mirror of
https://github.com/NVIDIA/TensorRT-LLM.git
synced 2026-01-14 06:27:45 +08:00
1485 lines
76 KiB
HTML
1485 lines
76 KiB
HTML
|
|
|
|
<!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>Overview — TensorRT-LLM</title>
|
|
|
|
|
|
|
|
<script data-cfasync="false">
|
|
document.documentElement.dataset.mode = localStorage.getItem("mode") || "";
|
|
document.documentElement.dataset.theme = localStorage.getItem("theme") || "";
|
|
</script>
|
|
<!--
|
|
this give us a css class that will be invisible only if js is disabled
|
|
-->
|
|
<noscript>
|
|
<style>
|
|
.pst-js-only { display: none !important; }
|
|
|
|
</style>
|
|
</noscript>
|
|
|
|
<!-- Loaded before other Sphinx assets -->
|
|
<link href="../_static/styles/theme.css?digest=8878045cc6db502f8baf" rel="stylesheet" />
|
|
<link href="../_static/styles/pydata-sphinx-theme.css?digest=8878045cc6db502f8baf" rel="stylesheet" />
|
|
|
|
<link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=8f2a1f02" />
|
|
<link rel="stylesheet" type="text/css" href="../_static/styles/nvidia-sphinx-theme.css?v=df3ac72c" />
|
|
<link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
|
|
<link rel="stylesheet" type="text/css" href="../_static/autodoc_pydantic.css" />
|
|
|
|
<!-- So that users can add custom icons -->
|
|
<script src="../_static/scripts/fontawesome.js?digest=8878045cc6db502f8baf"></script>
|
|
<!-- Pre-loaded scripts that we'll load fully later -->
|
|
<link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=8878045cc6db502f8baf" />
|
|
<link rel="preload" as="script" href="../_static/scripts/pydata-sphinx-theme.js?digest=8878045cc6db502f8baf" />
|
|
|
|
<script src="../_static/documentation_options.js?v=5929fcd5"></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=65e89d2a"></script>
|
|
<script>DOCUMENTATION_OPTIONS.pagename = 'performance/perf-overview';</script>
|
|
<script>
|
|
DOCUMENTATION_OPTIONS.theme_version = '0.16.1';
|
|
DOCUMENTATION_OPTIONS.theme_switcher_json_url = './_static/switcher.json';
|
|
DOCUMENTATION_OPTIONS.theme_switcher_version_match = '0.21.0rc2';
|
|
DOCUMENTATION_OPTIONS.show_version_warning_banner =
|
|
false;
|
|
</script>
|
|
<link rel="icon" href="../_static/favicon.png"/>
|
|
<link rel="index" title="Index" href="../genindex.html" />
|
|
<link rel="search" title="Search" href="../search.html" />
|
|
<link rel="next" title="TensorRT-LLM Benchmarking" href="perf-benchmarking.html" />
|
|
<link rel="prev" title="Disaggregated-Service (experimental)" href="../advanced/disaggregated-service.html" />
|
|
|
|
<meta name="viewport" content="width=device-width, initial-scale=1"/>
|
|
<meta name="docsearch:language" content="en"/>
|
|
<meta name="docsearch:version" content="0.21.0rc2" />
|
|
|
|
|
|
</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>
|
|
|
|
|
|
<dialog id="pst-search-dialog">
|
|
|
|
<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"
|
|
placeholder="Search the docs ..."
|
|
aria-label="Search the docs ..."
|
|
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>
|
|
</dialog>
|
|
|
|
<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">
|
|
<div class="bd-header__inner bd-page-width">
|
|
<button class="pst-navbar-icon sidebar-toggle primary-toggle" aria-label="Site navigation">
|
|
<span class="fa-solid fa-bars"></span>
|
|
</button>
|
|
|
|
|
|
<div class="col-lg-3 navbar-header-items__start">
|
|
|
|
<div class="navbar-item">
|
|
|
|
|
|
|
|
|
|
|
|
<a class="navbar-brand logo" href="../index.html">
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<img src="../_static/nvidia-logo-horiz-rgb-blk-for-screen.svg" class="logo__image only-light" alt="TensorRT-LLM - Home"/>
|
|
<img src="../_static/nvidia-logo-horiz-rgb-wht-for-screen.svg" class="logo__image only-dark pst-js-only" alt="TensorRT-LLM - Home"/>
|
|
|
|
|
|
<p class="title logo__title">TensorRT-LLM</p>
|
|
|
|
</a></div>
|
|
|
|
</div>
|
|
|
|
<div class="col-lg-9 navbar-header-items">
|
|
|
|
<div class="me-auto navbar-header-items__center">
|
|
|
|
<div class="navbar-item">
|
|
|
|
|
|
<div class="version-switcher__container dropdown pst-js-only">
|
|
<button id="pst-version-switcher-button-2"
|
|
type="button"
|
|
class="version-switcher__button btn btn-sm dropdown-toggle"
|
|
data-bs-toggle="dropdown"
|
|
aria-haspopup="listbox"
|
|
aria-controls="pst-version-switcher-list-2"
|
|
aria-label="Version switcher list"
|
|
>
|
|
Choose version <!-- this text may get changed later by javascript -->
|
|
<span class="caret"></span>
|
|
</button>
|
|
<div id="pst-version-switcher-list-2"
|
|
class="version-switcher__menu dropdown-menu list-group-flush py-0"
|
|
role="listbox" aria-labelledby="pst-version-switcher-button-2">
|
|
<!-- dropdown will be populated by javascript on page load -->
|
|
</div>
|
|
</div></div>
|
|
|
|
</div>
|
|
|
|
|
|
<div class="navbar-header-items__end">
|
|
|
|
<div class="navbar-item navbar-persistent--container">
|
|
|
|
|
|
<button class="btn search-button-field search-button__button pst-js-only" 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>
|
|
</div>
|
|
|
|
|
|
<div class="navbar-item">
|
|
|
|
<button class="btn btn-sm nav-link pst-navbar-icon theme-switch-button pst-js-only" aria-label="Color mode" data-bs-title="Color mode" data-bs-placement="bottom" data-bs-toggle="tooltip">
|
|
<i class="theme-switch fa-solid fa-sun fa-lg" data-mode="light" title="Light"></i>
|
|
<i class="theme-switch fa-solid fa-moon fa-lg" data-mode="dark" title="Dark"></i>
|
|
<i class="theme-switch fa-solid fa-circle-half-stroke fa-lg" data-mode="auto" title="System Settings"></i>
|
|
</button></div>
|
|
|
|
</div>
|
|
|
|
</div>
|
|
|
|
|
|
<div class="navbar-persistent--mobile">
|
|
|
|
<button class="btn search-button-field search-button__button pst-js-only" 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>
|
|
</div>
|
|
|
|
|
|
|
|
<button class="pst-navbar-icon sidebar-toggle secondary-toggle" aria-label="On this page">
|
|
<span class="fa-solid fa-outdent"></span>
|
|
</button>
|
|
|
|
</div>
|
|
|
|
</header>
|
|
|
|
|
|
<div class="bd-container">
|
|
<div class="bd-container__inner bd-page-width">
|
|
|
|
|
|
|
|
<dialog id="pst-primary-sidebar-modal"></dialog>
|
|
<div id="pst-primary-sidebar" class="bd-sidebar-primary bd-sidebar">
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<a class="navbar-brand logo" href="../index.html">
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<img src="../_static/nvidia-logo-horiz-rgb-blk-for-screen.svg" class="logo__image only-light" alt="TensorRT-LLM - Home"/>
|
|
<img src="../_static/nvidia-logo-horiz-rgb-wht-for-screen.svg" class="logo__image only-dark pst-js-only" alt="TensorRT-LLM - Home"/>
|
|
|
|
|
|
<p class="title logo__title">TensorRT-LLM</p>
|
|
|
|
</a>
|
|
|
|
|
|
|
|
<div class="sidebar-header-items sidebar-primary__section">
|
|
|
|
|
|
<div class="sidebar-header-items__center">
|
|
|
|
|
|
|
|
<div class="navbar-item">
|
|
|
|
|
|
<div class="version-switcher__container dropdown pst-js-only">
|
|
<button id="pst-version-switcher-button-3"
|
|
type="button"
|
|
class="version-switcher__button btn btn-sm dropdown-toggle"
|
|
data-bs-toggle="dropdown"
|
|
aria-haspopup="listbox"
|
|
aria-controls="pst-version-switcher-list-3"
|
|
aria-label="Version switcher list"
|
|
>
|
|
Choose version <!-- this text may get changed later by javascript -->
|
|
<span class="caret"></span>
|
|
</button>
|
|
<div id="pst-version-switcher-list-3"
|
|
class="version-switcher__menu dropdown-menu list-group-flush py-0"
|
|
role="listbox" aria-labelledby="pst-version-switcher-button-3">
|
|
<!-- dropdown will be populated by javascript on page load -->
|
|
</div>
|
|
</div></div>
|
|
|
|
|
|
</div>
|
|
|
|
|
|
|
|
<div class="sidebar-header-items__end">
|
|
|
|
<div class="navbar-item">
|
|
|
|
<button class="btn btn-sm nav-link pst-navbar-icon theme-switch-button pst-js-only" aria-label="Color mode" data-bs-title="Color mode" data-bs-placement="bottom" data-bs-toggle="tooltip">
|
|
<i class="theme-switch fa-solid fa-sun fa-lg" data-mode="light" title="Light"></i>
|
|
<i class="theme-switch fa-solid fa-moon fa-lg" data-mode="dark" title="Dark"></i>
|
|
<i class="theme-switch fa-solid fa-circle-half-stroke fa-lg" data-mode="auto" title="System Settings"></i>
|
|
</button></div>
|
|
|
|
</div>
|
|
|
|
</div>
|
|
|
|
<div class="sidebar-primary-items__start sidebar-primary__section">
|
|
<div class="sidebar-primary-item">
|
|
|
|
|
|
|
|
<nav class="bd-docs-nav bd-links"
|
|
aria-label="Table of Contents">
|
|
<p class="bd-links__title" role="heading" aria-level="1">Table of Contents</p>
|
|
<div class="bd-toc-item navbar-nav"><p aria-level="2" class="caption" role="heading"><span class="caption-text">Getting Started</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../overview.html">Overview</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../quick-start-guide.html">Quick Start Guide</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../key-features.html">Key Features</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../torch.html">PyTorch Backend</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../release-notes.html">Release Notes</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Installation</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../installation/linux.html">Installing on Linux</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../installation/build-from-source-linux.html">Building from Source Code on Linux</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../installation/grace-hopper.html">Installing on Grace Hopper</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">LLM API</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../llm-api/index.html">API Introduction</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../llm-api/reference.html">API Reference</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Examples</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1 has-children"><a class="reference internal" href="../examples/index.html">LLM Examples Introduction</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_medusa_decoding.html">Generate Text Using Medusa Decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_multilora.html">Generate text with multiple LoRA adapters</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_eagle_decoding.html">Generate Text Using Eagle Decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_async.html">Generate Text Asynchronously</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_distributed.html">Distributed LLM Generation</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_logits_processor.html">Control generated text using logits processor</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_eagle2_decoding.html">Generate Text Using Eagle2 Decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_kv_events.html">Get KV Cache Events</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_lookahead_decoding.html">Generate Text Using Lookahead Decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_quantization.html">Generation with Quantization</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_async_streaming.html">Generate Text in Streaming</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_guided_decoding.html">Generate text with guided decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference.html">Generate text</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_customize.html">Generate text with customization</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_auto_parallel.html">Automatic Parallelism with LLM</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_llm_distributed.html">Llm Mgmn Llm Distributed</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_trtllm_bench.html">Llm Mgmn Trtllm Bench</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_trtllm_serve.html">Llm Mgmn Trtllm Serve</a></li>
|
|
</ul>
|
|
</details></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../examples/customization.html">LLM Common Customizations</a></li>
|
|
<li class="toctree-l1 has-children"><a class="reference internal" href="../examples/llm_api_examples.html">LLM Examples</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_medusa_decoding.html">Generate Text Using Medusa Decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_multilora.html">Generate text with multiple LoRA adapters</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_eagle_decoding.html">Generate Text Using Eagle Decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_async.html">Generate Text Asynchronously</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_distributed.html">Distributed LLM Generation</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_logits_processor.html">Control generated text using logits processor</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_eagle2_decoding.html">Generate Text Using Eagle2 Decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_kv_events.html">Get KV Cache Events</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_lookahead_decoding.html">Generate Text Using Lookahead Decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_quantization.html">Generation with Quantization</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_async_streaming.html">Generate Text in Streaming</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_guided_decoding.html">Generate text with guided decoding</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference.html">Generate text</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_inference_customize.html">Generate text with customization</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_auto_parallel.html">Automatic Parallelism with LLM</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_llm_distributed.html">Llm Mgmn Llm Distributed</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_trtllm_bench.html">Llm Mgmn Trtllm Bench</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_trtllm_serve.html">Llm Mgmn Trtllm Serve</a></li>
|
|
</ul>
|
|
</details></li>
|
|
<li class="toctree-l1 has-children"><a class="reference internal" href="../examples/trtllm_serve_examples.html">Online Serving Examples</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/curl_chat_client.html">Curl Chat Client</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/curl_chat_client_for_multimodal.html">Curl Chat Client For Multimodal</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/curl_completion_client.html">Curl Completion Client</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/deepseek_r1_reasoning_parser.html">Deepseek R1 Reasoning Parser</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/genai_perf_client.html">Genai Perf Client</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/genai_perf_client_for_multimodal.html">Genai Perf Client For Multimodal</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/openai_chat_client.html">OpenAI Chat Client</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/openai_chat_client_for_multimodal.html">OpenAI Chat Client</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="../examples/openai_completion_client.html">OpenAI Completion Client</a></li>
|
|
</ul>
|
|
</details></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Model Definition API</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../python-api/tensorrt_llm.layers.html">Layers</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../python-api/tensorrt_llm.functional.html">Functionals</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../python-api/tensorrt_llm.models.html">Models</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../python-api/tensorrt_llm.plugin.html">Plugin</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../python-api/tensorrt_llm.quantization.html">Quantization</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../python-api/tensorrt_llm.runtime.html">Runtime</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">C++ API</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../_cpp_gen/executor.html">Executor</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../_cpp_gen/runtime.html">Runtime</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Command-Line Reference</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../commands/trtllm-build.html">trtllm-build</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../commands/trtllm-serve.html">trtllm-serve</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Architecture</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../architecture/overview.html">TensorRT-LLM Architecture</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../architecture/core-concepts.html">Model Definition</a></li>
|
|
|
|
|
|
|
|
<li class="toctree-l1"><a class="reference internal" href="../architecture/checkpoint.html">TensorRT-LLM Checkpoint</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../architecture/workflow.html">TensorRT-LLM Build Workflow</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../architecture/add-model.html">Adding a Model</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Advanced</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/gpt-attention.html">Multi-Head, Multi-Query, and Group-Query Attention</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/gpt-runtime.html">C++ GPT Runtime</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/executor.html">Executor API</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/graph-rewriting.html">Graph Rewriting Module</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/lora.html">Run gpt-2b + LoRA using Executor / cpp runtime</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/expert-parallelism.html">Expert Parallelism in TensorRT-LLM</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/kv-cache-management.html">KV Cache Management: Pools, Blocks, and Events</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/kv-cache-reuse.html">KV cache reuse</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/speculative-decoding.html">Speculative Sampling</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../advanced/disaggregated-service.html">Disaggregated-Service (experimental)</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Performance</span></p>
|
|
<ul class="current nav bd-sidenav">
|
|
<li class="toctree-l1 current active"><a class="current reference internal" href="#">Overview</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="perf-benchmarking.html">Benchmarking</a></li>
|
|
<li class="toctree-l1 has-children"><a class="reference internal" href="performance-tuning-guide/index.html">Performance Tuning Guide</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
|
|
<li class="toctree-l2"><a class="reference internal" href="performance-tuning-guide/benchmarking-default-performance.html">Benchmarking Default Performance</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="performance-tuning-guide/useful-build-time-flags.html">Useful Build-Time Flags</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="performance-tuning-guide/tuning-max-batch-size-and-max-num-tokens.html">Tuning Max Batch Size and Max Num Tokens</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="performance-tuning-guide/deciding-model-sharding-strategy.html">Deciding Model Sharding Strategy</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="performance-tuning-guide/fp8-quantization.html">FP8 Quantization</a></li>
|
|
<li class="toctree-l2"><a class="reference internal" href="performance-tuning-guide/useful-runtime-flags.html">Useful Runtime Options</a></li>
|
|
</ul>
|
|
</details></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="perf-analysis.html">Performance Analysis</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Reference</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../reference/troubleshooting.html">Troubleshooting</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../reference/support-matrix.html">Support Matrix</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../reference/precision.html">Numerical Precision</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../reference/memory.html">Memory Usage of TensorRT-LLM</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../reference/ci-overview.html">Continuous Integration Overview</a></li>
|
|
</ul>
|
|
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Blogs</span></p>
|
|
<ul class="nav bd-sidenav">
|
|
<li class="toctree-l1"><a class="reference internal" href="../blogs/H100vsA100.html">H100 has 4.6x A100 Performance in TensorRT-LLM, achieving 10,000 tok/s at 100ms to first token</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../blogs/H200launch.html">H200 achieves nearly 12,000 tokens/sec on Llama2-13B with TensorRT-LLM</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../blogs/Falcon180B-H200.html">Falcon-180B on a single H200 GPU with INT4 AWQ, and 6.7x faster Llama-70B over A100</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../blogs/quantization-in-TRT-LLM.html">Speed up inference with SOTA quantization techniques in TRT-LLM</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../blogs/XQA-kernel.html">New XQA-kernel provides 2.4x more Llama-70B throughput within the same latency budget</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog1_Pushing_Latency_Boundaries_Optimizing_DeepSeek-R1_Performance_on_NVIDIA_B200_GPUs.html">Pushing Latency Boundaries: Optimizing DeepSeek-R1 Performance on NVIDIA B200 GPUs</a></li>
|
|
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog2_DeepSeek_R1_MTP_Implementation_and_Optimization.html">DeepSeek R1 MTP Implementation and Optimization</a></li>
|
|
</ul>
|
|
</div>
|
|
</nav></div>
|
|
</div>
|
|
|
|
|
|
<div class="sidebar-primary-items__end sidebar-primary__section">
|
|
</div>
|
|
|
|
|
|
|
|
</div>
|
|
|
|
<main id="main-content" class="bd-main" role="main">
|
|
|
|
|
|
<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">
|
|
|
|
<nav aria-label="Breadcrumb" class="d-print-none">
|
|
<ul class="bd-breadcrumbs">
|
|
|
|
<li class="breadcrumb-item breadcrumb-home">
|
|
<a href="../index.html" class="nav-link" aria-label="Home">
|
|
<i class="fa-solid fa-home"></i>
|
|
</a>
|
|
</li>
|
|
<li class="breadcrumb-item active" aria-current="page"><span class="ellipsis">Overview</span></li>
|
|
</ul>
|
|
</nav>
|
|
</div>
|
|
|
|
</div>
|
|
|
|
|
|
</div>
|
|
</div>
|
|
|
|
|
|
|
|
|
|
<div id="searchbox"></div>
|
|
<article class="bd-article">
|
|
|
|
<section id="overview">
|
|
<span id="perf-overview"></span><h1>Overview<a class="headerlink" href="#overview" title="Link to this heading">#</a></h1>
|
|
<p>This document summarizes performance measurements of TensorRT-LLM on a number of GPUs across a set of key models.</p>
|
|
<p>The data in the following tables is provided as a reference point to help users validate observed performance.
|
|
It should <em>not</em> be considered as the peak performance that can be delivered by TensorRT-LLM.</p>
|
|
<p>We attempted to keep commands as simple as possible to ease reproducibility and left many options at their default settings.
|
|
Tuning batch sizes, parallelism configurations, and other options may lead to improved performance depending on your situaiton.</p>
|
|
<p>For DeepSeek R1 performance, please check out our <a class="reference internal" href="../blogs/Best_perf_practice_on_DeepSeek-R1_in_TensorRT-LLM.html"><span class="std std-doc">performance guide</span></a></p>
|
|
<section id="throughput-measurements">
|
|
<h2>Throughput Measurements<a class="headerlink" href="#throughput-measurements" title="Link to this heading">#</a></h2>
|
|
<p>The below table shows performance data where a local inference client is fed requests at an infinite rate (no delay between messages),
|
|
and shows the throughput scenario under maximum load. The reported metric is <code class="docutils literal notranslate"><span class="pre">Total</span> <span class="pre">Output</span> <span class="pre">Throughput</span> <span class="pre">(tokens/sec)</span></code>.</p>
|
|
<p>The performance numbers below were collected using the steps described in this document.</p>
|
|
<p>Testing was performed on models with weights quantized using <a class="reference external" href="https://nvidia.github.io/TensorRT-Model-Optimizer/#">ModelOpt</a> and published by NVIDIA on the <a class="reference external" href="https://huggingface.co/collections/nvidia/model-optimizer-66aa84f7966b3150262481a4">Model Optimizer HuggingFace Collection</a>.</p>
|
|
<section id="fp4-models">
|
|
<h3>FP4 Models:<a class="headerlink" href="#fp4-models" title="Link to this heading">#</a></h3>
|
|
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">nvidia</span><span class="o">/</span><span class="n">Llama</span><span class="o">-</span><span class="mf">3.3</span><span class="o">-</span><span class="mi">70</span><span class="n">B</span><span class="o">-</span><span class="n">Instruct</span><span class="o">-</span><span class="n">FP4</span>
|
|
<span class="n">nvidia</span><span class="o">/</span><span class="n">Llama</span><span class="o">-</span><span class="mf">3.1</span><span class="o">-</span><span class="mi">405</span><span class="n">B</span><span class="o">-</span><span class="n">Instruct</span><span class="o">-</span><span class="n">FP4</span>
|
|
</pre></div>
|
|
</div>
|
|
<section id="llama-3-3-70b-fp4">
|
|
<h4>Llama 3.3 70B FP4<a class="headerlink" href="#llama-3-3-70b-fp4" title="Link to this heading">#</a></h4>
|
|
<div class="pst-scrollable-table-container"><table class="table">
|
|
<thead>
|
|
<tr class="row-odd"><th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p>GPU</p></th>
|
|
<th class="head text-left"><p>B200</p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>TP Size</p></td>
|
|
<td class="text-left"><p>1</p></td>
|
|
<td class="text-left"><p>2</p></td>
|
|
<td class="text-left"><p>4</p></td>
|
|
<td class="text-left"><p>8</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>ISL, OSL</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>11,253.28</p></td>
|
|
<td class="text-left"><p>17,867.66</p></td>
|
|
<td class="text-left"><p>24,944.50</p></td>
|
|
<td class="text-left"><p>27,471.49</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>128, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>9,925.00</p></td>
|
|
<td class="text-left"><p>15,459.71</p></td>
|
|
<td class="text-left"><p>23,608.58</p></td>
|
|
<td class="text-left"><p>30,742.86</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 4096</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>6,318.92</p></td>
|
|
<td class="text-left"><p>8,711.88</p></td>
|
|
<td class="text-left"><p>17,659.74</p></td>
|
|
<td class="text-left"><p>24,947.05</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>500, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>7,559.88</p></td>
|
|
<td class="text-left"><p>10,602.27</p></td>
|
|
<td class="text-left"><p>20,910.23</p></td>
|
|
<td class="text-left"><p>28,182.34</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1000, 1000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>6,866.96</p></td>
|
|
<td class="text-left"><p>10,838.01</p></td>
|
|
<td class="text-left"><p>16,567.86</p></td>
|
|
<td class="text-left"><p>19,991.64</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>1000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>6,736.88</p></td>
|
|
<td class="text-left"><p>9,132.08</p></td>
|
|
<td class="text-left"><p>15,737.02</p></td>
|
|
<td class="text-left"><p>20,518.04</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1024, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>6,580.56</p></td>
|
|
<td class="text-left"><p>8,767.45</p></td>
|
|
<td class="text-left"><p>15,722.55</p></td>
|
|
<td class="text-left"><p>20,437.96</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>2048, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>1,375.49</p></td>
|
|
<td class="text-left"><p>1,610.69</p></td>
|
|
<td class="text-left"><p>2,707.58</p></td>
|
|
<td class="text-left"><p>3,717.82</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>2048, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>4,544.73</p></td>
|
|
<td class="text-left"><p>6,956.14</p></td>
|
|
<td class="text-left"><p>12,292.23</p></td>
|
|
<td class="text-left"><p>15,661.22</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>5000, 500</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>1,488.19</p></td>
|
|
<td class="text-left"><p>2,379.73</p></td>
|
|
<td class="text-left"><p>3,588.45</p></td>
|
|
<td class="text-left"><p>4,810.21</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>20000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>580.96</p></td>
|
|
<td class="text-left"><p>1,043.58</p></td>
|
|
<td class="text-left"><p>1,957.84</p></td>
|
|
<td class="text-left"><p>3,167.30</p></td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
</section>
|
|
<section id="llama-3-1-405b-fp4">
|
|
<h4>Llama 3.1 405B FP4<a class="headerlink" href="#llama-3-1-405b-fp4" title="Link to this heading">#</a></h4>
|
|
<div class="pst-scrollable-table-container"><table class="table">
|
|
<thead>
|
|
<tr class="row-odd"><th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p>GPU</p></th>
|
|
<th class="head text-left"><p>B200</p></th>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>TP Size</p></td>
|
|
<td class="text-left"><p>8</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>ISL, OSL</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>9,184.83</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>128, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>10,387.23</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 4096</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>8,741.80</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>500, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>9,242.34</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1000, 1000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>7,565.50</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>1000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>7,696.76</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1024, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>7,568.93</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>2048, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>953.57</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>2048, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>6,092.32</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>5000, 500</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>1,332.22</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>20000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>961.58</p></td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
</section>
|
|
</section>
|
|
<section id="fp8-models">
|
|
<h3>FP8 Models:<a class="headerlink" href="#fp8-models" title="Link to this heading">#</a></h3>
|
|
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">nvidia</span><span class="o">/</span><span class="n">Llama</span><span class="o">-</span><span class="mf">3.1</span><span class="o">-</span><span class="mi">8</span><span class="n">B</span><span class="o">-</span><span class="n">Instruct</span><span class="o">-</span><span class="n">FP8</span>
|
|
<span class="n">nvidia</span><span class="o">/</span><span class="n">Llama</span><span class="o">-</span><span class="mf">3.1</span><span class="o">-</span><span class="mi">70</span><span class="n">B</span><span class="o">-</span><span class="n">Instruct</span><span class="o">-</span><span class="n">FP8</span>
|
|
<span class="n">nvidia</span><span class="o">/</span><span class="n">Llama</span><span class="o">-</span><span class="mf">3.1</span><span class="o">-</span><span class="mi">405</span><span class="n">B</span><span class="o">-</span><span class="n">Instruct</span><span class="o">-</span><span class="n">FP8</span>
|
|
</pre></div>
|
|
</div>
|
|
<section id="llama-3-1-8b-fp8">
|
|
<h4>Llama 3.1 8B FP8<a class="headerlink" href="#llama-3-1-8b-fp8" title="Link to this heading">#</a></h4>
|
|
<div class="pst-scrollable-table-container"><table class="table">
|
|
<thead>
|
|
<tr class="row-odd"><th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p>GPU</p></th>
|
|
<th class="head text-left"><p>H200 141GB HBM3</p></th>
|
|
<th class="head text-left"><p>H100 80GB HBM3</p></th>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>TP Size</p></td>
|
|
<td class="text-left"><p>1</p></td>
|
|
<td class="text-left"><p>1</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>ISL, OSL</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>28,447.38</p></td>
|
|
<td class="text-left"><p>27,568.68</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>128, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>23,294.74</p></td>
|
|
<td class="text-left"><p>22,003.62</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 4096</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>17,481.48</p></td>
|
|
<td class="text-left"><p>13,640.35</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>500, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>21,462.57</p></td>
|
|
<td class="text-left"><p>17,794.39</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1000, 1000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>17,590.60</p></td>
|
|
<td class="text-left"><p>15,270.02</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>1000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>17,139.51</p></td>
|
|
<td class="text-left"><p>13,850.22</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1024, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>16,970.63</p></td>
|
|
<td class="text-left"><p>13,374.15</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>2048, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,531.33</p></td>
|
|
<td class="text-left"><p>3,495.05</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>2048, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>12,022.38</p></td>
|
|
<td class="text-left"><p>9,653.67</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>5000, 500</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,851.65</p></td>
|
|
<td class="text-left"><p>3,371.16</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>20000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>1,706.06</p></td>
|
|
<td class="text-left"><p>1,340.92</p></td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
</section>
|
|
<section id="llama-3-1-70b-fp8">
|
|
<h4>Llama 3.1 70B FP8<a class="headerlink" href="#llama-3-1-70b-fp8" title="Link to this heading">#</a></h4>
|
|
<div class="pst-scrollable-table-container"><table class="table">
|
|
<thead>
|
|
<tr class="row-odd"><th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p>GPU</p></th>
|
|
<th class="head text-left"><p>H200 141GB HBM3</p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p>H100 80GB HBM3</p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p></p></th>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>TP Size</p></td>
|
|
<td class="text-left"><p>1</p></td>
|
|
<td class="text-left"><p>2</p></td>
|
|
<td class="text-left"><p>4</p></td>
|
|
<td class="text-left"><p>8</p></td>
|
|
<td class="text-left"><p>1</p></td>
|
|
<td class="text-left"><p>2</p></td>
|
|
<td class="text-left"><p>4</p></td>
|
|
<td class="text-left"><p>8</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>ISL, OSL</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,657.58</p></td>
|
|
<td class="text-left"><p>6,477.50</p></td>
|
|
<td class="text-left"><p>10,466.04</p></td>
|
|
<td class="text-left"><p>15,554.57</p></td>
|
|
<td class="text-left"><p>3,191.27</p></td>
|
|
<td class="text-left"><p>6,183.41</p></td>
|
|
<td class="text-left"><p>10,260.68</p></td>
|
|
<td class="text-left"><p>14,686.01</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>128, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>4,351.07</p></td>
|
|
<td class="text-left"><p>8,450.31</p></td>
|
|
<td class="text-left"><p>13,438.71</p></td>
|
|
<td class="text-left"><p>20,750.58</p></td>
|
|
<td class="text-left"><p>745.19</p></td>
|
|
<td class="text-left"><p>5,822.02</p></td>
|
|
<td class="text-left"><p>11,442.01</p></td>
|
|
<td class="text-left"><p>17,463.99</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 4096</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>2,696.61</p></td>
|
|
<td class="text-left"><p>5,598.92</p></td>
|
|
<td class="text-left"><p>11,524.93</p></td>
|
|
<td class="text-left"><p>16,634.90</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,714.87</p></td>
|
|
<td class="text-left"><p>8,209.91</p></td>
|
|
<td class="text-left"><p>12,598.55</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>500, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,475.58</p></td>
|
|
<td class="text-left"><p>6,712.35</p></td>
|
|
<td class="text-left"><p>12,332.32</p></td>
|
|
<td class="text-left"><p>17,311.28</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>4,704.31</p></td>
|
|
<td class="text-left"><p>10,278.02</p></td>
|
|
<td class="text-left"><p>14,630.41</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1000, 1000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>2,727.42</p></td>
|
|
<td class="text-left"><p>5,097.36</p></td>
|
|
<td class="text-left"><p>8,698.15</p></td>
|
|
<td class="text-left"><p>12,794.92</p></td>
|
|
<td class="text-left"><p>734.67</p></td>
|
|
<td class="text-left"><p>4,191.26</p></td>
|
|
<td class="text-left"><p>7,427.35</p></td>
|
|
<td class="text-left"><p>11,082.48</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>1000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>2,913.54</p></td>
|
|
<td class="text-left"><p>5,841.15</p></td>
|
|
<td class="text-left"><p>9,016.49</p></td>
|
|
<td class="text-left"><p>13,174.68</p></td>
|
|
<td class="text-left"><p>526.31</p></td>
|
|
<td class="text-left"><p>3,920.44</p></td>
|
|
<td class="text-left"><p>7,590.35</p></td>
|
|
<td class="text-left"><p>11,108.11</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1024, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>2,893.02</p></td>
|
|
<td class="text-left"><p>5,565.28</p></td>
|
|
<td class="text-left"><p>9,017.72</p></td>
|
|
<td class="text-left"><p>13,117.34</p></td>
|
|
<td class="text-left"><p>525.43</p></td>
|
|
<td class="text-left"><p>3,896.14</p></td>
|
|
<td class="text-left"><p>7,557.32</p></td>
|
|
<td class="text-left"><p>11,028.32</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>2048, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>433.30</p></td>
|
|
<td class="text-left"><p>772.97</p></td>
|
|
<td class="text-left"><p>1,278.26</p></td>
|
|
<td class="text-left"><p>1,947.33</p></td>
|
|
<td class="text-left"><p>315.90</p></td>
|
|
<td class="text-left"><p>747.51</p></td>
|
|
<td class="text-left"><p>1,240.12</p></td>
|
|
<td class="text-left"><p>1,840.12</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>2048, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>1,990.25</p></td>
|
|
<td class="text-left"><p>3,822.83</p></td>
|
|
<td class="text-left"><p>7,068.68</p></td>
|
|
<td class="text-left"><p>10,529.06</p></td>
|
|
<td class="text-left"><p>357.98</p></td>
|
|
<td class="text-left"><p>2,732.86</p></td>
|
|
<td class="text-left"><p>5,640.31</p></td>
|
|
<td class="text-left"><p>8,772.88</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>5000, 500</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>543.88</p></td>
|
|
<td class="text-left"><p>1,005.81</p></td>
|
|
<td class="text-left"><p>1,714.77</p></td>
|
|
<td class="text-left"><p>2,683.22</p></td>
|
|
<td class="text-left"><p>203.27</p></td>
|
|
<td class="text-left"><p>866.77</p></td>
|
|
<td class="text-left"><p>1,571.92</p></td>
|
|
<td class="text-left"><p>2,399.78</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>20000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>276.99</p></td>
|
|
<td class="text-left"><p>618.01</p></td>
|
|
<td class="text-left"><p>1,175.35</p></td>
|
|
<td class="text-left"><p>2,021.08</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>408.43</p></td>
|
|
<td class="text-left"><p>910.77</p></td>
|
|
<td class="text-left"><p>1,568.84</p></td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
</section>
|
|
<section id="llama-3-1-405b-fp8">
|
|
<h4>Llama 3.1 405B FP8<a class="headerlink" href="#llama-3-1-405b-fp8" title="Link to this heading">#</a></h4>
|
|
<div class="pst-scrollable-table-container"><table class="table">
|
|
<thead>
|
|
<tr class="row-odd"><th class="head text-left"><p></p></th>
|
|
<th class="head text-left"><p>GPU</p></th>
|
|
<th class="head text-left"><p>H200 141GB HBM3</p></th>
|
|
<th class="head text-left"><p>H100 80GB HBM3</p></th>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>TP Size</p></td>
|
|
<td class="text-left"><p>8</p></td>
|
|
<td class="text-left"><p>8</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>ISL, OSL</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p></p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,800.11</p></td>
|
|
<td class="text-left"><p>3,732.40</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>128, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>5,661.13</p></td>
|
|
<td class="text-left"><p>4,572.23</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>128, 4096</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>5,167.18</p></td>
|
|
<td class="text-left"><p>2,911.42</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>500, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>4,854.29</p></td>
|
|
<td class="text-left"><p>3,661.85</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1000, 1000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,332.15</p></td>
|
|
<td class="text-left"><p>2,963.36</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>1000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,682.15</p></td>
|
|
<td class="text-left"><p>3,253.17</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>1024, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,685.56</p></td>
|
|
<td class="text-left"><p>3,089.16</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>2048, 128</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>453.42</p></td>
|
|
<td class="text-left"><p>448.89</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>2048, 2048</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>3,055.73</p></td>
|
|
<td class="text-left"><p>2,139.94</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p>5000, 500</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>656.11</p></td>
|
|
<td class="text-left"><p>579.14</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p>20000, 2000</p></td>
|
|
<td class="text-left"><p></p></td>
|
|
<td class="text-left"><p>514.02</p></td>
|
|
<td class="text-left"><p>370.26</p></td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
</section>
|
|
</section>
|
|
</section>
|
|
<section id="reproducing-benchmarked-results">
|
|
<h2>Reproducing Benchmarked Results<a class="headerlink" href="#reproducing-benchmarked-results" title="Link to this heading">#</a></h2>
|
|
<blockquote>
|
|
<div><p>[!NOTE] The only models supported in this workflow are those listed in the table above.</p>
|
|
</div></blockquote>
|
|
<p>The following tables are references for commands that are used as part of the benchmarking process. For a more detailed
|
|
description of this benchmarking workflow, see the <a class="reference external" href="https://nvidia.github.io/TensorRT-LLM/performance/perf-benchmarking.html">benchmarking suite documentation</a>.</p>
|
|
<section id="command-overview">
|
|
<h3>Command Overview<a class="headerlink" href="#command-overview" title="Link to this heading">#</a></h3>
|
|
<p>Starting with v0.19, testing was performed using the PyTorch backend - this workflow does not require an engine to be built.</p>
|
|
<div class="pst-scrollable-table-container"><table class="table">
|
|
<thead>
|
|
<tr class="row-odd"><th class="head text-left"><p>Stage</p></th>
|
|
<th class="head"><p>Description</p></th>
|
|
<th class="head"><p>Command</p></th>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr class="row-even"><td class="text-left"><p><a class="reference internal" href="#preparing-a-dataset">Dataset</a></p></td>
|
|
<td><p>Create a synthetic dataset</p></td>
|
|
<td><p><code class="docutils literal notranslate"><span class="pre">python</span> <span class="pre">benchmarks/cpp/prepare_dataset.py</span> <span class="pre">--tokenizer=$model_name</span> <span class="pre">--stdout</span> <span class="pre">token-norm-dist</span> <span class="pre">--num-requests=$num_requests</span> <span class="pre">--input-mean=$isl</span> <span class="pre">--output-mean=$osl</span> <span class="pre">--input-stdev=0</span> <span class="pre">--output-stdev=0</span> <span class="pre">></span> <span class="pre">$dataset_file</span></code></p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p><a class="reference internal" href="#running-the-benchmark">Run</a></p></td>
|
|
<td><p>Run a benchmark with a dataset</p></td>
|
|
<td><p><code class="docutils literal notranslate"><span class="pre">trtllm-bench</span> <span class="pre">--model</span> <span class="pre">$model_name</span> <span class="pre">throughput</span> <span class="pre">--dataset</span> <span class="pre">$dataset_file</span> <span class="pre">--backend</span> <span class="pre">pytorch</span> <span class="pre">--extra_llm_api_options</span> <span class="pre">$llm_options</span></code></p></td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
</section>
|
|
<section id="variables">
|
|
<h3>Variables<a class="headerlink" href="#variables" title="Link to this heading">#</a></h3>
|
|
<div class="pst-scrollable-table-container"><table class="table">
|
|
<thead>
|
|
<tr class="row-odd"><th class="head text-left"><p>Name</p></th>
|
|
<th class="head"><p>Description</p></th>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr class="row-even"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$isl</span></code></p></td>
|
|
<td><p>Benchmark input sequence length.</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$osl</span></code></p></td>
|
|
<td><p>Benchmark output sequence length.</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$tp_size</span></code></p></td>
|
|
<td><p>Tensor parallel mapping degree to run the benchmark with</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$pp_size</span></code></p></td>
|
|
<td><p>Pipeline parallel mapping degree to run the benchmark with</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$model_name</span></code></p></td>
|
|
<td><p>HuggingFace model name eg. meta-llama/Llama-2-7b-hf or use the path to a local weights directory</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$dataset_file</span></code></p></td>
|
|
<td><p>Location of the dataset file generated by <code class="docutils literal notranslate"><span class="pre">prepare_dataset.py</span></code></p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$num_requests</span></code></p></td>
|
|
<td><p>The number of requests to generate for dataset generation</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$seq_len</span></code></p></td>
|
|
<td><p>A sequence length of ISL + OSL</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td class="text-left"><p><code class="docutils literal notranslate"><span class="pre">$llm_options</span></code></p></td>
|
|
<td><p>(optional) A yaml file containing additional options for the LLM API</p></td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
</section>
|
|
<section id="preparing-a-dataset">
|
|
<h3>Preparing a Dataset<a class="headerlink" href="#preparing-a-dataset" title="Link to this heading">#</a></h3>
|
|
<p>In order to prepare a dataset, you can use the provided <a class="reference download internal" download="" href="../_downloads/ea8faa5e98124e92f96b66dc586fb429/prepare_dataset.py"><span class="xref download myst">script</span></a>.
|
|
To generate a synthetic dataset, run the following command:</p>
|
|
<div class="highlight-shell notranslate"><div class="highlight"><pre><span></span>python<span class="w"> </span>benchmarks/cpp/prepare_dataset.py<span class="w"> </span>--tokenizer<span class="o">=</span><span class="nv">$model_name</span><span class="w"> </span>--stdout<span class="w"> </span>token-norm-dist<span class="w"> </span>--num-requests<span class="o">=</span><span class="nv">$num_requests</span><span class="w"> </span>--input-mean<span class="o">=</span><span class="nv">$isl</span><span class="w"> </span>--output-mean<span class="o">=</span><span class="nv">$osl</span><span class="w"> </span>--input-stdev<span class="o">=</span><span class="m">0</span><span class="w"> </span>--output-stdev<span class="o">=</span><span class="m">0</span><span class="w"> </span>><span class="w"> </span><span class="nv">$dataset_file</span>
|
|
</pre></div>
|
|
</div>
|
|
<p>The command will generate a text file located at the path specified <code class="docutils literal notranslate"><span class="pre">$dataset_file</span></code> where all requests are of the same
|
|
input/output sequence length combinations. The script works by using the tokenizer to retrieve the vocabulary size and
|
|
randomly sample token IDs from it to create entirely random sequences. In the command above, all requests will be uniform
|
|
because the standard deviations for both input and output sequences are set to 0.</p>
|
|
<p>For each input and output sequence length combination, the table below details the <code class="docutils literal notranslate"><span class="pre">$num_requests</span></code> that were used. For
|
|
shorter input and output lengths, a larger number of messages were used to guarantee that the system hit a steady state
|
|
because requests enter and exit the system at a much faster rate. For longer input/output sequence lengths, requests
|
|
remain in the system longer and therefore require less requests to achieve steady state.</p>
|
|
<div class="pst-scrollable-table-container"><table class="table">
|
|
<thead>
|
|
<tr class="row-odd"><th class="head"><p>Input Length</p></th>
|
|
<th class="head"><p>Output Length</p></th>
|
|
<th class="head"><p>$seq_len</p></th>
|
|
<th class="head"><p>$num_requests</p></th>
|
|
</tr>
|
|
</thead>
|
|
<tbody>
|
|
<tr class="row-even"><td><p>128</p></td>
|
|
<td><p>128</p></td>
|
|
<td><p>256</p></td>
|
|
<td><p>30000</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td><p>128</p></td>
|
|
<td><p>2048</p></td>
|
|
<td><p>2176</p></td>
|
|
<td><p>3000</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td><p>128</p></td>
|
|
<td><p>4096</p></td>
|
|
<td><p>4224</p></td>
|
|
<td><p>1500</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td><p>1000</p></td>
|
|
<td><p>2000</p></td>
|
|
<td><p>3000</p></td>
|
|
<td><p>1500</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td><p>2048</p></td>
|
|
<td><p>128</p></td>
|
|
<td><p>2176</p></td>
|
|
<td><p>3000</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td><p>2048</p></td>
|
|
<td><p>2048</p></td>
|
|
<td><p>4096</p></td>
|
|
<td><p>1500</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td><p>5000</p></td>
|
|
<td><p>500</p></td>
|
|
<td><p>5500</p></td>
|
|
<td><p>1500</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td><p>1000</p></td>
|
|
<td><p>1000</p></td>
|
|
<td><p>2000</p></td>
|
|
<td><p>3000</p></td>
|
|
</tr>
|
|
<tr class="row-even"><td><p>500</p></td>
|
|
<td><p>2000</p></td>
|
|
<td><p>2500</p></td>
|
|
<td><p>3000</p></td>
|
|
</tr>
|
|
<tr class="row-odd"><td><p>20000</p></td>
|
|
<td><p>2000</p></td>
|
|
<td><p>22000</p></td>
|
|
<td><p>1000</p></td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
</section>
|
|
<section id="running-the-benchmark">
|
|
<h3>Running the Benchmark<a class="headerlink" href="#running-the-benchmark" title="Link to this heading">#</a></h3>
|
|
<p>To run the benchmark with the generated data set, simply use the <code class="docutils literal notranslate"><span class="pre">trtllm-bench</span> <span class="pre">throughput</span></code> subcommand. The benchmarker will
|
|
run an offline maximum throughput scenario such that all requests are queued in rapid succession. You simply need to provide
|
|
a model name (HuggingFace reference or path to a local model), a <a class="reference internal" href="#preparing-a-dataset">generated dataset</a>, and a file containing any desired extra options to the LLMApi (details in <a class="reference download internal" download="" href="../_downloads/cba6509356738d5d6b4dcb3b7f52cf39/llm_args.py"><span class="xref download myst">tensorrt_llm/llmapi/llm_args.py:LlmArgs</span></a>).</p>
|
|
<div class="highlight-shell notranslate"><div class="highlight"><pre><span></span>trtllm-bench<span class="w"> </span>--model<span class="w"> </span><span class="nv">$model_name</span><span class="w"> </span>throughput<span class="w"> </span>--dataset<span class="w"> </span><span class="nv">$dataset_file</span><span class="w"> </span>--backend<span class="w"> </span>pytorch<span class="w"> </span>--extra_llm_api_options<span class="w"> </span><span class="nv">$llm_options</span>
|
|
</pre></div>
|
|
</div>
|
|
<p><code class="docutils literal notranslate"><span class="pre">llm_options.yml</span></code></p>
|
|
<div class="highlight-yaml notranslate"><div class="highlight"><pre><span></span><span class="nt">use_cuda_graph</span><span class="p">:</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">true</span>
|
|
<span class="nt">cuda_graph_padding_enabled</span><span class="p">:</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">true</span>
|
|
<span class="nt">cuda_graph_batch_sizes</span><span class="p">:</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">1</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">2</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">4</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">8</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">16</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">32</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">64</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">128</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">256</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">384</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">512</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">1024</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">2048</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">4096</span>
|
|
<span class="w"> </span><span class="p p-Indicator">-</span><span class="w"> </span><span class="l l-Scalar l-Scalar-Plain">8192</span>
|
|
</pre></div>
|
|
</div>
|
|
<p>In majority of cases, we also use a higher KV cache percentage by setting <code class="docutils literal notranslate"><span class="pre">--kv_cache_free_gpu_mem_fraction</span> <span class="pre">0.95</span></code> in the benchmark command. This allows us to obtain better performance than the default setting of <code class="docutils literal notranslate"><span class="pre">0.90</span></code>. We fall back to <code class="docutils literal notranslate"><span class="pre">0.90</span></code> if we hit an out of memory issue.</p>
|
|
<p>The results will be printed to the terminal upon benchmark completion. For example,</p>
|
|
<div class="highlight-shell notranslate"><div class="highlight"><pre><span></span><span class="o">===========================================================</span>
|
|
<span class="o">=</span><span class="w"> </span>PERFORMANCE<span class="w"> </span><span class="nv">OVERVIEW</span>
|
|
<span class="o">===========================================================</span>
|
|
Request<span class="w"> </span>Throughput<span class="w"> </span><span class="o">(</span>req/sec<span class="o">)</span>:<span class="w"> </span><span class="m">43</span>.2089
|
|
Total<span class="w"> </span>Output<span class="w"> </span>Throughput<span class="w"> </span><span class="o">(</span>tokens/sec<span class="o">)</span>:<span class="w"> </span><span class="m">5530</span>.7382
|
|
Per<span class="w"> </span>User<span class="w"> </span>Output<span class="w"> </span>Throughput<span class="w"> </span><span class="o">(</span>tokens/sec/user<span class="o">)</span>:<span class="w"> </span><span class="m">2</span>.0563
|
|
Per<span class="w"> </span>GPU<span class="w"> </span>Output<span class="w"> </span>Throughput<span class="w"> </span><span class="o">(</span>tokens/sec/gpu<span class="o">)</span>:<span class="w"> </span><span class="m">5530</span>.7382
|
|
Total<span class="w"> </span>Token<span class="w"> </span>Throughput<span class="w"> </span><span class="o">(</span>tokens/sec<span class="o">)</span>:<span class="w"> </span><span class="m">94022</span>.5497
|
|
Total<span class="w"> </span>Latency<span class="w"> </span><span class="o">(</span>ms<span class="o">)</span>:<span class="w"> </span><span class="m">115716</span>.9214
|
|
Average<span class="w"> </span>request<span class="w"> </span>latency<span class="w"> </span><span class="o">(</span>ms<span class="o">)</span>:<span class="w"> </span><span class="m">75903</span>.4456
|
|
Per<span class="w"> </span>User<span class="w"> </span>Output<span class="w"> </span>Speed<span class="w"> </span><span class="o">[</span><span class="m">1</span>/TPOT<span class="o">]</span><span class="w"> </span><span class="o">(</span>tokens/sec/user<span class="o">)</span>:<span class="w"> </span><span class="m">5</span>.4656
|
|
Average<span class="w"> </span>time-to-first-token<span class="w"> </span><span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span><span class="o">(</span>ms<span class="o">)</span>:<span class="w"> </span><span class="m">52667</span>.0339
|
|
Average<span class="w"> </span>time-per-output-token<span class="w"> </span><span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span><span class="o">(</span>ms<span class="o">)</span>:<span class="w"> </span><span class="m">182</span>.9639
|
|
|
|
--<span class="w"> </span>Per-Request<span class="w"> </span>Time-per-Output-Token<span class="w"> </span><span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span>Breakdown<span class="w"> </span><span class="o">(</span>ms<span class="o">)</span>
|
|
|
|
<span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span>MINIMUM:<span class="w"> </span><span class="m">32</span>.8005
|
|
<span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span>MAXIMUM:<span class="w"> </span><span class="m">208</span>.4667
|
|
<span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span>AVERAGE:<span class="w"> </span><span class="m">182</span>.9639
|
|
<span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span>P50<span class="w"> </span>:<span class="w"> </span><span class="m">204</span>.0463
|
|
<span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span>P90<span class="w"> </span>:<span class="w"> </span><span class="m">206</span>.3863
|
|
<span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span>P95<span class="w"> </span>:<span class="w"> </span><span class="m">206</span>.5064
|
|
<span class="o">[</span>TPOT<span class="o">]</span><span class="w"> </span>P99<span class="w"> </span>:<span class="w"> </span><span class="m">206</span>.5821
|
|
|
|
--<span class="w"> </span>Per-Request<span class="w"> </span>Time-to-First-Token<span class="w"> </span><span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span>Breakdown<span class="w"> </span><span class="o">(</span>ms<span class="o">)</span>
|
|
|
|
<span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span>MINIMUM:<span class="w"> </span><span class="m">3914</span>.7621
|
|
<span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span>MAXIMUM:<span class="w"> </span><span class="m">107501</span>.2487
|
|
<span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span>AVERAGE:<span class="w"> </span><span class="m">52667</span>.0339
|
|
<span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span>P50<span class="w"> </span>:<span class="w"> </span><span class="m">52269</span>.7072
|
|
<span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span>P90<span class="w"> </span>:<span class="w"> </span><span class="m">96583</span>.7187
|
|
<span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span>P95<span class="w"> </span>:<span class="w"> </span><span class="m">101978</span>.4566
|
|
<span class="o">[</span>TTFT<span class="o">]</span><span class="w"> </span>P99<span class="w"> </span>:<span class="w"> </span><span class="m">106563</span>.4497
|
|
|
|
--<span class="w"> </span>Request<span class="w"> </span>Latency<span class="w"> </span>Breakdown<span class="w"> </span><span class="o">(</span>ms<span class="o">)</span><span class="w"> </span>-----------------------
|
|
|
|
<span class="o">[</span>Latency<span class="o">]</span><span class="w"> </span>P50<span class="w"> </span>:<span class="w"> </span><span class="m">78509</span>.2102
|
|
<span class="o">[</span>Latency<span class="o">]</span><span class="w"> </span>P90<span class="w"> </span>:<span class="w"> </span><span class="m">110804</span>.0017
|
|
<span class="o">[</span>Latency<span class="o">]</span><span class="w"> </span>P95<span class="w"> </span>:<span class="w"> </span><span class="m">111302</span>.9101
|
|
<span class="o">[</span>Latency<span class="o">]</span><span class="w"> </span>P99<span class="w"> </span>:<span class="w"> </span><span class="m">111618</span>.2158
|
|
<span class="o">[</span>Latency<span class="o">]</span><span class="w"> </span>MINIMUM:<span class="w"> </span><span class="m">24189</span>.0838
|
|
<span class="o">[</span>Latency<span class="o">]</span><span class="w"> </span>MAXIMUM:<span class="w"> </span><span class="m">111668</span>.0964
|
|
<span class="o">[</span>Latency<span class="o">]</span><span class="w"> </span>AVERAGE:<span class="w"> </span><span class="m">75903</span>.4456
|
|
</pre></div>
|
|
</div>
|
|
<blockquote>
|
|
<div><p>[!WARNING] In some cases, the benchmarker may not print anything at all. This behavior usually
|
|
means that the benchmark has hit an out of memory issue. Try reducing the KV cache percentage
|
|
using the <code class="docutils literal notranslate"><span class="pre">--kv_cache_free_gpu_mem_fraction</span></code> option to lower the percentage of used memory.</p>
|
|
</div></blockquote>
|
|
</section>
|
|
</section>
|
|
</section>
|
|
|
|
|
|
</article>
|
|
|
|
|
|
|
|
|
|
|
|
<footer class="prev-next-footer d-print-none">
|
|
|
|
<div class="prev-next-area">
|
|
<a class="left-prev"
|
|
href="../advanced/disaggregated-service.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">Disaggregated-Service (experimental)</p>
|
|
</div>
|
|
</a>
|
|
<a class="right-next"
|
|
href="perf-benchmarking.html"
|
|
title="next page">
|
|
<div class="prev-next-info">
|
|
<p class="prev-next-subtitle">next</p>
|
|
<p class="prev-next-title">TensorRT-LLM Benchmarking</p>
|
|
</div>
|
|
<i class="fa-solid fa-angle-right"></i>
|
|
</a>
|
|
</div>
|
|
</footer>
|
|
|
|
</div>
|
|
|
|
|
|
|
|
|
|
|
|
<dialog id="pst-secondary-sidebar-modal"></dialog>
|
|
<div id="pst-secondary-sidebar" class="bd-sidebar-secondary bd-toc"><div class="sidebar-secondary-items sidebar-secondary__inner">
|
|
|
|
|
|
<div class="sidebar-secondary-item">
|
|
<div
|
|
id="pst-page-navigation-heading-2"
|
|
class="page-toc tocsection onthispage">
|
|
<i class="fa-solid fa-list"></i> On this page
|
|
</div>
|
|
<nav class="bd-toc-nav page-toc" aria-labelledby="pst-page-navigation-heading-2">
|
|
<ul class="visible nav section-nav flex-column">
|
|
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#throughput-measurements">Throughput Measurements</a><ul class="nav section-nav flex-column">
|
|
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#fp4-models">FP4 Models:</a><ul class="nav section-nav flex-column">
|
|
<li class="toc-h4 nav-item toc-entry"><a class="reference internal nav-link" href="#llama-3-3-70b-fp4">Llama 3.3 70B FP4</a></li>
|
|
<li class="toc-h4 nav-item toc-entry"><a class="reference internal nav-link" href="#llama-3-1-405b-fp4">Llama 3.1 405B FP4</a></li>
|
|
</ul>
|
|
</li>
|
|
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#fp8-models">FP8 Models:</a><ul class="nav section-nav flex-column">
|
|
<li class="toc-h4 nav-item toc-entry"><a class="reference internal nav-link" href="#llama-3-1-8b-fp8">Llama 3.1 8B FP8</a></li>
|
|
<li class="toc-h4 nav-item toc-entry"><a class="reference internal nav-link" href="#llama-3-1-70b-fp8">Llama 3.1 70B FP8</a></li>
|
|
<li class="toc-h4 nav-item toc-entry"><a class="reference internal nav-link" href="#llama-3-1-405b-fp8">Llama 3.1 405B FP8</a></li>
|
|
</ul>
|
|
</li>
|
|
</ul>
|
|
</li>
|
|
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#reproducing-benchmarked-results">Reproducing Benchmarked Results</a><ul class="nav section-nav flex-column">
|
|
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#command-overview">Command Overview</a></li>
|
|
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#variables">Variables</a></li>
|
|
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#preparing-a-dataset">Preparing a Dataset</a></li>
|
|
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#running-the-benchmark">Running the Benchmark</a></li>
|
|
</ul>
|
|
</li>
|
|
</ul>
|
|
</nav></div>
|
|
|
|
</div></div>
|
|
|
|
|
|
|
|
</div>
|
|
<footer class="bd-footer-content">
|
|
|
|
</footer>
|
|
|
|
</main>
|
|
</div>
|
|
</div>
|
|
|
|
<!-- Scripts loaded after <body> so the DOM is not blocked -->
|
|
<script defer src="../_static/scripts/bootstrap.js?digest=8878045cc6db502f8baf"></script>
|
|
<script defer src="../_static/scripts/pydata-sphinx-theme.js?digest=8878045cc6db502f8baf"></script>
|
|
|
|
<footer class="bd-footer">
|
|
<div class="bd-footer__inner bd-page-width">
|
|
|
|
<div class="footer-items__start">
|
|
|
|
<div class="footer-item">
|
|
<a class="footer-brand logo" href="https://www.nvidia.com">
|
|
<img src="../_static/nvidia-logo-horiz-rgb-1c-blk-for-screen.svg" class="logo__image only-light" alt="NVIDIA"/>
|
|
<img src="../_static/nvidia-logo-horiz-rgb-1c-wht-for-screen.svg" class="logo__image only-dark" alt="NVIDIA"/>
|
|
</a></div>
|
|
|
|
<div class="footer-item">
|
|
|
|
<div class="footer-links">
|
|
|
|
|
|
<a class="external" href="https://www.nvidia.com/en-us/about-nvidia/privacy-policy/">Privacy Policy</a>
|
|
|
|
|
|
|
|
|
|
|
<a class="external" href="https://www.nvidia.com/en-us/about-nvidia/privacy-center/">Manage My Privacy</a>
|
|
|
|
|
|
|
|
|
|
|
<a class="external" href="https://www.nvidia.com/en-us/preferences/start/">Do Not Sell or Share My Data</a>
|
|
|
|
|
|
|
|
|
|
|
<a class="external" href="https://www.nvidia.com/en-us/about-nvidia/terms-of-service/">Terms of Service</a>
|
|
|
|
|
|
|
|
|
|
|
<a class="external" href="https://www.nvidia.com/en-us/about-nvidia/accessibility/">Accessibility</a>
|
|
|
|
|
|
|
|
|
|
|
<a class="external" href="https://www.nvidia.com/en-us/about-nvidia/company-policies/">Corporate Policies</a>
|
|
|
|
|
|
|
|
|
|
|
<a class="external" href="https://www.nvidia.com/en-us/product-security/">Product Security</a>
|
|
|
|
|
|
|
|
|
|
|
<a class="external" href="https://www.nvidia.com/en-us/contact/">Contact</a>
|
|
|
|
|
|
|
|
</div>
|
|
</div>
|
|
|
|
<div class="footer-item">
|
|
|
|
|
|
|
|
|
|
<p class="copyright">
|
|
|
|
Copyright © 2025, NVidia.
|
|
<br/>
|
|
|
|
</p>
|
|
</div>
|
|
|
|
<div class="footer-item">
|
|
<div class="extra_footer">
|
|
|
|
<p>Last updated on June 16, 2025.</p>
|
|
|
|
<p>This page is generated by TensorRT-LLM commit <a href="https://github.com/NVIDIA/TensorRT-LLM/tree/8445416">8445416</a>.</p>
|
|
|
|
</div></div>
|
|
|
|
</div>
|
|
|
|
|
|
|
|
</div>
|
|
|
|
</footer>
|
|
</body>
|
|
</html> |