		<br>

	<p align="center">
		<a href="index.html">Home</a> &nbsp | &nbsp
		<a href="Talks.html" target="_blank"> Talks </a> &nbsp | &nbsp
		<a href="gpututorial.html" target="_blank"> GPU Tutorial</a> &nbsp | &nbsp
		<a href="cgpddtutorial.html" target="_blank">CGPDD Tutorial</a> &nbsp | &nbsp
		<a href="graphtutorial.html" target="_blank">Graph Tutorial</a> &nbsp | &nbsp
		<a href="Travel.html" target="_blank">Travel Info</a> &nbsp | &nbsp
		<a href="http://www.sssihl.edu.in" target="_blank"> Official SSSIHL Website </a> &nbsp | &nbsp

</p>


		<div class="title"><h3 style="text-align:left;color:#333399;">Tutorial: Program Optimization on GPUs </h3> </div>


                            <dd><p style="TEXT-ALIGN: justify">Lectures with hand-on tutorial will be given by Prof. P Sadayappan, Ohio State University.</p><p style="TEXT-ALIGN: justify">
<p><strong>Objective of tutorial:</strong> The one day tutorial will focus on performance optimization &nbsp; techniques on GPUs.&nbsp;
</p>

<table border="2">
<th>Time</th>
<th>Sessions</th>
<th>Topic</th>

<tr>
<td>8:30 AM- 10:00 AM</td>
<td>Session I </td>
<td> <ul>
	<li>Quick review of parallel programming with OpenMP</li>
	<li>Introducing GPU architecture</li>
	<li>SIMT model </li>
	<li>Different types of memory on GPU Multidimensional thread space </li>
     </ul></td>
</tr>

<tr>
<td>10:00 AM - 10:30 AM</td>
<td>Tea Break </td>
<td> Tea Break</td>
</tr>

<tr>
<td>10:30 AM- 12:00 PM</td>
<td>Session II </td>
<td>
	<ul>
		<li>GPU performance aspects: </li>
		<li>Warp scheduling, Memory coalescing to efficient Global memory access</li>
		<li>Approaches to tolerate memory latency</li>
		<li>Emphasize the differences in loop optimizations between CPU and GPU </li>
		<li>(Forms of the loops in the CUDA reverse of OpenMP [stride 1 vs. stride N]), Filling the blanks variety (understand the variety) old exercises. (Different variants of reversing the arrays)</li>
		<li>Hands-on: simple examples – Hello World GPU example (hello world or some data parallel examples)</li>

	</ul>
</td>
</tr>

<tr>
<td>12:00 PM- 1:00 PM</td>
<td>Lunch Break</td>
<td> Lunch Break</td>
</tr>

<tr>
<td>1:00 PM- 2:30 PM</td>
<td>Session III</td>
<td> Optimizations for Shared memory. (Reuse in shared memory as much as possible). Are these optimizations still crucial considering the CPU like memory hierarchy in recent GPUs
<b>(1hour lecture + 1/2 hour exercise)</b>

 </td>
</tr>


<tr>
<td>2:30 PM- 3:00 PM</td>
<td>Tea Break</td>
<td> Tea Break</td>
</tr>


<tr>
<td>3:00 PM- 4:30 PM</td>
<td>Session IV </td>
<td> <ul> <li> Thread divergence. </li>
	  <li> 	simple exercise - stencils. </li>
	  <li>Do extra work in order to reduce the thread divergence.</li>
      </ul>

</td>
</tr>

</table>

<br> <br><br>
<strong>Lecture outline:</strong></p><p style="TEXT-ALIGN: justify">
</strong>

<strong> Instructor</strong> </br>
<img align = "left" src="https://cloud.githubusercontent.com/assets/16099374/11554165/8c423e20-99bb-11e5-8c9c-b67acf48c0ce.jpg" height = "125" width = "125" alt="Swami"> <br>
<strong><a href="http://web.cse.ohio-state.edu/~saday/" target="_blank">Prof. P Sadayappan </a> </strong>&nbsp;
<p style="TEXT-ALIGN: justify">Professor, Computer Science & Engineering  </p>
<p style="TEXT-ALIGN: justify">Ohio State University </p>

<strong> Email for communication:</strong> <p>iscsgpu@googlegroups.com </p>


<div id="footer">
	<div class="copyright">
		<p>Sri Sathya Sai Institute Of Higher Learning<br /></p>
		<address>Prasanthi Nilayam, Puttaparti 515134, India<br />
		</address>
	</div>

</div>

<!-- Page generated at 2015-11-17 09:37:27  -->
</div>


</body>
</html>