mirror of
https://github.com/NVIDIA/TensorRT-LLM.git
synced 2026-01-14 06:27:45 +08:00
775 lines
49 KiB
HTML
775 lines
49 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>KV Cache System — 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=933278ad" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/autodoc_pydantic.css" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/togglebutton.css?v=13237357" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/config_selector.css?v=e17d8078" />
|
||
<link rel="stylesheet" type="text/css" href="../_static/custom.css?v=19d20f17" />
|
||
|
||
<!-- 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 src="../_static/config_selector.js?v=aaf6cd4a"></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>var togglebuttonSelector = '.toggle, .admonition.dropdown';</script>
|
||
<script>DOCUMENTATION_OPTIONS.pagename = 'features/kvcache';</script>
|
||
<script>
|
||
DOCUMENTATION_OPTIONS.theme_version = '0.16.1';
|
||
DOCUMENTATION_OPTIONS.theme_switcher_json_url = './_static/switcher.json';
|
||
DOCUMENTATION_OPTIONS.theme_switcher_version_match = '1.2.0rc7';
|
||
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="Long Sequences" href="long-sequence.html" />
|
||
<link rel="prev" title="Disaggregated Serving" href="disagg-serving.html" />
|
||
|
||
|
||
<meta name="viewport" content="width=device-width, initial-scale=1"/>
|
||
<meta name="docsearch:language" content="en"/>
|
||
<meta name="docsearch:version" content="1.2.0rc7" />
|
||
|
||
|
||
</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 has-children"><a class="reference internal" href="../installation/index.html">Installation</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="../installation/containers.html">Pre-built release container images on NGC</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../installation/linux.html">Installing on Linux via <code class="docutils literal notranslate"><span class="pre">pip</span></code></a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../installation/build-from-source-linux.html">Building from Source Code on Linux</a></li>
|
||
</ul>
|
||
</details></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Deployment Guide</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<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_inference.html">Generate text</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_async_streaming.html">Generate text in streaming</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_guided_decoding.html">Generate text with guided decoding</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_multilora.html">Generate text with multiple LoRA adapters</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_sparse_attention.html">Sparse Attention</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_speculative_decoding.html">Speculative Decoding</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_kv_cache_connector.html">KV Cache Connector</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_kv_cache_offloading.html">KV Cache Offloading</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_runtime.html">Runtime Configuration Examples</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_sampling.html">Sampling Techniques Showcase</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_llm_distributed.html">Run LLM-API with pytorch backend on Slurm</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_trtllm_bench.html">Run trtllm-bench with pytorch backend on Slurm</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/llm_mgmn_trtllm_serve.html">Run trtllm-serve with pytorch backend on Slurm</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/aiperf_client.html">Aiperf Client</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/aiperf_client_for_multimodal.html">Aiperf Client For Multimodal</a></li>
|
||
<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/curl_responses_client.html">Curl Responses 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/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 for Multimodal</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/openai_completion_client.html">OpenAI Completion Client</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/openai_completion_client_for_lora.html">Openai Completion Client For Lora</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/openai_completion_client_json_schema.html">OpenAI Completion Client with JSON Schema</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../examples/openai_responses_client.html">OpenAI Responses Client</a></li>
|
||
</ul>
|
||
</details></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../examples/dynamo_k8s_example.html">Dynamo K8s Example</a></li>
|
||
<li class="toctree-l1 has-children"><a class="reference internal" href="../deployment-guide/index.html">Model Recipes</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="../deployment-guide/deployment-guide-for-deepseek-r1-on-trtllm.html">Deployment Guide for DeepSeek R1 on TensorRT LLM - Blackwell & Hopper Hardware</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../deployment-guide/deployment-guide-for-llama3.3-70b-on-trtllm.html">Deployment Guide for Llama3.3 70B on TensorRT LLM - Blackwell & Hopper Hardware</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../deployment-guide/deployment-guide-for-llama4-scout-on-trtllm.html">Deployment Guide for Llama4 Scout 17B on TensorRT LLM - Blackwell & Hopper Hardware</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../deployment-guide/deployment-guide-for-gpt-oss-on-trtllm.html">Deployment Guide for GPT-OSS on TensorRT-LLM - Blackwell Hardware</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../deployment-guide/deployment-guide-for-qwen3-on-trtllm.html">Deployment Guide for Qwen3 on TensorRT LLM - Blackwell & Hopper Hardware</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../deployment-guide/deployment-guide-for-qwen3-next-on-trtllm.html">Deployment Guide for Qwen3 Next on TensorRT LLM - Blackwell & Hopper Hardware</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../deployment-guide/deployment-guide-for-kimi-k2-thinking-on-trtllm.html">Deployment Guide for Kimi K2 Thinking on TensorRT LLM - Blackwell</a></li>
|
||
</ul>
|
||
</details></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Models</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="../models/supported-models.html">Supported Models</a></li>
|
||
|
||
<li class="toctree-l1"><a class="reference internal" href="../models/adding-new-model.html">Adding a New Model</a></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">CLI Reference</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="../commands/trtllm-bench.html">trtllm-bench</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../commands/trtllm-eval.html">trtllm-eval</a></li>
|
||
<li class="toctree-l1 has-children"><a class="reference internal" href="../commands/trtllm-serve/index.html">trtllm-serve</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="../commands/trtllm-serve/trtllm-serve.html">trtllm-serve</a></li>
|
||
<li class="toctree-l2"><a class="reference internal" href="../commands/trtllm-serve/run-benchmark-with-trtllm-serve.html">Run benchmarking with <code class="docutils literal notranslate"><span class="pre">trtllm-serve</span></code></a></li>
|
||
</ul>
|
||
</details></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">API Reference</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="../llm-api/index.html">LLM 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">Features</span></p>
|
||
<ul class="current nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="feature-combination-matrix.html">Feature Combination Matrix</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="attention.html">Multi-Head, Multi-Query, and Group-Query Attention</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="disagg-serving.html">Disaggregated Serving</a></li>
|
||
<li class="toctree-l1 current active"><a class="current reference internal" href="#">KV Cache System</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="long-sequence.html">Long Sequences</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="lora.html">LoRA (Low-Rank Adaptation)</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="multi-modality.html">Multimodal Support in TensorRT LLM</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="overlap-scheduler.html">Overlap Scheduler</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="paged-attention-ifb-scheduler.html">Paged Attention, IFB, and Request Scheduling</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="parallel-strategy.html">Parallelism in TensorRT LLM</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="quantization.html">Quantization</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="sampling.html">Sampling</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="additional-outputs.html">Additional Outputs</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="guided-decoding.html">Guided Decoding</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="speculative-decoding.html">Speculative Decoding</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="checkpoint-loading.html">Checkpoint Loading</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="auto_deploy/auto-deploy.html">AutoDeploy (Beta)</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="ray-orchestrator.html">Ray Orchestrator (Prototype)</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="torch_compile_and_piecewise_cuda_graph.html">Torch Compile & Piecewise CUDA Graph</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="helix.html">Helix Parallelism</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="kv-cache-connector.html">KV Cache Connector</a></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Developer Guide</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="../developer-guide/overview.html">Architecture Overview</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../developer-guide/perf-analysis.html">Performance Analysis</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../developer-guide/perf-benchmarking.html">TensorRT LLM Benchmarking</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../developer-guide/ci-overview.html">Continuous Integration Overview</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../developer-guide/dev-containers.html">Using Dev Containers</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../developer-guide/api-change.html">LLM API Change Guide</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../developer-guide/kv-transfer.html">Introduction to KV Cache Transmission</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/tech_blog/blog10_ADP_Balance_Strategy.html">ADP Balance Strategy</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog11_GPT_OSS_Eagle3.html">Running GPT-OSS-120B with Eagle3 Speculative Decoding on GB200/B200 (TensorRT LLM)</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog12_Combining_Guided_Decoding_and_Speculative_Decoding.html">Combining Guided Decoding and Speculative Decoding: Making CPU and GPU Cooperate Seamlessly</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog13_Inference_Time_Compute_Implementation_in_TensorRT-LLM.html">Inference Time Compute Implementation in TensorRT LLM</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog14_Scaling_Expert_Parallelism_in_TensorRT-LLM_part3.html">Scaling Expert Parallelism in TensorRT LLM (Part 3: Pushing the Performance Boundary)</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>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog3_Optimizing_DeepSeek_R1_Throughput_on_NVIDIA_Blackwell_GPUs.html">Optimizing DeepSeek R1 Throughput on NVIDIA Blackwell GPUs: A Deep Dive for Developers</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog4_Scaling_Expert_Parallelism_in_TensorRT-LLM.html">Scaling Expert Parallelism in TensorRT LLM (Part 1: Design and Implementation of Large-scale EP)</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog5_Disaggregated_Serving_in_TensorRT-LLM.html">Disaggregated Serving in TensorRT LLM</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog6_Llama4_maverick_eagle_guide.html">How to launch Llama4 Maverick + Eagle3 TensorRT LLM server</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog7_NGram_performance_Analysis_And_Auto_Enablement.html">N-Gram Speculative Decoding in TensorRT LLM</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog8_Scaling_Expert_Parallelism_in_TensorRT-LLM_part2.html">Scaling Expert Parallelism in TensorRT LLM (Part 2: Performance Status and Optimization)</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/tech_blog/blog9_Deploying_GPT_OSS_on_TRTLLM.html">Running a High Performance GPT-OSS-120B Inference Server with TensorRT LLM</a></li>
|
||
<li class="toctree-l1"><a class="reference internal" href="../blogs/Best_perf_practice_on_DeepSeek-R1_in_TensorRT-LLM.html">How to get best performance on DeepSeek-R1 in TensorRT LLM</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/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/H100vsA100.html">H100 has 4.6x A100 Performance in TensorRT LLM, achieving 10,000 tok/s at 100ms to first token</a></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Quick Links</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/TensorRT-LLM/releases">Releases</a></li>
|
||
<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/TensorRT-LLM">Github Code</a></li>
|
||
<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/TensorRT-LLM/issues?q=is%3Aissue%20state%3Aopen%20label%3Aroadmap">Roadmap</a></li>
|
||
</ul>
|
||
<p aria-level="2" class="caption" role="heading"><span class="caption-text">Use TensorRT Engine</span></p>
|
||
<ul class="nav bd-sidenav">
|
||
<li class="toctree-l1"><a class="reference internal" href="../legacy/tensorrt_quickstart.html">LLM API with TensorRT Engine</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">KV Cache System</span></li>
|
||
</ul>
|
||
</nav>
|
||
</div>
|
||
|
||
</div>
|
||
|
||
|
||
</div>
|
||
</div>
|
||
|
||
|
||
|
||
|
||
<div id="searchbox"></div>
|
||
<article class="bd-article">
|
||
|
||
<section class="tex2jax_ignore mathjax_ignore" id="kv-cache-system">
|
||
<h1>KV Cache System<a class="headerlink" href="#kv-cache-system" title="Link to this heading">#</a></h1>
|
||
<p>The KV cache stores previously computed key-value pairs for reuse during generation in order to avoid redundant calculations. The TensorRT LLM KV cache system also supports reuse across requests and uses a suite of tools like offloading and prioritized eviction to increase reuse. It supports variable attention window sizes and Multi-Head Attention (MHA) optimization techniques such as MQA and GQA.</p>
|
||
<section id="the-basics">
|
||
<h2>The Basics<a class="headerlink" href="#the-basics" title="Link to this heading">#</a></h2>
|
||
<p>The KV cache is a pool of blocks that can hold KV state for a fixed number of tokens. Multiple layers are packed within a single block, which requires all the layers to have the same number of heads and the same attention window size. A separate pool is created for each combination of attention window size and number of heads to support variable attention window size and optimization techniques like GQA.</p>
|
||
<p>The number of tokens that can be stored in a single block can be set by user when the model engine is created. It must be a power of two greater than 1. Blocks are assigned to requests as needed. Blocks are stored in a search structure as they are filled by requests, this allows later requests to reuse KV state if they have a matching prefix.</p>
|
||
<p>If more than one pool is created, available memory is divided among the pools. The fraction to assign to each pool is determined during initialization and is static. This is not optimal and we are working on providing a better solution.</p>
|
||
</section>
|
||
<section id="reuse-across-requests">
|
||
<h2>Reuse Across Requests<a class="headerlink" href="#reuse-across-requests" title="Link to this heading">#</a></h2>
|
||
<p>Blocks containing KV state computed for previous requests are stored in a radix search tree as soon as they are filled. A search is performed when a new request is added, and matched blocks are reused instead of calculated. Blocks that are reused can be shared among multiple requests, so reuse saves memory as well as computations.</p>
|
||
<p>Blocks remain reusable until they are evicted from the search tree. Eviction happens when a new (blank) block is needed. The core eviction scheme is prioritized LRU. All blocks are assigned a priority between 0 and 100 (100 being most important). All blocks of the lowest priority must be evicted before any blocks of the next priority can be evicted. If all blocks have the same priority, the least recently used block is evicted.</p>
|
||
<p>When a block is evicted from primary memory, its KV state is copied to a block in secondary memory. The secondary memory block remains in the search tree, so the block remains reusable until it is evicted from secondary memory. Eviction from secondary memory happens when a new block in secondary memory is needed to offload a primary block. The eviction scheme is the same for primary and secondary blocks.</p>
|
||
<p>One caveat in the current code is that only leaf blocks can be evicted (leaves are blocks with no descendants in the radix tree). This design works well for full attention layers, but not for limited attention layers. This will be fixed in a future version.</p>
|
||
<section id="retention-policy">
|
||
<h3>Retention Policy<a class="headerlink" href="#retention-policy" title="Link to this heading">#</a></h3>
|
||
<p>Blocks are assigned priority in line with the <a class="reference external" href="https://nvidia.github.io/TensorRT-LLM/llm-api/reference.html#tensorrt_llm.llmapi.KvCacheRetentionConfig">retention policy</a> of the request. Blocks with lower priority scores will be freed preferentially to blocks with higher priority. The retention policy is a list of <a class="reference external" href="https://nvidia.github.io/TensorRT-LLM/llm-api/reference.html#tensorrt_llm.llmapi.KvCacheRetentionConfig.TokenRangeRetentionConfig">TokenRangeRetentionConfig</a> objects, each specifying priority for a given range of tokens, such as “assign priority X to tokens 10 through 61”. You can also assign a duration in milliseconds for this to remain in effect. Priority reverts to the default of 35 after a period of <code class="docutils literal notranslate"><span class="pre">duration_ms</span></code> has elapsed from the first time the block was made available for reuse. TokenRangeRetentionConfig only applies to input (prompt) tokens. The property <code class="docutils literal notranslate"><span class="pre">decode_retention_policy</span></code> specifies what priority to assign to blocks with generated (decoded) tokens and <code class="docutils literal notranslate"><span class="pre">decode_duration_ms</span></code> specifies how long this should remain in effect. Priority reverts to the default after expiration. Any property that expects a duration can be set to None. This indicates that particular part of the retention policy never expires.</p>
|
||
<p>Not in use: <code class="docutils literal notranslate"><span class="pre">transfer_mode</span></code> is a debug option and should not be used.</p>
|
||
<p>See <a class="reference internal" href="../examples/kvcacheretentionconfig.html"><span class="std std-doc">this example</span></a> for an example of how to change block priorities of specific requests by altering their retention policy.</p>
|
||
</section>
|
||
<section id="speculative-decoding">
|
||
<h3>Speculative Decoding<a class="headerlink" href="#speculative-decoding" title="Link to this heading">#</a></h3>
|
||
<p>Reuse across requests is supported by all speculative decoding models. Please see <a class="reference internal" href="speculative-decoding.html"><span class="std std-doc">speculative decoding</span></a> for more details.</p>
|
||
</section>
|
||
</section>
|
||
<section id="limited-attention-window-size">
|
||
<h2>Limited Attention Window Size<a class="headerlink" href="#limited-attention-window-size" title="Link to this heading">#</a></h2>
|
||
<p>TensorRT LLM takes advantage of layers with limited attention window size in order to reduce computations and memory usage. Blocks that leave the attention window are freed and placed on the radix search tree so they can be reused.</p>
|
||
</section>
|
||
<section id="mqa-gqa">
|
||
<h2>MQA / GQA<a class="headerlink" href="#mqa-gqa" title="Link to this heading">#</a></h2>
|
||
<p>TensorRT LLM takes advantage of grouped query attention in order to save memory. KV cache will create blocks with only enough space to store state for the discrete query head groups. For MHA, there is one group per head, for MQA there is a single group for all the heads. GQA strikes a balance between these two.</p>
|
||
</section>
|
||
<section id="controlling-kv-cache-behavior">
|
||
<h2>Controlling KV Cache Behavior<a class="headerlink" href="#controlling-kv-cache-behavior" title="Link to this heading">#</a></h2>
|
||
<p>Many of the features in the KV cache system are optional or have user defined properties that alter how they work. Users can control KV cache features through class <a class="reference external" href="https://nvidia.github.io/TensorRT-LLM/llm-api/reference.html#tensorrt_llm.llmapi.KvCacheConfig">KVCacheConfig</a>. The remainder of this section describes how to change the most important behaviors of the KV cache system.</p>
|
||
<p>See <a class="reference internal" href="../examples/kvcacheconfig.html"><span class="std std-doc">this example</span></a> for an example of how to use KvCacheConfig to control KV cache behavior.</p>
|
||
<section id="datatype">
|
||
<h3>Datatype<a class="headerlink" href="#datatype" title="Link to this heading">#</a></h3>
|
||
<p>Perhaps the most important property is <code class="docutils literal notranslate"><span class="pre">dtype</span></code> which specifies what data type is held in KV cache. The default ‘auto’ specifies that data type should be inferred from model config.</p>
|
||
</section>
|
||
<section id="how-much-memory-is-allocated-to-kv-cache">
|
||
<h3>How Much Memory is Allocated to KV Cache<a class="headerlink" href="#how-much-memory-is-allocated-to-kv-cache" title="Link to this heading">#</a></h3>
|
||
<p>Property <code class="docutils literal notranslate"><span class="pre">free_gpu_memory_fraction</span></code> is a ratio > 0 and < 1 that specifies how much of free GPU memory should be allocated to KV cache. The default is 90% (ratio of 0.9). If <code class="docutils literal notranslate"><span class="pre">max_tokens</span></code> is also set, KV cache will determine how much memory is needed to hold <code class="docutils literal notranslate"><span class="pre">max_tokens</span></code> and will allocate the lesser of <code class="docutils literal notranslate"><span class="pre">max_tokens</span></code> and <code class="docutils literal notranslate"><span class="pre">free_gpu_memory_fraction</span></code>.</p>
|
||
</section>
|
||
<section id="enable-disable-cross-request-reuse">
|
||
<h3>Enable/Disable Cross Request Reuse<a class="headerlink" href="#enable-disable-cross-request-reuse" title="Link to this heading">#</a></h3>
|
||
<p>Block reuse across requests is enabled by default, but can be disabled by setting <code class="docutils literal notranslate"><span class="pre">enable_block_reuse</span></code> to False.</p>
|
||
</section>
|
||
<section id="kv-cache-salting-for-secure-reuse">
|
||
<h3>KV Cache Salting for Secure Reuse<a class="headerlink" href="#kv-cache-salting-for-secure-reuse" title="Link to this heading">#</a></h3>
|
||
<p>KV cache salting provides a security mechanism to control which requests can reuse cached KV states. When a <code class="docutils literal notranslate"><span class="pre">cache_salt</span></code> parameter is provided with a request, the KV cache system will only allow reuse of cached blocks given the same cache salt value. This prevents potential security issues such as prompt theft attacks, where malicious users might try to infer information from cached states of other users’ requests.</p>
|
||
<p>To use cache salting, specify the <code class="docutils literal notranslate"><span class="pre">cache_salt</span></code> parameter as a string when creating requests. Only requests with matching cache salt values can share cached KV blocks. The salt value can be any non-empty string, such as a user ID, tenant ID, or hash string.</p>
|
||
</section>
|
||
<section id="enable-offloading-to-host-memory">
|
||
<h3>Enable Offloading to Host Memory<a class="headerlink" href="#enable-offloading-to-host-memory" title="Link to this heading">#</a></h3>
|
||
<p>Before a block is evicted from GPU memory, it can optionally be offloaded to host (CPU) memory. The block remains reusable until it is evicted from host memory. When an offloaded block is reused, it is first copied back into GPU memory. Offloading is controlled with property <code class="docutils literal notranslate"><span class="pre">host_cache_size</span></code> which specifies how much host memory (in bytes) should be allocated for offloading. The default is 0.</p>
|
||
<p>When offloading is enabled, the client can prevent specific blocks from being offloaded by toggling block priority. Blocks with lower priority than a certain threshold are not offloaded; they are evicted directly from GPU memory to reduce traffic between GPU and host. This priority is set with <code class="docutils literal notranslate"><span class="pre">secondary_offload_min_priority</span></code>. Default value is 35, meaning any block with lower priority than 35 will not be offloaded.</p>
|
||
<p>Here is an <a class="reference download internal" download="" href="../_downloads/289baa7eb9a76c4944f3e2800f51f4c4/llm_kv_cache_offloading.py"><span class="xref download myst">example</span></a> to show how to enable host offloading.</p>
|
||
</section>
|
||
<section id="partial-reuse">
|
||
<h3>Partial Reuse<a class="headerlink" href="#partial-reuse" title="Link to this heading">#</a></h3>
|
||
<p>Partial reuse of a block can happen when some but not all tokens are matched. It is enabled by default, but can be disabled by setting <code class="docutils literal notranslate"><span class="pre">enable_partial_reuse</span></code> to False.</p>
|
||
<p>The property <code class="docutils literal notranslate"><span class="pre">copy_on_partial_reuse</span></code> specifies whether a block should be copied or not in order to allow partial reuse. If copying is disabled, a partially matched block can only be reused if no other request is using it. If copying is enabled, partially matched blocks are not reused directly, instead a new block is created and the matched tokens are copied into the new block. This allows multiple requests to partially reuse a block.</p>
|
||
</section>
|
||
<section id="attention-window-size">
|
||
<h3>Attention Window Size<a class="headerlink" href="#attention-window-size" title="Link to this heading">#</a></h3>
|
||
<p>Property <code class="docutils literal notranslate"><span class="pre">max_attention_window</span></code> specifies the maximum attention window size for each layer in the model as a list of integer values. If the length of this list is less than number of layers, the list is repeated as many times as necessary. For instance, if the model has only full attention layers and maximum sequence length is 4096, you can specify this as <code class="docutils literal notranslate"><span class="pre">max_attention_window</span> <span class="pre">=</span> <span class="pre">[4096]</span></code>. If the first layer is full attention, the second layer is limited attention with window size 256 and then this repeats for the remaining layers, you specify this as <code class="docutils literal notranslate"><span class="pre">max_attention_window</span> <span class="pre">=</span> <span class="pre">[4096,256]</span></code>. This means first layer is full attention, second layer is limited attention, third layer is full attention, fourth layer is limited attention and so on.</p>
|
||
</section>
|
||
<section id="deprecated-properties">
|
||
<h3>Deprecated Properties<a class="headerlink" href="#deprecated-properties" title="Link to this heading">#</a></h3>
|
||
<p>Properties <code class="docutils literal notranslate"><span class="pre">use_uvm</span></code> and <code class="docutils literal notranslate"><span class="pre">sink_token_length</span></code> have been deprecated and will be removed in a future release.</p>
|
||
</section>
|
||
</section>
|
||
</section>
|
||
|
||
|
||
</article>
|
||
|
||
|
||
|
||
|
||
|
||
<footer class="prev-next-footer d-print-none">
|
||
|
||
<div class="prev-next-area">
|
||
<a class="left-prev"
|
||
href="disagg-serving.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 Serving</p>
|
||
</div>
|
||
</a>
|
||
<a class="right-next"
|
||
href="long-sequence.html"
|
||
title="next page">
|
||
<div class="prev-next-info">
|
||
<p class="prev-next-subtitle">next</p>
|
||
<p class="prev-next-title">Long Sequences</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="#the-basics">The Basics</a></li>
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#reuse-across-requests">Reuse Across Requests</a><ul class="nav section-nav flex-column">
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#retention-policy">Retention Policy</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#speculative-decoding">Speculative Decoding</a></li>
|
||
</ul>
|
||
</li>
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#limited-attention-window-size">Limited Attention Window Size</a></li>
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#mqa-gqa">MQA / GQA</a></li>
|
||
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#controlling-kv-cache-behavior">Controlling KV Cache Behavior</a><ul class="nav section-nav flex-column">
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#datatype">Datatype</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#how-much-memory-is-allocated-to-kv-cache">How Much Memory is Allocated to KV Cache</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#enable-disable-cross-request-reuse">Enable/Disable Cross Request Reuse</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#kv-cache-salting-for-secure-reuse">KV Cache Salting for Secure Reuse</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#enable-offloading-to-host-memory">Enable Offloading to Host Memory</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#partial-reuse">Partial Reuse</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#attention-window-size">Attention Window Size</a></li>
|
||
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#deprecated-properties">Deprecated Properties</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/">Your Privacy Choices</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 January 04, 2026.</p>
|
||
|
||
<p>This page is generated by TensorRT-LLM commit <a href="https://github.com/NVIDIA/TensorRT-LLM/tree/a65b0d4">a65b0d4</a>.</p>
|
||
|
||
</div></div>
|
||
|
||
</div>
|
||
|
||
|
||
|
||
</div>
|
||
|
||
</footer>
|
||
</body>
|
||
</html> |