cpp-taskflow/docs/CUDASTDForEach.html at master · ezhangle/cpp-taskflow · GitHub

History

138 lines (134 loc) · 14.1 KB

Raw

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

<!DOCTYPE html>

<head>

<title>CUDA Standard Algorithms » Parallel Iterations | Taskflow QuickStart</title>

</head>

<body>

<a href="https://taskflow.github.io"><img src="taskflow_logo.png" alt="" />Taskflow</a> | <a href="index.html" class="m-thin">QuickStart</a>

</svg></a>

</div>

<li><a href="pages.html">Handbook</a></li>

<li><a href="namespaces.html">Namespaces</a></li>

</ol>

<li><a href="annotated.html">Classes</a></li>

<li><a href="files.html">Files</a></li>

</svg></a></li>

</ol>

</div>

</nav></header>

<h1>

<a href="cudaStandardAlgorithms.html">CUDA Standard Algorithms</a> »

Parallel Iterations

</h1>

<h3>Contents</h3>

<ul>

<li><a href="#CUDASTDParallelIterationIncludeTheHeader">Include the Header</a></li>

<li><a href="#CUDASTDIndexBasedParallelFor">Index-based Parallel Iterations</a></li>

<li><a href="#CUDASTDIteratorBasedParallelFor">Iterator-based Parallel Iterations</a></li>

</ul>

</nav>

Taskflow provides standard template methods for performing parallel iterations over a range of items a CUDA GPU.<section id="CUDASTDParallelIterationIncludeTheHeader"><h2><a href="#CUDASTDParallelIterationIncludeTheHeader">Include the Header</a></h2>You need to include the header file, <code>taskflow/cuda/algorithm/for_each.hpp</code>, for using the parallel-iteration algorithm.<pre class="m-code">#include <taskflow/cuda/algorithm/for_each.hpp></pre></section><section id="CUDASTDIndexBasedParallelFor"><h2><a href="#CUDASTDIndexBasedParallelFor">Index-based Parallel Iterations</a></h2>Index-based parallel-for performs parallel iterations over a range <code>[first, last)</code> with the given <code>step</code> size. The task created by <a href="namespacetf.html#a01ad7ce62fa6f42f2f2fbff3659b7884" class="m-doc">tf::cuda_for_each_index</a> represents a kernel of parallel execution for the following loop:<pre class="m-code">// positive step: first, first+step, first+2*step, ...

for(auto i=first; i<last; i+=step) {

callable(i);

}

// negative step: first, first-step, first-2*step, ...

for(auto i=first; i>last; i+=step) {

callable(i);

}</pre>Each iteration <code>i</code> is independent of each other and is assigned one kernel thread to run the callable. The following example creates a kernel that assigns each entry of <code>data</code> to 1 over the range [0, 100) with step size 1.<pre class="m-code">tf::cudaDefaultExecutionPolicy policy;

auto data = tf::cuda_malloc_shared<int>(100);

// assigns each element in data to 1 over the range [0, 100) with step size 1

tf::cuda_for_each_index(

policy, 0, 100, 1, [data] __device__ (int idx) { data[idx] = 1; }

);

// synchronize the execution

policy.synchronize();</pre>The parallel-iteration algorithm runs asynchronously through the stream specified in the execution policy. You need to synchronize the stream to obtain correct results.</section><section id="CUDASTDIteratorBasedParallelFor"><h2><a href="#CUDASTDIteratorBasedParallelFor">Iterator-based Parallel Iterations</a></h2>Iterator-based parallel-for performs parallel iterations over a range specified by two STL-styled iterators, <code>first</code> and <code>last</code>. The task created by <a href="namespacetf.html#a7c449cec0b93503b8280d05add35e9f4" class="m-doc">tf::cuda_for_each</a> represents a parallel execution of the following loop:<pre class="m-code">for(auto i=first; i<last; i++) {

callable(*i);

}</pre>The two iterators, <code>first</code> and <code>last</code>, are typically two raw pointers to the first element and the next to the last element in the range in GPU memory space. The following example creates a <code>for_each</code> kernel that assigns each element in <code>gpu_data</code> to 1 over the range <code>[data, data + 1000)</code>.<pre class="m-code">tf::cudaDefaultExecutionPolicy policy;

// assigns each element in data to 1 over the range [0, 1000) with step size 1

tf::cuda_for_each(

policy, data, data + 1000, [] __device__ (int& item) { item = 1; }

);

// synchronize the execution

policy.synchronize();</pre>Each iteration is independent of each other and is assigned one kernel thread to run the callable. Since the callable runs on GPU, it must be declared with a <code>__device__</code> specifier.</section>

</div>

</article></main>

<div>Tab / T to search, Esc to close</div>

</div>

<form>

</form>

Search for symbols, directories, files, pages or

modules. You can omit any prefix from the symbol or file path; adding a

<code>:</code> or <code>/</code> suffix lists all members of given symbol or

directory.

Use ↓

/ ↑ to navigate through the list,

Enter to go.

Tab autocompletes common prefix, you can

copy a link to the result using ⌘

L while ⌘

M produces a Markdown link.

</div>

<div id="search-notfound" class="m-text m-warning m-text-center">Sorry, nothing was found.</div>

</div>

Taskflow handbook is part of the <a href="https://taskflow.github.io">Taskflow project</a>, copyright © <a href="https://tsung-wei-huang.github.io/">Dr. Tsung-Wei Huang</a>, 2018–2024. Generated by <a href="https://doxygen.org/">Doxygen</a> 1.9.1 and <a href="https://mcss.mosra.cz/">m.css</a>.

</div>

</nav></footer>

</body>

</html>

cpp-taskflow/docs/CUDASTDForEach.html at master · ezhangle/cpp-taskflow · GitHub

Expand file tree

Breadcrumbs

CUDASTDForEach.html

Latest commit

History

Breadcrumbs

CUDASTDForEach.html

File metadata and controls