mesytec-mnode/external/taskflow-3.8.0/docs/xml/CUDASTDFind.xml

<?xml version='1.0' encoding='UTF-8' standalone='no'?>
<doxygen xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="compound.xsd" version="1.9.1" xml:lang="en-US">
  <compounddef id="CUDASTDFind" kind="page">
    <compoundname>CUDASTDFind</compoundname>
    <title>Parallel Find</title>
    <tableofcontents>
      <tocsect>
        <name>Include the Header</name>
        <reference>CUDASTDFind_1CUDASTDFindIncludeTheHeader</reference>
    </tocsect>
      <tocsect>
        <name>Find an Element in a Range</name>
        <reference>CUDASTDFind_1CUDASTDFindItems</reference>
    </tocsect>
      <tocsect>
        <name>Find the Minimum Element in a Range</name>
        <reference>CUDASTDFind_1CUDASTDFindMinItems</reference>
    </tocsect>
      <tocsect>
        <name>Find the Maximum Element in a Range</name>
        <reference>CUDASTDFind_1CUDASTDFindMaxItems</reference>
    </tocsect>
    </tableofcontents>
    <briefdescription>
    </briefdescription>
    <detaileddescription>
<para>Taskflow provides standalone template methods for finding elements in the given ranges using GPU.</para>
<sect1 id="CUDASTDFind_1CUDASTDFindIncludeTheHeader">
<title>Include the Header</title>
<para>You need to include the header file, <computeroutput>taskflow/cuda/algorithm/find.hpp</computeroutput>, for using the parallel-find algorithm.</para>
<para><programlisting filename=".cpp"><codeline><highlight class="preprocessor">#include<sp/>&lt;<ref refid="find_8hpp" kindref="compound">taskflow/cuda/algorithm/find.hpp</ref>&gt;</highlight></codeline>
</programlisting></para>
</sect1>
<sect1 id="CUDASTDFind_1CUDASTDFindItems">
<title>Find an Element in a Range</title>
<para><ref refid="namespacetf_1a5f9dabd7c5d0fa5166cf76d9fa5a038e" kindref="member">tf::cuda_find_if</ref> finds the index of the first element in the range <computeroutput>[first, last)</computeroutput> that satisfies the given criteria. This is equivalent to the parallel execution of the following loop:</para>
<para><programlisting filename=".cpp"><codeline><highlight class="keywordtype">unsigned</highlight><highlight class="normal"><sp/>idx<sp/>=<sp/>0;</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">for</highlight><highlight class="normal">(;<sp/>first<sp/>!=<sp/>last;<sp/>++first,<sp/>++idx)<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/></highlight><highlight class="keywordflow">if</highlight><highlight class="normal"><sp/>(p(*first))<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/><sp/><sp/></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/>idx;</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/>}</highlight></codeline>
<codeline><highlight class="normal">}</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/>idx;</highlight></codeline>
</programlisting></para>
<para>If no such an element is found, the size of the range is returned. The following code finds the index of the first element that is dividable by <computeroutput>17</computeroutput> over a range of one million elements.</para>
<para><programlisting filename=".cpp"><codeline><highlight class="keyword">const</highlight><highlight class="normal"><sp/></highlight><highlight class="keywordtype">size_t</highlight><highlight class="normal"><sp/>N<sp/>=<sp/>1000000;</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>vec<sp/>=<sp/>tf::cuda_malloc_shared&lt;int&gt;(N);<sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight class="comment">//<sp/>vector</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>idx<sp/>=<sp/>tf::cuda_malloc_shared&lt;unsigned&gt;(1);<sp/><sp/></highlight><highlight class="comment">//<sp/>index</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>initializes<sp/>the<sp/>data</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">for</highlight><highlight class="normal">(</highlight><highlight class="keywordtype">size_t</highlight><highlight class="normal"><sp/>i=0;<sp/>i&lt;N;<sp/>vec[i++]<sp/>=<sp/><ref refid="cpp/numeric/random/rand" kindref="compound" external="/home/thuang295/Code/taskflow/doxygen/cppreference-doxygen-web.tag.xml">rand</ref>());</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>create<sp/>an<sp/>execution<sp/>policy</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"><ref refid="classtf_1_1cudaExecutionPolicy" kindref="compound">tf::cudaDefaultExecutionPolicy</ref><sp/>policy;</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>finds<sp/>the<sp/>index<sp/>of<sp/>the<sp/>first<sp/>element<sp/>that<sp/>is<sp/>a<sp/>multiple<sp/>of<sp/>17</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"><ref refid="namespacetf_1a5f9dabd7c5d0fa5166cf76d9fa5a038e" kindref="member">tf::cuda_find_if</ref>(</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/>policy,<sp/>vec,<sp/>vec+N,<sp/>idx,<sp/>[]<sp/>__device__<sp/>(</highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>v)<sp/>{<sp/></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/>v%17<sp/>==<sp/>0;<sp/>}</highlight></codeline>
<codeline><highlight class="normal">);</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>wait<sp/>for<sp/>the<sp/>find<sp/>operation<sp/>to<sp/>complete</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal">stream.synchronize();</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>verifies<sp/>the<sp/>result</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">if</highlight><highlight class="normal">(*idx<sp/>!=<sp/>N)<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/>assert(vec[*idx]<sp/>%17<sp/>==<sp/>0);</highlight></codeline>
<codeline><highlight class="normal">}</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>deletes<sp/>the<sp/>memory</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal">cudaFree(vec);</highlight></codeline>
<codeline><highlight class="normal">cudaFree(idx);</highlight></codeline>
</programlisting></para>
<para>The find-if algorithm runs <emphasis>asynchronously</emphasis> through the stream specified in the execution policy. You need to synchronize the stream to obtain the correct result.</para>
</sect1>
<sect1 id="CUDASTDFind_1CUDASTDFindMinItems">
<title>Find the Minimum Element in a Range</title>
<para><ref refid="namespacetf_1a572c13198191c46765264f8afabe2e9f" kindref="member">tf::cuda_min_element</ref> finds the index of the minimum element in the given range <computeroutput>[first, last)</computeroutput> using the given comparison function object. This is equivalent to a parallel execution of the following loop:</para>
<para><programlisting filename=".cpp"><codeline><highlight class="keywordflow">if</highlight><highlight class="normal">(first<sp/>==<sp/>last)<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/>0;</highlight></codeline>
<codeline><highlight class="normal">}</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>smallest<sp/>=<sp/>first;</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">for</highlight><highlight class="normal"><sp/>(++first;<sp/>first<sp/>!=<sp/>last;<sp/>++first)<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/></highlight><highlight class="keywordflow">if</highlight><highlight class="normal"><sp/>(op(*first,<sp/>*smallest))<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/><sp/><sp/>smallest<sp/>=<sp/>first;</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/>}</highlight></codeline>
<codeline><highlight class="normal">}</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/><ref refid="cpp/iterator/distance" kindref="compound" external="/home/thuang295/Code/taskflow/doxygen/cppreference-doxygen-web.tag.xml">std::distance</ref>(first,<sp/>smallest);</highlight></codeline>
</programlisting></para>
<para>The following code finds the index of the minimum element in a range of one millions elements using GPU computing:</para>
<para><programlisting filename=".cpp"><codeline><highlight class="keyword">const</highlight><highlight class="normal"><sp/></highlight><highlight class="keywordtype">size_t</highlight><highlight class="normal"><sp/>N<sp/>=<sp/>1000000;</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>vec<sp/>=<sp/>tf::cuda_malloc_shared&lt;int&gt;(N);<sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight class="comment">//<sp/>vector</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>idx<sp/>=<sp/>tf::cuda_malloc_shared&lt;unsigned&gt;(1);<sp/><sp/></highlight><highlight class="comment">//<sp/>index</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>initializes<sp/>the<sp/>data</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">for</highlight><highlight class="normal">(</highlight><highlight class="keywordtype">size_t</highlight><highlight class="normal"><sp/>i=0;<sp/>i&lt;N;<sp/>vec[i++]<sp/>=<sp/><ref refid="cpp/numeric/random/rand" kindref="compound" external="/home/thuang295/Code/taskflow/doxygen/cppreference-doxygen-web.tag.xml">rand</ref>());</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>create<sp/>an<sp/>execution<sp/>policy</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"><ref refid="classtf_1_1cudaStream" kindref="compound">tf::cudaStream</ref><sp/>stream;</highlight></codeline>
<codeline><highlight class="normal"><ref refid="classtf_1_1cudaExecutionPolicy" kindref="compound">tf::cudaDefaultExecutionPolicy</ref><sp/>policy(stream);</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>queries<sp/>the<sp/>required<sp/>buffer<sp/>size<sp/>to<sp/>find<sp/>the<sp/>minimum<sp/>element<sp/>over<sp/>N<sp/>element</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>bytes<sp/><sp/>=<sp/>policy.<ref refid="classtf_1_1cudaExecutionPolicy_1abcafb001cd68c1135392f4bcda5a2a05" kindref="member">min_element_bufsz</ref>&lt;</highlight><highlight class="keywordtype">int</highlight><highlight class="normal">&gt;(N);</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>buffer<sp/>=<sp/>tf::cuda_malloc_device&lt;std::byte&gt;(bytes);</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>finds<sp/>the<sp/>minimum<sp/>element<sp/>using<sp/>the<sp/>less<sp/>comparator</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"><ref refid="namespacetf_1a572c13198191c46765264f8afabe2e9f" kindref="member">tf::cuda_min_element</ref>(</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/>policy,<sp/>vec,<sp/>vec+N,<sp/>idx,<sp/>[]<sp/>__device__<sp/>(</highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>a,<sp/></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>b)<sp/>{<sp/></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/>a&lt;b;<sp/>},<sp/>buffer</highlight></codeline>
<codeline><highlight class="normal">);</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>wait<sp/>for<sp/>the<sp/>min-element<sp/>operation<sp/>completes</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal">stream.<ref refid="classtf_1_1cudaStream_1a1a81d6005e8d60ad082dba2303a8aa30" kindref="member">synchronize</ref>();</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>verifies<sp/>the<sp/>result</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal">assert(vec[*idx]<sp/>==<sp/>*std::min_element(vec,<sp/>vec+N,<sp/><ref refid="cpp/utility/functional/less" kindref="compound" external="/home/thuang295/Code/taskflow/doxygen/cppreference-doxygen-web.tag.xml">std::less&lt;int&gt;</ref>{}));</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>deletes<sp/>the<sp/>memory</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal">cudaFree(vec);</highlight></codeline>
<codeline><highlight class="normal">cudaFree(idx);</highlight></codeline>
<codeline><highlight class="normal">cudaFree(buffer);</highlight></codeline>
</programlisting></para>
<para>Since the GPU min-element algorithm may require extra buffer to store the temporary results, you need to provide a buffer of size at least larger or equal to the value returned from <computeroutput><ref refid="classtf_1_1cudaExecutionPolicy_1abcafb001cd68c1135392f4bcda5a2a05" kindref="member">tf::cudaDefaultExecutionPolicy::min_element_bufsz</ref></computeroutput>.</para>
<para><simplesect kind="attention"><para>You must keep the buffer alive before the <ref refid="namespacetf_1a572c13198191c46765264f8afabe2e9f" kindref="member">tf::cuda_min_element</ref> completes.</para>
</simplesect>
</para>
</sect1>
<sect1 id="CUDASTDFind_1CUDASTDFindMaxItems">
<title>Find the Maximum Element in a Range</title>
<para>Similar to <ref refid="namespacetf_1a572c13198191c46765264f8afabe2e9f" kindref="member">tf::cuda_min_element</ref>, <ref refid="namespacetf_1a3fc577fd0a8f127770bcf68bc56c073e" kindref="member">tf::cuda_max_element</ref> finds the index of the maximum element in the given range <computeroutput>[first, last)</computeroutput> using the given comparison function object. This is equivalent to a parallel execution of the following loop:</para>
<para><programlisting filename=".cpp"><codeline><highlight class="keywordflow">if</highlight><highlight class="normal">(first<sp/>==<sp/>last)<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/>0;</highlight></codeline>
<codeline><highlight class="normal">}</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>largest<sp/>=<sp/>first;</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">for</highlight><highlight class="normal"><sp/>(++first;<sp/>first<sp/>!=<sp/>last;<sp/>++first)<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/></highlight><highlight class="keywordflow">if</highlight><highlight class="normal"><sp/>(op(*largest,<sp/>*first))<sp/>{</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/><sp/><sp/>largest<sp/>=<sp/>first;</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/>}</highlight></codeline>
<codeline><highlight class="normal">}</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/><ref refid="cpp/iterator/distance" kindref="compound" external="/home/thuang295/Code/taskflow/doxygen/cppreference-doxygen-web.tag.xml">std::distance</ref>(first,<sp/>largest);</highlight></codeline>
</programlisting></para>
<para>The following code finds the index of the maximum element in a range of one millions elements using GPU computing:</para>
<para><programlisting filename=".cpp"><codeline><highlight class="keyword">const</highlight><highlight class="normal"><sp/></highlight><highlight class="keywordtype">size_t</highlight><highlight class="normal"><sp/>N<sp/>=<sp/>1000000;</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>vec<sp/>=<sp/>tf::cuda_malloc_shared&lt;int&gt;(N);<sp/><sp/><sp/><sp/><sp/><sp/><sp/></highlight><highlight class="comment">//<sp/>vector</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>idx<sp/>=<sp/>tf::cuda_malloc_shared&lt;unsigned&gt;(1);<sp/><sp/></highlight><highlight class="comment">//<sp/>index</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>initializes<sp/>the<sp/>data</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keywordflow">for</highlight><highlight class="normal">(</highlight><highlight class="keywordtype">size_t</highlight><highlight class="normal"><sp/>i=0;<sp/>i&lt;N;<sp/>vec[i++]<sp/>=<sp/><ref refid="cpp/numeric/random/rand" kindref="compound" external="/home/thuang295/Code/taskflow/doxygen/cppreference-doxygen-web.tag.xml">rand</ref>());</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>create<sp/>an<sp/>execution<sp/>policy</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"><ref refid="classtf_1_1cudaStream" kindref="compound">tf::cudaStream</ref><sp/>stream;</highlight></codeline>
<codeline><highlight class="normal"><ref refid="classtf_1_1cudaExecutionPolicy" kindref="compound">tf::cudaDefaultExecutionPolicy</ref><sp/>policy(stream);</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>queries<sp/>the<sp/>required<sp/>buffer<sp/>size<sp/>to<sp/>find<sp/>the<sp/>maximum<sp/>element<sp/>over<sp/>N<sp/>element</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>bytes<sp/><sp/>=<sp/>policy.<ref refid="classtf_1_1cudaExecutionPolicy_1a31fe75c4b0765df3035e12be49af88aa" kindref="member">max_element_bufsz</ref>&lt;</highlight><highlight class="keywordtype">int</highlight><highlight class="normal">&gt;(N);</highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>buffer<sp/>=<sp/>tf::cuda_malloc_device&lt;std::byte&gt;(bytes);</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>finds<sp/>the<sp/>maximum<sp/>element<sp/>using<sp/>the<sp/>less<sp/>comparator</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"><ref refid="namespacetf_1a3fc577fd0a8f127770bcf68bc56c073e" kindref="member">tf::cuda_max_element</ref>(</highlight></codeline>
<codeline><highlight class="normal"><sp/><sp/>policy,<sp/>vec,<sp/>vec+N,<sp/>idx,<sp/>[]<sp/>__device__<sp/>(</highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>a,<sp/></highlight><highlight class="keyword">auto</highlight><highlight class="normal"><sp/>b)<sp/>{<sp/></highlight><highlight class="keywordflow">return</highlight><highlight class="normal"><sp/>a&lt;b;<sp/>},<sp/>buffer</highlight></codeline>
<codeline><highlight class="normal">);</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>wait<sp/>for<sp/>the<sp/>max-element<sp/>operation<sp/>to<sp/>complete</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal">stream.<ref refid="classtf_1_1cudaStream_1a1a81d6005e8d60ad082dba2303a8aa30" kindref="member">synchronize</ref>();</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>verifies<sp/>the<sp/>result</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal">assert(vec[*idx]<sp/>==<sp/>*std::max_element(vec,<sp/>vec+N,<sp/><ref refid="cpp/utility/functional/less" kindref="compound" external="/home/thuang295/Code/taskflow/doxygen/cppreference-doxygen-web.tag.xml">std::less&lt;int&gt;</ref>{}));</highlight></codeline>
<codeline><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal"></highlight><highlight class="comment">//<sp/>deletes<sp/>the<sp/>memory</highlight><highlight class="normal"></highlight></codeline>
<codeline><highlight class="normal">cudaFree(vec);</highlight></codeline>
<codeline><highlight class="normal">cudaFree(idx);</highlight></codeline>
<codeline><highlight class="normal">cudaFree(buffer);</highlight></codeline>
</programlisting></para>
<para>Since the GPU max-element algorithm may require extra buffer to store the temporary results, you need to provide a buffer of size at least larger or equal to the value returned from <computeroutput><ref refid="classtf_1_1cudaExecutionPolicy_1a31fe75c4b0765df3035e12be49af88aa" kindref="member">tf::cudaDefaultExecutionPolicy::max_element_bufsz</ref></computeroutput>.</para>
<para><simplesect kind="attention"><para>You must keep the buffer alive before <ref refid="namespacetf_1a3fc577fd0a8f127770bcf68bc56c073e" kindref="member">tf::cuda_max_element</ref> completes. </para>
</simplesect>
</para>
</sect1>
    </detaileddescription>
    <location file="doxygen/cuda_std_algorithms/cuda_std_find.dox"/>
  </compounddef>
</doxygen>