fluid_compiler.html



<!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>PaddlePaddle Fluid: Towards a Compiled Programming Language &mdash; PaddlePaddle  documentation</title>
  

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

        <link rel="index" title="Index"
              href="../genindex.html"/>
        <link rel="search" title="Search" href="../search.html"/>
    <link rel="top" title="PaddlePaddle  documentation" href="../index.html"/> 
<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">

  <div class="wy-grid-for-nav">

    
    <nav data-toggle="wy-nav-shift" class="wy-nav-side">
      <div class="wy-side-scroll">
        <div class="wy-side-nav-search">
          

            <a href="../index_en.html" class="icon icon-home"> PaddlePaddle
          

          </a>

          
<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 class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
          
<nav class="doc-menu-vertical" role="navigation">

<ul>
<li class="toctree-l1"><a class="reference internal" href="../getstarted/index_en.html">GET STARTED</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../getstarted/quickstart_en.html">Quick Start</a></li>
<li class="toctree-l2"><a class="reference internal" href="../getstarted/concepts/use_concepts_en.html">Basic Concept</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../build_and_install/index_en.html">Install and Build</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../build_and_install/pip_install_en.html">Install using pip</a></li>
<li class="toctree-l2"><a class="reference internal" href="../build_and_install/docker_install_en.html">Run in Docker Containers</a></li>
<li class="toctree-l2"><a class="reference internal" href="../build_and_install/build_from_source_en.html">Build from Sources</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../howto/index_en.html">HOW TO</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../howto/cmd_parameter/index_en.html">Set Command-line Parameters</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../howto/cmd_parameter/use_case_en.html">Use Case</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/cmd_parameter/arguments_en.html">Argument Outline</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/cmd_parameter/detail_introduction_en.html">Detail Description</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../howto/cluster/index_en.html">Distributed Training</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../howto/cluster/preparations_en.html">Preparations</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/cluster/cmd_argument_en.html">Command-line arguments</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/cluster/multi_cluster/index_en.html">Use different clusters</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../howto/cluster/multi_cluster/fabric_en.html">Fabric</a></li>
<li class="toctree-l4"><a class="reference internal" href="../howto/cluster/multi_cluster/openmpi_en.html">OpenMPI</a></li>
<li class="toctree-l4"><a class="reference internal" href="../howto/cluster/multi_cluster/k8s_en.html">Kubernetes</a></li>
<li class="toctree-l4"><a class="reference internal" href="../howto/cluster/multi_cluster/k8s_aws_en.html">Kubernetes on AWS</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../howto/rnn/index_en.html">RNN Models</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../howto/rnn/rnn_config_en.html">RNN Configuration</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/rnn/recurrent_group_en.html">Recurrent Group Tutorial</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/rnn/hierarchical_layer_en.html">Layers supporting hierarchical sequence as input</a></li>
<li class="toctree-l3"><a class="reference internal" href="../howto/rnn/hrnn_rnn_api_compare_en.html">API comparision between RNN and hierarchical RNN</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../howto/optimization/gpu_profiling_en.html">Tune GPU Performance</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../dev/index_en.html">Development</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../dev/contribute_to_paddle_en.html">Contribute Code</a></li>
<li class="toctree-l2"><a class="reference internal" href="../dev/write_docs_en.html">Contribute Documentation</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="../faq/index_en.html">FAQ</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../faq/build_and_install/index_en.html">Install, Build and Unit test</a></li>
<li class="toctree-l2"><a class="reference internal" href="../faq/model/index_en.html">Model Configuration</a></li>
<li class="toctree-l2"><a class="reference internal" href="../faq/parameter/index_en.html">Parameter Setting</a></li>
<li class="toctree-l2"><a class="reference internal" href="../faq/local/index_en.html">Local Training and Prediction</a></li>
<li class="toctree-l2"><a class="reference internal" href="../faq/cluster/index_en.html">Cluster Training and Prediction</a></li>
</ul>
</li>
</ul>

</nav>

        </div>
      </div>
    </nav>

    <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">

      
      <nav class="wy-nav-top" role="navigation" aria-label="top navigation">
        <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
        <a href="../index_en.html">PaddlePaddle</a>
      </nav>


      <div class="wy-nav-content">
        <div class="rst-content">
          

<div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
    <li><a href="../index_en.html">Docs</a> &raquo;</li>
      
    <li>PaddlePaddle Fluid: Towards a Compiled Programming Language</li>
      <li class="wy-breadcrumbs-aside">
        
          
            <a href="../_sources/design/fluid_compiler.md.txt" rel="nofollow"> View page source</a>
          
        
      </li>
  </ul>
  <hr/>
</div>
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
            
  <div class="section" id="paddlepaddle-fluid-towards-a-compiled-programming-language">
<span id="paddlepaddle-fluid-towards-a-compiled-programming-language"></span><h1>PaddlePaddle Fluid: Towards a Compiled Programming Language<a class="headerlink" href="#paddlepaddle-fluid-towards-a-compiled-programming-language" title="Permalink to this headline">¶</a></h1>
<p>As described in <a class="reference internal" href="fluid.html"><span class="doc">fluid.md</span></a>, when a Fluid application program
runs, it generates a <code class="docutils literal"><span class="pre">ProgramDesc</span></code> protobuf message as an intermediate
representation of itself.  The C++ class <code class="docutils literal"><span class="pre">Executor</span></code> can run this
protobuf message as an interpreter.  This article describes the Fluid
compiler.</p>
<p><img alt="" src="../_images/fluid-compiler.png" /></p>
<div class="section" id="programdesc">
<span id="programdesc"></span><h2>ProgramDesc<a class="headerlink" href="#programdesc" title="Permalink to this headline">¶</a></h2>
<p>Before we go deeper into the idea of compiled language, let us take a
look at a simple example Fluid application.</p>
<div class="highlight-python"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="s2">&quot;fluid&quot;</span>

<span class="n">func</span> <span class="n">paddlepaddle</span><span class="p">()</span> <span class="p">{</span>
  <span class="n">X</span> <span class="o">=</span> <span class="n">fluid</span><span class="o">.</span><span class="n">read</span><span class="p">(</span><span class="o">...</span><span class="p">)</span>
  <span class="n">W</span> <span class="o">=</span> <span class="n">fluid</span><span class="o">.</span><span class="n">Tensor</span><span class="p">(</span><span class="o">...</span><span class="p">)</span>
  <span class="n">Y</span> <span class="o">=</span> <span class="n">fluid</span><span class="o">.</span><span class="n">mult</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">W</span><span class="p">)</span>
<span class="p">}</span>
</pre></div>
</div>
<p>This program consists of a <a class="reference internal" href="block.html"><span class="doc">block</span></a> of three operators &#8211;
<code class="docutils literal"><span class="pre">read</span></code>, <code class="docutils literal"><span class="pre">assign</span></code>, and <code class="docutils literal"><span class="pre">mult</span></code>.  Its <code class="docutils literal"><span class="pre">ProgramDesc</span></code> message looks like
the following</p>
<div class="highlight-protobuf"><div class="highlight"><pre><span></span><span class="kd">message</span> <span class="nc">ProgramDesc</span> <span class="p">{</span>
  <span class="n">block</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">Block</span> <span class="p">{</span>
    <span class="na">vars</span> <span class="o">=</span> <span class="p">[</span><span class="n">X</span><span class="p">,</span> <span class="n">W</span><span class="p">,</span> <span class="n">Y</span><span class="p">],</span>
    <span class="na">ops</span> <span class="o">=</span> <span class="p">[</span>
      <span class="n">read</span><span class="p">(</span><span class="na">output</span> <span class="o">=</span> <span class="n">X</span><span class="p">)</span>
      <span class="n">assign</span><span class="p">(</span><span class="na">input</span> <span class="o">=</span> <span class="o">...</span><span class="p">,</span> <span class="na">output</span> <span class="o">=</span> <span class="n">W</span><span class="p">)</span>
      <span class="n">mult</span><span class="p">(</span><span class="na">input</span> <span class="o">=</span> <span class="p">{</span><span class="n">X</span><span class="p">,</span> <span class="n">W</span><span class="p">},</span> <span class="na">output</span> <span class="o">=</span> <span class="n">Y</span><span class="p">)</span>
    <span class="p">],</span>
  <span class="p">}</span>
<span class="p">}</span>
</pre></div>
</div>
</div>
<div class="section" id="transpilers">
<span id="transpilers"></span><h2>Transpilers<a class="headerlink" href="#transpilers" title="Permalink to this headline">¶</a></h2>
<p>We can write a transpiler program that takes a <code class="docutils literal"><span class="pre">ProgramDesc</span></code>, e.g.,
the above one, and outputs another <code class="docutils literal"><span class="pre">ProgramDesc</span></code>.  Let us take some
examples:</p>
<ol class="simple">
<li><em>Memory optimization transpiler</em>: We can write a transpiler that
inserts some <code class="docutils literal"><span class="pre">FreeMemoryOp</span></code>s in the above example <code class="docutils literal"><span class="pre">ProgramDesc</span></code> so
to free memory early, before the end of an iteration, so to keep a
small memory footprint.</li>
<li><em>Distributed training transpiler</em>: We can write a transpiler that
converts a<code class="docutils literal"><span class="pre">ProgramDesc</span></code> into its distributed version of two
<code class="docutils literal"><span class="pre">ProgramDesc</span></code>s &#8211; one for running by the trainer processes and the
other for the parameter server.</li>
</ol>
<p>In the rest of this article, we talk about a special kind of
transpiler, <em>Native code generator</em>, which takes a <code class="docutils literal"><span class="pre">ProgramDesc</span></code> and
generates a <code class="docutils literal"><span class="pre">.cu</span></code> (or <code class="docutils literal"><span class="pre">.cc</span></code>) file, which could be built by C++
compilers (gcc, nvcc, icc) into binaries.</p>
</div>
<div class="section" id="native-code-generator">
<span id="native-code-generator"></span><h2>Native Code Generator<a class="headerlink" href="#native-code-generator" title="Permalink to this headline">¶</a></h2>
<p>For the above example, the native code generator transpiler, say, the
CUDA code generator, should generate a <code class="docutils literal"><span class="pre">main</span></code> function:</p>
<div class="highlight-c++"><div class="highlight"><pre><span></span><span class="kt">void</span> <span class="nf">main</span><span class="p">()</span> <span class="p">{</span>
  <span class="k">auto</span> <span class="n">X</span> <span class="o">=</span> <span class="n">fluid_cuda_read</span><span class="p">(...);</span>
  <span class="k">auto</span> <span class="n">W</span> <span class="o">=</span> <span class="n">fluid_cuda_create_tensor</span><span class="p">(...);</span>
  <span class="k">auto</span> <span class="n">Y</span> <span class="o">=</span> <span class="n">fluid_cuda_mult</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">W</span><span class="p">);</span>
<span class="p">}</span>
</pre></div>
</div>
<p>and the definitions of functions <code class="docutils literal"><span class="pre">fluid_cuda_read</span></code>,
<code class="docutils literal"><span class="pre">fluid_cuda_create_tensor</span></code>, and <code class="docutils literal"><span class="pre">fluid_cuda_mult</span></code>.  Please be aware
that each function could just define a C++ instance of an operator and
run it.  For example</p>
<div class="highlight-c++"><div class="highlight"><pre><span></span><span class="n">paddle</span><span class="o">::</span><span class="n">Tensor</span> <span class="n">fluid_cuda_read</span><span class="p">(...)</span> <span class="p">{</span>
  <span class="n">paddle</span><span class="o">::</span><span class="n">Tensor</span> <span class="n">t</span><span class="p">;</span>
  <span class="n">paddle</span><span class="o">::</span><span class="k">operator</span><span class="o">::</span><span class="n">Read</span> <span class="n">r</span><span class="p">(</span><span class="o">&amp;</span><span class="n">t</span><span class="p">,</span> <span class="p">...);</span>
  <span class="n">r</span><span class="p">.</span><span class="n">Run</span><span class="p">();</span>
  <span class="k">return</span> <span class="n">t</span><span class="p">;</span>
<span class="p">}</span>
</pre></div>
</div>
<p>For computational operators that have multiple <em>kernels</em>, each for a
specific hardware platform, for example, the <code class="docutils literal"><span class="pre">mult</span></code> operator, the
generated code should call its CUDA kernel:</p>
<div class="highlight-c++"><div class="highlight"><pre><span></span><span class="n">paddle</span><span class="o">::</span><span class="n">Tensor</span> <span class="n">fluid_cuda_mult</span><span class="p">(</span><span class="k">const</span> <span class="n">paddle</span><span class="o">::</span><span class="n">Tensor</span><span class="o">&amp;</span> <span class="n">a</span><span class="p">,</span> 
                               <span class="k">const</span> <span class="n">paddle</span><span class="o">::</span><span class="n">Tensor</span><span class="o">&amp;</span> <span class="n">b</span><span class="p">)</span> <span class="p">{</span>
  <span class="n">paddle</span><span class="o">::</span><span class="n">Tensor</span> <span class="n">t</span><span class="p">;</span>
  <span class="n">paddle</span><span class="o">::</span><span class="k">operator</span><span class="o">::</span><span class="n">Mult</span> <span class="n">m</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="p">...);</span>
  <span class="n">Mult</span><span class="p">.</span><span class="n">Run</span><span class="p">(</span><span class="n">cuda_context</span><span class="p">);</span>
<span class="p">}</span>
</pre></div>
</div>
<p>where <code class="docutils literal"><span class="pre">cuda_context</span></code> could be a global variable of type
<code class="docutils literal"><span class="pre">paddle::CUDADeviceContext</span></code>.</p>
</div>
<div class="section" id="multi-block-code-generation">
<span id="multi-block-code-generation"></span><h2>Multi-Block Code Generation<a class="headerlink" href="#multi-block-code-generation" title="Permalink to this headline">¶</a></h2>
<p>Most Fluid application programs may have more than one blocks.  To
execute them, we need to trace <a class="reference internal" href="scope.html"><span class="doc">scopes</span></a>.</p>
</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
        };
    </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="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>

  
    <script type="text/javascript" src="../_static/js/theme.js"></script>
  

  <script type="text/javascript">
      jQuery(function () {
          SphinxRtdTheme.StickyNav.enable();
      });
  </script>
   

</body>
</html>