mirror of https://github.com/inclusionAI/AReaL
641 lines
26 KiB
HTML
Executable File
641 lines
26 KiB
HTML
Executable File
|
||
<!DOCTYPE html>
|
||
|
||
|
||
<html lang="en" data-content_root="../" >
|
||
|
||
<head>
|
||
<meta charset="utf-8" />
|
||
<meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="viewport" content="width=device-width, initial-scale=1" />
|
||
|
||
<title>Model Worker — AReaL Documentation</title>
|
||
|
||
|
||
|
||
<script data-cfasync="false">
|
||
document.documentElement.dataset.mode = localStorage.getItem("mode") || "";
|
||
document.documentElement.dataset.theme = localStorage.getItem("theme") || "";
|
||
</script>
|
||
|
||
<!-- Loaded before other Sphinx assets -->
|
||
<link href="../_static/styles/theme.css?digest=dfe6caa3a7d634c4db9b" rel="stylesheet" />
|
||
<link href="../_static/styles/bootstrap.css?digest=dfe6caa3a7d634c4db9b" rel="stylesheet" />
|
||
<link href="../_static/styles/pydata-sphinx-theme.css?digest=dfe6caa3a7d634c4db9b" rel="stylesheet" />
|
||
|
||
|
||
<link href="../_static/vendor/fontawesome/6.5.2/css/all.min.css?digest=dfe6caa3a7d634c4db9b" rel="stylesheet" />
|
||
<link rel="preload" as="font" type="font/woff2" crossorigin href="../_static/vendor/fontawesome/6.5.2/webfonts/fa-solid-900.woff2" />
|
||
<link rel="preload" as="font" type="font/woff2" crossorigin href="../_static/vendor/fontawesome/6.5.2/webfonts/fa-brands-400.woff2" />
|
||
<link rel="preload" as="font" type="font/woff2" crossorigin href="../_static/vendor/fontawesome/6.5.2/webfonts/fa-regular-400.woff2" />
|
||
|
||
<link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=fa44fd50" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/styles/sphinx-book-theme.css?v=eba8b062" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/togglebutton.css?v=13237357" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/mystnb.4510f1fc1dee50b3e5859aac5469c37c29e427902b24a333a5f9fcb2f0b3ac41.css?v=be8a1c11" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/sphinx-thebe.css?v=4fa983c6" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/sphinx-design.min.css?v=95c83b7e" />
|
||
|
||
<!-- Pre-loaded scripts that we'll load fully later -->
|
||
<link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=dfe6caa3a7d634c4db9b" />
|
||
<link rel="preload" as="script" href="../_static/scripts/pydata-sphinx-theme.js?digest=dfe6caa3a7d634c4db9b" />
|
||
<script src="../_static/vendor/fontawesome/6.5.2/js/all.min.js?digest=dfe6caa3a7d634c4db9b"></script>
|
||
|
||
<script src="../_static/documentation_options.js?v=9eb32ce0"></script>
|
||
<script src="../_static/doctools.js?v=9a2dae69"></script>
|
||
<script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
|
||
<script src="../_static/clipboard.min.js?v=a7894cd8"></script>
|
||
<script src="../_static/copybutton.js?v=f281be69"></script>
|
||
<script src="../_static/scripts/sphinx-book-theme.js?v=887ef09a"></script>
|
||
<script>let toggleHintShow = 'Click to show';</script>
|
||
<script>let toggleHintHide = 'Click to hide';</script>
|
||
<script>let toggleOpenOnPrint = 'true';</script>
|
||
<script src="../_static/togglebutton.js?v=4a39c7ea"></script>
|
||
<script>var togglebuttonSelector = '.toggle, .admonition.dropdown';</script>
|
||
<script src="../_static/design-tabs.js?v=f930bc37"></script>
|
||
<script>const THEBE_JS_URL = "https://unpkg.com/thebe@0.8.2/lib/index.js"; const thebe_selector = ".thebe,.cell"; const thebe_selector_input = "pre"; const thebe_selector_output = ".output, .cell_output"</script>
|
||
<script async="async" src="../_static/sphinx-thebe.js?v=c100c467"></script>
|
||
<script>var togglebuttonSelector = '.toggle, .admonition.dropdown';</script>
|
||
<script>const THEBE_JS_URL = "https://unpkg.com/thebe@0.8.2/lib/index.js"; const thebe_selector = ".thebe,.cell"; const thebe_selector_input = "pre"; const thebe_selector_output = ".output, .cell_output"</script>
|
||
<script>DOCUMENTATION_OPTIONS.pagename = 'developer/model_worker';</script>
|
||
<link rel="index" title="Index" href="../genindex.html" />
|
||
<link rel="search" title="Search" href="../search.html" />
|
||
<link rel="next" title="Algorithm, Interface & Backends" href="algo_interface.html" />
|
||
<link rel="prev" title="Master Worker" href="master_worker.html" />
|
||
<meta name="viewport" content="width=device-width, initial-scale=1"/>
|
||
<meta name="docsearch:language" content="en"/>
|
||
</head>
|
||
|
||
|
||
<body data-bs-spy="scroll" data-bs-target=".bd-toc-nav" data-offset="180" data-bs-root-margin="0px 0px -60%" data-default-mode="">
|
||
|
||
|
||
|
||
<div id="pst-skip-link" class="skip-link d-print-none"><a href="#main-content">Skip to main content</a></div>
|
||
|
||
<div id="pst-scroll-pixel-helper"></div>
|
||
|
||
<button type="button" class="btn rounded-pill" id="pst-back-to-top">
|
||
<i class="fa-solid fa-arrow-up"></i>Back to top</button>
|
||
|
||
|
||
<input type="checkbox"
|
||
class="sidebar-toggle"
|
||
id="pst-primary-sidebar-checkbox"/>
|
||
<label class="overlay overlay-primary" for="pst-primary-sidebar-checkbox"></label>
|
||
|
||
<input type="checkbox"
|
||
class="sidebar-toggle"
|
||
id="pst-secondary-sidebar-checkbox"/>
|
||
<label class="overlay overlay-secondary" for="pst-secondary-sidebar-checkbox"></label>
|
||
|
||
<div class="search-button__wrapper">
|
||
<div class="search-button__overlay"></div>
|
||
<div class="search-button__search-container">
|
||
<form class="bd-search d-flex align-items-center"
|
||
action="../search.html"
|
||
method="get">
|
||
<i class="fa-solid fa-magnifying-glass"></i>
|
||
<input type="search"
|
||
class="form-control"
|
||
name="q"
|
||
id="search-input"
|
||
placeholder="Search this book..."
|
||
aria-label="Search this book..."
|
||
autocomplete="off"
|
||
autocorrect="off"
|
||
autocapitalize="off"
|
||
spellcheck="false"/>
|
||
<span class="search-button__kbd-shortcut"><kbd class="kbd-shortcut__modifier">Ctrl</kbd>+<kbd>K</kbd></span>
|
||
</form></div>
|
||
</div>
|
||
|
||
<div class="pst-async-banner-revealer d-none">
|
||
<aside id="bd-header-version-warning" class="d-none d-print-none" aria-label="Version warning"></aside>
|
||
</div>
|
||
|
||
|
||
<header class="bd-header navbar navbar-expand-lg bd-navbar d-print-none">
|
||
</header>
|
||
|
||
|
||
<div class="bd-container">
|
||
<div class="bd-container__inner bd-page-width">
|
||
|
||
|
||
|
||
<div class="bd-sidebar-primary bd-sidebar">
|
||
|
||
|
||
|
||
<div class="sidebar-header-items sidebar-primary__section">
|
||
|
||
|
||
|
||
|
||
</div>
|
||
|
||
<div class="sidebar-primary-items__start sidebar-primary__section">
|
||
<div class="sidebar-primary-item">
|
||
|
||
|
||
|
||
|
||
|
||
<a class="navbar-brand logo" href="../intro.html">
|
||
|
||
|
||
|
||
|
||
|
||
|
||
|
||
|
||
|
||
|
||
<img src="../_static/logo.png" class="logo__image only-light" alt="AReaL Documentation - Home"/>
|
||
<script>document.write(`<img src="../_static/logo.png" class="logo__image only-dark" alt="AReaL Documentation - Home"/>`);</script>
|
||
|
||
|
||
</a></div>
|
||
<div class="sidebar-primary-item">
|
||
|
||
<script>
|
||
document.write(`
|
||
<button class="btn search-button-field search-button__button" title="Search" aria-label="Search" data-bs-placement="bottom" data-bs-toggle="tooltip">
|
||
<i class="fa-solid fa-magnifying-glass"></i>
|
||
<span class="search-button__default-text">Search</span>
|
||
<span class="search-button__kbd-shortcut"><kbd class="kbd-shortcut__modifier">Ctrl</kbd>+<kbd class="kbd-shortcut__modifier">K</kbd></span>
|
||
</button>
|
||
`);
|
||
</script></div>
|
||
<div class="sidebar-primary-item"><nav class="bd-links bd-docs-nav" aria-label="Main">
|
||
<div class="bd-toc-item navbar-nav active">
|
||
|
||
<ul class="nav bd-sidenav bd-sidenav__home-link">
|
||
<li class="toctree-l1">
|
||
<a class="reference internal" href="../intro.html">
|
||
Overview
|
||
</a>
|
||
</li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Tutorial</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../training.html">Training</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../eval.html">Evaluation</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../troubleshooting.html">Troubleshooting</a></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Developer Manual</span></p>
|
||
<ul class="current nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="exp_launch.html">Launching Procedure</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="master_worker.html">Master Worker</a></li>
|
||
<li class="toctree-l1 current active"><a class="current reference internal" href="#">Model Worker</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="algo_interface.html">Algorithm, Interface & Backends</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="allocation_parallel.html">Allocation & Parallelism</a></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Contributing</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="../contrib.html">Contribution Guide</a></li>
|
||
</ul>
|
||
|
||
</div>
|
||
</nav></div>
|
||
</div>
|
||
|
||
|
||
<div class="sidebar-primary-items__end sidebar-primary__section">
|
||
</div>
|
||
|
||
<div id="rtd-footer-container"></div>
|
||
|
||
|
||
</div>
|
||
|
||
<main id="main-content" class="bd-main" role="main">
|
||
|
||
|
||
|
||
<div class="sbt-scroll-pixel-helper"></div>
|
||
|
||
<div class="bd-content">
|
||
<div class="bd-article-container">
|
||
|
||
<div class="bd-header-article d-print-none">
|
||
<div class="header-article-items header-article__inner">
|
||
|
||
<div class="header-article-items__start">
|
||
|
||
<div class="header-article-item"><button class="sidebar-toggle primary-toggle btn btn-sm" title="Toggle primary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
|
||
<span class="fa-solid fa-bars"></span>
|
||
</button></div>
|
||
|
||
</div>
|
||
|
||
|
||
<div class="header-article-items__end">
|
||
|
||
<div class="header-article-item">
|
||
|
||
<div class="article-header-buttons">
|
||
|
||
|
||
|
||
|
||
|
||
<div class="dropdown dropdown-source-buttons">
|
||
<button class="btn dropdown-toggle" type="button" data-bs-toggle="dropdown" aria-expanded="false" aria-label="Source repositories">
|
||
<i class="fab fa-github"></i>
|
||
</button>
|
||
<ul class="dropdown-menu">
|
||
|
||
|
||
|
||
<li><a href="https://github.com/inclusionAI/AReaL" target="_blank"
|
||
class="btn btn-sm btn-source-repository-button dropdown-item"
|
||
title="Source repository"
|
||
data-bs-placement="left" data-bs-toggle="tooltip"
|
||
>
|
||
|
||
|
||
<span class="btn__icon-container">
|
||
<i class="fab fa-github"></i>
|
||
</span>
|
||
<span class="btn__text-container">Repository</span>
|
||
</a>
|
||
</li>
|
||
|
||
|
||
|
||
|
||
<li><a href="https://github.com/inclusionAI/AReaL/issues/new?title=Issue%20on%20page%20%2Fdeveloper/model_worker.html&body=Your%20issue%20content%20here." target="_blank"
|
||
class="btn btn-sm btn-source-issues-button dropdown-item"
|
||
title="Open an issue"
|
||
data-bs-placement="left" data-bs-toggle="tooltip"
|
||
>
|
||
|
||
|
||
<span class="btn__icon-container">
|
||
<i class="fas fa-lightbulb"></i>
|
||
</span>
|
||
<span class="btn__text-container">Open issue</span>
|
||
</a>
|
||
</li>
|
||
|
||
</ul>
|
||
</div>
|
||
|
||
|
||
|
||
|
||
|
||
|
||
<div class="dropdown dropdown-download-buttons">
|
||
<button class="btn dropdown-toggle" type="button" data-bs-toggle="dropdown" aria-expanded="false" aria-label="Download this page">
|
||
<i class="fas fa-download"></i>
|
||
</button>
|
||
<ul class="dropdown-menu">
|
||
|
||
|
||
|
||
<li><a href="../_sources/developer/model_worker.md" target="_blank"
|
||
class="btn btn-sm btn-download-source-button dropdown-item"
|
||
title="Download source file"
|
||
data-bs-placement="left" data-bs-toggle="tooltip"
|
||
>
|
||
|
||
|
||
<span class="btn__icon-container">
|
||
<i class="fas fa-file"></i>
|
||
</span>
|
||
<span class="btn__text-container">.md</span>
|
||
</a>
|
||
</li>
|
||
|
||
|
||
|
||
|
||
<li>
|
||
<button onclick="window.print()"
|
||
class="btn btn-sm btn-download-pdf-button dropdown-item"
|
||
title="Print to PDF"
|
||
data-bs-placement="left" data-bs-toggle="tooltip"
|
||
>
|
||
|
||
|
||
<span class="btn__icon-container">
|
||
<i class="fas fa-file-pdf"></i>
|
||
</span>
|
||
<span class="btn__text-container">.pdf</span>
|
||
</button>
|
||
</li>
|
||
|
||
</ul>
|
||
</div>
|
||
|
||
|
||
|
||
|
||
<button onclick="toggleFullScreen()"
|
||
class="btn btn-sm btn-fullscreen-button"
|
||
title="Fullscreen mode"
|
||
data-bs-placement="bottom" data-bs-toggle="tooltip"
|
||
>
|
||
|
||
|
||
<span class="btn__icon-container">
|
||
<i class="fas fa-expand"></i>
|
||
</span>
|
||
|
||
</button>
|
||
|
||
|
||
|
||
<script>
|
||
document.write(`
|
||
<button class="btn btn-sm nav-link pst-navbar-icon theme-switch-button" title="light/dark" aria-label="light/dark" data-bs-placement="bottom" data-bs-toggle="tooltip">
|
||
<i class="theme-switch fa-solid fa-sun fa-lg" data-mode="light"></i>
|
||
<i class="theme-switch fa-solid fa-moon fa-lg" data-mode="dark"></i>
|
||
<i class="theme-switch fa-solid fa-circle-half-stroke fa-lg" data-mode="auto"></i>
|
||
</button>
|
||
`);
|
||
</script>
|
||
|
||
|
||
<script>
|
||
document.write(`
|
||
<button class="btn btn-sm pst-navbar-icon search-button search-button__button" title="Search" aria-label="Search" data-bs-placement="bottom" data-bs-toggle="tooltip">
|
||
<i class="fa-solid fa-magnifying-glass fa-lg"></i>
|
||
</button>
|
||
`);
|
||
</script>
|
||
<button class="sidebar-toggle secondary-toggle btn btn-sm" title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
|
||
<span class="fa-solid fa-list"></span>
|
||
</button>
|
||
</div></div>
|
||
|
||
</div>
|
||
|
||
</div>
|
||
</div>
|
||
|
||
|
||
|
||
<div id="jb-print-docs-body" class="onlyprint">
|
||
<h1>Model Worker</h1>
|
||
<!-- Table of contents -->
|
||
<div id="print-main-content">
|
||
<div id="jb-print-toc">
|
||
|
||
<div>
|
||
<h2> Contents </h2>
|
||
</div>
|
||
<nav aria-label="Page">
|
||
<ul class="visible nav section-nav flex-column">
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#master-model-worker-interaction">Master-Model Worker Interaction</a></li>
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#communication-protocol">Communication Protocol</a><ul class="nav section-nav flex-column">
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#request-reply-pattern">Request-Reply Pattern</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#request-hooks">Request Hooks</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#request-types">Request Types</a></li>
|
||
</ul>
|
||
</li>
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#data-management">Data Management</a><ul class="nav section-nav flex-column">
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#distributed-dataset-storage">Distributed Dataset Storage</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#data-transfer-process">Data Transfer Process</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#mfc-output-handling">MFC Output Handling</a></li>
|
||
</ul>
|
||
</li>
|
||
</ul>
|
||
</nav>
|
||
</div>
|
||
</div>
|
||
</div>
|
||
|
||
|
||
|
||
<div id="searchbox"></div>
|
||
<article class="bd-article">
|
||
|
||
<section class="tex2jax_ignore mathjax_ignore" id="model-worker">
|
||
<h1>Model Worker<a class="headerlink" href="#model-worker" title="Link to this heading">#</a></h1>
|
||
<section id="master-model-worker-interaction">
|
||
<h2>Master-Model Worker Interaction<a class="headerlink" href="#master-model-worker-interaction" title="Link to this heading">#</a></h2>
|
||
<p>The master worker sends remote procedure calls (RPCs) to model workers to execute actual computations like <code class="docutils literal notranslate"><span class="pre">actor_gen</span></code> and <code class="docutils literal notranslate"><span class="pre">actor_train</span></code>. The figure below illustrates their interaction throughout an experiment:</p>
|
||
<p><img alt="" src="../_images/master-model-interaction.png" /></p>
|
||
<p>Model worker “compute” involves running a model interface with a specific backend (covered in detail later). For PPO algorithms, model workers sequentially execute:</p>
|
||
<ul class="simple">
|
||
<li><p><code class="docutils literal notranslate"><span class="pre">actor_gen</span></code>: <code class="docutils literal notranslate"><span class="pre">actor</span></code> model with SGlang backend + <code class="docutils literal notranslate"><span class="pre">PPOActorInterface.generate</span></code></p></li>
|
||
<li><p><code class="docutils literal notranslate"><span class="pre">rew_inf</span></code>: <code class="docutils literal notranslate"><span class="pre">reward</span></code> model (can be null for RLVR) + <code class="docutils literal notranslate"><span class="pre">MultiTaskRewardInterface.inference</span></code></p></li>
|
||
<li><p><code class="docutils literal notranslate"><span class="pre">actor_train</span></code>: <code class="docutils literal notranslate"><span class="pre">actor</span></code> model with Megatron backend + <code class="docutils literal notranslate"><span class="pre">PPOActorInterface.train_step</span></code></p></li>
|
||
</ul>
|
||
</section>
|
||
<section id="communication-protocol">
|
||
<h2>Communication Protocol<a class="headerlink" href="#communication-protocol" title="Link to this heading">#</a></h2>
|
||
<section id="request-reply-pattern">
|
||
<h3>Request-Reply Pattern<a class="headerlink" href="#request-reply-pattern" title="Link to this heading">#</a></h3>
|
||
<p>The master worker and model workers communicate through a <code class="docutils literal notranslate"><span class="pre">request_reply_stream</span></code> channel that handles requests and metadata responses (actual data like <code class="docutils literal notranslate"><span class="pre">input_ids</span></code> transfers through other channels).</p>
|
||
<p>Master (client) can send these requests to model workers (servers):</p>
|
||
<ul class="simple">
|
||
<li><p><strong>fetch</strong>: Worker loads local dataset data and sends metadata (e.g., sequence length) to master for buffer storage</p></li>
|
||
<li><p><strong>spec</strong>: Worker returns dataset specifications for master to calculate experiment steps</p></li>
|
||
<li><p><strong>model_config</strong>: Worker provides transformer model configuration</p></li>
|
||
<li><p><strong>clear_data_cache</strong>: Worker clears data transfer and GPU caches</p></li>
|
||
<li><p><strong>initialize</strong>: Worker initializes parameters, gradient buffers, and optimizer states</p></li>
|
||
<li><p><strong>generate/inference/train_step</strong>: Worker executes corresponding computation (note: “inference” refers to single forward pass)</p></li>
|
||
</ul>
|
||
</section>
|
||
<section id="request-hooks">
|
||
<h3>Request Hooks<a class="headerlink" href="#request-hooks" title="Link to this heading">#</a></h3>
|
||
<p>Computation requests (“generate”/”inference”/”train_step”) support pre- and post-hooks for:</p>
|
||
<ul class="simple">
|
||
<li><p>Data transfer (pre-hook)</p></li>
|
||
<li><p>Evaluation</p></li>
|
||
<li><p>Offloading</p></li>
|
||
<li><p>Parameter reallocation</p></li>
|
||
<li><p>Checkpointing (post-hooks)</p></li>
|
||
</ul>
|
||
<p>These hooks often require NCCL communication/synchronization between workers. Implementing them as dedicated hooks prevents deadlocks that could occur if these operations interleaved with other NCCL communications.</p>
|
||
</section>
|
||
<section id="request-types">
|
||
<h3>Request Types<a class="headerlink" href="#request-types" title="Link to this heading">#</a></h3>
|
||
<ul class="simple">
|
||
<li><p><strong>Blocking requests</strong>: Long-running operations requiring NCCL synchronization. Workers can’t execute immediately since concurrent blocking requests may need coordinated data transfers. Master sends a “flush” request to indicate all concurrent requests have been sent.</p></li>
|
||
<li><p><strong>Non-blocking requests</strong>: Shorter operations without NCCL requirements that can execute immediately.</p></li>
|
||
</ul>
|
||
</section>
|
||
</section>
|
||
<section id="data-management">
|
||
<h2>Data Management<a class="headerlink" href="#data-management" title="Link to this heading">#</a></h2>
|
||
<section id="distributed-dataset-storage">
|
||
<h3>Distributed Dataset Storage<a class="headerlink" href="#distributed-dataset-storage" title="Link to this heading">#</a></h3>
|
||
<p>Datasets distribute across model workers without overlap. For each model:</p>
|
||
<ul class="simple">
|
||
<li><p>Processes with PP rank = -1 and TP rank = 0 serve as DP heads</p></li>
|
||
<li><p>Data stores on DP heads of the model used in the first MFC (e.g., actor model DP heads for PPO)</p></li>
|
||
</ul>
|
||
<p>During “fetch” requests:</p>
|
||
<ol class="arabic simple">
|
||
<li><p>DP head worker loads data into local buffer</p></li>
|
||
<li><p>Sends metadata to master</p></li>
|
||
<li><p>Master tracks metadata and later instructs workers which data to use for each MFC via computation request hooks</p></li>
|
||
</ol>
|
||
</section>
|
||
<section id="data-transfer-process">
|
||
<h3>Data Transfer Process<a class="headerlink" href="#data-transfer-process" title="Link to this heading">#</a></h3>
|
||
<p>For each MFC, the master:</p>
|
||
<ol class="arabic simple">
|
||
<li><p>Specifies which data to use</p></li>
|
||
<li><p>Provides data locations across workers</p></li>
|
||
<li><p>Workers redistribute data using:</p>
|
||
<ul class="simple">
|
||
<li><p><code class="docutils literal notranslate"><span class="pre">Redistributor</span></code>: Generates NCCL broadcast/gather/scatter communication plan</p></li>
|
||
<li><p><code class="docutils literal notranslate"><span class="pre">DataManager</span></code>: Executes the plan</p></li>
|
||
</ul>
|
||
</li>
|
||
</ol>
|
||
<p>After redistribution, workers with same DP rank receive identical input data.</p>
|
||
</section>
|
||
<section id="mfc-output-handling">
|
||
<h3>MFC Output Handling<a class="headerlink" href="#mfc-output-handling" title="Link to this heading">#</a></h3>
|
||
<p>Only workers with PP rank=-1 and TP rank=0 produce output data. These workers:</p>
|
||
<ol class="arabic simple">
|
||
<li><p>Store data locally</p></li>
|
||
<li><p>Notify master of data locations</p></li>
|
||
<li><p>Master generates new redistribution plans for subsequent MFCs based on this layout information</p></li>
|
||
</ol>
|
||
</section>
|
||
</section>
|
||
</section>
|
||
|
||
<script type="text/x-thebe-config">
|
||
{
|
||
requestKernel: true,
|
||
binderOptions: {
|
||
repo: "binder-examples/jupyter-stacks-datascience",
|
||
ref: "master",
|
||
},
|
||
codeMirrorConfig: {
|
||
theme: "abcdef",
|
||
mode: "python"
|
||
},
|
||
kernelOptions: {
|
||
name: "python3",
|
||
path: "./developer"
|
||
},
|
||
predefinedOutput: true
|
||
}
|
||
</script>
|
||
<script>kernelName = 'python3'</script>
|
||
|
||
</article>
|
||
|
||
|
||
|
||
|
||
|
||
|
||
<footer class="prev-next-footer d-print-none">
|
||
|
||
<div class="prev-next-area">
|
||
<a class="left-prev"
|
||
href="master_worker.html"
|
||
title="previous page">
|
||
<i class="fa-solid fa-angle-left"></i>
|
||
<div class="prev-next-info">
|
||
<p class="prev-next-subtitle">previous</p>
|
||
<p class="prev-next-title">Master Worker</p>
|
||
</div>
|
||
</a>
|
||
<a class="right-next"
|
||
href="algo_interface.html"
|
||
title="next page">
|
||
<div class="prev-next-info">
|
||
<p class="prev-next-subtitle">next</p>
|
||
<p class="prev-next-title">Algorithm, Interface & Backends</p>
|
||
</div>
|
||
<i class="fa-solid fa-angle-right"></i>
|
||
</a>
|
||
</div>
|
||
</footer>
|
||
|
||
</div>
|
||
|
||
|
||
|
||
<div class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
|
||
|
||
|
||
<div class="sidebar-secondary-item">
|
||
<div class="page-toc tocsection onthispage">
|
||
<i class="fa-solid fa-list"></i> Contents
|
||
</div>
|
||
<nav class="bd-toc-nav page-toc">
|
||
<ul class="visible nav section-nav flex-column">
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#master-model-worker-interaction">Master-Model Worker Interaction</a></li>
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#communication-protocol">Communication Protocol</a><ul class="nav section-nav flex-column">
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#request-reply-pattern">Request-Reply Pattern</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#request-hooks">Request Hooks</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#request-types">Request Types</a></li>
|
||
</ul>
|
||
</li>
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#data-management">Data Management</a><ul class="nav section-nav flex-column">
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#distributed-dataset-storage">Distributed Dataset Storage</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#data-transfer-process">Data Transfer Process</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#mfc-output-handling">MFC Output Handling</a></li>
|
||
</ul>
|
||
</li>
|
||
</ul>
|
||
</nav></div>
|
||
|
||
</div></div>
|
||
|
||
|
||
</div>
|
||
<footer class="bd-footer-content">
|
||
|
||
<div class="bd-footer-content__inner container">
|
||
|
||
<div class="footer-item">
|
||
|
||
<p class="component-author">
|
||
By Wei Fu
|
||
</p>
|
||
|
||
</div>
|
||
|
||
<div class="footer-item">
|
||
|
||
|
||
<p class="copyright">
|
||
|
||
© Copyright 2023.
|
||
<br/>
|
||
|
||
</p>
|
||
|
||
</div>
|
||
|
||
<div class="footer-item">
|
||
|
||
</div>
|
||
|
||
<div class="footer-item">
|
||
|
||
</div>
|
||
|
||
</div>
|
||
</footer>
|
||
|
||
|
||
</main>
|
||
</div>
|
||
</div>
|
||
|
||
<!-- Scripts loaded after <body> so the DOM is not blocked -->
|
||
<script src="../_static/scripts/bootstrap.js?digest=dfe6caa3a7d634c4db9b"></script>
|
||
<script src="../_static/scripts/pydata-sphinx-theme.js?digest=dfe6caa3a7d634c4db9b"></script>
|
||
|
||
<footer class="bd-footer">
|
||
</footer>
|
||
</body>
|
||
</html> |