<!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="Execution Control"></meta>
<meta name="abstract" content=""></meta>
<meta name="description" content=""></meta>
<meta name="DC.Format" content="XHTML"></meta>
<meta name="DC.Identifier" content="group__CUDART__EXECUTION"></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 Runtime 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="../common/formatting/jquery.min.js"></script>
<script type="text/javascript" charset="utf-8" src="../common/formatting/jquery.ba-hashchange.min.js"></script>
<link rel="canonical" href="http://docs.nvidia.com/cuda/cuda-runtime-api/index.html"></link>
<link rel="stylesheet" type="text/css" href="../common/formatting/qwcode.highlight.css"></link>
</head>
<body>
<article id="contents">
<div id="breadcrumbs"><a href="group__CUDART__EVENT.html" shape="rect">< Previous</a> | <a href="group__CUDART__MEMORY.html" shape="rect">Next ></a></div>
<div id="release-info">CUDA Runtime API
(<a href="../../pdf/CUDA_Runtime_API.pdf">PDF</a>)
-
CUDA Toolkit v5.5
(<a href="https://developer.nvidia.com/cuda-toolkit-archive">older</a>)
-
Last updated
July 19, 2013
-
<a href="mailto:cudatools@nvidia.com?subject=CUDA Tools Documentation Feedback: cuda-runtime-api">Send Feedback</a></div>
<div class="topic reference apiRef apiPackage cppModule" id="group__CUDART__EXECUTION"><a name="group__CUDART__EXECUTION" shape="rect">
<!-- --></a><h2 class="topictitle2 cppModule">2.6. Execution Control</h2>
<div class="section">
<p>This section describes the execution control functions of the CUDA runtime application programming interface.</p>
<p class="p">Some functions have overloaded C++ API template versions documented separately in the <a class="xref" href="group__CUDART__HIGHLEVEL.html#group__CUDART__HIGHLEVEL" title="C++-style interface built on top of CUDA runtime API." shape="rect">C++ API Routines</a> module.
</p>
</div>
<h3 class="fake_sectiontitle member_header">Functions</h3>
<dl class="members">
<dt><span class="member_type"><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> </span><span class="member_name"><a href="#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" shape="rect">cudaConfigureCall</a> ( dim3<span> </span><span class="keyword keyword apiItemName">gridDim</span>, dim3<span> </span><span class="keyword keyword apiItemName">blockDim</span>, size_t<span> </span><span class="keyword keyword apiItemName">sharedMem</span> = <span class="ph ph apiData">0</span>, <a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1ge15d9c8b7a240312b533d6122558085a" title="" shape="rect">cudaStream_t</a><span> </span><span class="keyword keyword apiItemName">stream</span> = <span class="ph ph apiData">0</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Configure a device-launch. </span></dd>
<dt><span class="member_type"><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> </span><span class="member_name"><a href="#group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" shape="rect">cudaFuncGetAttributes</a> ( <a href="structcudaFuncAttributes.html#structcudaFuncAttributes" title="" shape="rect">cudaFuncAttributes</a>*<span> </span><span class="keyword keyword apiItemName">attr</span>, const void*<span> </span><span class="keyword keyword apiItemName">func</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Find out attributes for a given function. </span></dd>
<dt><span class="member_type"><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> </span><span class="member_name"><a href="#group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" shape="rect">cudaFuncSetCacheConfig</a> ( const void*<span> </span><span class="keyword keyword apiItemName">func</span>, <a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gb980f35ed69ee7991704de29a13de49b" title="" shape="rect">cudaFuncCache</a><span> </span><span class="keyword keyword apiItemName">cacheConfig</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Sets the preferred cache configuration for a device function. </span></dd>
<dt><span class="member_type"><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> </span><span class="member_name"><a href="#group__CUDART__EXECUTION_1gcb2e02dc6c75e4f705dfd2252748fe99" shape="rect">cudaFuncSetSharedMemConfig</a> ( const void*<span> </span><span class="keyword keyword apiItemName">func</span>, <a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1g6e62d15f3c224625e8c9aa946f1709a6" title="" shape="rect">cudaSharedMemConfig</a><span> </span><span class="keyword keyword apiItemName">config</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Sets the shared memory configuration for a device function. </span></dd>
<dt><span class="member_type"><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> </span><span class="member_name"><a href="#group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" shape="rect">cudaLaunch</a> ( const void*<span> </span><span class="keyword keyword apiItemName">func</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Launches a device function. </span></dd>
<dt><span class="member_type"><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> </span><span class="member_name"><a href="#group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" shape="rect">cudaSetDoubleForDevice</a> ( double*<span> </span><span class="keyword keyword apiItemName">d</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Converts a double argument to be executed on a device. </span></dd>
<dt><span class="member_type"><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> </span><span class="member_name"><a href="#group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" shape="rect">cudaSetDoubleForHost</a> ( double*<span> </span><span class="keyword keyword apiItemName">d</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Converts a double argument after execution on a device. </span></dd>
<dt><span class="member_type"><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> </span><span class="member_name"><a href="#group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" shape="rect">cudaSetupArgument</a> ( const void*<span> </span><span class="keyword keyword apiItemName">arg</span>, size_t<span> </span><span class="keyword keyword apiItemName">size</span>, size_t<span> </span><span class="keyword keyword apiItemName">offset</span> ) </span></dt>
<dd class="shortdesc"><span></span><span class="desc">Configure a device launch. </span></dd>
</dl>
<div class="description">
<h3 class="sectiontitle">Functions</h3>
<dl class="description">
<dt class="description"><a name="group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" id="group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" shape="rect">
<!-- --></a><span><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> cudaConfigureCall ( dim3<span> </span><span class="keyword keyword apiItemName">gridDim</span>, dim3<span> </span><span class="keyword keyword apiItemName">blockDim</span>, size_t<span> </span><span class="keyword keyword apiItemName">sharedMem</span> = <span class="ph ph apiData">0</span>, <a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1ge15d9c8b7a240312b533d6122558085a" title="" shape="rect">cudaStream_t</a><span> </span><span class="keyword keyword apiItemName">stream</span> = <span class="ph ph apiData">0</span> ) </span></dt>
<dd class="description">
<div class="section">Configure a device-launch. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">gridDim</span></tt></dt>
<dd>- Grid dimensions </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">blockDim</span></tt></dt>
<dd>- Block dimensions </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">sharedMem</span></tt></dt>
<dd>- Shared memory </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">stream</span></tt></dt>
<dd>- Stream identifier</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e355f04607d824883b4a50662830d591" shape="rect">cudaSuccess</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038d3934977826b2e1488dfeb7e0614e776" shape="rect">cudaErrorInvalidConfiguration</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Specifies the grid and block dimensions for the device call to be executed similar to the execution configuration syntax.
<a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall()</a> is stack based. Each call pushes data on top of an execution stack. This data contains the dimension for the grid and thread
blocks, together with any arguments for the call.
</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__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" title="Sets the preferred cache configuration for a device function." shape="rect">cudaFuncSetCacheConfig ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" title="Find out attributes for a given function." shape="rect">cudaFuncGetAttributes ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" title="Launches a device function." shape="rect">cudaLaunch ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" title="Converts a double argument to be executed on a device." shape="rect">cudaSetDoubleForDevice</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" title="Converts a double argument after execution on a device." shape="rect">cudaSetDoubleForHost</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" title="Configure a device launch." shape="rect">cudaSetupArgument ( C API)</a>,
</p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" id="group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" shape="rect">
<!-- --></a><span><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> cudaFuncGetAttributes ( <a href="structcudaFuncAttributes.html#structcudaFuncAttributes" title="" shape="rect">cudaFuncAttributes</a>*<span> </span><span class="keyword keyword apiItemName">attr</span>, const void*<span> </span><span class="keyword keyword apiItemName">func</span> ) </span></dt>
<dd class="description">
<div class="section">Find out attributes for a given function. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">attr</span></tt></dt>
<dd>- Return pointer to function's attributes </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">func</span></tt></dt>
<dd>- Device function symbol</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e355f04607d824883b4a50662830d591" shape="rect">cudaSuccess</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038ce7993a88ecf2c57b8102d55d997a18c" shape="rect">cudaErrorInitializationError</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038b6af535e7e53d3f21e2437e8977b8c2e" shape="rect">cudaErrorInvalidDeviceFunction</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>This function obtains the attributes of a function specified via <tt class="ph tt code">func</tt>. <tt class="ph tt code">func</tt> is a device function symbol and must be declared as a <tt class="ph tt code">__global__</tt> function. The fetched attributes are placed in <tt class="ph tt code">attr</tt>. If the specified function does not exist, then <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038b6af535e7e53d3f21e2437e8977b8c2e" shape="rect">cudaErrorInvalidDeviceFunction</a> is returned.
</p>
<p class="p">Note that some function attributes such as <a class="xref" href="structcudaFuncAttributes.html#structcudaFuncAttributes_1b8350f491105f159ff684f5c59167af0" shape="rect">maxThreadsPerBlock</a> may vary based on the device that is currently being used.
</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">Use of a string naming a function as the <tt class="ph tt code">func</tt> paramater was deprecated in CUDA 4.1 and removed in CUDA 5.0.
</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__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" title="Sets the preferred cache configuration for a device function." shape="rect">cudaFuncSetCacheConfig ( C API)</a>, cudaFuncGetAttributes (C++ API), <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" title="Launches a device function." shape="rect">cudaLaunch ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" title="Converts a double argument to be executed on a device." shape="rect">cudaSetDoubleForDevice</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" title="Converts a double argument after execution on a device." shape="rect">cudaSetDoubleForHost</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" title="Configure a device launch." shape="rect">cudaSetupArgument ( C API)</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" id="group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" shape="rect">
<!-- --></a><span><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> cudaFuncSetCacheConfig ( const void*<span> </span><span class="keyword keyword apiItemName">func</span>, <a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gb980f35ed69ee7991704de29a13de49b" title="" shape="rect">cudaFuncCache</a><span> </span><span class="keyword keyword apiItemName">cacheConfig</span> ) </span></dt>
<dd class="description">
<div class="section">Sets the preferred cache configuration for a device function. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">func</span></tt></dt>
<dd>- Device function symbol </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">cacheConfig</span></tt></dt>
<dd>- Requested cache configuration</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e355f04607d824883b4a50662830d591" shape="rect">cudaSuccess</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038ce7993a88ecf2c57b8102d55d997a18c" shape="rect">cudaErrorInitializationError</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038b6af535e7e53d3f21e2437e8977b8c2e" shape="rect">cudaErrorInvalidDeviceFunction</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>On devices where the L1 cache and shared memory use the same hardware resources, this sets through <tt class="ph tt code">cacheConfig</tt> the preferred cache configuration for the function specified via <tt class="ph tt code">func</tt>. This is only a preference. The runtime will use the requested configuration if possible, but it is free to choose a different
configuration if required to execute <tt class="ph tt code">func</tt>.
</p>
<p class="p"><tt class="ph tt code">func</tt> is a device function symbol and must be declared as a <tt class="ph tt code">__global__</tt> function. If the specified function does not exist, then <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038b6af535e7e53d3f21e2437e8977b8c2e" shape="rect">cudaErrorInvalidDeviceFunction</a> is returned.
</p>
<p class="p">This setting does nothing on devices where the size of the L1 cache and shared memory are fixed.</p>
<p class="p">Launching a kernel with a different preference than the most recent preference setting may insert a device-side synchronization
point.
</p>
<p class="p">The supported cache configurations are:
<ul class="ul">
<li class="li">
<p class="p"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1ggb980f35ed69ee7991704de29a13de49b3b4b8c65376ce1ca107be413e15981bc" shape="rect">cudaFuncCachePreferNone</a>: no preference for shared memory or L1 (default)
</p>
</li>
<li class="li">
<p class="p"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1ggb980f35ed69ee7991704de29a13de49b84725d25c531f9bafc61ae329afe5b2b" shape="rect">cudaFuncCachePreferShared</a>: prefer larger shared memory and smaller L1 cache
</p>
</li>
<li class="li">
<p class="p"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1ggb980f35ed69ee7991704de29a13de49b8ecb48ccbc2230c81528a2c7c695100e" shape="rect">cudaFuncCachePreferL1</a>: prefer larger L1 cache and smaller shared memory
</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">Note that this function may also return error codes from previous, asynchronous launches. </p>
</li>
<li class="li">
<p class="p">Use of a string naming a function as the <tt class="ph tt code">func</tt> paramater was deprecated in CUDA 4.1 and removed in CUDA 5.0.
</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__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall</a>, cudaFuncSetCacheConfig (C++ API), <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" title="Find out attributes for a given function." shape="rect">cudaFuncGetAttributes ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" title="Launches a device function." shape="rect">cudaLaunch ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" title="Converts a double argument to be executed on a device." shape="rect">cudaSetDoubleForDevice</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" title="Converts a double argument after execution on a device." shape="rect">cudaSetDoubleForHost</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" title="Configure a device launch." shape="rect">cudaSetupArgument ( C API)</a>, <a class="xref" href="group__CUDART__THREAD__DEPRECATED.html#group__CUDART__THREAD__DEPRECATED_1g7a82ef85e4f7e0cff0a84b2f7f6bc63a" title="Returns the preferred cache configuration for the current device." shape="rect">cudaThreadGetCacheConfig</a>, <a class="xref" href="group__CUDART__THREAD__DEPRECATED.html#group__CUDART__THREAD__DEPRECATED_1g27d0f538b3018142bf04deae7f02c49e" title="Sets the preferred cache configuration for the current device." shape="rect">cudaThreadSetCacheConfig</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDART__EXECUTION_1gcb2e02dc6c75e4f705dfd2252748fe99" id="group__CUDART__EXECUTION_1gcb2e02dc6c75e4f705dfd2252748fe99" shape="rect">
<!-- --></a><span><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> cudaFuncSetSharedMemConfig ( const void*<span> </span><span class="keyword keyword apiItemName">func</span>, <a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1g6e62d15f3c224625e8c9aa946f1709a6" title="" shape="rect">cudaSharedMemConfig</a><span> </span><span class="keyword keyword apiItemName">config</span> ) </span></dt>
<dd class="description">
<div class="section">Sets the shared memory configuration for a device function. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">func</span></tt></dt>
<dd>- Device function symbol </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">config</span></tt></dt>
<dd>- Requested shared memory configuration</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e355f04607d824883b4a50662830d591" shape="rect">cudaSuccess</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038ce7993a88ecf2c57b8102d55d997a18c" shape="rect">cudaErrorInitializationError</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038b6af535e7e53d3f21e2437e8977b8c2e" shape="rect">cudaErrorInvalidDeviceFunction</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e00383e8aef5398ee38e28ed41e357b48917c" shape="rect">cudaErrorInvalidValue</a>,
</p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>On devices with configurable shared memory banks, this function will force all subsequent launches of the specified device
function to have the given shared memory bank size configuration. On any given launch of the function, the shared memory configuration
of the device will be temporarily changed if needed to suit the function's preferred configuration. Changes in shared memory
configuration between subsequent launches of functions, may introduce a device side synchronization point.
</p>
<p class="p">Any per-function setting of shared memory bank size set via <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gcb2e02dc6c75e4f705dfd2252748fe99" title="Sets the shared memory configuration for a device function." shape="rect">cudaFuncSetSharedMemConfig</a> will override the device wide setting set by <a class="xref" href="group__CUDART__DEVICE.html#group__CUDART__DEVICE_1g1a4789fb687cc36dccc98f25c96f0cd8" title="Sets the shared memory configuration for the current device." shape="rect">cudaDeviceSetSharedMemConfig</a>.
</p>
<p class="p">Changing the shared memory bank size will not increase shared memory usage or affect occupancy of kernels, but may have major
effects on performance. Larger bank sizes will allow for greater potential bandwidth to shared memory, but will change what
kinds of accesses to shared memory will result in bank conflicts.
</p>
<p class="p">This function will do nothing on devices with fixed shared memory bank size.</p>
<p class="p">The supported bank configurations are:
<ul class="ul">
<li class="li">
<p class="p">cudaSharedMemBankSizeDefault: use the device's shared memory configuration when launching this function.</p>
</li>
<li class="li">
<p class="p">cudaSharedMemBankSizeFourByte: set shared memory bank width to be four bytes natively when launching this function.</p>
</li>
<li class="li">
<p class="p">cudaSharedMemBankSizeEightByte: set shared memory bank width to be eight bytes natively when launching this function.</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">Note that this function may also return error codes from previous, asynchronous launches. </p>
</li>
<li class="li">
<p class="p">Use of a string naming a function as the <tt class="ph tt code">func</tt> paramater was deprecated in CUDA 4.1 and removed in CUDA 5.0.
</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__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall</a>, <a class="xref" href="group__CUDART__DEVICE.html#group__CUDART__DEVICE_1g1a4789fb687cc36dccc98f25c96f0cd8" title="Sets the shared memory configuration for the current device." shape="rect">cudaDeviceSetSharedMemConfig</a>, <a class="xref" href="group__CUDART__DEVICE.html#group__CUDART__DEVICE_1ge2a985ebc36d53e84f37f6482b059aaa" title="Returns the shared memory configuration for the current device." shape="rect">cudaDeviceGetSharedMemConfig</a>, <a class="xref" href="group__CUDART__DEVICE.html#group__CUDART__DEVICE_1gac27b566beee1aa9175373bb9e29b8d1" title="Sets the preferred cache configuration for the current device." shape="rect">cudaDeviceSetCacheConfig</a>, <a class="xref" href="group__CUDART__DEVICE.html#group__CUDART__DEVICE_1g91d4f0ad53af9cef174b764598444349" title="Returns the preferred cache configuration for the current device." shape="rect">cudaDeviceGetCacheConfig</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" title="Sets the preferred cache configuration for a device function." shape="rect">cudaFuncSetCacheConfig</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" id="group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" shape="rect">
<!-- --></a><span><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> cudaLaunch ( const void*<span> </span><span class="keyword keyword apiItemName">func</span> ) </span></dt>
<dd class="description">
<div class="section">Launches a device function. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">func</span></tt></dt>
<dd>- Device function symbol</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e355f04607d824883b4a50662830d591" shape="rect">cudaSuccess</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038b6af535e7e53d3f21e2437e8977b8c2e" shape="rect">cudaErrorInvalidDeviceFunction</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038d3934977826b2e1488dfeb7e0614e776" shape="rect">cudaErrorInvalidConfiguration</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038099def426efea2b2366d8d7ad09f974a" shape="rect">cudaErrorLaunchFailure</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e00384ea7f1214c348f64f565ca860933e5a5" shape="rect">cudaErrorLaunchTimeout</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038dfddff5cba4a975a67b298ee0752f714" shape="rect">cudaErrorLaunchOutOfResources</a>, <a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e00383be0ba00f56924404620f89b7d405362" shape="rect">cudaErrorSharedObjectInitFailed</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Launches the function <tt class="ph tt code">func</tt> on the device. The parameter <tt class="ph tt code">func</tt> must be a device function symbol. The parameter specified by <tt class="ph tt code">func</tt> must be declared as a <tt class="ph tt code">__global__</tt> function. <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" title="Launches a device function." shape="rect">cudaLaunch()</a> must be preceded by a call to <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall()</a> since it pops the data that was pushed by <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall()</a> from the execution stack.
</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">Use of a string naming a variable as the <tt class="ph tt code">symbol</tt> paramater was removed in CUDA 5.0.
</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__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" title="Sets the preferred cache configuration for a device function." shape="rect">cudaFuncSetCacheConfig ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" title="Find out attributes for a given function." shape="rect">cudaFuncGetAttributes ( C API)</a>, cudaLaunch (C++ API), <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" title="Converts a double argument to be executed on a device." shape="rect">cudaSetDoubleForDevice</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" title="Converts a double argument after execution on a device." shape="rect">cudaSetDoubleForHost</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" title="Configure a device launch." shape="rect">cudaSetupArgument ( C API)</a>, <a class="xref" href="group__CUDART__THREAD__DEPRECATED.html#group__CUDART__THREAD__DEPRECATED_1g7a82ef85e4f7e0cff0a84b2f7f6bc63a" title="Returns the preferred cache configuration for the current device." shape="rect">cudaThreadGetCacheConfig</a>, <a class="xref" href="group__CUDART__THREAD__DEPRECATED.html#group__CUDART__THREAD__DEPRECATED_1g27d0f538b3018142bf04deae7f02c49e" title="Sets the preferred cache configuration for the current device." shape="rect">cudaThreadSetCacheConfig</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" id="group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" shape="rect">
<!-- --></a><span><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> cudaSetDoubleForDevice ( double*<span> </span><span class="keyword keyword apiItemName">d</span> ) </span></dt>
<dd class="description">
<div class="section">Converts a double argument to be executed on a device. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">d</span></tt></dt>
<dd>- Double to convert</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e355f04607d824883b4a50662830d591" shape="rect">cudaSuccess</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>
Converts the double value of <tt class="ph tt code">d</tt> to an internal float representation if the device does not support double arithmetic. If the device does natively support
doubles, then this function does nothing.
</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__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" title="Sets the preferred cache configuration for a device function." shape="rect">cudaFuncSetCacheConfig ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" title="Find out attributes for a given function." shape="rect">cudaFuncGetAttributes ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" title="Launches a device function." shape="rect">cudaLaunch ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" title="Converts a double argument after execution on a device." shape="rect">cudaSetDoubleForHost</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" title="Configure a device launch." shape="rect">cudaSetupArgument ( C API)</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" id="group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" shape="rect">
<!-- --></a><span><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> cudaSetDoubleForHost ( double*<span> </span><span class="keyword keyword apiItemName">d</span> ) </span></dt>
<dd class="description">
<div class="section">Converts a double argument after execution on a device. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">d</span></tt></dt>
<dd>- Double to convert</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e355f04607d824883b4a50662830d591" shape="rect">cudaSuccess</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Converts the double value of <tt class="ph tt code">d</tt> from a potentially internal float representation if the device does not support double arithmetic. If the device does natively
support doubles, then this function does nothing.
</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__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" title="Sets the preferred cache configuration for a device function." shape="rect">cudaFuncSetCacheConfig ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" title="Find out attributes for a given function." shape="rect">cudaFuncGetAttributes ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" title="Launches a device function." shape="rect">cudaLaunch ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" title="Converts a double argument to be executed on a device." shape="rect">cudaSetDoubleForDevice</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" title="Configure a device launch." shape="rect">cudaSetupArgument ( C API)</a></p>
<p class="p"></p>
</div>
</dd>
<dt class="description"><a name="group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" id="group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" shape="rect">
<!-- --></a><span><a href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gf599e5b8b829ce7db0f5216928f6ecb6" title="" shape="rect">cudaError_t</a> cudaSetupArgument ( const void*<span> </span><span class="keyword keyword apiItemName">arg</span>, size_t<span> </span><span class="keyword keyword apiItemName">size</span>, size_t<span> </span><span class="keyword keyword apiItemName">offset</span> ) </span></dt>
<dd class="description">
<div class="section">Configure a device launch. </div>
<div class="section">
<h6 class="parameter_header">
Parameters
</h6>
<dl class="table-display-params">
<dt><tt class="code"><span class="keyword keyword apiItemName">arg</span></tt></dt>
<dd>- Argument to push for a kernel launch </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">size</span></tt></dt>
<dd>- Size of argument </dd>
<dt><tt class="code"><span class="keyword keyword apiItemName">offset</span></tt></dt>
<dd>- Offset in argument stack to push new arg</dd>
</dl>
</div>
<div class="section">
<h6 class="return_header">Returns</h6>
<p class="return"><a class="xref" href="group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e355f04607d824883b4a50662830d591" shape="rect">cudaSuccess</a></p>
</div>
<div class="section">
<h6 class="description_header">Description</h6>
<p>Pushes <tt class="ph tt code">size</tt> bytes of the argument pointed to by <tt class="ph tt code">arg</tt> at <tt class="ph tt code">offset</tt> bytes from the start of the parameter passing area, which starts at offset 0. The arguments are stored in the top of the
execution stack. <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gad1ff3ec13bd21524425779542eca8ef" title="Configure a device launch." shape="rect">cudaSetupArgument()</a> must be preceded by a call to <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall()</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__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g19a7dd5a102b499c39f6a7648bec757a" title="Configure a device-launch." shape="rect">cudaConfigureCall</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g4f35d04be20a41c5df96613a748eecc1" title="Sets the preferred cache configuration for a device function." shape="rect">cudaFuncSetCacheConfig ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g2c1ba377d147198dcc0b631f0193eddc" title="Find out attributes for a given function." shape="rect">cudaFuncGetAttributes ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gf4f3215d06b6d90849a3d977f9440002" title="Launches a device function." shape="rect">cudaLaunch ( C API)</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1gb9574eae787e2c3c91f4954542b17515" title="Converts a double argument to be executed on a device." shape="rect">cudaSetDoubleForDevice</a>, <a class="xref" href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION_1g857c928b6e8a9e4c4c2ec4d4cc5c4063" title="Converts a double argument after execution on a device." shape="rect">cudaSetDoubleForHost</a>, cudaSetupArgument (C++ API),
</p>
<p class="p"></p>
</div>
</dd>
</dl>
</div>
</div>
<hr id="contents-end"></hr>
<div id="breadcrumbs"><a href="group__CUDART__EVENT.html" shape="rect">< Previous</a> | <a href="group__CUDART__MEMORY.html" shape="rect">Next ></a></div>
<div id="release-info">CUDA Runtime API
(<a href="../../pdf/CUDA_Runtime_API.pdf">PDF</a>)
-
CUDA Toolkit v5.5
(<a href="https://developer.nvidia.com/cuda-toolkit-archive">older</a>)
-
Last updated
July 19, 2013
-
<a href="mailto:cudatools@nvidia.com?subject=CUDA Tools Documentation Feedback: cuda-runtime-api">Send Feedback</a></div>
</article>
<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>
<nav id="site-nav">
<div class="category closed"><span class="twiddle">▷</span><a href="../index.html" title="The root of the site.">CUDA Toolkit</a></div>
<ul class="closed">
<li><a href="../cuda-toolkit-release-notes/index.html" title="The Release Notes for the CUDA Toolkit from v4.0 to today.">Release Notes</a></li>
<li><a href="../eula/index.html" title="The End User License Agreements for the NVIDIA CUDA Toolkit, the NVIDIA CUDA Samples, the NVIDIA Display Driver, and NVIDIA NSight (Visual Studio Edition).">EULA</a></li>
<li><a href="../cuda-getting-started-guide-for-linux/index.html" title="This guide discusses how to install and check for correct operation of the CUDA Development Tools on GNU/Linux systems.">Getting Started Linux</a></li>
<li><a href="../cuda-getting-started-guide-for-mac-os-x/index.html" title="This guide discusses how to install and check for correct operation of the CUDA Development Tools on Mac OS X systems.">Getting Started Mac OS X</a></li>
<li><a href="../cuda-getting-started-guide-for-microsoft-windows/index.html" title="This guide discusses how to install and check for correct operation of the CUDA Development Tools on Microsoft Windows systems.">Getting Started Windows</a></li>
<li><a href="../cuda-c-programming-guide/index.html" title="This guide provides a detailed discussion of the CUDA programming model and programming interface. It then describes the hardware implementation, and provides guidance on how to achieve maximum performance. The Appendixes include a list of all CUDA-enabled devices, detailed description of all extensions to the C language, listings of supported mathematical functions, C++ features supported in host and device code, details on texture fetching, technical specifications of various devices, and concludes by introducing the low-level driver API.">Programming Guide</a></li>
<li><a href="../cuda-c-best-practices-guide/index.html" title="This guide presents established parallelization and optimization techniques and explains coding metaphors and idioms that can greatly simplify programming for CUDA-capable GPU architectures. The intent is to provide guidelines for obtaining the best performance from NVIDIA GPUs using the CUDA Toolkit.">Best Practices Guide</a></li>
<li><a href="../kepler-compatibility-guide/index.html" title="This application note is intended to help developers ensure that their NVIDIA CUDA applications will run effectively on GPUs based on the NVIDIA Kepler Architecture. This document provides guidance to ensure that your software applications are compatible with Kepler.">Kepler Compatibility Guide</a></li>
<li><a href="../kepler-tuning-guide/index.html" title="Kepler is NVIDIA's next-generation architecture for CUDA compute applications. Applications that follow the best practices for the Fermi architecture should typically see speedups on the Kepler architecture without any code changes. This guide summarizes the ways that an application can be fine-tuned to gain additional speedups by leveraging Kepler architectural features.">Kepler Tuning Guide</a></li>
<li><a href="../parallel-thread-execution/index.html" title="This guide provides detailed instructions on the use of PTX, a low-level parallel thread execution virtual machine and instruction set architecture (ISA). PTX exposes the GPU as a data-parallel computing device.">PTX ISA</a></li>
<li><a href="../optimus-developer-guide/index.html" title="This document explains how CUDA APIs can be used to query for GPU capabilities in NVIDIA Optimus systems.">Developer Guide for Optimus</a></li>
<li><a href="../video-decoder/index.html" title="This document provides the video decoder API specification and the format conversion and display using DirectX or OpenGL following decode.">Video Decoder</a></li>
<li><a href="../video-encoder/index.html" title="This document provides the CUDA video encoder specifications, including the C-library API functions and encoder query parameters.">Video Encoder</a></li>
<li><a href="../inline-ptx-assembly/index.html" title="This document shows how to inline PTX (parallel thread execution) assembly language statements into CUDA code. It describes available assembler statement parameters and constraints, and the document also provides a list of some pitfalls that you may encounter.">Inline PTX Assembly</a></li>
<li><a href="../cuda-runtime-api/index.html" title="The CUDA runtime API.">CUDA Runtime API</a></li>
<li><a href="../cuda-driver-api/index.html" title="The CUDA driver API.">CUDA Driver API</a></li>
<li><a href="../cuda-math-api/index.html" title="The CUDA math API.">CUDA Math API</a></li>
<li><a href="../cublas/index.html" title="The CUBLAS library is an implementation of BLAS (Basic Linear Algebra Subprograms) on top of the NVIDIA CUDA runtime. It allows the user to access the computational resources of NVIDIA Graphical Processing Unit (GPU), but does not auto-parallelize across multiple GPUs.">CUBLAS</a></li>
<li><a href="../cufft/index.html" title="The CUFFT library user guide.">CUFFT</a></li>
<li><a href="../curand/index.html" title="The CURAND library user guide.">CURAND</a></li>
<li><a href="../cusparse/index.html" title="The CUSPARSE library user guide.">CUSPARSE</a></li>
<li><a href="../npp/index.html" title="NVIDIA NPP is a library of functions for performing CUDA accelerated processing. The initial set of functionality in the library focuses on imaging and video processing and is widely applicable for developers in these areas. NPP will evolve over time to encompass more of the compute heavy tasks in a variety of problem domains. The NPP library is written to maximize flexibility, while maintaining high performance.">NPP</a></li>
<li><a href="../thrust/index.html" title="The Thrust getting started guide.">Thrust</a></li>
<li><a href="../cuda-samples/index.html" title="This document contains a complete listing of the code samples that are included with the NVIDIA CUDA Toolkit. It describes each code sample, lists the minimum GPU specification, and provides links to the source code and white papers if available.">CUDA Samples</a></li>
<li><a href="../cuda-compiler-driver-nvcc/index.html" title="This document is a reference guide on the use of the CUDA compiler driver nvcc. Instead of being a specific CUDA compilation driver, nvcc mimics the behavior of the GNU compiler gcc, accepting a range of conventional compiler options, such as for defining macros and include/library paths, and for steering the compilation process.">NVCC</a></li>
<li><a href="../cuda-gdb/index.html" title="The NVIDIA tool for debugging CUDA applications running on Linux and Mac, providing developers with a mechanism for debugging CUDA applications running on actual hardware. CUDA-GDB is an extension to the x86-64 port of GDB, the GNU Project debugger.">CUDA-GDB</a></li>
<li><a href="../cuda-memcheck/index.html" title="CUDA-MEMCHECK is a suite of run time tools capable of precisely detecting out of bounds and misaligned memory access errors, checking device allocation leaks, reporting hardware errors and identifying shared memory data access hazards.">CUDA-MEMCHECK</a></li>
<li><a href="../nsight-eclipse-edition-getting-started-guide/index.html" title="Nsight Eclipse Edition getting started guide">Nsight Eclipse Edition</a></li>
<li><a href="../profiler-users-guide/index.html" title="This is the guide to the Profiler.">Profiler</a></li>
<li><a href="../cuda-binary-utilities/index.html" title="The application notes for cuobjdump and nvdisasm.">CUDA Binary Utilities</a></li>
<li><a href="../floating-point/index.html" title="A number of issues related to floating point accuracy and compliance are a frequent source of confusion on both CPUs and GPUs. The purpose of this white paper is to discuss the most common issues related to NVIDIA GPUs and to supplement the documentation in the CUDA C Programming Guide.">Floating Point and IEEE 754</a></li>
<li><a href="../incomplete-lu-cholesky/index.html" title="In this white paper we show how to use the CUSPARSE and CUBLAS libraries to achieve a 2x speedup over CPU in the incomplete-LU and Cholesky preconditioned iterative methods. We focus on the Bi-Conjugate Gradient Stabilized and Conjugate Gradient iterative methods, that can be used to solve large sparse nonsymmetric and symmetric positive definite linear systems, respectively. Also, we comment on the parallel sparse triangular solve, which is an essential building block in these algorithms.">Incomplete-LU and Cholesky Preconditioned Iterative Methods</a></li>
<li><a href="../libnvvm-api/index.html" title="The libNVVM API.">libNVVM API</a></li>
<li><a href="../libdevice-users-guide/index.html" title="The libdevice library is an LLVM bitcode library that implements common functions for GPU kernels.">libdevice User's Guide</a></li>
<li><a href="../nvvm-ir-spec/index.html" title="NVVM IR is a compiler IR (internal representation) based on the LLVM IR. The NVVM IR is designed to represent GPU compute kernels (for example, CUDA kernels). High-level language front-ends, like the CUDA C compiler front-end, can generate NVVM IR.">NVVM IR</a></li>
<li><a href="../cupti/index.html" title="The CUPTI API.">CUPTI</a></li>
<li><a href="../debugger-api/index.html" title="The CUDA debugger API.">Debugger API</a></li>
<li><a href="../gpudirect-rdma/index.html" title="A tool for Kepler-class GPUs and CUDA 5.0 enabling a direct path for communication between the GPU and a peer device on the PCI Express bus when the devices share the same upstream root complex using standard features of PCI Express. This document introduces the technology and describes the steps necessary to enable a RDMA for GPUDirect connection to NVIDIA GPUs within the Linux device driver model.">RDMA for GPUDirect</a></li>
</ul>
<div class="category"><span class="twiddle">▼</span><a href="index.html" title="CUDA Runtime API">CUDA Runtime API</a></div>
<ul>
<li><a href="api-sync-behavior.html#api-sync-behavior">1. API synchronization behavior </a></li>
<li><a href="modules.html#modules">2. Modules</a><ul>
<li><a href="group__CUDART__DEVICE.html#group__CUDART__DEVICE">2.1. Device Management</a></li>
<li><a href="group__CUDART__THREAD__DEPRECATED.html#group__CUDART__THREAD__DEPRECATED">2.2. Thread Management [DEPRECATED]</a></li>
<li><a href="group__CUDART__ERROR.html#group__CUDART__ERROR">2.3. Error Handling</a></li>
<li><a href="group__CUDART__STREAM.html#group__CUDART__STREAM">2.4. Stream Management</a></li>
<li><a href="group__CUDART__EVENT.html#group__CUDART__EVENT">2.5. Event Management</a></li>
<li><a href="group__CUDART__EXECUTION.html#group__CUDART__EXECUTION">2.6. Execution Control</a></li>
<li><a href="group__CUDART__MEMORY.html#group__CUDART__MEMORY">2.7. Memory Management</a></li>
<li><a href="group__CUDART__UNIFIED.html#group__CUDART__UNIFIED">2.8. Unified Addressing</a></li>
<li><a href="group__CUDART__PEER.html#group__CUDART__PEER">2.9. Peer Device Memory Access</a></li>
<li><a href="group__CUDART__OPENGL.html#group__CUDART__OPENGL">2.10. OpenGL Interoperability</a></li>
<li><a href="group__CUDART__OPENGL__DEPRECATED.html#group__CUDART__OPENGL__DEPRECATED">2.11. OpenGL Interoperability [DEPRECATED]</a></li>
<li><a href="group__CUDART__D3D9.html#group__CUDART__D3D9">2.12. Direct3D 9 Interoperability</a></li>
<li><a href="group__CUDART__D3D9__DEPRECATED.html#group__CUDART__D3D9__DEPRECATED">2.13. Direct3D 9 Interoperability [DEPRECATED]</a></li>
<li><a href="group__CUDART__D3D10.html#group__CUDART__D3D10">2.14. Direct3D 10 Interoperability</a></li>
<li><a href="group__CUDART__D3D10__DEPRECATED.html#group__CUDART__D3D10__DEPRECATED">2.15. Direct3D 10 Interoperability [DEPRECATED]</a></li>
<li><a href="group__CUDART__D3D11.html#group__CUDART__D3D11">2.16. Direct3D 11 Interoperability</a></li>
<li><a href="group__CUDART__D3D11__DEPRECATED.html#group__CUDART__D3D11__DEPRECATED">2.17. Direct3D 11 Interoperability [DEPRECATED]</a></li>
<li><a href="group__CUDART__VDPAU.html#group__CUDART__VDPAU">2.18. VDPAU Interoperability</a></li>
<li><a href="group__CUDART__INTEROP.html#group__CUDART__INTEROP">2.19. Graphics Interoperability</a></li>
<li><a href="group__CUDART__TEXTURE.html#group__CUDART__TEXTURE">2.20. Texture Reference Management</a></li>
<li><a href="group__CUDART__SURFACE.html#group__CUDART__SURFACE">2.21. Surface Reference Management</a></li>
<li><a href="group__CUDART__TEXTURE__OBJECT.html#group__CUDART__TEXTURE__OBJECT">2.22. Texture Object Management</a></li>
<li><a href="group__CUDART__SURFACE__OBJECT.html#group__CUDART__SURFACE__OBJECT">2.23. Surface Object Management</a></li>
<li><a href="group__CUDART____VERSION.html#group__CUDART____VERSION">2.24. Version Management</a></li>
<li><a href="group__CUDART__HIGHLEVEL.html#group__CUDART__HIGHLEVEL">2.25. C++ API Routines</a></li>
<li><a href="group__CUDART__DRIVER.html#group__CUDART__DRIVER">2.26. Interactions with the CUDA Driver API</a></li>
<li><a href="group__CUDART__PROFILER.html#group__CUDART__PROFILER">2.27. Profiler Control</a></li>
<li><a href="group__CUDART__TYPES.html#group__CUDART__TYPES">2.28. Data types used by CUDA Runtime</a></li>
</ul>
</li>
<li><a href="annotated.html#annotated">3. Data Structures</a><ul>
<li><a href="structcudaChannelFormatDesc.html#structcudaChannelFormatDesc">3.1. cudaChannelFormatDesc</a></li>
<li><a href="structcudaDeviceProp.html#structcudaDeviceProp">3.2. cudaDeviceProp</a></li>
<li><a href="structcudaExtent.html#structcudaExtent">3.3. cudaExtent</a></li>
<li><a href="structcudaFuncAttributes.html#structcudaFuncAttributes">3.4. cudaFuncAttributes</a></li>
<li><a href="structcudaIpcEventHandle__t.html#structcudaIpcEventHandle__t">3.5. cudaIpcEventHandle_t</a></li>
<li><a href="structcudaIpcMemHandle__t.html#structcudaIpcMemHandle__t">3.6. cudaIpcMemHandle_t</a></li>
<li><a href="structcudaMemcpy3DParms.html#structcudaMemcpy3DParms">3.7. cudaMemcpy3DParms</a></li>
<li><a href="structcudaMemcpy3DPeerParms.html#structcudaMemcpy3DPeerParms">3.8. cudaMemcpy3DPeerParms</a></li>
<li><a href="structcudaPitchedPtr.html#structcudaPitchedPtr">3.9. cudaPitchedPtr</a></li>
<li><a href="structcudaPointerAttributes.html#structcudaPointerAttributes">3.10. cudaPointerAttributes</a></li>
<li><a href="structcudaPos.html#structcudaPos">3.11. cudaPos</a></li>
<li><a href="structcudaResourceDesc.html#structcudaResourceDesc">3.12. cudaResourceDesc</a></li>
<li><a href="structcudaResourceViewDesc.html#structcudaResourceViewDesc">3.13. cudaResourceViewDesc</a></li>
<li><a href="structcudaTextureDesc.html#structcudaTextureDesc">3.14. cudaTextureDesc</a></li>
<li><a href="structsurfaceReference.html#structsurfaceReference">3.15. surfaceReference</a></li>
<li><a href="structtextureReference.html#structtextureReference">3.16. textureReference</a></li>
</ul>
</li>
<li><a href="functions.html#functions">4. Data Fields</a></li>
<li><a href="deprecated.html#deprecated">5. Deprecated List</a></li>
<li><a href="notices-header.html#notices-header">Notices</a><ul></ul>
</li>
</ul>
</nav>
<nav id="search-results">
<h2>Search Results</h2>
<ol></ol>
</nav>
<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/omniture/s_code_us_dev_aut1-nolinktrackin.js"></script>
<script language="JavaScript" type="text/javascript" charset="utf-8" src="../common/scripts/omniture/omniture.js"></script>
<noscript><a href="http://www.omniture.com" title="Web Analytics"><img src="http://omniture.nvidia.com/b/ss/nvidiacudadocs/1/H.17--NS/0" height="1" width="1" border="0" alt=""></img></a></noscript>
<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>
</body>
</html>
