-
-
Notifications
You must be signed in to change notification settings - Fork 1.4k
Expand file tree
/
Copy pathParallelTransforms.html
More file actions
275 lines (273 loc) · 19.5 KB
/
Copy pathParallelTransforms.html
File metadata and controls
275 lines (273 loc) · 19.5 KB
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
<!-- HTML header for doxygen 1.13.1-->
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "https://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" lang="en-US">
<head>
<meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/>
<meta http-equiv="X-UA-Compatible" content="IE=11"/>
<meta name="generator" content="Doxygen 1.13.1"/>
<meta name="viewport" content="width=device-width, initial-scale=1"/>
<title>Taskflow: A General-purpose Task-parallel Programming System: Parallel Transforms</title>
<link href="tabs.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="jquery.js"></script>
<script type="text/javascript" src="dynsections.js"></script>
<script type="text/javascript" src="clipboard.js"></script>
<link href="navtree.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="navtreedata.js"></script>
<script type="text/javascript" src="navtree.js"></script>
<script type="text/javascript" src="resize.js"></script>
<script type="text/javascript" src="cookie.js"></script>
<link href="search/search.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="search/searchdata.js"></script>
<script type="text/javascript" src="search/search.js"></script>
<link href="doxygen.css" rel="stylesheet" type="text/css" />
<link href="custom.css" rel="stylesheet" type="text/css"/>
</head>
<body>
<div id="top"><!-- do not remove this div, it is closed by doxygen! -->
<div id="titlearea">
<table cellspacing="0" cellpadding="0">
<tbody>
<tr id="projectrow">
<td id="projectlogo"><img alt="Logo" src="taskflow_logo.png"/></td>
<td id="projectalign">
<div id="projectname"><a href="https://github.com/taskflow/taskflow" style="color:inherit; text-decoration:none;">Taskflow: A General-purpose Task-parallel Programming System</a>
</div>
</td>
</tr>
</tbody>
</table>
</div>
<!-- end header part -->
<!-- Generated by Doxygen 1.13.1 -->
<script type="text/javascript">
/* @license magnet:?xt=urn:btih:d3d9a9a6595521f9666a5e94cc830dab83b65699&dn=expat.txt MIT */
var searchBox = new SearchBox("searchBox", "search/",'.html');
/* @license-end */
</script>
<script type="text/javascript">
/* @license magnet:?xt=urn:btih:d3d9a9a6595521f9666a5e94cc830dab83b65699&dn=expat.txt MIT */
$(function() { codefold.init(0); });
/* @license-end */
</script>
<script type="text/javascript" src="menudata.js"></script>
<script type="text/javascript" src="menu.js"></script>
<script type="text/javascript">
/* @license magnet:?xt=urn:btih:d3d9a9a6595521f9666a5e94cc830dab83b65699&dn=expat.txt MIT */
$(function() {
initMenu('',true,false,'search.php','Search',true);
$(function() { init_search(); });
});
/* @license-end */
</script>
<div id="main-nav"></div>
</div><!-- top -->
<div id="side-nav" class="ui-resizable side-nav-resizable">
<div id="nav-tree">
<div id="nav-tree-contents">
<div id="nav-sync" class="sync"></div>
</div>
</div>
<div id="splitbar" style="-moz-user-select:none;"
class="ui-resizable-handle">
</div>
</div>
<script type="text/javascript">
/* @license magnet:?xt=urn:btih:d3d9a9a6595521f9666a5e94cc830dab83b65699&dn=expat.txt MIT */
$(function(){initNavTree('ParallelTransforms.html',''); initResizable(true); });
/* @license-end */
</script>
<div id="doc-content">
<!-- window showing the filter options -->
<div id="MSearchSelectWindow"
onmouseover="return searchBox.OnSearchSelectShow()"
onmouseout="return searchBox.OnSearchSelectHide()"
onkeydown="return searchBox.OnSearchSelectKey(event)">
</div>
<!-- iframe showing the search results (closed by default) -->
<div id="MSearchResultsWindow">
<div id="MSearchResults">
<div class="SRPage">
<div id="SRIndex">
<div id="SRResults"></div>
<div class="SRStatus" id="Loading">Loading...</div>
<div class="SRStatus" id="Searching">Searching...</div>
<div class="SRStatus" id="NoMatches">No Matches</div>
</div>
</div>
</div>
</div>
<div><div class="header">
<div class="headertitle"><div class="title">Parallel Transforms</div></div>
</div><!--header-->
<div class="contents">
<div class="toc"><h3>Table of Contents</h3>
<ul>
<li class="level1">
<a href="#ParallelTransformsInclude">Include the Header</a>
</li>
<li class="level1">
<a href="#ParallelTransformsUnary">Create a Unary Parallel-Transform Task</a>
<ul>
<li class="level2">
<a href="#ParallelTransformsCaptureIteratorsByReference">Capture Iterators by Reference</a>
</li>
</ul>
</li>
<li class="level1">
<a href="#ParallelTransformsBinary">Create a Binary Parallel-Transform Task</a>
<ul>
<li class="level2">
<a href="#ParallelBinaryTransformsCaptureIteratorsByReference">Capture Iterators by Reference</a>
</li>
</ul>
</li>
<li class="level1">
<a href="#ParallelTransformsConfigureAPartitioner">Configure a Partitioner</a>
</li>
</ul>
</div>
<div class="textblock"><p>Taskflow provides template functions for constructing tasks to perform parallel transforms over ranges of items.</p>
<h1><a class="anchor" id="ParallelTransformsInclude"></a>
Include the Header</h1>
<p>You need to include the header file, <code>taskflow/algorithm/transform.hpp</code>, for creating a parallel-transform task.</p>
<div class="fragment"><div class="line"><span class="preprocessor">#include <taskflow/algorithm/transform.hpp></span></div>
</div><!-- fragment --><h1><a class="anchor" id="ParallelTransformsUnary"></a>
Create a Unary Parallel-Transform Task</h1>
<p>A unary parallel-transform applies a callable to every element in a source range and writes the result to a destination range. The task created by <a class="el" href="classtf_1_1FlowBuilder.html#a058c250de62b9d1e4305b8ddf03906ee" title="constructs a parallel-transform task">tf::Taskflow::transform(B first1, E last1, O d_first, C c, P part)</a> represents parallel execution of the following loop:</p>
<div class="fragment"><div class="line"><span class="keywordflow">while</span> (first1 != last1) {</div>
<div class="line"> *d_first++ = c(*first1++);</div>
<div class="line">}</div>
</div><!-- fragment --><p><a class="el" href="classtf_1_1FlowBuilder.html#a058c250de62b9d1e4305b8ddf03906ee" title="constructs a parallel-transform task">tf::Taskflow::transform</a> simultaneously applies the callable <code>c</code> to the object obtained by dereferencing every iterator in the range <code>[first1, last1)</code> and stores the result in another range beginning at <code>d_first</code>. It is the user's responsibility to ensure the range is valid within the execution of the parallel-transform task.</p>
<div class="fragment"><div class="line">std::vector<int> src = {1, 2, 3, 4, 5};</div>
<div class="line">std::vector<int> tgt(src.size());</div>
<div class="line"> </div>
<div class="line">taskflow.transform(src.begin(), src.end(), tgt.begin(), [](<span class="keywordtype">int</span> i) {</div>
<div class="line"> return i + 1;</div>
<div class="line">});</div>
</div><!-- fragment --><h2><a class="anchor" id="ParallelTransformsCaptureIteratorsByReference"></a>
Capture Iterators by Reference</h2>
<p>You can pass iterators by reference using <a href="https://en.cppreference.com/w/cpp/utility/functional/ref">std::ref</a> to marshal parameter updates between dependent tasks. This is useful when the range is not known at task-graph construction time but is initialized by an upstream task.</p>
<div class="fragment"><div class="line">std::vector<int> src, tgt;</div>
<div class="line">std::vector<int>::iterator first, last, d_first;</div>
<div class="line"> </div>
<div class="line"><a class="code hl_class" href="classtf_1_1Task.html">tf::Task</a> init = taskflow.emplace([&]() {</div>
<div class="line"> src.resize(1000);</div>
<div class="line"> tgt.resize(1000);</div>
<div class="line"> first = src.begin();</div>
<div class="line"> last = src.end();</div>
<div class="line"> d_first = tgt.begin();</div>
<div class="line">});</div>
<div class="line"> </div>
<div class="line"><a class="code hl_class" href="classtf_1_1Task.html">tf::Task</a> transform = taskflow.transform(</div>
<div class="line"> std::ref(first), std::ref(last), std::ref(d_first),</div>
<div class="line"> [](<span class="keywordtype">int</span> i) {</div>
<div class="line"> <span class="keywordflow">return</span> i + 1;</div>
<div class="line"> }</div>
<div class="line">);</div>
<div class="line"> </div>
<div class="line"><span class="comment">// wrong! first, last, and d_first are captured by copy at construction time</span></div>
<div class="line"><span class="comment">// tf::Task transform = taskflow.transform(first, last, d_first, [](int i) {</span></div>
<div class="line"><span class="comment">// return i + 1;</span></div>
<div class="line"><span class="comment">// });</span></div>
<div class="line"> </div>
<div class="line">init.<a class="code hl_function" href="classtf_1_1Task.html#a8c78c453295a553c1c016e4062da8588">precede</a>(transform);</div>
<div class="ttc" id="aclasstf_1_1Task_html"><div class="ttname"><a href="classtf_1_1Task.html">tf::Task</a></div><div class="ttdoc">class to create a task handle over a taskflow node</div><div class="ttdef"><b>Definition</b> task.hpp:569</div></div>
<div class="ttc" id="aclasstf_1_1Task_html_a8c78c453295a553c1c016e4062da8588"><div class="ttname"><a href="classtf_1_1Task.html#a8c78c453295a553c1c016e4062da8588">tf::Task::precede</a></div><div class="ttdeci">Task & precede(Ts &&... tasks)</div><div class="ttdoc">adds precedence links from this to other tasks</div><div class="ttdef"><b>Definition</b> task.hpp:1258</div></div>
</div><!-- fragment --><p>When <code>init</code> finishes, the parallel-transform task <code>transform</code> will see <code>first</code> pointing to the beginning of <code>src</code> and <code>last</code> pointing to the end of <code>src</code>, and performs parallel transforms over the 1000 items storing results starting at <code>d_first</code>.</p>
<h1><a class="anchor" id="ParallelTransformsBinary"></a>
Create a Binary Parallel-Transform Task</h1>
<p>A binary parallel-transform applies a callable to pairs of elements drawn from two source ranges and writes each result to a destination range. The overload <a class="el" href="classtf_1_1FlowBuilder.html#a18d263a5c043a216380441c9d0c72a60" title="constructs a parallel-transform task">tf::Taskflow::transform(B1 first1, E1 last1, B2 first2, O d_first, C c, P part)</a> represents parallel execution of the following loop:</p>
<div class="fragment"><div class="line"><span class="keywordflow">while</span> (first1 != last1) {</div>
<div class="line"> *d_first++ = c(*first1++, *first2++);</div>
<div class="line">}</div>
</div><!-- fragment --><p>The following example creates a parallel-transform task that adds two ranges element-wise and stores the result in a target range:</p>
<div class="fragment"><div class="line">std::vector<int> src1 = {1, 2, 3, 4, 5};</div>
<div class="line">std::vector<int> src2 = {5, 4, 3, 2, 1};</div>
<div class="line">std::vector<int> tgt(src1.size());</div>
<div class="line"> </div>
<div class="line">taskflow.transform(</div>
<div class="line"> src1.begin(), src1.end(), src2.begin(), tgt.begin(),</div>
<div class="line"> [](<span class="keywordtype">int</span> i, <span class="keywordtype">int</span> j) {</div>
<div class="line"> return i + j;</div>
<div class="line"> }</div>
<div class="line">);</div>
</div><!-- fragment --><h2><a class="anchor" id="ParallelBinaryTransformsCaptureIteratorsByReference"></a>
Capture Iterators by Reference</h2>
<p>As with the unary overload, all iterators can be passed by reference using <a href="https://en.cppreference.com/w/cpp/utility/functional/ref">std::ref</a> so that an upstream task can set up the ranges before the parallel-transform runs.</p>
<div class="fragment"><div class="line">std::vector<int> src1, src2, tgt;</div>
<div class="line">std::vector<int>::iterator first1, last1, first2, d_first;</div>
<div class="line"> </div>
<div class="line"><a class="code hl_class" href="classtf_1_1Task.html">tf::Task</a> init = taskflow.emplace([&]() {</div>
<div class="line"> src1.resize(1000);</div>
<div class="line"> src2.resize(1000);</div>
<div class="line"> tgt.resize(1000);</div>
<div class="line"> first1 = src1.begin();</div>
<div class="line"> last1 = src1.end();</div>
<div class="line"> first2 = src2.begin();</div>
<div class="line"> d_first = tgt.begin();</div>
<div class="line">});</div>
<div class="line"> </div>
<div class="line"><a class="code hl_class" href="classtf_1_1Task.html">tf::Task</a> transform = taskflow.transform(</div>
<div class="line"> std::ref(first1), std::ref(last1), std::ref(first2), std::ref(d_first),</div>
<div class="line"> [](<span class="keywordtype">int</span> i, <span class="keywordtype">int</span> j) {</div>
<div class="line"> <span class="keywordflow">return</span> i + j;</div>
<div class="line"> }</div>
<div class="line">);</div>
<div class="line"> </div>
<div class="line"><span class="comment">// wrong! all iterators are captured by copy at construction time</span></div>
<div class="line"><span class="comment">// tf::Task transform = taskflow.transform(</span></div>
<div class="line"><span class="comment">// first1, last1, first2, d_first, [](int i, int j) { return i + j; }</span></div>
<div class="line"><span class="comment">// );</span></div>
<div class="line"> </div>
<div class="line">init.<a class="code hl_function" href="classtf_1_1Task.html#a8c78c453295a553c1c016e4062da8588">precede</a>(transform);</div>
</div><!-- fragment --><p>When <code>init</code> finishes, the parallel-transform task <code>transform</code> will see all four iterators updated and performs parallel transforms over the 1000 item pairs, storing each result in <code>tgt</code>.</p>
<h1><a class="anchor" id="ParallelTransformsConfigureAPartitioner"></a>
Configure a Partitioner</h1>
<p>A partitioner controls how the iteration space is divided among workers. Taskflow provides four partitioners, each suited to different workload characteristics:</p>
<ul>
<li><a class="el" href="classtf_1_1StaticPartitioner.html" title="class to construct a static partitioner for scheduling parallel algorithms">tf::StaticPartitioner</a> divides the range into equal-sized chunks ahead of execution and assigns them to workers in order. It has the lowest scheduling overhead and delivers the best performance when every element costs roughly the same amount of work to transform.</li>
<li><a class="el" href="classtf_1_1DynamicPartitioner.html" title="class to create a dynamic partitioner for scheduling parallel algorithms">tf::DynamicPartitioner</a> distributes fixed-sized chunks to workers on demand as they become available. It adapts well to workloads where transform cost varies per element, at the expense of slightly higher coordination overhead.</li>
<li><a class="el" href="classtf_1_1GuidedPartitioner.html" title="class to create a guided partitioner for scheduling parallel algorithms">tf::GuidedPartitioner</a> distributes chunks whose size decreases adaptively as work is consumed — large chunks early to reduce overhead, smaller chunks late to balance the tail. This is the default partitioner and delivers stable, near-optimal performance across a wide range of workloads.</li>
<li><a class="el" href="classtf_1_1RandomPartitioner.html" title="class to construct a random partitioner for scheduling parallel algorithms">tf::RandomPartitioner</a> distributes chunks of randomly sampled sizes, which can help avoid systematic load imbalances caused by data-dependent cost patterns.</li>
</ul>
<p>The following example creates two parallel-transform tasks using different partitioners:</p>
<div class="fragment"><div class="line">std::vector<int> src1 = {1, 2, 3, 4, 5};</div>
<div class="line">std::vector<int> src2 = {5, 4, 3, 2, 1};</div>
<div class="line">std::vector<int> tgt1(src1.size());</div>
<div class="line">std::vector<int> tgt2(src1.size());</div>
<div class="line"> </div>
<div class="line"><a class="code hl_class" href="classtf_1_1StaticPartitioner.html">tf::StaticPartitioner</a> static_partitioner(0); <span class="comment">// chunk size auto-determined</span></div>
<div class="line"><a class="code hl_class" href="classtf_1_1GuidedPartitioner.html">tf::GuidedPartitioner</a> guided_partitioner(0); <span class="comment">// minimum chunk size auto-determined</span></div>
<div class="line"> </div>
<div class="line"><span class="comment">// parallel-transform with static partitioner</span></div>
<div class="line">taskflow.transform(</div>
<div class="line"> src1.begin(), src1.end(), src2.begin(), tgt1.begin(),</div>
<div class="line"> [](<span class="keywordtype">int</span> i, <span class="keywordtype">int</span> j) { return i + j; },</div>
<div class="line"> static_partitioner</div>
<div class="line">);</div>
<div class="line"> </div>
<div class="line"><span class="comment">// parallel-transform with guided partitioner</span></div>
<div class="line">taskflow.transform(</div>
<div class="line"> src1.begin(), src1.end(), src2.begin(), tgt2.begin(),</div>
<div class="line"> [](<span class="keywordtype">int</span> i, <span class="keywordtype">int</span> j) { return i + j; },</div>
<div class="line"> guided_partitioner</div>
<div class="line">);</div>
<div class="ttc" id="aclasstf_1_1GuidedPartitioner_html"><div class="ttname"><a href="classtf_1_1GuidedPartitioner.html">tf::GuidedPartitioner</a></div><div class="ttdoc">class to create a guided partitioner for scheduling parallel algorithms</div><div class="ttdef"><b>Definition</b> partitioner.hpp:417</div></div>
<div class="ttc" id="aclasstf_1_1StaticPartitioner_html"><div class="ttname"><a href="classtf_1_1StaticPartitioner.html">tf::StaticPartitioner</a></div><div class="ttdoc">class to construct a static partitioner for scheduling parallel algorithms</div><div class="ttdef"><b>Definition</b> partitioner.hpp:262</div></div>
</div><!-- fragment --><p>As a rule of thumb, prefer <a class="el" href="classtf_1_1StaticPartitioner.html" title="class to construct a static partitioner for scheduling parallel algorithms">tf::StaticPartitioner</a> when every element costs the same to transform (e.g., element-wise arithmetic) and <a class="el" href="classtf_1_1GuidedPartitioner.html" title="class to create a guided partitioner for scheduling parallel algorithms">tf::GuidedPartitioner</a> for irregular workloads (e.g., transforms whose cost depends on the element value). <a class="el" href="classtf_1_1DynamicPartitioner.html" title="class to create a dynamic partitioner for scheduling parallel algorithms">tf::DynamicPartitioner</a> is a good choice when chunks must be kept small and strictly equal in size.</p>
<dl class="section note"><dt>Note</dt><dd>By default, parallel-transform tasks use <a class="el" href="namespacetf.html#ace2c5adcd5039483eebb6dbdbb6f33e3" title="default partitioner set to tf::GuidedPartitioner">tf::DefaultPartitioner</a> (currently <a class="el" href="classtf_1_1GuidedPartitioner.html" title="class to create a guided partitioner for scheduling parallel algorithms">tf::GuidedPartitioner</a>) if no partitioner is specified. </dd></dl>
</div></div><!-- contents -->
</div><!-- PageDoc -->
</div><!-- doc-content -->
<!-- HTML footer for doxygen 1.13.1-->
<!-- start footer part -->
<div id="nav-path" class="navpath"><!-- id is needed for treeview function! -->
<ul>
<li class="navelem"><a class="el" href="Algorithms.html">Taskflow Algorithms</a></li>
<li class="footer">
Maintained by <a href="https://tsung-wei-huang.github.io/">Dr. Tsung-Wei Huang</a>
—
Generated by <a href="https://www.doxygen.org/index.html"><img class="footer" src="doxygen.svg" width="104" height="31" alt="doxygen"/></a> 1.13.1
</li>
</ul>
</div>