Skip to content

P
Paddle

PaddlePaddle / Paddle
大约 1 年 前同步成功

通知 2298
Star 20931
Fork 5422
P
Paddle
切换分支/标签
查找文件 普通视图历史永久链接Permalink
mkldnn_fluid.html 39.4 KB
Newer Older
T
Deploy to GitHub Pages: fac25fb53f526c34f23f0c6045f2b33b9ba18fb2  
Travis CI 已提交
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 


<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
  <meta charset="utf-8">
  
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  
  <title>Design Doc: Add MKLDNN Kernel in Fluid Operator &mdash; PaddlePaddle  文档</title>
  

  
  

  

  
  
    

  

  
  
    <link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
  

  
  
        <link rel="index" title="索引"
              href="../../genindex.html"/>
        <link rel="search" title="搜索" href="../../search.html"/>
    <link rel="top" title="PaddlePaddle  文档" href="../../index.html"/> 

  <link rel="stylesheet" href="https://cdn.jsdelivr.net/perfect-scrollbar/0.6.14/css/perfect-scrollbar.min.css" type="text/css" />
  <link rel="stylesheet" href="../../_static/css/override.css" type="text/css" />
  <script>
  var _hmt = _hmt || [];
  (function() {
    var hm = document.createElement("script");
    hm.src = "//hm.baidu.com/hm.js?b9a314ab40d04d805655aab1deee08ba";
    var s = document.getElementsByTagName("script")[0]; 
    s.parentNode.insertBefore(hm, s);
  })();
  </script>

  

  
  <script src="../../_static/js/modernizr.min.js"></script>

</head>

<body class="wy-body-for-nav" role="document">

  
  <header class="site-header">
    <div class="site-logo">
      <a href="/"><img src="../../_static/images/PP_w.png"></a>
    </div>
    <div class="site-nav-links">
      <div class="site-menu">
        <a class="fork-on-github" href="https://github.com/PaddlePaddle/Paddle" target="_blank"><i class="fa fa-github"></i>Fork me on Github</a>
        <div class="language-switcher dropdown">
          <a type="button" data-toggle="dropdown">
            <span>English</span>
            <i class="fa fa-angle-up"></i>
            <i class="fa fa-angle-down"></i>
          </a>
          <ul class="dropdown-menu">
            <li><a href="/doc_cn">中文</a></li>
            <li><a href="/doc">English</a></li>
          </ul>
        </div>
        <ul class="site-page-links">
          <li><a href="/">Home</a></li>
        </ul>
      </div>
      <div class="doc-module">
        
        <ul>
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_cn.html">新手入门</a></li>
T
Deploy to GitHub Pages: 7fa1725b04c7f1668e575665c8144225ad8d97a4  
Travis CI 已提交
85 86 87 
<li class="toctree-l1"><a class="reference internal" href="../../build_and_install/index_cn.html">安装与编译</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../howto/index_cn.html">进阶使用</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../dev/index_cn.html">开发标准</a></li>
T
Deploy to GitHub Pages: fac25fb53f526c34f23f0c6045f2b33b9ba18fb2  
Travis CI 已提交
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_cn.html">FAQ</a></li>
</ul>

        
<div role="search">
  <form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
    <input type="text" name="q" placeholder="Search docs" />
    <input type="hidden" name="check_keywords" value="yes" />
    <input type="hidden" name="area" value="default" />
  </form>
</div>        
      </div>
    </div>
  </header>
  
  <div class="main-content-wrap">

    
    <nav class="doc-menu-vertical" role="navigation">
        
          
          <ul>
<li class="toctree-l1"><a class="reference internal" href="../../getstarted/index_cn.html">新手入门</a><ul>
T
Deploy to GitHub Pages: 7fa1725b04c7f1668e575665c8144225ad8d97a4  
Travis CI 已提交
111 112 
<li class="toctree-l2"><a class="reference internal" href="../../getstarted/quickstart_cn.html">快速开始</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../getstarted/concepts/use_concepts_cn.html">基本使用概念</a></li>
T
Deploy to GitHub Pages: fac25fb53f526c34f23f0c6045f2b33b9ba18fb2  
Travis CI 已提交
113 114 
</ul>
</li>
T
Deploy to GitHub Pages: 7fa1725b04c7f1668e575665c8144225ad8d97a4  
Travis CI 已提交
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 
<li class="toctree-l1"><a class="reference internal" href="../../build_and_install/index_cn.html">安装与编译</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/pip_install_cn.html">使用pip安装</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/docker_install_cn.html">使用Docker安装运行</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../build_and_install/build_from_source_cn.html">从源码编译</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../howto/index_cn.html">进阶使用</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../howto/cmd_parameter/index_cn.html">命令行参数设置</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../howto/cmd_parameter/use_case_cn.html">使用案例</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/cmd_parameter/arguments_cn.html">参数概述</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/cmd_parameter/detail_introduction_cn.html">细节描述</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../howto/cluster/index_cn.html">分布式训练</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../howto/cluster/preparations_cn.html">环境准备</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/cluster/cmd_argument_cn.html">启动参数说明</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/cluster/multi_cluster/index_cn.html">在不同集群中运行</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../howto/cluster/multi_cluster/k8s_cn.html">Kubernetes单机训练</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../howto/cluster/multi_cluster/k8s_distributed_cn.html">Kubernetes分布式训练</a></li>
T
Deploy to GitHub Pages: 9344e4e3006057b733116cdcfa57dd18bccd1ed9  
Travis CI 已提交
134 
<li class="toctree-l4"><a class="reference internal" href="../../howto/cluster/multi_cluster/openmpi_cn.html">在OpenMPI集群中启动训练</a></li>
T
Deploy to GitHub Pages: 00595f759152d22c0e8aabf8c6007838577ed00c  
Travis CI 已提交
135 
<li class="toctree-l4"><a class="reference internal" href="../../howto/cluster/multi_cluster/fabric_cn.html">使用fabric启动集群训练</a></li>
T
Deploy to GitHub Pages: 9344e4e3006057b733116cdcfa57dd18bccd1ed9  
Travis CI 已提交
136 
<li class="toctree-l4"><a class="reference internal" href="../../howto/cluster/multi_cluster/k8s_aws_cn.html">Kubernetes on AWS</a></li>
T
Deploy to GitHub Pages: fac25fb53f526c34f23f0c6045f2b33b9ba18fb2  
Travis CI 已提交
137 138 139 140 
</ul>
</li>
</ul>
</li>
T
Deploy to GitHub Pages: 7fa1725b04c7f1668e575665c8144225ad8d97a4  
Travis CI 已提交
141 142 143 144 
<li class="toctree-l2"><a class="reference internal" href="../../howto/capi/index_cn.html">C-API预测库</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../howto/capi/compile_paddle_lib_cn.html">安装与编译C-API预测库</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/capi/organization_of_the_inputs_cn.html">输入/输出数据组织</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/capi/workflow_of_capi_cn.html">C-API使用流程</a></li>
T
Deploy to GitHub Pages: fac25fb53f526c34f23f0c6045f2b33b9ba18fb2  
Travis CI 已提交
145 146 
</ul>
</li>
T
Deploy to GitHub Pages: 100dd4c5416f4f39a142f510a112b104d19f2028  
Travis CI 已提交
147 
<li class="toctree-l2"><a class="reference internal" href="../../howto/rnn/index_cn.html">RNN模型</a><ul>
T
Deploy to GitHub Pages: 7fa1725b04c7f1668e575665c8144225ad8d97a4  
Travis CI 已提交
148 149 150 151 
<li class="toctree-l3"><a class="reference internal" href="../../howto/rnn/rnn_config_cn.html">RNN配置</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/rnn/recurrent_group_cn.html">Recurrent Group教程</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/rnn/hierarchical_layer_cn.html">支持双层序列作为输入的Layer</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../howto/rnn/hrnn_rnn_api_compare_cn.html">单双层RNN API对比介绍</a></li>
T
Deploy to GitHub Pages: 03c858a9bee98dbbcbedd7a6be886402a2e97609  
Travis CI 已提交
152 153 
</ul>
</li>
T
Deploy to GitHub Pages: 7fa1725b04c7f1668e575665c8144225ad8d97a4  
Travis CI 已提交
154 
<li class="toctree-l2"><a class="reference internal" href="../../howto/optimization/gpu_profiling_cn.html">GPU性能调优</a></li>
T
Deploy to GitHub Pages: fac25fb53f526c34f23f0c6045f2b33b9ba18fb2  
Travis CI 已提交
155 156 
</ul>
</li>
T
Deploy to GitHub Pages: 7fa1725b04c7f1668e575665c8144225ad8d97a4  
Travis CI 已提交
157 158 159 
<li class="toctree-l1"><a class="reference internal" href="../../dev/index_cn.html">开发标准</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../dev/contribute_to_paddle_cn.html">如何贡献代码</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../dev/write_docs_cn.html">如何贡献文档</a></li>
T
Deploy to GitHub Pages: fac25fb53f526c34f23f0c6045f2b33b9ba18fb2  
Travis CI 已提交
160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../faq/index_cn.html">FAQ</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../faq/build_and_install/index_cn.html">编译安装与单元测试</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/model/index_cn.html">模型配置</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/parameter/index_cn.html">参数设置</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/local/index_cn.html">本地训练与预测</a></li>
<li class="toctree-l2"><a class="reference internal" href="../../faq/cluster/index_cn.html">集群训练与预测</a></li>
</ul>
</li>
</ul>

        
    </nav>
    
    <section class="doc-content-wrap">

      

 







<div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
      
    <li>Design Doc: Add MKLDNN Kernel in Fluid Operator</li>
  </ul>
</div>
      
      <div class="wy-nav-content" id="doc-content">
        <div class="rst-content">
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
            
  <div class="section" id="design-doc-add-mkldnn-kernel-in-fluid-operator">
<span id="design-doc-add-mkldnn-kernel-in-fluid-operator"></span><h1>Design Doc: Add MKLDNN Kernel in Fluid Operator<a class="headerlink" href="#design-doc-add-mkldnn-kernel-in-fluid-operator" title="永久链接至标题"></a></h1>
<div class="section" id="principles">
<span id="principles"></span><h2>Principles<a class="headerlink" href="#principles" title="永久链接至标题"></a></h2>
<p>First of all, we should follow some basical principles like:</p>
<ol class="simple">
<li><a class="reference external" href="https://github.com/PaddlePaddle/Paddle/blob/develop/doc/howto/dev/new_op_en.md">How to write a new operator</a>. We are trying to add a new kind of kernel into operators, so basically we should follow this doc.</li>
<li><a class="reference external" href="https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/support_new_device.md">Supporting new Device/Library</a>. Since MKLDNN is a new library to fluid, we should add <code class="docutils literal"><span class="pre">MKLDNNDeviceContext</span></code> and maybe <code class="docutils literal"><span class="pre">mkldnn_helper.h</span></code>, just like <a class="reference external" href="https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/cudnn_helper.h">cudnn_helper.h</a>.</li>
<li><a class="reference external" href="https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/switch_kernel.md">Switch Kernel</a>. Another important point is that we should ensure the data synchronization between different kernel types, which is this <a class="reference external" href="https://github.com/PaddlePaddle/Paddle/issues/6549">topic</a>. So basically we should override <code class="docutils literal"><span class="pre">GetExpectedKernelType</span></code> and <code class="docutils literal"><span class="pre">trans</span></code> functions to support switching kernels.</li>
<li><a class="reference external" href="https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/operator_kernel_type.md">The Keys of Operator Kernel Type</a>. Kernel Type is a pivotal conception which can record the <code class="docutils literal"><span class="pre">Place</span></code>, <code class="docutils literal"><span class="pre">Library</span></code>, <code class="docutils literal"><span class="pre">DataType</span></code> and <code class="docutils literal"><span class="pre">Layout</span></code>.</li>
</ol>
</div>
<div class="section" id="sulution">
<span id="sulution"></span><h2>Sulution<a class="headerlink" href="#sulution" title="永久链接至标题"></a></h2>
<p>In general, there are four parts we should follow to run a MKL-DNN primitive.</p>
<ul class="simple">
<li>Create a primitive descriptor that describe this operator</li>
<li>Create a primitive itself by primitive descriptor and the engine</li>
<li>Create all memory buffers that primitive needed</li>
<li>Launch a stream to execute the primitive created
More details can refer to <a class="reference external" href="http://01org.github.io/mkl-dnn">here</a>.</li>
</ul>
<p>It&#8217;s better to avoid reinitialization of primitives and memory handles in the first three stages in every iteration. So we plan to create a map to record all the <code class="docutils literal"><span class="pre">primitive</span></code> and <code class="docutils literal"><span class="pre">memory</span></code>, which should not take too much memories as discussed <a class="reference external" href="https://github.com/PaddlePaddle/Paddle/issues/6822">here</a>.</p>
<p>It&#8217;s assumed that following three conditions should be satisfied.</p>
<ol class="simple">
<li>there is a unique key for each operator instance. May be the actual name of <code class="docutils literal"><span class="pre">Output</span> <span class="pre">Tensor</span></code>.</li>
<li>the <code class="docutils literal"><span class="pre">Input</span> <span class="pre">Tensor</span></code> inside <code class="docutils literal"><span class="pre">Compute</span></code> function is the one after converted.</li>
<li>we can get the phase(eg. <code class="docutils literal"><span class="pre">is_test</span></code>) inside <code class="docutils literal"><span class="pre">Compute</span></code> function, otherwise we need to expose this attribue to user.</li>
</ol>
<div class="section" id="compute">
<span id="compute"></span><h3>Compute<a class="headerlink" href="#compute" title="永久链接至标题"></a></h3>
<p>The algorithm of <code class="docutils literal"><span class="pre">Compute</span></code> would be described as follow, let&#8217;s take conv like an example.</p>
<div class="highlight-c++"><div class="highlight"><pre><span></span>  <span class="n">PADDLE_ENFORCE</span><span class="p">(</span><span class="n">platform</span><span class="o">::</span><span class="n">is_cpu_place</span><span class="p">(</span><span class="n">ctx</span><span class="p">.</span><span class="n">GetPlace</span><span class="p">()),</span> <span class="s">&quot;It must use CPUPlace.&quot;</span><span class="p">);</span>
  <span class="n">PADDLE_ENFORCE</span><span class="p">(</span><span class="n">platform</span><span class="o">::</span><span class="n">is_mkldnn_library</span><span class="p">(</span><span class="n">ctx</span><span class="p">.</span><span class="n">GetLibrary</span><span class="p">()),</span> <span class="s">&quot;It must use MKLDNN Library.&quot;</span><span class="p">);</span>

  <span class="k">auto</span><span class="o">&amp;</span> <span class="n">dev_ctx</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="k">template</span> <span class="n">device_context</span><span class="o">&lt;</span><span class="n">platform</span><span class="o">::</span><span class="n">MKLDNNDeviceContext</span><span class="o">&gt;</span><span class="p">();</span>

  <span class="c1">// find primitive by unique key from mkldnn context</span>
  <span class="c1">// the op_key should be a unique name of this op instance</span>
  <span class="k">auto</span><span class="o">&amp;</span> <span class="n">p</span> <span class="o">=</span> <span class="n">dev_ctx</span><span class="p">.</span><span class="n">findPrimitive</span><span class="p">(</span><span class="n">op_key</span> <span class="o">+</span> <span class="s">&quot;_fwd&quot;</span><span class="p">);</span>

  <span class="c1">// assuming the input tensor inside this compute function is the one after converted</span>
  <span class="c1">// this point should be guarantee by another mechanism</span>
  <span class="k">auto</span><span class="o">&amp;</span> <span class="n">i</span> <span class="o">=</span> <span class="n">dev_ctx</span><span class="p">.</span><span class="n">findMemory</span><span class="p">(</span><span class="n">op_key</span> <span class="o">+</span> <span class="s">&quot;_input&quot;</span><span class="p">);</span>
  
  <span class="k">if</span> <span class="p">(</span><span class="n">p</span> <span class="o">==</span> <span class="k">nullptr</span> <span class="o">||</span> <span class="n">i</span> <span class="o">==</span> <span class="k">nullptr</span> <span class="o">||</span> <span class="n">inputSizeChanged</span><span class="p">(</span><span class="n">p</span><span class="p">,</span> <span class="n">i</span><span class="p">))</span>  <span class="p">{</span>
    <span class="k">auto</span> <span class="n">fwd_primitive_desc</span> <span class="o">=</span> <span class="n">createPrimitiveDesc</span><span class="p">(</span><span class="n">ctx</span><span class="p">);</span>
    <span class="k">auto</span><span class="o">*</span> <span class="n">input</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Input</span><span class="o">&lt;</span><span class="n">Tensor</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;Input&quot;</span><span class="p">);</span>
    <span class="k">auto</span><span class="o">*</span> <span class="n">filter</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Input</span><span class="o">&lt;</span><span class="n">Tensor</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;Filter&quot;</span><span class="p">);</span>
    <span class="k">auto</span><span class="o">*</span> <span class="n">output</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Output</span><span class="o">&lt;</span><span class="n">Tensor</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;Output&quot;</span><span class="p">);</span>
    <span class="n">shared_ptr</span><span class="o">&lt;</span><span class="n">mkldnn</span><span class="o">::</span><span class="n">memory</span><span class="o">&gt;</span> <span class="n">in</span><span class="p">(</span><span class="k">new</span> <span class="n">mkldnn</span><span class="o">::</span><span class="n">memory</span><span class="p">(</span><span class="n">fwd_primitive_desc</span><span class="o">-&gt;</span><span class="n">src_primitive_desc</span><span class="p">(),</span> <span class="n">input</span><span class="o">-&gt;</span><span class="n">data</span><span class="o">&lt;</span><span class="n">T</span><span class="o">&gt;</span><span class="p">()));</span>
    <span class="n">shared_ptr</span><span class="o">&lt;</span><span class="n">mkldnn</span><span class="o">::</span><span class="n">memory</span><span class="o">&gt;</span> <span class="n">wgt</span><span class="p">(</span><span class="k">new</span> <span class="n">mkldnn</span><span class="o">::</span><span class="n">memory</span><span class="p">(</span><span class="n">fwd_primitive_desc</span><span class="o">-&gt;</span><span class="n">weights_primitive_desc</span><span class="p">(),</span> <span class="n">filter</span><span class="o">-&gt;</span><span class="n">data</span><span class="o">&lt;</span><span class="n">T</span><span class="o">&gt;</span><span class="p">()));</span>
    <span class="n">shared_ptr</span><span class="o">&lt;</span><span class="n">mkldnn</span><span class="o">::</span><span class="n">memory</span><span class="o">&gt;</span> <span class="n">out</span><span class="p">(</span><span class="k">new</span> <span class="n">mkldnn</span><span class="o">::</span><span class="n">memory</span><span class="p">(</span><span class="n">fwd_primitive_desc</span><span class="o">-&gt;</span><span class="n">dst_primitive_desc</span><span class="p">(),</span> <span class="n">output</span><span class="o">-&gt;</span><span class="n">mutable_data</span><span class="o">&lt;</span><span class="n">T</span><span class="o">&gt;</span><span class="p">(</span><span class="n">ctx</span><span class="p">.</span><span class="n">GetPlace</span><span class="p">())));</span>
    <span class="n">shared_ptr</span><span class="o">&lt;</span><span class="n">mkldnn</span><span class="o">::</span><span class="n">conv_fwd</span><span class="o">&gt;</span> <span class="n">fwd_primitive</span><span class="p">(</span><span class="k">new</span> <span class="n">mkldnn</span><span class="o">::</span><span class="n">conv_fwd</span><span class="p">(</span><span class="o">*</span><span class="n">fwd_primitive_desc</span><span class="p">,</span> <span class="o">*</span><span class="n">in</span><span class="p">,</span> <span class="o">*</span><span class="n">wgt</span><span class="p">,</span> <span class="o">*</span><span class="n">out</span><span class="p">));</span>

    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">addMemory</span><span class="p">(</span><span class="n">op_key</span><span class="o">+</span><span class="s">&quot;_input&quot;</span><span class="p">,</span> <span class="n">in</span><span class="p">);</span>
    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">addMemory</span><span class="p">(</span><span class="n">op_key</span><span class="o">+</span><span class="s">&quot;_output&quot;</span><span class="p">,</span> <span class="n">out</span><span class="p">);</span>
    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">addMemory</span><span class="p">(</span><span class="n">op_key</span><span class="o">+</span><span class="s">&quot;_filer&quot;</span><span class="p">,</span> <span class="n">wgt</span><span class="p">);</span>
    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">addPrimitive</span><span class="p">(</span><span class="n">op_key</span><span class="o">+</span><span class="s">&quot;_fwd&quot;</span><span class="p">,</span> <span class="n">fwd_primitive</span><span class="p">);</span>
    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">addPrimitiveDesc</span><span class="p">(</span><span class="n">op_key</span><span class="o">+</span><span class="s">&quot;_fwd_PD&quot;</span><span class="p">,</span> <span class="n">fwd_primitive_desc</span><span class="p">);</span>
  <span class="p">}</span>

  <span class="n">p</span> <span class="o">=</span> <span class="n">dev_ctx</span><span class="p">.</span><span class="n">findPrimitive</span><span class="p">(</span><span class="n">op_key</span> <span class="o">+</span> <span class="s">&quot;_fwd&quot;</span><span class="p">);</span>

  <span class="n">PADDLE_ENFORCE</span><span class="p">(</span><span class="n">p</span><span class="p">,</span> <span class="s">&quot;Should have forward Primitive&quot;</span><span class="p">);</span>
  <span class="n">PADDLE_ENFORCE</span><span class="p">(</span><span class="n">dev_ctx</span><span class="p">.</span><span class="n">findMemory</span><span class="p">(</span><span class="n">op_unique_key</span><span class="o">+</span><span class="s">&quot;_input&quot;</span><span class="p">),</span> <span class="s">&quot;Should have input memory&quot;</span><span class="p">);</span>
  <span class="n">PADDLE_ENFORCE</span><span class="p">(</span><span class="n">dev_ctx</span><span class="p">.</span><span class="n">findMemory</span><span class="p">(</span><span class="n">op_unique_key</span><span class="o">+</span><span class="s">&quot;_output&quot;</span><span class="p">),</span> <span class="s">&quot;Should have output memory&quot;</span><span class="p">);</span>
  <span class="n">PADDLE_ENFORCE</span><span class="p">(</span><span class="n">dev_ctx</span><span class="p">.</span><span class="n">findMemory</span><span class="p">(</span><span class="n">op_unique_key</span><span class="o">+</span><span class="s">&quot;_filter&quot;</span><span class="p">),</span> <span class="s">&quot;Should have filter memory&quot;</span><span class="p">);</span>
  <span class="n">PADDLE_ENFORCE</span><span class="p">(</span><span class="n">dev_ctx</span><span class="p">.</span><span class="n">findPrimitiveDesc</span><span class="p">(</span><span class="n">op_unique_key</span><span class="o">+</span><span class="s">&quot;_fwd_PD&quot;</span><span class="p">),</span> <span class="s">&quot;Should have forward PrimitiveDesc&quot;</span><span class="p">);</span>
  <span class="n">dev_ctx</span><span class="p">.</span><span class="n">submit</span><span class="p">(</span><span class="n">p</span><span class="p">);</span>
  <span class="n">dev_ctx</span><span class="p">.</span><span class="n">execute</span><span class="p">();</span>  <span class="c1">// the convert primitive should have already contained.</span>
</pre></div>
</div>
<p>The <code class="docutils literal"><span class="pre">createPrimitiveDesc</span></code> returns the primitive descripotor of this operator, would be like this:</p>
<div class="highlight-c++"><div class="highlight"><pre><span></span>  <span class="k">auto</span><span class="o">*</span> <span class="n">input</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Input</span><span class="o">&lt;</span><span class="n">Tensor</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;Input&quot;</span><span class="p">);</span>
  <span class="k">auto</span><span class="o">*</span> <span class="n">filter</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Input</span><span class="o">&lt;</span><span class="n">Tensor</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;Filter&quot;</span><span class="p">);</span>
  <span class="k">auto</span><span class="o">*</span> <span class="n">output</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Output</span><span class="o">&lt;</span><span class="n">Tensor</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;Output&quot;</span><span class="p">);</span>
  <span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o">&lt;</span><span class="kt">int</span><span class="o">&gt;</span> <span class="n">strides</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Attr</span><span class="o">&lt;</span><span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o">&lt;</span><span class="kt">int</span><span class="o">&gt;&gt;</span><span class="p">(</span><span class="s">&quot;strides&quot;</span><span class="p">);</span>
  <span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o">&lt;</span><span class="kt">int</span><span class="o">&gt;</span> <span class="n">paddings</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Attr</span><span class="o">&lt;</span><span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o">&lt;</span><span class="kt">int</span><span class="o">&gt;&gt;</span><span class="p">(</span><span class="s">&quot;paddings&quot;</span><span class="p">);</span>
  <span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o">&lt;</span><span class="kt">int</span><span class="o">&gt;</span> <span class="n">dilations</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Attr</span><span class="o">&lt;</span><span class="n">std</span><span class="o">::</span><span class="n">vector</span><span class="o">&lt;</span><span class="kt">int</span><span class="o">&gt;&gt;</span><span class="p">(</span><span class="s">&quot;dilations&quot;</span><span class="p">);</span>
  <span class="kt">int</span> <span class="n">groups</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Attr</span><span class="o">&lt;</span><span class="kt">int</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;groups&quot;</span><span class="p">);</span>
  <span class="n">algorithm</span> <span class="n">algo</span> <span class="o">=</span> <span class="k">static_cast</span><span class="o">&lt;</span><span class="n">algorithm</span><span class="o">&gt;</span><span class="p">(</span><span class="n">ctx</span><span class="p">.</span><span class="n">Attr</span><span class="o">&lt;</span><span class="kt">int</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;convolution_algorithm_option&quot;</span><span class="p">));</span>
  <span class="n">prop_kind</span> <span class="n">pk</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="n">Attr</span><span class="o">&lt;</span><span class="kt">bool</span><span class="o">&gt;</span><span class="p">(</span><span class="s">&quot;is_test&quot;</span><span class="p">)</span> <span class="o">?</span> <span class="n">prop_kind</span><span class="o">::</span><span class="nl">forward_inference</span> <span class="p">:</span> <span class="n">prop_kind</span><span class="o">::</span><span class="n">forward_training</span><span class="p">;</span>
    
  <span class="k">auto</span> <span class="n">fwd_desc</span> <span class="o">=</span> <span class="n">mkldnn</span><span class="o">::</span><span class="n">conv_fwd</span><span class="o">::</span><span class="n">desc</span><span class="p">(</span><span class="cm">/* all the setting above*/</span><span class="p">);</span>
  <span class="n">shared_ptr</span><span class="o">&lt;</span><span class="n">mkldnn</span><span class="o">::</span><span class="n">conv_fwd</span><span class="o">::</span><span class="n">primitive_desc</span><span class="o">&gt;</span> <span class="n">fwd_primitive_desc</span><span class="p">(</span><span class="k">new</span> <span class="n">mkldnn</span><span class="o">::</span><span class="n">conv_fwd</span><span class="o">::</span><span class="n">primitive_desc</span><span class="p">(</span><span class="n">fwd_desc</span><span class="p">,</span> <span class="n">ctx</span><span class="p">.</span><span class="n">getEngine</span><span class="p">()));</span>

  <span class="k">return</span> <span class="n">fwd_primitive_desc</span><span class="p">;</span>
  <span class="p">}</span>
</pre></div>
</div>
</div>
<div class="section" id="mkldnndevicecontext">
<span id="mkldnndevicecontext"></span><h3>MKLDNNDeviceContext<a class="headerlink" href="#mkldnndevicecontext" title="永久链接至标题"></a></h3>
<p><code class="docutils literal"><span class="pre">MKLDNNDeviceContext</span></code>, which is very straightforward, should contain some base information like: <code class="docutils literal"><span class="pre">stream</span></code>, <code class="docutils literal"><span class="pre">engine</span></code> and the map needed.</p>
</div>
<div class="section" id="mkldnn-helper">
<span id="mkldnn-helper"></span><h3>mkldnn_helper<a class="headerlink" href="#mkldnn-helper" title="永久链接至标题"></a></h3>
<p>Some functions would be put in <code class="docutils literal"><span class="pre">paddle/platform/mkldnn_helper.h</span></code>.</p>
<ul class="simple">
<li>create MKLDNN memories</li>
<li>create MKLDNN primitives</li>
<li>error check function</li>
<li>etc</li>
</ul>
</div>
<div class="section" id="kernel-switch">
<span id="kernel-switch"></span><h3>Kernel Switch<a class="headerlink" href="#kernel-switch" title="永久链接至标题"></a></h3>
<p>We should <code class="docutils literal"><span class="pre">reorder</span></code> the different Layout from other device or to other device. <code class="docutils literal"><span class="pre">GetExpectedKernelType</span></code> and <code class="docutils literal"><span class="pre">trans</span></code> functions can help us to implement it.</p>
<p><code class="docutils literal"><span class="pre">GetExpectedKernelType</span></code> should get the context, and this operator can return the best <code class="docutils literal"><span class="pre">KernelType</span></code>.
<code class="docutils literal"><span class="pre">trans</span></code> would be like this:</p>
<div class="highlight-c++"><div class="highlight"><pre><span></span><span class="kt">void</span> <span class="nf">trans</span><span class="p">(</span><span class="n">inputs</span><span class="p">,</span> <span class="n">ctx</span><span class="p">)</span> <span class="k">override</span> <span class="p">{</span>
  <span class="k">if</span> <span class="p">(</span><span class="n">NoNeedTrans</span><span class="p">())</span> <span class="p">{</span>
    <span class="k">return</span><span class="p">;</span>
  <span class="p">}</span>
  <span class="c1">// find reorder primitive by op_key from context</span>
  <span class="k">auto</span><span class="o">&amp;</span> <span class="n">dev_ctx</span> <span class="o">=</span> <span class="n">ctx</span><span class="p">.</span><span class="k">template</span> <span class="n">device_context</span><span class="o">&lt;</span><span class="n">platform</span><span class="o">::</span><span class="n">MKLDNNDeviceContext</span><span class="o">&gt;</span><span class="p">();</span>
  <span class="k">auto</span><span class="o">&amp;</span> <span class="n">p</span> <span class="o">=</span> <span class="n">dev_ctx</span><span class="p">.</span><span class="n">findPrimitive</span><span class="p">(</span><span class="n">op_key</span> <span class="o">+</span> <span class="s">&quot;_reorder_input&quot;</span><span class="p">);</span>
  <span class="k">auto</span><span class="o">&amp;</span> <span class="n">i</span> <span class="o">=</span> <span class="n">dev_ctx</span><span class="p">.</span><span class="n">findMemory</span><span class="p">(</span><span class="n">op_key</span> <span class="o">+</span> <span class="s">&quot;_src_input&quot;</span><span class="p">);</span>

  <span class="k">if</span> <span class="p">(</span><span class="n">p</span> <span class="o">==</span> <span class="k">nullptr</span> <span class="o">||</span> <span class="n">i</span> <span class="o">==</span> <span class="k">nullptr</span> <span class="o">||</span> <span class="n">changeSized</span><span class="p">(</span><span class="n">i</span><span class="p">,</span> <span class="n">input</span><span class="p">))</span> <span class="p">{</span>
    <span class="k">auto</span> <span class="n">prim</span> <span class="o">=</span> <span class="n">createPrimitiveDesc</span><span class="p">(</span><span class="n">ctx</span><span class="p">);</span>
    <span class="k">auto</span> <span class="n">src</span> <span class="o">=</span> <span class="n">createMemory</span><span class="p">(</span><span class="n">memoryDesc</span><span class="p">(</span><span class="n">input</span><span class="o">-&gt;</span><span class="n">dims</span><span class="p">(),</span> <span class="n">actual_layout</span><span class="p">),</span> <span class="n">input</span><span class="o">-&gt;</span><span class="n">data</span><span class="p">);</span>
    <span class="k">auto</span> <span class="n">newbuffer</span> <span class="o">=</span> <span class="n">paddle</span><span class="o">::</span><span class="n">memory</span><span class="o">::</span><span class="n">Alloc</span><span class="p">(</span><span class="n">ctx</span><span class="p">.</span><span class="n">GetPlace</span><span class="p">(),</span> <span class="n">input</span><span class="o">-&gt;</span><span class="n">size_in_bytes</span><span class="p">());</span>
    <span class="k">auto</span> <span class="n">dst</span> <span class="o">=</span> <span class="n">createMemory</span><span class="p">(</span><span class="n">p</span><span class="o">-&gt;</span><span class="n">expected_desc</span><span class="p">(),</span> <span class="n">newbuffer</span><span class="o">-&gt;</span><span class="n">data</span><span class="p">);</span>
    <span class="k">auto</span> <span class="n">reorder_primitive</span><span class="p">(</span><span class="k">new</span> <span class="n">mkldnn</span><span class="o">::</span><span class="n">reorder</span><span class="p">(</span><span class="n">src</span><span class="p">,</span> <span class="n">dst</span><span class="p">));</span>

    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">addMemory</span><span class="p">(</span><span class="n">op_key</span><span class="o">+</span><span class="s">&quot;_src_input&quot;</span><span class="p">,</span> <span class="n">src</span><span class="p">);</span>
    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">addMemory</span><span class="p">(</span><span class="n">op_key</span><span class="o">+</span><span class="s">&quot;_input&quot;</span><span class="p">,</span> <span class="n">dst</span><span class="p">);</span>
    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">addPrimitive</span><span class="p">(</span><span class="n">op_key</span><span class="o">+</span><span class="s">&quot;_reorder_input&quot;</span><span class="p">,</span> <span class="n">reorder_primitive</span><span class="p">);</span>
  <span class="p">}</span>

  <span class="n">p</span> <span class="o">=</span> <span class="n">dev_ctx</span><span class="p">.</span><span class="n">findPrimitive</span><span class="p">(</span><span class="n">op_key</span> <span class="o">+</span> <span class="s">&quot;_reorder_input&quot;</span><span class="p">);</span>
  <span class="n">PADDLE_ENFORCE</span><span class="p">(</span><span class="n">p</span><span class="p">,</span> <span class="s">&quot;Should have Reorder Primitive&quot;</span><span class="p">);</span>
  <span class="n">dev_ctx</span><span class="p">.</span><span class="n">submit</span><span class="p">(</span><span class="n">p</span><span class="p">);</span>
  <span class="k">if</span> <span class="p">(</span><span class="o">!</span> <span class="k">this</span><span class="o">-&gt;</span><span class="n">isMKLDNNKernel</span><span class="p">())</span> <span class="p">{</span>
    <span class="c1">// execute immediately only if this is not mkldnn kernel function.</span>
    <span class="c1">// otherwise, it can be executed with the operator primitive in Compute</span>
    <span class="n">dev_ctx</span><span class="p">.</span><span class="n">stream</span><span class="p">();</span>
  <span class="p">}</span>
  <span class="c1">// after submit, the input tensor in ExecutionContext should be changed as the converted one</span>
  <span class="c1">// there should be another mechanism to ensure this</span>
<span class="p">}</span>
</pre></div>
</div>
</div>
<div class="section" id="unit-test">
<span id="unit-test"></span><h3>Unit Test<a class="headerlink" href="#unit-test" title="永久链接至标题"></a></h3>
<p>All the functions should be tested corresponding.
TBD</p>
</div>
</div>
</div>


           </div>
          </div>
          <footer>
  

  <hr/>

  <div role="contentinfo">
    <p>
        &copy; Copyright 2016, PaddlePaddle developers.

    </p>
  </div>
  Built with <a href="http://sphinx-doc.org/">Sphinx</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>. 

</footer>

        </div>
      </div>

    </section>

  </div>
  


  

    <script type="text/javascript">
        var DOCUMENTATION_OPTIONS = {
            URL_ROOT:'../../',
            VERSION:'',
            COLLAPSE_INDEX:false,
            FILE_SUFFIX:'.html',
            HAS_SOURCE:  true,
            SOURCELINK_SUFFIX: ".txt",
        };
    </script>
      <script type="text/javascript" src="../../_static/jquery.js"></script>
      <script type="text/javascript" src="../../_static/underscore.js"></script>
      <script type="text/javascript" src="../../_static/doctools.js"></script>
      <script type="text/javascript" src="../../_static/translations.js"></script>
      <script type="text/javascript" src="https://cdn.bootcss.com/mathjax/2.7.0/MathJax.js"></script>
       
  

  
  
    <script type="text/javascript" src="../../_static/js/theme.js"></script>
  
  
  <script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/js/bootstrap.min.js" integrity="sha384-Tc5IQib027qvyjSMfHjOMaLkfuWVxZxUPnCJA7l2mCWNIpG9mGCD8wGNIcPD7Txa" crossorigin="anonymous"></script>
  <script src="https://cdn.jsdelivr.net/perfect-scrollbar/0.6.14/js/perfect-scrollbar.jquery.min.js"></script>
  <script src="../../_static/js/paddle_doc_init.js"></script> 

</body>
</html>