<!DOCTYPE html
PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" lang="en-us" xml:lang="en-us">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8"></meta>
<meta http-equiv="X-UA-Compatible" content="IE=edge"></meta>
<meta name="copyright" content="(C) Copyright 2005"></meta>
<meta name="DC.rights.owner" content="(C) Copyright 2005"></meta>
<meta name="DC.Type" content="cppModule"></meta>
<meta name="DC.Title" content="Stream Management"></meta>
<meta name="abstract" content=""></meta>
<meta name="description" content=""></meta>
<meta name="DC.Format" content="XHTML"></meta>
<meta name="DC.Identifier" content="group__CUDA__STREAM"></meta>
<link rel="stylesheet" type="text/css" href="../common/formatting/commonltr.css"></link>
<link rel="stylesheet" type="text/css" href="../common/formatting/site.css"></link>
<title>CUDA Driver API :: CUDA Toolkit Documentation</title>
<!--[if lt IE 9]>
<script src="../common/formatting/html5shiv-printshiv.min.js"></script>
<![endif]-->
<script type="text/javascript" charset="utf-8" src="//assets.adobedtm.com/b92787824f2e0e9b68dc2e993f9bd995339fe417/satelliteLib-7ba51e58dc61bcb0e9311aadd02a0108ab24cc6c.js"></script>
<script type="text/javascript" charset="utf-8" src="../common/formatting/jquery.min.js"></script>
<script type="text/javascript" charset="utf-8" src="../common/formatting/jquery.ba-hashchange.min.js"></script>
<script type="text/javascript" charset="utf-8" src="../common/formatting/jquery.scrollintoview.min.js"></script>
<script type="text/javascript" src="../search/htmlFileList.js"></script>
<script type="text/javascript" src="../search/htmlFileInfoList.js"></script>
<script type="text/javascript" src="../search/nwSearchFnt.min.js"></script>
<script type="text/javascript" src="../search/stemmers/en_stemmer.min.js"></script>
<script type="text/javascript" src="../search/index-1.js"></script>
<script type="text/javascript" src="../search/index-2.js"></script>
<script type="text/javascript" src="../search/index-3.js"></script>
<link rel="canonical" href="http://docs.nvidia.com/cuda/cuda-driver-api/index.html"></link>
<link rel="stylesheet" type="text/css" href="../common/formatting/qwcode.highlight.css"></link>
</head>
<body>
<header id="header"><span id="company">NVIDIA</span><span id="site-title">CUDA Toolkit Documentation</span><form id="search" method="get" action="search">
<input type="text" name="search-text"></input><fieldset id="search-location">
<legend>Search In:</legend>
<label><input type="radio" name="search-type" value="site"></input>Entire Site</label>
<label><input type="radio" name="search-type" value="document"></input>Just This Document</label></fieldset>
<button type="reset">clear search</button>
<button id="submit" type="submit">search</button></form>
</header>
<div id="site-content">
<nav id="site-nav">
<div class="category closed"><a href="../index.html" title="The root of the site.">CUDA Toolkit
v10.1.168</a></div>
<div class="category"><a href="index.html" title="CUDA Driver API">CUDA Driver API</a></div>
<ul>
<li>
<div class="section-link"><a href="driver-vs-runtime-api.html#driver-vs-runtime-api">1. Difference between the driver and runtime APIs </a></div>
</li>
<li>
<div class="section-link"><a href="api-sync-behavior.html#api-sync-behavior">2. API synchronization behavior </a></div>
</li>
<li>
<div class="section-link"><a href="stream-sync-behavior.html#stream-sync-behavior">3. Stream synchronization behavior</a></div>
</li>
<li>
<div class="section-link"><a href="graphs-thread-safety.html#graphs-thread-safety">4. Graph object thread safety</a></div>
</li>
<li>
<div class="section-link"><a href="modules.html#modules">5. Modules</a></div>
<ul>
<li>
<div class="section-link"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES">5.1. Data types used by CUDA driver</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__ERROR.html#group__CUDA__ERROR">5.2. Error Handling</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__INITIALIZE.html#group__CUDA__INITIALIZE">5.3. Initialization</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__VERSION.html#group__CUDA__VERSION">5.4. Version Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__DEVICE.html#group__CUDA__DEVICE">5.5. Device Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__DEVICE__DEPRECATED.html#group__CUDA__DEVICE__DEPRECATED">5.6. Device Management [DEPRECATED]</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__PRIMARY__CTX.html#group__CUDA__PRIMARY__CTX">5.7. Primary Context Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__CTX.html#group__CUDA__CTX">5.8. Context Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__CTX__DEPRECATED.html#group__CUDA__CTX__DEPRECATED">5.9. Context Management [DEPRECATED]</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__MODULE.html#group__CUDA__MODULE">5.10. Module Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__MEM.html#group__CUDA__MEM">5.11. Memory Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__UNIFIED.html#group__CUDA__UNIFIED">5.12. Unified Addressing</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__STREAM.html#group__CUDA__STREAM">5.13. Stream Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__EVENT.html#group__CUDA__EVENT">5.14. Event Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__EXTRES__INTEROP.html#group__CUDA__EXTRES__INTEROP">5.15. External Resource Interoperability</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__MEMOP.html#group__CUDA__MEMOP">5.16. Stream memory operations</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__EXEC.html#group__CUDA__EXEC">5.17. Execution Control</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__EXEC__DEPRECATED.html#group__CUDA__EXEC__DEPRECATED">5.18. Execution Control [DEPRECATED]</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__GRAPH.html#group__CUDA__GRAPH">5.19. Graph Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__OCCUPANCY.html#group__CUDA__OCCUPANCY">5.20. Occupancy</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__TEXREF__DEPRECATED.html#group__CUDA__TEXREF__DEPRECATED">5.21. Texture Reference Management [DEPRECATED]</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__SURFREF__DEPRECATED.html#group__CUDA__SURFREF__DEPRECATED">5.22. Surface Reference Management [DEPRECATED]</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__TEXOBJECT.html#group__CUDA__TEXOBJECT">5.23. Texture Object Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__SURFOBJECT.html#group__CUDA__SURFOBJECT">5.24. Surface Object Management</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__PEER__ACCESS.html#group__CUDA__PEER__ACCESS">5.25. Peer Context Memory Access</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__GRAPHICS.html#group__CUDA__GRAPHICS">5.26. Graphics Interoperability</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__PROFILER.html#group__CUDA__PROFILER">5.27. Profiler Control</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__GL.html#group__CUDA__GL">5.28. OpenGL Interoperability</a></div>
<ul>
<li>
<div class="section-link"><a href="group__CUDA__GL__DEPRECATED.html#group__CUDA__GL__DEPRECATED">5.28.1. OpenGL Interoperability [DEPRECATED]</a></div>
</li>
</ul>
</li>
<li>
<div class="section-link"><a href="group__CUDA__D3D9.html#group__CUDA__D3D9">5.29. Direct3D 9 Interoperability</a></div>
<ul>
<li>
<div class="section-link"><a href="group__CUDA__D3D9__DEPRECATED.html#group__CUDA__D3D9__DEPRECATED">5.29.1. Direct3D 9 Interoperability [DEPRECATED]</a></div>
</li>
</ul>
</li>
<li>
<div class="section-link"><a href="group__CUDA__D3D10.html#group__CUDA__D3D10">5.30. Direct3D 10 Interoperability</a></div>
<ul>
<li>
<div class="section-link"><a href="group__CUDA__D3D10__DEPRECATED.html#group__CUDA__D3D10__DEPRECATED">5.30.1. Direct3D 10 Interoperability [DEPRECATED]</a></div>
</li>
</ul>
</li>
<li>
<div class="section-link"><a href="group__CUDA__D3D11.html#group__CUDA__D3D11">5.31. Direct3D 11 Interoperability</a></div>
<ul>
<li>
<div class="section-link"><a href="group__CUDA__D3D11__DEPRECATED.html#group__CUDA__D3D11__DEPRECATED">5.31.1. Direct3D 11 Interoperability [DEPRECATED]</a></div>
</li>
</ul>
</li>
<li>
<div class="section-link"><a href="group__CUDA__VDPAU.html#group__CUDA__VDPAU">5.32. VDPAU Interoperability</a></div>
</li>
<li>
<div class="section-link"><a href="group__CUDA__EGL.html#group__CUDA__EGL">5.33. EGL Interoperability</a></div>
</li>
</ul>
</li>
<li>
<div class="section-link"><a href="annotated.html#annotated">6. Data Structures</a></div>
<ul>
<li>
<div class="section-link"><a href="structCUDA__ARRAY3D__DESCRIPTOR.html#structCUDA__ARRAY3D__DESCRIPTOR">6.1. CUDA_ARRAY3D_DESCRIPTOR</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__ARRAY__DESCRIPTOR.html#structCUDA__ARRAY__DESCRIPTOR">6.2. CUDA_ARRAY_DESCRIPTOR</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__EXTERNAL__MEMORY__BUFFER__DESC.html#structCUDA__EXTERNAL__MEMORY__BUFFER__DESC">6.3. CUDA_EXTERNAL_MEMORY_BUFFER_DESC</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__EXTERNAL__MEMORY__HANDLE__DESC.html#structCUDA__EXTERNAL__MEMORY__HANDLE__DESC">6.4. CUDA_EXTERNAL_MEMORY_HANDLE_DESC</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__EXTERNAL__MEMORY__MIPMAPPED__ARRAY__DESC.html#structCUDA__EXTERNAL__MEMORY__MIPMAPPED__ARRAY__DESC">6.5. CUDA_EXTERNAL_MEMORY_MIPMAPPED_ARRAY_DESC</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__EXTERNAL__SEMAPHORE__HANDLE__DESC.html#structCUDA__EXTERNAL__SEMAPHORE__HANDLE__DESC">6.6. CUDA_EXTERNAL_SEMAPHORE_HANDLE_DESC</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__EXTERNAL__SEMAPHORE__SIGNAL__PARAMS.html#structCUDA__EXTERNAL__SEMAPHORE__SIGNAL__PARAMS">6.7. CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__EXTERNAL__SEMAPHORE__WAIT__PARAMS.html#structCUDA__EXTERNAL__SEMAPHORE__WAIT__PARAMS">6.8. CUDA_EXTERNAL_SEMAPHORE_WAIT_PARAMS</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__HOST__NODE__PARAMS.html#structCUDA__HOST__NODE__PARAMS">6.9. CUDA_HOST_NODE_PARAMS</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__KERNEL__NODE__PARAMS.html#structCUDA__KERNEL__NODE__PARAMS">6.10. CUDA_KERNEL_NODE_PARAMS</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__LAUNCH__PARAMS.html#structCUDA__LAUNCH__PARAMS">6.11. CUDA_LAUNCH_PARAMS</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__MEMCPY2D.html#structCUDA__MEMCPY2D">6.12. CUDA_MEMCPY2D</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__MEMCPY3D.html#structCUDA__MEMCPY3D">6.13. CUDA_MEMCPY3D</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__MEMCPY3D__PEER.html#structCUDA__MEMCPY3D__PEER">6.14. CUDA_MEMCPY3D_PEER</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__MEMSET__NODE__PARAMS.html#structCUDA__MEMSET__NODE__PARAMS">6.15. CUDA_MEMSET_NODE_PARAMS</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__POINTER__ATTRIBUTE__P2P__TOKENS.html#structCUDA__POINTER__ATTRIBUTE__P2P__TOKENS">6.16. CUDA_POINTER_ATTRIBUTE_P2P_TOKENS</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__RESOURCE__DESC.html#structCUDA__RESOURCE__DESC">6.17. CUDA_RESOURCE_DESC</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__RESOURCE__VIEW__DESC.html#structCUDA__RESOURCE__VIEW__DESC">6.18. CUDA_RESOURCE_VIEW_DESC</a></div>
</li>
<li>
<div class="section-link"><a href="structCUDA__TEXTURE__DESC.html#structCUDA__TEXTURE__DESC">6.19. CUDA_TEXTURE_DESC</a></div>
</li>
<li>
<div class="section-link"><a href="structCUdevprop.html#structCUdevprop">6.20. CUdevprop</a></div>
</li>
<li>
<div class="section-link"><a href="structCUeglFrame.html#structCUeglFrame">6.21. CUeglFrame</a></div>
</li>
<li>
<div class="section-link"><a href="structCUipcEventHandle.html#structCUipcEventHandle">6.22. CUipcEventHandle</a></div>
</li>
<li>
<div class="section-link"><a href="structCUipcMemHandle.html#structCUipcMemHandle">6.23. CUipcMemHandle</a></div>
</li>
<li>
<div class="section-link"><a href="unionCUstreamBatchMemOpParams.html#unionCUstreamBatchMemOpParams">6.24. CUstreamBatchMemOpParams</a></div>
</li>
</ul>
</li>
<li>
<div class="section-link"><a href="functions.html#functions">7. Data Fields</a></div>
</li>
<li>
<div class="section-link"><a href="deprecated.html#deprecated">8. Deprecated List</a></div>
</li>
<li>
<div class="section-link"><a href="notices-header.html#notices-header">Notices</a></div>
<ul></ul>
</li>
</ul>
</nav>
<div id="resize-nav"></div>
<nav id="search-results">
<h2>Search Results</h2>
<ol></ol>
</nav>
<div id="contents-container">
<div id="breadcrumbs-container">
<div id="breadcrumbs"><a href="group__CUDA__UNIFIED.html" shape="rect">< Previous</a> | <a href="group__CUDA__EVENT.html" shape="rect">Next ></a></div>
<div id="release-info">CUDA Driver API
(<a href="../../pdf/CUDA_Driver_API.pdf">PDF</a>)
-
v10.1.168
(<a href="https://developer.nvidia.com/cuda-toolkit-archive">older</a>)
-
Last updated April 24, 2019
-
<a href="mailto:CUDAIssues@nvidia.com?subject=CUDA Toolkit Documentation Feedback: CUDA Driver API">Send Feedback</a></div>
</div>
<article id="contents">
<div class="topic reference apiRef apiPackage cppModule" id="group__CUDA__STREAM"><a name="group__CUDA__STREAM" shape="rect">
<!-- --></a><h2 class="topictitle2 cppModule">5.13. Stream Management</h2>
<div class="section">
<p>This section describes the stream management functions of the low-level CUDA driver application programming interface. </p>
</div>
<h3 class="fake_sectiontitle member_header">Functions</h3>
<dl class="members">
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" shape="rect">cuStreamAddCallback</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ge5743a8c48527f1040107a68205c5ba9" title="" shape="rect">CUstreamCallback</a><span> </span><span class="keyword keyword apiItemName">callback</span>, void*<span> </span><span class="keyword keyword apiItemName">userData</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Add a callback to a compute stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g6e468d680e263e7eba02a56643c50533" shape="rect">cuStreamAttachMemAsync</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">dptr</span>, size_t<span> </span><span class="keyword keyword apiItemName">length</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Attach memory to a stream asynchronously. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" shape="rect">cuStreamBeginCapture</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gd984ba65031d63f1ed11ec76728c2669" title="" shape="rect">CUstreamCaptureMode</a><span> </span><span class="keyword keyword apiItemName">mode</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Begins graph capture on a stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" shape="rect">cuStreamCreate</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a>*<span> </span><span class="keyword keyword apiItemName">phStream</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">Flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Create a stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" shape="rect">cuStreamCreateWithPriority</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a>*<span> </span><span class="keyword keyword apiItemName">phStream</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span>, int <span> </span><span class="keyword keyword apiItemName">priority</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Create a stream with the given priority. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" shape="rect">cuStreamDestroy</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Destroys a stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" shape="rect">cuStreamEndCapture</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g69f555c38df5b3fa1ed25efef794739a" title="" shape="rect">CUgraph</a>*<span> </span><span class="keyword keyword apiItemName">phGraph</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Ends capture on a stream, returning the captured graph. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g13145ece1d79a1d79a1d22abb9663216" shape="rect">cuStreamGetCaptureInfo</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g7480c0f2bd19894e54fcd2c04d6efb91" title="" shape="rect">CUstreamCaptureStatus</a>*<span> </span><span class="keyword keyword apiItemName">captureStatus</span>, cuuint64_t*<span> </span><span class="keyword keyword apiItemName">id</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Query capture status of a stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g1107907025eaa3387fdc590a9379a681" shape="rect">cuStreamGetCtx</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gf9f5bd81658f866613785b3a0bb7d7d9" title="" shape="rect">CUcontext</a>*<span> </span><span class="keyword keyword apiItemName">pctx</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Query the context associated with a stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g4d39786855a6bed01215c1907fbbfbb7" shape="rect">cuStreamGetFlags</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, unsigned int*<span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Query the flags of a given stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g5bd5cb26915a2ecf1921807339488484" shape="rect">cuStreamGetPriority</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, int*<span> </span><span class="keyword keyword apiItemName">priority</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Query the priority of a given stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g37823c49206e3704ae23c7ad78560bca" shape="rect">cuStreamIsCapturing</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g7480c0f2bd19894e54fcd2c04d6efb91" title="" shape="rect">CUstreamCaptureStatus</a>*<span> </span><span class="keyword keyword apiItemName">captureStatus</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Returns a stream's capture status. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" shape="rect">cuStreamQuery</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Determine status of a compute stream. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" shape="rect">cuStreamSynchronize</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Wait until a stream's tasks are completed. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" shape="rect">cuStreamWaitEvent</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g6d740185cf0953636d4ae37f68d7559b" title="" shape="rect">CUevent</a><span> </span><span class="keyword keyword apiItemName">hEvent</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">Flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Make a compute stream wait on an event. </span></dd>
<dt><span class="member_type"><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> </span><span class="member_name"><a href="#group__CUDA__STREAM_1g378135b262f02a43a7caeab239ae493d" shape="rect">cuThreadExchangeStreamCaptureMode</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gd984ba65031d63f1ed11ec76728c2669" title="" shape="rect">CUstreamCaptureMode</a>*<span> </span><span class="keyword keyword apiItemName">mode</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Swaps the stream capture interaction mode for a thread. </span></dd>
</dl>
<div class="description">
<h3 class="sectiontitle">Functions</h3>
<dl class="description">
<dt class="description"><a name="group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" id="group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamAddCallback ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ge5743a8c48527f1040107a68205c5ba9" title="" shape="rect">CUstreamCallback</a><span> </span><span class="keyword keyword apiItemName">callback</span>, void*<span> </span><span class="keyword keyword apiItemName">userData</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="description">
<div class="section">Add a callback to a compute stream. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream to add callback to </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">callback</span></tt></dt>
<dd>- The function to call once preceding stream operations are complete </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">userData</span></tt></dt>
<dd>- User specified data to be passed to the callback function </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>- Reserved for future use, must be 0</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e954756ae7ade0dfd09faeccb513dd831b" shape="rect">CUDA_ERROR_NOT_SUPPORTED</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>
<div class="note note"><span class="notetitle">Note:</span><p class="p">This function is slated for eventual deprecation and removal. If you do not require the callback to execute in case of a device
error, consider using <a class="xref" href="group__CUDA__EXEC.html#group__CUDA__EXEC_1gab95a78143bae7f21eebb978f91e7f3f" title="Enqueues a host function call in a stream." shape="rect">cuLaunchHostFunc</a>. Additionally, this function is not supported with <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a> and <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" title="Ends capture on a stream, returning the captured graph." shape="rect">cuStreamEndCapture</a>, unlike <a class="xref" href="group__CUDA__EXEC.html#group__CUDA__EXEC_1gab95a78143bae7f21eebb978f91e7f3f" title="Enqueues a host function call in a stream." shape="rect">cuLaunchHostFunc</a>.
</p>
</div>
</p>
<p class="p">
Adds a callback to be called on the host after all currently enqueued items in the stream have completed. For each cuStreamAddCallback
call, the callback will be executed exactly once. The callback will block later work in the stream until it is finished.
</p>
<p class="p">The callback may be passed <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a> or an error code. In the event of a device error, all subsequently executed callbacks will receive an appropriate <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" shape="rect">CUresult</a>.
</p>
<p class="p">Callbacks must not make any CUDA API calls. Attempting to use a CUDA API will result in <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9e23d0197c490ec332a43e55b167968a3" shape="rect">CUDA_ERROR_NOT_PERMITTED</a>. Callbacks must not perform any synchronization that may depend on outstanding device work or other callbacks that are not
mandated to run earlier. Callbacks without a mandated order (in independent streams) execute in undefined order and may be
serialized.
</p>
<p class="p">For the purposes of Unified Memory, callback execution makes a number of guarantees:
<ul class="ul">
<li class="li">
<p class="p">The callback stream is considered idle for the duration of the callback. Thus, for example, a callback may always use memory
attached to the callback stream.
</p>
</li>
<li class="li">
<p class="p">The start of execution of a callback has the same effect as synchronizing an event recorded in the same stream immediately
prior to the callback. It thus synchronizes streams which have been "joined" prior to the callback.
</p>
</li>
<li class="li">
<p class="p">Adding device work to any stream does not have the effect of making the stream active until all preceding host functions and
stream callbacks have executed. Thus, for example, a callback might use global attached memory even if work has been added
to another stream, if the work has been ordered behind the callback with an event.
</p>
</li>
<li class="li">
<p class="p">Completion of a callback does not cause a stream to become active except as described above. The callback stream will remain
idle if no device work follows the callback, and will remain idle across consecutive callbacks without device work in between.
Thus, for example, stream synchronization can be done by signaling from a callback at the end of the stream.
</p>
</li>
</ul>
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><ul class="ul">
<li class="li">
<p class="p">This function uses standard <a class="xref xmlonly" href="stream-sync-behavior.html#stream-sync-behavior__default-stream" shape="rect">default stream</a> semantics.
</p>
</li>
<li class="li">
<p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</li>
</ul>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" title="Determine status of a compute stream." shape="rect">cuStreamQuery</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" title="Make a compute stream wait on an event." shape="rect">cuStreamWaitEvent</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gb347ded34dc326af404aa02af5388a32" title="Allocates memory that will be automatically managed by the Unified Memory system." shape="rect">cuMemAllocManaged</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6e468d680e263e7eba02a56643c50533" title="Attach memory to a stream asynchronously." shape="rect">cuStreamAttachMemAsync</a>, cuStreamLaunchHostFunc, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1g74aa9f4b1c2f12d994bf13876a5a2498" target="_blank" shape="rect">cudaStreamAddCallback</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g6e468d680e263e7eba02a56643c50533" id="group__CUDA__STREAM_1g6e468d680e263e7eba02a56643c50533" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamAttachMemAsync ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">dptr</span>, size_t<span> </span><span class="keyword keyword apiItemName">length</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="description">
<div class="section">Attach memory to a stream asynchronously. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream in which to enqueue the attach operation </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">dptr</span></tt></dt>
<dd>- Pointer to memory (must be a pointer to managed memory or to a valid host-accessible region of system-allocated pageable
memory)
</dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">length</span></tt></dt>
<dd>- Length of memory </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>- Must be one of <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g17c5d5f9b585aa2d6f121847d1a78f4c" shape="rect">CUmemAttach_flags</a></dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e954756ae7ade0dfd09faeccb513dd831b" shape="rect">CUDA_ERROR_NOT_SUPPORTED</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Enqueues an operation in <tt class="ph tt code">hStream</tt> to specify stream association of <tt class="ph tt code">length</tt> bytes of memory starting from <tt class="ph tt code">dptr</tt>. This function is a stream-ordered operation, meaning that it is dependent on, and will only take effect when, previous work
in stream has completed. Any previous association is automatically replaced.
</p>
<p class="p"><tt class="ph tt code">dptr</tt> must point to one of the following types of memories:
<ul class="ul">
<li class="li">
<p class="p">managed memory declared using the __managed__ keyword or allocated with <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gb347ded34dc326af404aa02af5388a32" title="Allocates memory that will be automatically managed by the Unified Memory system." shape="rect">cuMemAllocManaged</a>.
</p>
</li>
<li class="li">
<p class="p">a valid host-accessible region of system-allocated pageable memory. This type of memory may only be specified if the device
associated with the stream reports a non-zero value for the device attribute <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a35fdcdbe1dfc3ad5ec428c279e0efb9cd" shape="rect">CU_DEVICE_ATTRIBUTE_PAGEABLE_MEMORY_ACCESS</a>.
</p>
</li>
</ul>
</p>
<p class="p">For managed allocations, <tt class="ph tt code">length</tt> must be either zero or the entire allocation's size. Both indicate that the entire allocation's stream association is being
changed. Currently, it is not possible to change stream association for a portion of a managed allocation.
</p>
<p class="p">For pageable host allocations, <tt class="ph tt code">length</tt> must be non-zero.
</p>
<p class="p">The stream association is specified using <tt class="ph tt code">flags</tt> which must be one of <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g17c5d5f9b585aa2d6f121847d1a78f4c" shape="rect">CUmemAttach_flags</a>. If the <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg17c5d5f9b585aa2d6f121847d1a78f4c0b42aae6a29b41b734d4c0dea6c33313" shape="rect">CU_MEM_ATTACH_GLOBAL</a> flag is specified, the memory can be accessed by any stream on any device. If the <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg17c5d5f9b585aa2d6f121847d1a78f4c8b59c62cab9c7a762338e5fae92e2e9c" shape="rect">CU_MEM_ATTACH_HOST</a> flag is specified, the program makes a guarantee that it won't access the memory on the device from any stream on a device
that has a zero value for the device attribute <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a333110e44c9cb6ead02f03ff6f6fd495e" shape="rect">CU_DEVICE_ATTRIBUTE_CONCURRENT_MANAGED_ACCESS</a>. If the <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg17c5d5f9b585aa2d6f121847d1a78f4c4b70b6a5e039f61eccc6b8db3dfac442" shape="rect">CU_MEM_ATTACH_SINGLE</a> flag is specified and <tt class="ph tt code">hStream</tt> is associated with a device that has a zero value for the device attribute <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a333110e44c9cb6ead02f03ff6f6fd495e" shape="rect">CU_DEVICE_ATTRIBUTE_CONCURRENT_MANAGED_ACCESS</a>, the program makes a guarantee that it will only access the memory on the device from <tt class="ph tt code">hStream</tt>. It is illegal to attach singly to the NULL stream, because the NULL stream is a virtual global stream and not a specific
stream. An error will be returned in this case.
</p>
<p class="p">When memory is associated with a single stream, the Unified Memory system will allow CPU access to this memory region so long
as all operations in <tt class="ph tt code">hStream</tt> have completed, regardless of whether other streams are active. In effect, this constrains exclusive ownership of the managed
memory region by an active GPU to per-stream activity instead of whole-GPU activity.
</p>
<p class="p">Accessing memory on the device from streams that are not associated with it will produce undefined results. No error checking
is performed by the Unified Memory system to ensure that kernels launched into other streams do not access this region.
</p>
<p class="p">It is a program's responsibility to order calls to <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6e468d680e263e7eba02a56643c50533" title="Attach memory to a stream asynchronously." shape="rect">cuStreamAttachMemAsync</a> via events, synchronization or other means to ensure legal access to memory at all times. Data visibility and coherency will
be changed appropriately for all kernels which follow a stream-association change.
</p>
<p class="p">If <tt class="ph tt code">hStream</tt> is destroyed while data is associated with it, the association is removed and the association reverts to the default visibility
of the allocation as specified at <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gb347ded34dc326af404aa02af5388a32" title="Allocates memory that will be automatically managed by the Unified Memory system." shape="rect">cuMemAllocManaged</a>. For __managed__ variables, the default association is always <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg17c5d5f9b585aa2d6f121847d1a78f4c0b42aae6a29b41b734d4c0dea6c33313" shape="rect">CU_MEM_ATTACH_GLOBAL</a>. Note that destroying a stream is an asynchronous operation, and as a result, the change to default association won't happen
until all work in the stream has completed.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><ul class="ul">
<li class="li">
<p class="p">This function uses standard <a class="xref xmlonly" href="stream-sync-behavior.html#stream-sync-behavior__default-stream" shape="rect">default stream</a> semantics.
</p>
</li>
<li class="li">
<p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</li>
</ul>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" title="Determine status of a compute stream." shape="rect">cuStreamQuery</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" title="Make a compute stream wait on an event." shape="rect">cuStreamWaitEvent</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gb347ded34dc326af404aa02af5388a32" title="Allocates memory that will be automatically managed by the Unified Memory system." shape="rect">cuMemAllocManaged</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__HIGHLEVEL.html#group__CUDART__HIGHLEVEL_1g496353d630c29c44a2e33f531a3944d1" target="_blank" shape="rect">cudaStreamAttachMemAsync</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" id="group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamBeginCapture ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gd984ba65031d63f1ed11ec76728c2669" title="" shape="rect">CUstreamCaptureMode</a><span> </span><span class="keyword keyword apiItemName">mode</span> ) </span></dt>
<dd class="description">
<div class="section">Begins graph capture on a stream. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream in which to initiate capture </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">mode</span></tt></dt>
<dd>- Controls the interaction of this capture sequence with other API calls that are potentially unsafe. For more details see
<a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g378135b262f02a43a7caeab239ae493d" title="Swaps the stream capture interaction mode for a thread." shape="rect">cuThreadExchangeStreamCaptureMode</a>.
</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Begin graph capture on <tt class="ph tt code">hStream</tt>. When a stream is in capture mode, all operations pushed into the stream will not be executed, but will instead be captured
into a graph, which will be returned via <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" title="Ends capture on a stream, returning the captured graph." shape="rect">cuStreamEndCapture</a>. Capture may not be initiated if <tt class="ph tt code">stream</tt> is CU_STREAM_LEGACY. Capture must be ended on the same stream in which it was initiated, and it may only be initiated if
the stream is not already in capture mode. The capture mode may be queried via <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g37823c49206e3704ae23c7ad78560bca" title="Returns a stream's capture status." shape="rect">cuStreamIsCapturing</a>. A unique id representing the capture sequence may be queried via <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g13145ece1d79a1d79a1d22abb9663216" title="Query capture status of a stream." shape="rect">cuStreamGetCaptureInfo</a>.
</p>
<p class="p">If <tt class="ph tt code">mode</tt> is not CU_STREAM_CAPTURE_MODE_RELAXED, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" title="Ends capture on a stream, returning the captured graph." shape="rect">cuStreamEndCapture</a> must be called on this stream from the same thread.
</p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Kernels captured using this API must not use texture and surface references. Reading or writing through any texture or surface
reference is undefined behavior. This restriction does not apply to texture and surface objects.
</p>
</div>
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g37823c49206e3704ae23c7ad78560bca" title="Returns a stream's capture status." shape="rect">cuStreamIsCapturing</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" title="Ends capture on a stream, returning the captured graph." shape="rect">cuStreamEndCapture</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g378135b262f02a43a7caeab239ae493d" title="Swaps the stream capture interaction mode for a thread." shape="rect">cuThreadExchangeStreamCaptureMode</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" id="group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamCreate ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a>*<span> </span><span class="keyword keyword apiItemName">phStream</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">Flags</span> ) </span></dt>
<dd class="description">
<div class="section">Create a stream. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">phStream</span></tt></dt>
<dd>- Returned newly created stream </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">Flags</span></tt></dt>
<dd>- Parameters for stream creation</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9264c50688ed110e8476b591befe60c02" shape="rect">CUDA_ERROR_OUT_OF_MEMORY</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Creates a stream and returns a handle in <tt class="ph tt code">phStream</tt>. The <tt class="ph tt code">Flags</tt> argument determines behaviors of the stream. Valid values for <tt class="ph tt code">Flags</tt> are:
<ul class="ul">
<li class="li">
<p class="p"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg775cb4ffbb7adf91e190067d9ad1752aaa5df0ec96f491f1be1124fdf265a066" shape="rect">CU_STREAM_DEFAULT</a>: Default stream creation flag.
</p>
</li>
<li class="li">
<p class="p"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg775cb4ffbb7adf91e190067d9ad1752a89727d1d315214a6301abe98b419aff6" shape="rect">CU_STREAM_NON_BLOCKING</a>: Specifies that work running in the created stream may run concurrently with work in stream 0 (the NULL stream), and that
the created stream should perform no implicit synchronization with stream 0.
</p>
</li>
</ul>
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" title="Create a stream with the given priority." shape="rect">cuStreamCreateWithPriority</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g5bd5cb26915a2ecf1921807339488484" title="Query the priority of a given stream." shape="rect">cuStreamGetPriority</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g4d39786855a6bed01215c1907fbbfbb7" title="Query the flags of a given stream." shape="rect">cuStreamGetFlags</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" title="Make a compute stream wait on an event." shape="rect">cuStreamWaitEvent</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" title="Determine status of a compute stream." shape="rect">cuStreamQuery</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" title="Add a callback to a compute stream." shape="rect">cuStreamAddCallback</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1g6a3c4b819e6a994c26d0c4824a4c80da" target="_blank" shape="rect">cudaStreamCreate</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1gb1e32aff9f59119e4d0a9858991c4ad3" target="_blank" shape="rect">cudaStreamCreateWithFlags</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" id="group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamCreateWithPriority ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a>*<span> </span><span class="keyword keyword apiItemName">phStream</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span>, int <span> </span><span class="keyword keyword apiItemName">priority</span> ) </span></dt>
<dd class="description">
<div class="section">Create a stream with the given priority. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">phStream</span></tt></dt>
<dd>- Returned newly created stream </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>- Flags for stream creation. See <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a> for a list of valid flags
</dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">priority</span></tt></dt>
<dd>- Stream priority. Lower numbers represent higher priorities. See <a class="xref" href="group__CUDA__CTX.html#group__CUDA__CTX_1g137920ab61a71be6ce67605b9f294091" title="Returns numerical values that correspond to the least and greatest stream priorities." shape="rect">cuCtxGetStreamPriorityRange</a> for more information about meaningful stream priorities that can be passed.
</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9264c50688ed110e8476b591befe60c02" shape="rect">CUDA_ERROR_OUT_OF_MEMORY</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Creates a stream with the specified priority and returns a handle in <tt class="ph tt code">phStream</tt>. This API alters the scheduler priority of work in the stream. Work in a higher priority stream may preempt work already
executing in a low priority stream.
</p>
<p class="p"><tt class="ph tt code">priority</tt> follows a convention where lower numbers represent higher priorities. '0' represents default priority. The range of meaningful
numerical priorities can be queried using <a class="xref" href="group__CUDA__CTX.html#group__CUDA__CTX_1g137920ab61a71be6ce67605b9f294091" title="Returns numerical values that correspond to the least and greatest stream priorities." shape="rect">cuCtxGetStreamPriorityRange</a>. If the specified priority is outside the numerical range returned by <a class="xref" href="group__CUDA__CTX.html#group__CUDA__CTX_1g137920ab61a71be6ce67605b9f294091" title="Returns numerical values that correspond to the least and greatest stream priorities." shape="rect">cuCtxGetStreamPriorityRange</a>, it will automatically be clamped to the lowest or the highest number in the range.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><ul class="ul">
<li class="li">
<p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</li>
<li class="li">
<p class="p">Stream priorities are supported only on GPUs with compute capability 3.5 or higher.</p>
</li>
<li class="li">
<p class="p">In the current implementation, only compute kernels launched in priority streams are affected by the stream's priority. Stream
priorities have no effect on host-to-device and device-to-host memory operations.
</p>
</li>
</ul>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g5bd5cb26915a2ecf1921807339488484" title="Query the priority of a given stream." shape="rect">cuStreamGetPriority</a>, <a class="xref" href="group__CUDA__CTX.html#group__CUDA__CTX_1g137920ab61a71be6ce67605b9f294091" title="Returns numerical values that correspond to the least and greatest stream priorities." shape="rect">cuCtxGetStreamPriorityRange</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g4d39786855a6bed01215c1907fbbfbb7" title="Query the flags of a given stream." shape="rect">cuStreamGetFlags</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" title="Make a compute stream wait on an event." shape="rect">cuStreamWaitEvent</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" title="Determine status of a compute stream." shape="rect">cuStreamQuery</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" title="Add a callback to a compute stream." shape="rect">cuStreamAddCallback</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1ge2be9e9858849bf62ba4a8b66d1c3540" target="_blank" shape="rect">cudaStreamCreateWithPriority</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" id="group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamDestroy ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span> ) </span></dt>
<dd class="description">
<div class="section">Destroys a stream. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream to destroy</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Destroys the stream specified by <tt class="ph tt code">hStream</tt>.
</p>
<p class="p">In case the device is still doing work in the stream <tt class="ph tt code">hStream</tt> when <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy()</a> is called, the function will return immediately and the resources associated with <tt class="ph tt code">hStream</tt> will be released automatically once the device has completed all work in <tt class="ph tt code">hStream</tt>.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" title="Make a compute stream wait on an event." shape="rect">cuStreamWaitEvent</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" title="Determine status of a compute stream." shape="rect">cuStreamQuery</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" title="Add a callback to a compute stream." shape="rect">cuStreamAddCallback</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1gfda584f1788ca983cb21c5f4d2033a62" target="_blank" shape="rect">cudaStreamDestroy</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" id="group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamEndCapture ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g69f555c38df5b3fa1ed25efef794739a" title="" shape="rect">CUgraph</a>*<span> </span><span class="keyword keyword apiItemName">phGraph</span> ) </span></dt>
<dd class="description">
<div class="section">Ends capture on a stream, returning the captured graph. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream to query </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">phGraph</span></tt></dt>
<dd>- The captured graph</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e979282fa9b0bd6a56167b5ddf44391440" shape="rect">CUDA_ERROR_STREAM_CAPTURE_WRONG_THREAD</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>End capture on <tt class="ph tt code">hStream</tt>, returning the captured graph via <tt class="ph tt code">phGraph</tt>. Capture must have been initiated on <tt class="ph tt code">hStream</tt> via a call to <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a>. If capture was invalidated, due to a violation of the rules of stream capture, then a NULL graph will be returned.
</p>
<p class="p">If the <tt class="ph tt code">mode</tt> argument to <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a> was not CU_STREAM_CAPTURE_MODE_RELAXED, this call must be from the same thread as <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a>.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g37823c49206e3704ae23c7ad78560bca" title="Returns a stream's capture status." shape="rect">cuStreamIsCapturing</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g13145ece1d79a1d79a1d22abb9663216" id="group__CUDA__STREAM_1g13145ece1d79a1d79a1d22abb9663216" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamGetCaptureInfo ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g7480c0f2bd19894e54fcd2c04d6efb91" title="" shape="rect">CUstreamCaptureStatus</a>*<span> </span><span class="keyword keyword apiItemName">captureStatus</span>, cuuint64_t*<span> </span><span class="keyword keyword apiItemName">id</span> ) </span></dt>
<dd class="description">
<div class="section">Query capture status of a stream. </div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9960a55453736ec87ca941f9bc2d80abe" shape="rect">CUDA_ERROR_STREAM_CAPTURE_IMPLICIT</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Query the capture status of a stream and and get an id for the capture sequence, which is unique over the lifetime of the
process.
</p>
<p class="p">If called on <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ga53e8210837f039dd6434a3a4c3324aa" shape="rect">CU_STREAM_LEGACY</a> (the "null stream") while a stream not created with <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg775cb4ffbb7adf91e190067d9ad1752a89727d1d315214a6301abe98b419aff6" shape="rect">CU_STREAM_NON_BLOCKING</a> is capturing, returns <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9960a55453736ec87ca941f9bc2d80abe" shape="rect">CUDA_ERROR_STREAM_CAPTURE_IMPLICIT</a>.
</p>
<p class="p">A valid id is returned only if both of the following are true:
<ul class="ul">
<li class="li">
<p class="p">the call returns CUDA_SUCCESS</p>
</li>
<li class="li">
<p class="p">captureStatus is set to <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg7480c0f2bd19894e54fcd2c04d6efb91c799fa3d867e2b300dfc45a6e90bc15d" shape="rect">CU_STREAM_CAPTURE_STATUS_ACTIVE</a></p>
</li>
</ul>
</p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g37823c49206e3704ae23c7ad78560bca" title="Returns a stream's capture status." shape="rect">cuStreamIsCapturing</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g1107907025eaa3387fdc590a9379a681" id="group__CUDA__STREAM_1g1107907025eaa3387fdc590a9379a681" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamGetCtx ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gf9f5bd81658f866613785b3a0bb7d7d9" title="" shape="rect">CUcontext</a>*<span> </span><span class="keyword keyword apiItemName">pctx</span> ) </span></dt>
<dd class="description">
<div class="section">Query the context associated with a stream. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Handle to the stream to be queried </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">pctx</span></tt></dt>
<dd>- Returned context associated with the stream</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a>,
</p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Returns the CUDA context that the stream is associated with.</p>
<p class="p">The stream handle <tt class="ph tt code">hStream</tt> can refer to any of the following:
<ul class="ul">
<li class="li">
<p class="p">a stream created via any of the CUDA driver APIs such as <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a> and <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" title="Create a stream with the given priority." shape="rect">cuStreamCreateWithPriority</a>, or their runtime API equivalents such as <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1g6a3c4b819e6a994c26d0c4824a4c80da" target="_blank" shape="rect">cudaStreamCreate</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1gb1e32aff9f59119e4d0a9858991c4ad3" target="_blank" shape="rect">cudaStreamCreateWithFlags</a> and <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1ge2be9e9858849bf62ba4a8b66d1c3540" target="_blank" shape="rect">cudaStreamCreateWithPriority</a>. The returned context is the context that was active in the calling thread when the stream was created. Passing an invalid
handle will result in undefined behavior.
</p>
</li>
<li class="li">
<p class="p">any of the special streams such as the NULL stream, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ga53e8210837f039dd6434a3a4c3324aa" shape="rect">CU_STREAM_LEGACY</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g02e40b82600f62c42ed29abb150f857c" shape="rect">CU_STREAM_PER_THREAD</a>. The runtime API equivalents of these are also accepted, which are NULL, <a class="xref" href="../cuda-runtime-api/group__CUDART__TYPES.html#group__CUDART__TYPES_1g4e62d09dde16ba457b0a97f3a5262246" target="_blank" shape="rect">cudaStreamLegacy</a> and <a class="xref" href="../cuda-runtime-api/group__CUDART__TYPES.html#group__CUDART__TYPES_1g7b7129befd6f52708309acafd1c46197" target="_blank" shape="rect">cudaStreamPerThread</a> respectively. Specifying any of the special handles will return the context current to the calling thread. If no context
is current to the calling thread, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a> is returned.
</p>
</li>
</ul>
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" title="Create a stream with the given priority." shape="rect">cuStreamCreateWithPriority</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g5bd5cb26915a2ecf1921807339488484" title="Query the priority of a given stream." shape="rect">cuStreamGetPriority</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g4d39786855a6bed01215c1907fbbfbb7" title="Query the flags of a given stream." shape="rect">cuStreamGetFlags</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" title="Make a compute stream wait on an event." shape="rect">cuStreamWaitEvent</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" title="Determine status of a compute stream." shape="rect">cuStreamQuery</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" title="Add a callback to a compute stream." shape="rect">cuStreamAddCallback</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1g6a3c4b819e6a994c26d0c4824a4c80da" target="_blank" shape="rect">cudaStreamCreate</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1gb1e32aff9f59119e4d0a9858991c4ad3" target="_blank" shape="rect">cudaStreamCreateWithFlags</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g4d39786855a6bed01215c1907fbbfbb7" id="group__CUDA__STREAM_1g4d39786855a6bed01215c1907fbbfbb7" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamGetFlags ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, unsigned int*<span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="description">
<div class="section">Query the flags of a given stream. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Handle to the stream to be queried </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>- Pointer to an unsigned integer in which the stream's flags are returned The value returned in <tt class="ph tt code">flags</tt> is a logical 'OR' of all flags that were used while creating this stream. See <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a> for the list of valid flags
</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9264c50688ed110e8476b591befe60c02" shape="rect">CUDA_ERROR_OUT_OF_MEMORY</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Query the flags of a stream created using <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a> or <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" title="Create a stream with the given priority." shape="rect">cuStreamCreateWithPriority</a> and return the flags in <tt class="ph tt code">flags</tt>.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g5bd5cb26915a2ecf1921807339488484" title="Query the priority of a given stream." shape="rect">cuStreamGetPriority</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1ga7f311f88126d751b9a7d3302ad6d0f8" target="_blank" shape="rect">cudaStreamGetFlags</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g5bd5cb26915a2ecf1921807339488484" id="group__CUDA__STREAM_1g5bd5cb26915a2ecf1921807339488484" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamGetPriority ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, int*<span> </span><span class="keyword keyword apiItemName">priority</span> ) </span></dt>
<dd class="description">
<div class="section">Query the priority of a given stream. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Handle to the stream to be queried </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">priority</span></tt></dt>
<dd>- Pointer to a signed integer in which the stream's priority is returned </dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9264c50688ed110e8476b591befe60c02" shape="rect">CUDA_ERROR_OUT_OF_MEMORY</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Query the priority of a stream created using <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a> or <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" title="Create a stream with the given priority." shape="rect">cuStreamCreateWithPriority</a> and return the priority in <tt class="ph tt code">priority</tt>. Note that if the stream was created with a priority outside the numerical range returned by <a class="xref" href="group__CUDA__CTX.html#group__CUDA__CTX_1g137920ab61a71be6ce67605b9f294091" title="Returns numerical values that correspond to the least and greatest stream priorities." shape="rect">cuCtxGetStreamPriorityRange</a>, this function returns the clamped priority. See <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" title="Create a stream with the given priority." shape="rect">cuStreamCreateWithPriority</a> for details about priority clamping.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g95c1a8c7c3dacb13091692dd9c7f7471" title="Create a stream with the given priority." shape="rect">cuStreamCreateWithPriority</a>, <a class="xref" href="group__CUDA__CTX.html#group__CUDA__CTX_1g137920ab61a71be6ce67605b9f294091" title="Returns numerical values that correspond to the least and greatest stream priorities." shape="rect">cuCtxGetStreamPriorityRange</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g4d39786855a6bed01215c1907fbbfbb7" title="Query the flags of a given stream." shape="rect">cuStreamGetFlags</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1g192bb727d15c4407c119747de7d198a6" target="_blank" shape="rect">cudaStreamGetPriority</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g37823c49206e3704ae23c7ad78560bca" id="group__CUDA__STREAM_1g37823c49206e3704ae23c7ad78560bca" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamIsCapturing ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g7480c0f2bd19894e54fcd2c04d6efb91" title="" shape="rect">CUstreamCaptureStatus</a>*<span> </span><span class="keyword keyword apiItemName">captureStatus</span> ) </span></dt>
<dd class="description">
<div class="section">Returns a stream's capture status. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream to query </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">captureStatus</span></tt></dt>
<dd>- Returns the stream's capture status</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9960a55453736ec87ca941f9bc2d80abe" shape="rect">CUDA_ERROR_STREAM_CAPTURE_IMPLICIT</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Return the capture status of <tt class="ph tt code">hStream</tt> via <tt class="ph tt code">captureStatus</tt>. After a successful call, <tt class="ph tt code">*captureStatus</tt> will contain one of the following:
<ul class="ul">
<li class="li">
<p class="p"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg7480c0f2bd19894e54fcd2c04d6efb91e4023001f651dbdd3e3f55a1afc87fb3" shape="rect">CU_STREAM_CAPTURE_STATUS_NONE</a>: The stream is not capturing.
</p>
</li>
<li class="li">
<p class="p"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg7480c0f2bd19894e54fcd2c04d6efb91c799fa3d867e2b300dfc45a6e90bc15d" shape="rect">CU_STREAM_CAPTURE_STATUS_ACTIVE</a>: The stream is capturing.
</p>
</li>
<li class="li">
<p class="p"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg7480c0f2bd19894e54fcd2c04d6efb916b8a69837a782cd52243d481a2c6f51a" shape="rect">CU_STREAM_CAPTURE_STATUS_INVALIDATED</a>: The stream was capturing but an error has invalidated the capture sequence. The capture sequence must be terminated with
<a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" title="Ends capture on a stream, returning the captured graph." shape="rect">cuStreamEndCapture</a> on the stream where it was initiated in order to continue using <tt class="ph tt code">hStream</tt>.
</p>
</li>
</ul>
</p>
<p class="p">Note that, if this is called on <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ga53e8210837f039dd6434a3a4c3324aa" shape="rect">CU_STREAM_LEGACY</a> (the "null stream") while a blocking stream in the same context is capturing, it will return <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9960a55453736ec87ca941f9bc2d80abe" shape="rect">CUDA_ERROR_STREAM_CAPTURE_IMPLICIT</a> and <tt class="ph tt code">*captureStatus</tt> is unspecified after the call. The blocking stream capture is not invalidated.
</p>
<p class="p">When a blocking stream is capturing, the legacy stream is in an unusable state until the blocking stream capture is terminated.
The legacy stream is not supported for stream capture, but attempted use would have an implicit dependency on the capturing
stream(s).
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" title="Ends capture on a stream, returning the captured graph." shape="rect">cuStreamEndCapture</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" id="group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamQuery ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span> ) </span></dt>
<dd class="description">
<div class="section">Determine status of a compute stream. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream to query status of</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9edd9cef666ce620352e619a36b6c3f34" shape="rect">CUDA_ERROR_NOT_READY</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Returns <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a> if all operations in the stream specified by <tt class="ph tt code">hStream</tt> have completed, or <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9edd9cef666ce620352e619a36b6c3f34" shape="rect">CUDA_ERROR_NOT_READY</a> if not.
</p>
<p class="p">For the purposes of Unified Memory, a return value of <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a> is equivalent to having called <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize()</a>.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><ul class="ul">
<li class="li">
<p class="p">This function uses standard <a class="xref xmlonly" href="stream-sync-behavior.html#stream-sync-behavior__default-stream" shape="rect">default stream</a> semantics.
</p>
</li>
<li class="li">
<p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</li>
</ul>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" title="Make a compute stream wait on an event." shape="rect">cuStreamWaitEvent</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" title="Add a callback to a compute stream." shape="rect">cuStreamAddCallback</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1g2021adeb17905c7ec2a3c1bf125c5435" target="_blank" shape="rect">cudaStreamQuery</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" id="group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamSynchronize ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span> ) </span></dt>
<dd class="description">
<div class="section">Wait until a stream's tasks are completed. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream to wait for</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Waits until the device has completed all operations in the stream specified by <tt class="ph tt code">hStream</tt>. If the context was created with the <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg9f889e28a45a295b5c8ce13aa05f6cd462aebfe6432ade3feb32f1a409027852" shape="rect">CU_CTX_SCHED_BLOCKING_SYNC</a> flag, the CPU thread will block until the stream is finished with all of its tasks.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><ul class="ul">
<li class="li">
<p class="p">This function uses standard <a class="xref xmlonly" href="stream-sync-behavior.html#stream-sync-behavior__default-stream" shape="rect">default stream</a> semantics.
</p>
</li>
<li class="li">
<p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</li>
</ul>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" title="Make a compute stream wait on an event." shape="rect">cuStreamWaitEvent</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" title="Determine status of a compute stream." shape="rect">cuStreamQuery</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" title="Add a callback to a compute stream." shape="rect">cuStreamAddCallback</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1g82b5784f674c17c6df64affe618bf45e" target="_blank" shape="rect">cudaStreamSynchronize</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" id="group__CUDA__STREAM_1g6a898b652dfc6aa1d5c8d97062618b2f" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamWaitEvent ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">hStream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g6d740185cf0953636d4ae37f68d7559b" title="" shape="rect">CUevent</a><span> </span><span class="keyword keyword apiItemName">hEvent</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">Flags</span> ) </span></dt>
<dd class="description">
<div class="section">Make a compute stream wait on an event. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">hStream</span></tt></dt>
<dd>- Stream to wait </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">hEvent</span></tt></dt>
<dd>- Event to wait on (may not be NULL) </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">Flags</span></tt></dt>
<dd>- Parameters for the operation (must be 0)</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a484e9af32c1e9893ff21f0e0191a12d" shape="rect">CUDA_ERROR_INVALID_CONTEXT</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e90fd2e63260c6317ba943af0f7e4b8d21" shape="rect">CUDA_ERROR_INVALID_HANDLE</a>,
</p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Makes all future work submitted to <tt class="ph tt code">hStream</tt> wait for all work captured in <tt class="ph tt code">hEvent</tt>. See <a class="xref" href="group__CUDA__EVENT.html#group__CUDA__EVENT_1g95424d3be52c4eb95d83861b70fb89d1" title="Records an event." shape="rect">cuEventRecord()</a> for details on what is captured by an event. The synchronization will be performed efficiently on the device when applicable.
<tt class="ph tt code">hEvent</tt> may be from a different context or device than <tt class="ph tt code">hStream</tt>.
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><ul class="ul">
<li class="li">
<p class="p">This function uses standard <a class="xref xmlonly" href="stream-sync-behavior.html#stream-sync-behavior__default-stream" shape="rect">default stream</a> semantics.
</p>
</li>
<li class="li">
<p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</li>
</ul>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1ga581f0c5833e21ded8b5a56594e243f4" title="Create a stream." shape="rect">cuStreamCreate</a>, <a class="xref" href="group__CUDA__EVENT.html#group__CUDA__EVENT_1g95424d3be52c4eb95d83861b70fb89d1" title="Records an event." shape="rect">cuEventRecord</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g1b0d24bbe97fa68e4bc511fb6adfeb0b" title="Determine status of a compute stream." shape="rect">cuStreamQuery</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g15e49dd91ec15991eb7c0a741beb7dad" title="Wait until a stream's tasks are completed." shape="rect">cuStreamSynchronize</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g613d97a277d7640f4cb1c03bd51c2483" title="Add a callback to a compute stream." shape="rect">cuStreamAddCallback</a>, <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g244c8833de4596bcd31a06cdf21ee758" title="Destroys a stream." shape="rect">cuStreamDestroy</a>, <a class="xref" href="../cuda-runtime-api/group__CUDART__STREAM.html#group__CUDART__STREAM_1gc301fd024e6fd4a17074d229d4504077" target="_blank" shape="rect">cudaStreamWaitEvent</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__STREAM_1g378135b262f02a43a7caeab239ae493d" id="group__CUDA__STREAM_1g378135b262f02a43a7caeab239ae493d" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuThreadExchangeStreamCaptureMode ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gd984ba65031d63f1ed11ec76728c2669" title="" shape="rect">CUstreamCaptureMode</a>*<span> </span><span class="keyword keyword apiItemName">mode</span> ) </span></dt>
<dd class="description">
<div class="section">Swaps the stream capture interaction mode for a thread. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">mode</span></tt></dt>
<dd>- Pointer to mode value to swap with the current mode</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9a0eed720f8a87cd1c5fd1c453bc7a03d" shape="rect">CUDA_SUCCESS</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e9acf52f132faf29b473cdda6061f0f44a" shape="rect">CUDA_ERROR_DEINITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e98feb999f0af99b4a25ab26b3866f4df8" shape="rect">CUDA_ERROR_NOT_INITIALIZED</a>, <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Sets the calling thread's stream capture interaction mode to the value contained in <tt class="ph tt code">*mode</tt>, and overwrites <tt class="ph tt code">*mode</tt> with the previous mode for the thread. To facilitate deterministic behavior across function or module boundaries, callers
are encouraged to use this API in a push-pop fashion: <pre xml:space="preserve"> <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gd984ba65031d63f1ed11ec76728c2669" shape="rect">CUstreamCaptureMode</a> mode = desiredMode;
<a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g378135b262f02a43a7caeab239ae493d" title="Swaps the stream capture interaction mode for a thread." shape="rect">cuThreadExchangeStreamCaptureMode</a>(&mode);
...
<a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g378135b262f02a43a7caeab239ae493d" title="Swaps the stream capture interaction mode for a thread." shape="rect">cuThreadExchangeStreamCaptureMode</a>(&mode); // restore previous mode</pre></p>
<p class="p">During stream capture (see <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a>), some actions, such as a call to <a class="xref" href="../cuda-runtime-api/group__CUDART__MEMORY.html#group__CUDART__MEMORY_1g37d37965bfb4803b6d4e59ff26856356" target="_blank" shape="rect">cudaMalloc</a>, may be unsafe. In the case of <a class="xref" href="../cuda-runtime-api/group__CUDART__MEMORY.html#group__CUDART__MEMORY_1g37d37965bfb4803b6d4e59ff26856356" target="_blank" shape="rect">cudaMalloc</a>, the operation is not enqueued asynchronously to a stream, and is not observed by stream capture. Therefore, if the sequence
of operations captured via <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a> depended on the allocation being replayed whenever the graph is launched, the captured graph would be invalid.
</p>
<p class="p">Therefore, stream capture places restrictions on API calls that can be made within or concurrently to a <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a>-<a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g03dab8b2ba76b00718955177a929970c" title="Ends capture on a stream, returning the captured graph." shape="rect">cuStreamEndCapture</a> sequence. This behavior can be controlled via this API and flags to <a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a>.
</p>
<p class="p">A thread's mode is one of the following:
<ul class="ul">
<li class="li">
<p class="p"><tt class="ph tt code">CU_STREAM_CAPTURE_MODE_GLOBAL:</tt> This is the default mode. If the local thread has an ongoing capture sequence that was not initiated with <tt class="ph tt code">CU_STREAM_CAPTURE_MODE_RELAXED</tt> at <tt class="ph tt code">cuStreamBeginCapture</tt>, or if any other thread has a concurrent capture sequence initiated with <tt class="ph tt code">CU_STREAM_CAPTURE_MODE_GLOBAL</tt>, this thread is prohibited from potentially unsafe API calls.
</p>
</li>
<li class="li">
<p class="p"><tt class="ph tt code">CU_STREAM_CAPTURE_MODE_THREAD_LOCAL:</tt> If the local thread has an ongoing capture sequence not initiated with <tt class="ph tt code">CU_STREAM_CAPTURE_MODE_RELAXED</tt>, it is prohibited from potentially unsafe API calls. Concurrent capture sequences in other threads are ignored.
</p>
</li>
<li class="li">
<p class="p"><tt class="ph tt code">CU_STREAM_CAPTURE_MODE_RELAXED:</tt> The local thread is not prohibited from potentially unsafe API calls. Note that the thread is still prohibited from API calls
which necessarily conflict with stream capture, for example, attempting <a class="xref" href="group__CUDA__EVENT.html#group__CUDA__EVENT_1g6f0704d755066b0ee705749ae911deef" title="Queries an event's status." shape="rect">cuEventQuery</a> on an event that was last recorded inside a capture sequence.
</p>
</li>
</ul>
</p>
<p class="p"></p>
<p class="p"></p>
<p class="p">
<div class="note note"><span class="notetitle">Note:</span><p class="p">Note that this function may also return error codes from previous, asynchronous launches.</p>
</div>
</p>
<p class="p"></p>
<p class="p apiDesc_subtitle"><strong class="ph b">See also:</strong></p>
<p class="p see_subsection"><a class="xref" href="group__CUDA__STREAM.html#group__CUDA__STREAM_1g767167da0bbf07157dc20b6c258a2143" title="Begins graph capture on a stream." shape="rect">cuStreamBeginCapture</a></p>
<p class="p"></p>
</div>
</dd>
</dl>
</div>
</div>
<hr id="contents-end"></hr>
</article>
</div>
</div>
<script language="JavaScript" type="text/javascript" charset="utf-8" src="../common/formatting/common.min.js"></script>
<script language="JavaScript" type="text/javascript" charset="utf-8" src="../common/scripts/google-analytics/google-analytics-write.js"></script>
<script language="JavaScript" type="text/javascript" charset="utf-8" src="../common/scripts/google-analytics/google-analytics-tracker.js"></script>
<script type="text/javascript">var switchTo5x=true;</script><script type="text/javascript" src="http://w.sharethis.com/button/buttons.js"></script><script type="text/javascript">stLight.options({publisher: "998dc202-a267-4d8e-bce9-14debadb8d92", doNotHash: false, doNotCopy: false, hashAddressBar: false});</script><script type="text/javascript">_satellite.pageBottom();</script></body>
</html>
distrib
>
Mageia
>
7
>
x86_64
>
media
>
nonfree-updates
>
by-pkgid
>
b86a85131cc739c1c53d0b55840a4328
>
files
>
1399
