<!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 memory operations"></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__MEMOP"></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__EXTRES__INTEROP.html" shape="rect">< Previous</a> | <a href="group__CUDA__EXEC.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__MEMOP"><a name="group__CUDA__MEMOP" shape="rect">
<!-- --></a><h2 class="topictitle2 cppModule">5.16. Stream memory operations</h2>
<div class="section">
<p>This section describes the stream memory operations of the low-level CUDA driver application programming interface.</p>
<p class="p">The whole set of operations is disabled by default. Users are required to explicitly enable them, e.g. on Linux by passing
the kernel module parameter shown below: modprobe nvidia NVreg_EnableStreamMemOPs=1 There is currently no way to enable these
operations on other operating systems.
</p>
<p class="p">Users can programmatically query whether the device supports these operations with <a class="xref" href="group__CUDA__DEVICE.html#group__CUDA__DEVICE_1g9c3e1414f0ad901d3278a4d6645fc266" title="Returns information about the device." shape="rect">cuDeviceGetAttribute()</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a33fcdc07d0e356a84b6c02730df6d43db" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_MEM_OPS</a>.
</p>
<p class="p">Support for the <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggf16864e8693d888f8178067470001b212f10c357c1882ae7b0b947225c1a0d55" shape="rect">CU_STREAM_WAIT_VALUE_NOR</a> flag can be queried with <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a3be06f9c31c264e0b29be6cc6330b5047" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_WAIT_VALUE_NOR</a>.
</p>
<p class="p">Support for the <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" title="Write a value to memory." shape="rect">cuStreamWriteValue64()</a> and <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" title="Wait on a memory location." shape="rect">cuStreamWaitValue64()</a> functions, as well as for the <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggb257b534afdb704b6ebdb99c16a5b2927e4ff0402540ee028226c47858cf8a99" shape="rect">CU_STREAM_MEM_OP_WAIT_VALUE_64</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggb257b534afdb704b6ebdb99c16a5b2926e576dc8791fa4abf7a3b36e428cd8de" shape="rect">CU_STREAM_MEM_OP_WRITE_VALUE_64</a> flags, can be queried with <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a320db4c9e8d81eb05a937d4ed5c132575" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS</a>.
</p>
<p class="p">Support for both <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggf16864e8693d888f8178067470001b215dd93e7173619c943fae495568f4d771" shape="rect">CU_STREAM_WAIT_VALUE_FLUSH</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1ggb257b534afdb704b6ebdb99c16a5b292d4e126a89b57ebf4a213ee1774ed103f" shape="rect">CU_STREAM_MEM_OP_FLUSH_REMOTE_WRITES</a> requires dedicated platform hardware features and can be queried with <a class="xref" href="group__CUDA__DEVICE.html#group__CUDA__DEVICE_1g9c3e1414f0ad901d3278a4d6645fc266" title="Returns information about the device." shape="rect">cuDeviceGetAttribute()</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a34c702e528f71bf513ee0f3f32878ce9d" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_FLUSH_REMOTE_WRITES</a>.
</p>
<p class="p">Note that all memory pointers passed as parameters to these operations are device pointers. Where necessary a device pointer
should be obtained, for example with <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1g57a39e5cba26af4d06be67fc77cc62f0" title="Passes back device pointer of mapped pinned memory." shape="rect">cuMemHostGetDevicePointer()</a>.
</p>
<p class="p">None of the operations accepts pointers to managed memory buffers (<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>
</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__MEMOP_1g764c442de9b671f9dec856e8ae531ed1" shape="rect">cuStreamBatchMemOp</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">count</span>, <a href="unionCUstreamBatchMemOpParams.html#unionCUstreamBatchMemOpParams" title="" shape="rect">CUstreamBatchMemOpParams</a>*<span> </span><span class="keyword keyword apiItemName">paramArray</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Batch operations to synchronize the stream via memory operations. </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__MEMOP_1g629856339de7bc6606047385addbb398" shape="rect">cuStreamWaitValue32</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">addr</span>, cuuint32_t<span> </span><span class="keyword keyword apiItemName">value</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Wait on a memory location. </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__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" shape="rect">cuStreamWaitValue64</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">addr</span>, cuuint64_t<span> </span><span class="keyword keyword apiItemName">value</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Wait on a memory location. </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__MEMOP_1g091455366d56dc2f1f69726aafa369b0" shape="rect">cuStreamWriteValue32</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">addr</span>, cuuint32_t<span> </span><span class="keyword keyword apiItemName">value</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Write a value to memory. </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__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" shape="rect">cuStreamWriteValue64</a> ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">addr</span>, cuuint64_t<span> </span><span class="keyword keyword apiItemName">value</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Write a value to memory. </span></dd>
</dl>
<div class="description">
<h3 class="sectiontitle">Functions</h3>
<dl class="description">
<dt class="description"><a name="group__CUDA__MEMOP_1g764c442de9b671f9dec856e8ae531ed1" id="group__CUDA__MEMOP_1g764c442de9b671f9dec856e8ae531ed1" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamBatchMemOp ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">count</span>, <a href="unionCUstreamBatchMemOpParams.html#unionCUstreamBatchMemOpParams" title="" shape="rect">CUstreamBatchMemOpParams</a>*<span> </span><span class="keyword keyword apiItemName">paramArray</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="description">
<div class="section">Batch operations to synchronize the stream via memory operations. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">stream</span></tt></dt>
<dd>The stream to enqueue the operations in. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">count</span></tt></dt>
<dd>The number of operations in the array. Must be less than 256. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">paramArray</span></tt></dt>
<dd>The types and parameters of the individual operations. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>Reserved for future expansion; 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_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</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>This is a batch version of <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g629856339de7bc6606047385addbb398" title="Wait on a memory location." shape="rect">cuStreamWaitValue32()</a> and <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g091455366d56dc2f1f69726aafa369b0" title="Write a value to memory." shape="rect">cuStreamWriteValue32()</a>. Batching operations may avoid some performance overhead in both the API call and the device execution versus adding them
to the stream in separate API calls. The operations are enqueued in the order they appear in the array.
</p>
<p class="p">See <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb257b534afdb704b6ebdb99c16a5b292" shape="rect">CUstreamBatchMemOpType</a> for the full set of supported operations, and <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g629856339de7bc6606047385addbb398" title="Wait on a memory location." shape="rect">cuStreamWaitValue32()</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" title="Wait on a memory location." shape="rect">cuStreamWaitValue64()</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g091455366d56dc2f1f69726aafa369b0" title="Write a value to memory." shape="rect">cuStreamWriteValue32()</a>, and <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" title="Write a value to memory." shape="rect">cuStreamWriteValue64()</a> for details of specific operations.
</p>
<p class="p">Basic support for this can be queried with <a class="xref" href="group__CUDA__DEVICE.html#group__CUDA__DEVICE_1g9c3e1414f0ad901d3278a4d6645fc266" title="Returns information about the device." shape="rect">cuDeviceGetAttribute()</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a33fcdc07d0e356a84b6c02730df6d43db" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_MEM_OPS</a>. See related APIs for details on querying support for specific operations.
</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__MEMOP.html#group__CUDA__MEMOP_1g629856339de7bc6606047385addbb398" title="Wait on a memory location." shape="rect">cuStreamWaitValue32</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" title="Wait on a memory location." shape="rect">cuStreamWaitValue64</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g091455366d56dc2f1f69726aafa369b0" title="Write a value to memory." shape="rect">cuStreamWriteValue32</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" title="Write a value to memory." shape="rect">cuStreamWriteValue64</a>, <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__MEMOP_1g629856339de7bc6606047385addbb398" id="group__CUDA__MEMOP_1g629856339de7bc6606047385addbb398" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamWaitValue32 ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">addr</span>, cuuint32_t<span> </span><span class="keyword keyword apiItemName">value</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="description">
<div class="section">Wait on a memory location. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">stream</span></tt></dt>
<dd>The stream to synchronize on the memory location. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">addr</span></tt></dt>
<dd>The memory location to wait on. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">value</span></tt></dt>
<dd>The value to compare with the memory location. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>See <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gf16864e8693d888f8178067470001b21" shape="rect">CUstreamWaitValue_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_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</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 a synchronization of the stream on the given memory location. Work ordered after the operation will block until the
given condition on the memory is satisfied. By default, the condition is to wait for (int32_t)(*addr - value) >= 0, a cyclic
greater-or-equal. Other condition types can be specified via <tt class="ph tt code">flags</tt>.
</p>
<p class="p">If the memory was registered via <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister()</a>, the device pointer should be obtained with <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1g57a39e5cba26af4d06be67fc77cc62f0" title="Passes back device pointer of mapped pinned memory." shape="rect">cuMemHostGetDevicePointer()</a>. This function cannot be used with managed memory (<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>
<p class="p">Support for this can be queried with <a class="xref" href="group__CUDA__DEVICE.html#group__CUDA__DEVICE_1g9c3e1414f0ad901d3278a4d6645fc266" title="Returns information about the device." shape="rect">cuDeviceGetAttribute()</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a33fcdc07d0e356a84b6c02730df6d43db" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_MEM_OPS</a>.
</p>
<p class="p">Support for CU_STREAM_WAIT_VALUE_NOR can be queried with <a class="xref" href="group__CUDA__DEVICE.html#group__CUDA__DEVICE_1g9c3e1414f0ad901d3278a4d6645fc266" title="Returns information about the device." shape="rect">cuDeviceGetAttribute()</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a3be06f9c31c264e0b29be6cc6330b5047" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_WAIT_VALUE_NOR</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__MEMOP.html#group__CUDA__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" title="Wait on a memory location." shape="rect">cuStreamWaitValue64</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g091455366d56dc2f1f69726aafa369b0" title="Write a value to memory." shape="rect">cuStreamWriteValue32</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" title="Write a value to memory." shape="rect">cuStreamWriteValue64</a><a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g764c442de9b671f9dec856e8ae531ed1" title="Batch operations to synchronize the stream via memory operations." shape="rect">cuStreamBatchMemOp</a>, <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister</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></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" id="group__CUDA__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamWaitValue64 ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">addr</span>, cuuint64_t<span> </span><span class="keyword keyword apiItemName">value</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="description">
<div class="section">Wait on a memory location. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">stream</span></tt></dt>
<dd>The stream to synchronize on the memory location. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">addr</span></tt></dt>
<dd>The memory location to wait on. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">value</span></tt></dt>
<dd>The value to compare with the memory location. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>See <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gf16864e8693d888f8178067470001b21" shape="rect">CUstreamWaitValue_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_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</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 a synchronization of the stream on the given memory location. Work ordered after the operation will block until the
given condition on the memory is satisfied. By default, the condition is to wait for (int64_t)(*addr - value) >= 0, a cyclic
greater-or-equal. Other condition types can be specified via <tt class="ph tt code">flags</tt>.
</p>
<p class="p">If the memory was registered via <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister()</a>, the device pointer should be obtained with <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1g57a39e5cba26af4d06be67fc77cc62f0" title="Passes back device pointer of mapped pinned memory." shape="rect">cuMemHostGetDevicePointer()</a>.
</p>
<p class="p">Support for this can be queried with <a class="xref" href="group__CUDA__DEVICE.html#group__CUDA__DEVICE_1g9c3e1414f0ad901d3278a4d6645fc266" title="Returns information about the device." shape="rect">cuDeviceGetAttribute()</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a320db4c9e8d81eb05a937d4ed5c132575" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS</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__MEMOP.html#group__CUDA__MEMOP_1g629856339de7bc6606047385addbb398" title="Wait on a memory location." shape="rect">cuStreamWaitValue32</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g091455366d56dc2f1f69726aafa369b0" title="Write a value to memory." shape="rect">cuStreamWriteValue32</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" title="Write a value to memory." shape="rect">cuStreamWriteValue64</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g764c442de9b671f9dec856e8ae531ed1" title="Batch operations to synchronize the stream via memory operations." shape="rect">cuStreamBatchMemOp</a>, <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister</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></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__MEMOP_1g091455366d56dc2f1f69726aafa369b0" id="group__CUDA__MEMOP_1g091455366d56dc2f1f69726aafa369b0" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamWriteValue32 ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">addr</span>, cuuint32_t<span> </span><span class="keyword keyword apiItemName">value</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="description">
<div class="section">Write a value to memory. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">stream</span></tt></dt>
<dd>The stream to do the write in. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">addr</span></tt></dt>
<dd>The device address to write to. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">value</span></tt></dt>
<dd>The value to write. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>See <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g57787c039734758648d92109c7373141" shape="rect">CUstreamWriteValue_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_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</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>Write a value to memory. Unless the <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg57787c039734758648d92109c737314152598852c1a097794bc7d08576f94995" shape="rect">CU_STREAM_WRITE_VALUE_NO_MEMORY_BARRIER</a> flag is passed, the write is preceded by a system-wide memory fence, equivalent to a __threadfence_system() but scoped to
the stream rather than a CUDA thread.
</p>
<p class="p">If the memory was registered via <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister()</a>, the device pointer should be obtained with <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1g57a39e5cba26af4d06be67fc77cc62f0" title="Passes back device pointer of mapped pinned memory." shape="rect">cuMemHostGetDevicePointer()</a>. This function cannot be used with managed memory (<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>
<p class="p">Support for this can be queried with <a class="xref" href="group__CUDA__DEVICE.html#group__CUDA__DEVICE_1g9c3e1414f0ad901d3278a4d6645fc266" title="Returns information about the device." shape="rect">cuDeviceGetAttribute()</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a33fcdc07d0e356a84b6c02730df6d43db" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_STREAM_MEM_OPS</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__MEMOP.html#group__CUDA__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" title="Write a value to memory." shape="rect">cuStreamWriteValue64</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g629856339de7bc6606047385addbb398" title="Wait on a memory location." shape="rect">cuStreamWaitValue32</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" title="Wait on a memory location." shape="rect">cuStreamWaitValue64</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g764c442de9b671f9dec856e8ae531ed1" title="Batch operations to synchronize the stream via memory operations." shape="rect">cuStreamBatchMemOp</a>, <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister</a>, <a class="xref" href="group__CUDA__EVENT.html#group__CUDA__EVENT_1g95424d3be52c4eb95d83861b70fb89d1" title="Records an event." shape="rect">cuEventRecord</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDA__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" id="group__CUDA__MEMOP_1gc8af1e8b96d7561840affd5217dd6830" shape="rect">
<!-- --></a><span><a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gc6c391505e117393cc2558fff6bfc2e9" title="" shape="rect">CUresult</a> cuStreamWriteValue64 ( <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gb946c7f02e09efd788a204718015d88a" title="" shape="rect">CUstream</a><span> </span><span class="keyword keyword apiItemName">stream</span>, <a href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g5e264ce2ad6a38761e7e04921ef771de" title="" shape="rect">CUdeviceptr</a><span> </span><span class="keyword keyword apiItemName">addr</span>, cuuint64_t<span> </span><span class="keyword keyword apiItemName">value</span>, unsigned int <span> </span><span class="keyword keyword apiItemName">flags</span> ) </span></dt>
<dd class="description">
<div class="section">Write a value to memory. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">stream</span></tt></dt>
<dd>The stream to do the write in. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">addr</span></tt></dt>
<dd>The device address to write to. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">value</span></tt></dt>
<dd>The value to write. </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">flags</span></tt></dt>
<dd>See <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1g57787c039734758648d92109c7373141" shape="rect">CUstreamWriteValue_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_1ggc6c391505e117393cc2558fff6bfc2e990696c86fcee1f536a1ec7d25867feeb" shape="rect">CUDA_ERROR_INVALID_VALUE</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>Write a value to memory. Unless the <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gg57787c039734758648d92109c737314152598852c1a097794bc7d08576f94995" shape="rect">CU_STREAM_WRITE_VALUE_NO_MEMORY_BARRIER</a> flag is passed, the write is preceded by a system-wide memory fence, equivalent to a __threadfence_system() but scoped to
the stream rather than a CUDA thread.
</p>
<p class="p">If the memory was registered via <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister()</a>, the device pointer should be obtained with <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1g57a39e5cba26af4d06be67fc77cc62f0" title="Passes back device pointer of mapped pinned memory." shape="rect">cuMemHostGetDevicePointer()</a>.
</p>
<p class="p">Support for this can be queried with <a class="xref" href="group__CUDA__DEVICE.html#group__CUDA__DEVICE_1g9c3e1414f0ad901d3278a4d6645fc266" title="Returns information about the device." shape="rect">cuDeviceGetAttribute()</a> and <a class="xref" href="group__CUDA__TYPES.html#group__CUDA__TYPES_1gge12b8a782bebe21b1ac0091bf9f4e2a320db4c9e8d81eb05a937d4ed5c132575" shape="rect">CU_DEVICE_ATTRIBUTE_CAN_USE_64_BIT_STREAM_MEM_OPS</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__MEMOP.html#group__CUDA__MEMOP_1g091455366d56dc2f1f69726aafa369b0" title="Write a value to memory." shape="rect">cuStreamWriteValue32</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g629856339de7bc6606047385addbb398" title="Wait on a memory location." shape="rect">cuStreamWaitValue32</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g6910c1258c5f15aa5d699f0fd60d6933" title="Wait on a memory location." shape="rect">cuStreamWaitValue64</a>, <a class="xref" href="group__CUDA__MEMOP.html#group__CUDA__MEMOP_1g764c442de9b671f9dec856e8ae531ed1" title="Batch operations to synchronize the stream via memory operations." shape="rect">cuStreamBatchMemOp</a>, <a class="xref" href="group__CUDA__MEM.html#group__CUDA__MEM_1gf0a9fe11544326dabd743b7aa6b54223" title="Registers an existing host memory range for use by CUDA." shape="rect">cuMemHostRegister</a>, <a class="xref" href="group__CUDA__EVENT.html#group__CUDA__EVENT_1g95424d3be52c4eb95d83861b70fb89d1" title="Records an event." shape="rect">cuEventRecord</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
>
1393
