forked from codeplaysoftware/syclacademy
-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
370 lines (363 loc) · 11.7 KB
/
index.html
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
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<link rel="stylesheet" href="../common-revealjs/css/reveal.css">
<link rel="stylesheet" href="../common-revealjs/css/theme/white.css">
<link rel="stylesheet" href="../common-revealjs/css/custom.css">
<script>
// This is needed when printing the slides to pdf
var link = document.createElement( 'link' );
link.rel = 'stylesheet';
link.type = 'text/css';
link.href = window.location.search.match( /print-pdf/gi ) ? '../common-revealjs/css/print/pdf.css' : '../common-revealjs/css/print/paper.css';
document.getElementsByTagName( 'head' )[0].appendChild( link );
</script>
<script>
// This is used to display the static images on each slide,
// See global-images in this html file and custom.css
(function() {
if(window.addEventListener) {
window.addEventListener('load', () => {
let slides = document.getElementsByClassName("slide-background");
if (slides.length === 0) {
slides = document.getElementsByClassName("pdf-page")
}
// Insert global images on each slide
for(let i = 0, max = slides.length; i < max; i++) {
let cln = document.getElementById("global-images").cloneNode(true);
cln.removeAttribute("id");
slides[i].appendChild(cln);
}
// Remove top level global images
let elem = document.getElementById("global-images");
elem.parentElement.removeChild(elem);
}, false);
}
})();
</script>
</head>
<body>
<div class="reveal">
<div class="slides">
<div id="global-images" class="global-images">
<img src="../common-revealjs/images/sycl_academy.png" />
<img src="../common-revealjs/images/sycl_logo.png" />
<img src="../common-revealjs/images/trademarks.png" />
<img src="../common-revealjs/images/codeplay.png" />
</div>
<!--Slide 1-->
<section class="hbox">
<div class="hbox" data-markdown>
## Coalesced Global Memory
</div>
</section>
<!--Slide 2-->
<section class="hbox" data-markdown>
## Learning Objectives
* Learn about coalesced global memory access
* Learn about the performance impact
* Learn about row-major vs column-major
* Learn about SoA vs AoS
</section>
<!--Slide 3-->
<section>
<div class="hbox" data-markdown>
#### Coalesced global memory
</div>
<div class="container" data-markdown>
* Reading from and writing to global memory is generally very expensive.
* It often involves copying data across an off-chip bus.
* This means you generally want to avoid unnecessary accesses.
* Memory access operations is done in chunks.
* This means accessing data this is physically close together in memory is more efficient.
</div>
</section>
<!--Slide 4-->
<section>
<div class="hbox" data-markdown>
#### Coalesced global memory
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 5-->
<section>
<div class="hbox" data-markdown>
#### Coalesced global memory
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 6-->
<section>
<div class="hbox" data-markdown>
#### Coalesced global memory
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 7-->
<section>
<div class="hbox" data-markdown>
#### Coalesced global memory
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 8-->
<section>
<div class="hbox" data-markdown>
#### Coalesced global memory
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 9-->
<section>
<div class="hbox" data-markdown>
#### Coalesced global memory
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 10-->
<section>
<div class="hbox" data-markdown>
#### Row-major vs Column-major
</div>
<div class="container" data-markdown>
* Coalescing global memory access is particularly important when working in multiple dimensions.
* This is because when doing so you have to convert from a position in 2d space to a linear memory space.
* There are two ways to do this; generally referred to as row-major and column-major.
</div>
</section>
<!--Slide 11-->
<section>
<div class="hbox" data-markdown>
#### Row-major vs Column-major
</div>
</section>
<section>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 12-->
<section>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 13-->
<section>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 16-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
* Another area this is a factor is when composing data structures.
* It's often instinctive to have struct representing a collection of data and then have an array of this - often referred to as Array of Structs (AoS).
* But for data parallel architectures such as a GPU it's more efficient to have sequential elements of the same type stored contiguously in memory - often referred to as Struct of Arrays (SoA).
</div>
</section>
<!--Slide 17-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 18-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 19-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 20-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 21-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 22-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 23-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 24-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 25-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 26-->
<section>
<div class="hbox" data-markdown>
#### AoS vs SoA
</div>
<div class="container" data-markdown>
</div>
</section>
<!--Slide 27-->
<section>
<div class="hbox" data-markdown>
#### Coalesced image convolution performance
</div>
<div class="container"data-markdown>
</div>
</section>
<!--Slide 28-->
<section>
<div class="hbox" data-markdown>
#### Vec types
</div>
<div class="container">
<code class="code-100pc"><pre>
auto f4 = sycl::float4{1.0f, 2.0f, 3.0f, 4.0f}; // {1.0f, 2.0f, 3.0f, 4.0f}
</code></pre>
</div>
<div class="container">
<code class="code-100pc"><pre>
auto f2 = sycl::float2{2.0f, 3.0f}; // {2.0f, 3.0f}
auto f4 = sycl::float4{1.0f, f2, 4.0f}; // {1.0f, 2.0f, 3.0f, 4.0f}
</code></pre>
</div>
<div class="container">
<code class="code-100pc"><pre>
auto f4 = sycl::float4{0.0f}; // {0.0f, 0.0f, 0.0f, 0.0f}
</code></pre>
</div>
<div class="container" data-markdown>
* A `vec` can be constructed with any combination of scalar and vector values which add up to the correct number of elements.
* A `vec` can also be constructed from a single scalar in which case it will initialize every element to that value.
</div>
</section>
<!--Slide 11-->
<section>
<div class="hbox" data-markdown>
#### Vec operators
</div>
<div class="container">
<code class="code-100pc"><pre>
auto f4a = sycl::float4{1.0f, 2.0f, 3.0f, 4.0f}; // {1.0f, 2.0f, 3.0f, 4.0f}
auto f4b = sycl::float4{2.0f}; // {2.0f, 2.0f, 2.0f, 2.0f}
auto f4r = f4a * f4b; // {2.0f, 4.0f, 6.0f, 8.0f}
</code></pre>
</div>
<div class="container" data-markdown>
* The `vec` class provides a number of operators such as `+`, `-`, `*`, `/` and many more, which perform the operation elemeent-wise.
</div>
</section>
<section>
<div class="hbox" data-markdown>
#### Vec sizes
</div>
<div class="container">
<code class="code-100pc"><pre>
sycl::int2
sycl::int3 (N.B sizeof(int3) == sizeof(int4))
sycl::int4
sycl::int8
sycl::int16</code></pre>
</div>
<div class="container" data-markdown>
* Vectors can be made from all char, integer or floating point types.
* Using vector types:
* Can make code more readable
* Can give better memory access patterns.
</div>
</section>
<section>
<div class="hbox" data-markdown>
## Questions
</div>
</section>
<!--Slide 29-->
<section>
<div class="hbox" data-markdown>
#### Exercise
</div>
<div class="container" data-markdown>
Code_Exercises/Exercise_16_Coalesced_Global_Memory/source
</div>
<div class="container" data-markdown>
Try inverting the dimensions when calculating the linear address in memory and measure the performance.
</div>
</section>
</div>
</div>
<script src="../common-revealjs/js/reveal.js"></script>
<script src="../common-revealjs/plugin/markdown/marked.js"></script>
<script src="../common-revealjs/plugin/markdown/markdown.js"></script>
<script src="../common-revealjs/plugin/notes/notes.js"></script>
<script>
Reveal.initialize({mouseWheel: true, defaultNotes: true});
Reveal.configure({ slideNumber: true });
</script>
</body>
</html>