cutlass/docs/namespacecutlass_1_1referen...

1179 lines
62 KiB
HTML

<!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">
<head>
<meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/>
<meta http-equiv="X-UA-Compatible" content="IE=9"/>
<meta name="generator" content="Doxygen 1.8.11"/>
<title>CUTLASS: cutlass::reference::device Namespace Reference</title>
<link href="tabs.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="jquery.js"></script>
<script type="text/javascript" src="dynsections.js"></script>
<link href="search/search.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="search/searchdata.js"></script>
<script type="text/javascript" src="search/search.js"></script>
<script type="text/javascript">
$(document).ready(function() { init_search(); });
</script>
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
extensions: ["tex2jax.js"],
jax: ["input/TeX","output/HTML-CSS"],
});
</script><script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js"></script>
<link href="doxygen.css" rel="stylesheet" type="text/css" />
</head>
<body>
<div id="top"><!-- do not remove this div, it is closed by doxygen! -->
<div id="titlearea">
<table cellspacing="0" cellpadding="0">
<tbody>
<tr style="height: 56px;">
<td id="projectlogo"><img alt="Logo" src="cutlass-logo-small.png"/></td>
<td id="projectalign" style="padding-left: 0.5em;">
<div id="projectname">CUTLASS
</div>
<div id="projectbrief">CUDA Templates for Linear Algebra Subroutines and Solvers</div>
</td>
</tr>
</tbody>
</table>
</div>
<!-- end header part -->
<!-- Generated by Doxygen 1.8.11 -->
<script type="text/javascript">
var searchBox = new SearchBox("searchBox", "search",false,'Search');
</script>
<div id="navrow1" class="tabs">
<ul class="tablist">
<li><a href="index.html"><span>Main&#160;Page</span></a></li>
<li><a href="modules.html"><span>Modules</span></a></li>
<li class="current"><a href="namespaces.html"><span>Namespaces</span></a></li>
<li><a href="annotated.html"><span>Classes</span></a></li>
<li><a href="files.html"><span>Files</span></a></li>
<li>
<div id="MSearchBox" class="MSearchBoxInactive">
<span class="left">
<img id="MSearchSelect" src="search/mag_sel.png"
onmouseover="return searchBox.OnSearchSelectShow()"
onmouseout="return searchBox.OnSearchSelectHide()"
alt=""/>
<input type="text" id="MSearchField" value="Search" accesskey="S"
onfocus="searchBox.OnSearchFieldFocus(true)"
onblur="searchBox.OnSearchFieldFocus(false)"
onkeyup="searchBox.OnSearchFieldChange(event)"/>
</span><span class="right">
<a id="MSearchClose" href="javascript:searchBox.CloseResultsWindow()"><img id="MSearchCloseImg" border="0" src="search/close.png" alt=""/></a>
</span>
</div>
</li>
</ul>
</div>
<div id="navrow2" class="tabs2">
<ul class="tablist">
<li><a href="namespaces.html"><span>Namespace&#160;List</span></a></li>
<li><a href="namespacemembers.html"><span>Namespace&#160;Members</span></a></li>
</ul>
</div>
<!-- window showing the filter options -->
<div id="MSearchSelectWindow"
onmouseover="return searchBox.OnSearchSelectShow()"
onmouseout="return searchBox.OnSearchSelectHide()"
onkeydown="return searchBox.OnSearchSelectKey(event)">
</div>
<!-- iframe showing the search results (closed by default) -->
<div id="MSearchResultsWindow">
<iframe src="javascript:void(0)" frameborder="0"
name="MSearchResults" id="MSearchResults">
</iframe>
</div>
<div id="nav-path" class="navpath">
<ul>
<li class="navelem"><a class="el" href="namespacecutlass.html">cutlass</a></li><li class="navelem"><a class="el" href="namespacecutlass_1_1reference.html">reference</a></li><li class="navelem"><a class="el" href="namespacecutlass_1_1reference_1_1device.html">device</a></li> </ul>
</div>
</div><!-- top -->
<div class="header">
<div class="summary">
<a href="#namespaces">Namespaces</a> &#124;
<a href="#nested-classes">Classes</a> &#124;
<a href="#func-members">Functions</a> </div>
<div class="headertitle">
<div class="title">cutlass::reference::device Namespace Reference</div> </div>
</div><!--header-->
<div class="contents">
<table class="memberdecls">
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="namespaces"></a>
Namespaces</h2></td></tr>
<tr class="memitem:namespacecutlass_1_1reference_1_1device_1_1detail"><td class="memItemLeft" align="right" valign="top"> &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device_1_1detail.html">detail</a></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:namespacecutlass_1_1reference_1_1device_1_1kernel"><td class="memItemLeft" align="right" valign="top"> &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device_1_1kernel.html">kernel</a></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:namespacecutlass_1_1reference_1_1device_1_1thread"><td class="memItemLeft" align="right" valign="top"> &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device_1_1thread.html">thread</a></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
</table><table class="memberdecls">
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="nested-classes"></a>
Classes</h2></td></tr>
<tr class="memitem:"><td class="memItemLeft" align="right" valign="top">struct &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1reference_1_1device_1_1BlockForEach.html">BlockForEach</a></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:"><td class="memItemLeft" align="right" valign="top">struct &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1reference_1_1device_1_1Gemm.html">Gemm</a></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:"><td class="memItemLeft" align="right" valign="top">struct &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout4e016ab7cfc644acd7cb4ae770339773.html">Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, AccumulatorType, arch::OpMultiplyAdd &gt;</a></td></tr>
<tr class="memdesc:"><td class="mdescLeft">&#160;</td><td class="mdescRight">Partial specialization for multiply-add. <a href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout4e016ab7cfc644acd7cb4ae770339773.html#details">More...</a><br /></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:"><td class="memItemLeft" align="right" valign="top">struct &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout30b72addd464a2ca4a26785cbfd77a8e.html">Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, AccumulatorType, arch::OpMultiplyAddSaturate &gt;</a></td></tr>
<tr class="memdesc:"><td class="mdescLeft">&#160;</td><td class="mdescRight">Partial specialization for multiply-add-saturate. <a href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout30b72addd464a2ca4a26785cbfd77a8e.html#details">More...</a><br /></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:"><td class="memItemLeft" align="right" valign="top">struct &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout660562b232f408218828ca5915b7e73a.html">Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, AccumulatorType, arch::OpXorPopc &gt;</a></td></tr>
<tr class="memdesc:"><td class="mdescLeft">&#160;</td><td class="mdescRight">Partial specialization for XOR-popc. <a href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout660562b232f408218828ca5915b7e73a.html#details">More...</a><br /></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:"><td class="memItemLeft" align="right" valign="top">struct &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1reference_1_1device_1_1TensorDiagonalForEach.html">TensorDiagonalForEach</a></td></tr>
<tr class="memdesc:"><td class="mdescLeft">&#160;</td><td class="mdescRight">Launches a kernel calling a functor for each element along a tensor's diagonal. <a href="structcutlass_1_1reference_1_1device_1_1TensorDiagonalForEach.html#details">More...</a><br /></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:"><td class="memItemLeft" align="right" valign="top">struct &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1reference_1_1device_1_1TensorForEach.html">TensorForEach</a></td></tr>
<tr class="memdesc:"><td class="mdescLeft">&#160;</td><td class="mdescRight">Launches a kernel calling a functor for each element in a tensor's index space. <a href="structcutlass_1_1reference_1_1device_1_1TensorForEach.html#details">More...</a><br /></td></tr>
<tr class="separator:"><td class="memSeparator" colspan="2">&#160;</td></tr>
</table><table class="memberdecls">
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="func-members"></a>
Functions</h2></td></tr>
<tr class="memitem:a4b872e5b16985b2cf31530a9090a8423"><td class="memTemplParams" colspan="2">template&lt;typename ElementA , typename LayoutA , typename ElementB , typename LayoutB , typename ElementC , typename LayoutC , typename ScalarType , typename AccumulatorType , typename InnerProductOp = multiply_add&lt;AccumulatorType&gt;, typename ConvertOp = NumericConverter&lt;ElementC, ScalarType&gt;&gt; </td></tr>
<tr class="memitem:a4b872e5b16985b2cf31530a9090a8423"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a4b872e5b16985b2cf31530a9090a8423">compute_gemm</a> (<a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html">gemm::GemmCoord</a> problem_size, ScalarType alpha, <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementA, LayoutA &gt; tensor_a, <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementB, LayoutB &gt; tensor_b, ScalarType beta, <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementC, LayoutC &gt; tensor_c, <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementC, LayoutC &gt; tensor_d, AccumulatorType initial_accum)</td></tr>
<tr class="separator:a4b872e5b16985b2cf31530a9090a8423"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:aa1b04f721cb13fb3f110acf6b29dc53b"><td class="memTemplParams" colspan="2">template&lt;typename ElementA , typename LayoutA , typename ElementB , typename LayoutB , typename ElementC , typename LayoutC , typename ScalarType , typename AccumulatorType , typename InnerProductOp = multiply_add&lt;AccumulatorType&gt;, typename ConvertOp = NumericConverter&lt;ElementC, ScalarType&gt;&gt; </td></tr>
<tr class="memitem:aa1b04f721cb13fb3f110acf6b29dc53b"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#aa1b04f721cb13fb3f110acf6b29dc53b">compute_gemm</a> (<a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html">gemm::GemmCoord</a> problem_size, ScalarType alpha, <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementA, LayoutA &gt; tensor_a, <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementB, LayoutB &gt; tensor_b, ScalarType beta, <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementC, LayoutC &gt; tensor_c, AccumulatorType initial_accum)</td></tr>
<tr class="separator:aa1b04f721cb13fb3f110acf6b29dc53b"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:aaa524d4e141cc8934eb9a981e1c89fc5"><td class="memTemplParams" colspan="2">template&lt;typename TensorRefCollectionA , typename TensorRefCollectionB , typename TensorRefCollectionC , typename ScalarType , typename AccumulatorType , typename InnerProductOp , typename ConvertOp &gt; </td></tr>
<tr class="memitem:aaa524d4e141cc8934eb9a981e1c89fc5"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#aaa524d4e141cc8934eb9a981e1c89fc5">BatchedGemm</a> (<a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html">gemm::GemmCoord</a> problem_size, int batch_count, ScalarType alpha, TensorRefCollectionA const &amp;tensor_a, TensorRefCollectionB const &amp;tensor_b, ScalarType beta, TensorRefCollectionC &amp;tensor_c, AccumulatorType initial_accum)</td></tr>
<tr class="memdesc:aaa524d4e141cc8934eb9a981e1c89fc5"><td class="mdescLeft">&#160;</td><td class="mdescRight">Computes a batch of GEMMs over a set of matrices of common dimension. <a href="#aaa524d4e141cc8934eb9a981e1c89fc5">More...</a><br /></td></tr>
<tr class="separator:aaa524d4e141cc8934eb9a981e1c89fc5"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:abbb24b1a372b793bf35320443c179875"><td class="memTemplParams" colspan="2">template&lt;typename TensorRefCollectionA , typename TensorRefCollectionB , typename TensorRefCollectionC , typename ScalarType , typename AccumulatorType &gt; </td></tr>
<tr class="memitem:abbb24b1a372b793bf35320443c179875"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#abbb24b1a372b793bf35320443c179875">BatchedGemm</a> (<a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html">gemm::GemmCoord</a> problem_size, int batch_count, ScalarType alpha, TensorRefCollectionA const &amp;tensor_a, TensorRefCollectionB const &amp;tensor_b, ScalarType beta, TensorRefCollectionC &amp;tensor_c)</td></tr>
<tr class="separator:abbb24b1a372b793bf35320443c179875"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:aad19927d67f15b89e66560cb77f2a813"><td class="memTemplParams" colspan="2">template&lt;typename Element &gt; </td></tr>
<tr class="memitem:aad19927d67f15b89e66560cb77f2a813"><td class="memTemplItemLeft" align="right" valign="top">bool&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#aad19927d67f15b89e66560cb77f2a813">BlockCompareEqual</a> (Element const *ptr_A, Element const *ptr_B, size_t capacity, int grid_size=0, int block_size=0)</td></tr>
<tr class="memdesc:aad19927d67f15b89e66560cb77f2a813"><td class="mdescLeft">&#160;</td><td class="mdescRight">Performs a bit-level equality check between two blocks. <a href="#aad19927d67f15b89e66560cb77f2a813">More...</a><br /></td></tr>
<tr class="separator:aad19927d67f15b89e66560cb77f2a813"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a286d24a9faabc0be18f96e1069dca23e"><td class="memTemplParams" colspan="2">template&lt;typename Element &gt; </td></tr>
<tr class="memitem:a286d24a9faabc0be18f96e1069dca23e"><td class="memTemplItemLeft" align="right" valign="top">bool&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a286d24a9faabc0be18f96e1069dca23e">BlockCompareRelativelyEqual</a> (Element const *ptr_A, Element const *ptr_B, size_t capacity, Element epsilon, Element nonzero_floor, int grid_size=0, int block_size=0)</td></tr>
<tr class="memdesc:a286d24a9faabc0be18f96e1069dca23e"><td class="mdescLeft">&#160;</td><td class="mdescRight">Performs a bit-level equality check between two blocks. <a href="#a286d24a9faabc0be18f96e1069dca23e">More...</a><br /></td></tr>
<tr class="separator:a286d24a9faabc0be18f96e1069dca23e"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:ad71c8103c1f6a2d46a9ba6877844a69a"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:ad71c8103c1f6a2d46a9ba6877844a69a"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#ad71c8103c1f6a2d46a9ba6877844a69a">TensorFillRandomGaussian</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view, uint64_t seed, Element mean=Element(0), Element stddev=Element(1), int bits=-1)</td></tr>
<tr class="memdesc:ad71c8103c1f6a2d46a9ba6877844a69a"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a tensor with random values with a Gaussian distribution. <a href="#ad71c8103c1f6a2d46a9ba6877844a69a">More...</a><br /></td></tr>
<tr class="separator:ad71c8103c1f6a2d46a9ba6877844a69a"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a478e311bfbe901d167090032b6c28732"><td class="memTemplParams" colspan="2">template&lt;typename Element &gt; </td></tr>
<tr class="memitem:a478e311bfbe901d167090032b6c28732"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a478e311bfbe901d167090032b6c28732">BlockFillRandomGaussian</a> (Element *ptr, size_t capacity, uint64_t seed, Element mean=Element(0), Element stddev=Element(1), int bits=-1)</td></tr>
<tr class="memdesc:a478e311bfbe901d167090032b6c28732"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a tensor with random values with a Gaussian distribution. <a href="#a478e311bfbe901d167090032b6c28732">More...</a><br /></td></tr>
<tr class="separator:a478e311bfbe901d167090032b6c28732"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a448cf6f610939c95615ab66d7ca18b4c"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:a448cf6f610939c95615ab66d7ca18b4c"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a448cf6f610939c95615ab66d7ca18b4c">TensorFillRandomUniform</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view, uint64_t seed, Element max=Element(1), Element min=Element(0), int bits=-1)</td></tr>
<tr class="memdesc:a448cf6f610939c95615ab66d7ca18b4c"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a tensor with random values with a uniform random distribution. <a href="#a448cf6f610939c95615ab66d7ca18b4c">More...</a><br /></td></tr>
<tr class="separator:a448cf6f610939c95615ab66d7ca18b4c"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a6f7f618350cf975e261a4ee758650c66"><td class="memTemplParams" colspan="2">template&lt;typename Element &gt; </td></tr>
<tr class="memitem:a6f7f618350cf975e261a4ee758650c66"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a6f7f618350cf975e261a4ee758650c66">BlockFillRandomUniform</a> (Element *ptr, size_t capacity, uint64_t seed, Element max=Element(1), Element min=Element(0), int bits=-1)</td></tr>
<tr class="memdesc:a6f7f618350cf975e261a4ee758650c66"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a tensor with random values with a uniform random distribution. <a href="#a6f7f618350cf975e261a4ee758650c66">More...</a><br /></td></tr>
<tr class="separator:a6f7f618350cf975e261a4ee758650c66"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:aee20536c8ac0a5adcbb162c76eb89c00"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:aee20536c8ac0a5adcbb162c76eb89c00"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#aee20536c8ac0a5adcbb162c76eb89c00">TensorFillDiagonal</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view, Element diag=Element(1), Element other=Element(0))</td></tr>
<tr class="memdesc:aee20536c8ac0a5adcbb162c76eb89c00"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a tensor everywhere with a unique value for its diagonal. <a href="#aee20536c8ac0a5adcbb162c76eb89c00">More...</a><br /></td></tr>
<tr class="separator:aee20536c8ac0a5adcbb162c76eb89c00"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a6e23d479ebb3760d5846ed1b67e450e4"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:a6e23d479ebb3760d5846ed1b67e450e4"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a6e23d479ebb3760d5846ed1b67e450e4">TensorFill</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view, Element val=Element(0))</td></tr>
<tr class="memdesc:a6e23d479ebb3760d5846ed1b67e450e4"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a tensor with a uniform value. <a href="#a6e23d479ebb3760d5846ed1b67e450e4">More...</a><br /></td></tr>
<tr class="separator:a6e23d479ebb3760d5846ed1b67e450e4"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a6b0f21995c4fd5c33617550e6905c78e"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:a6b0f21995c4fd5c33617550e6905c78e"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a6b0f21995c4fd5c33617550e6905c78e">TensorFillIdentity</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view)</td></tr>
<tr class="memdesc:a6b0f21995c4fd5c33617550e6905c78e"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a tensor's diagonal with 1 and 0 everywhere else. <a href="#a6b0f21995c4fd5c33617550e6905c78e">More...</a><br /></td></tr>
<tr class="separator:a6b0f21995c4fd5c33617550e6905c78e"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:aaff3d7919a2f2dce14eb254c17eead9a"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:aaff3d7919a2f2dce14eb254c17eead9a"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#aaff3d7919a2f2dce14eb254c17eead9a">TensorUpdateDiagonal</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view, Element diag=Element(1))</td></tr>
<tr class="memdesc:aaff3d7919a2f2dce14eb254c17eead9a"><td class="mdescLeft">&#160;</td><td class="mdescRight">Writes a uniform value to the diagonal of a tensor without modifying off-diagonal elements. <a href="#aaff3d7919a2f2dce14eb254c17eead9a">More...</a><br /></td></tr>
<tr class="separator:aaff3d7919a2f2dce14eb254c17eead9a"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a8ab743402a5664eb255b08efd0da3481"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:a8ab743402a5664eb255b08efd0da3481"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a8ab743402a5664eb255b08efd0da3481">TensorUpdateOffDiagonal</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view, Element other=Element(1))</td></tr>
<tr class="memdesc:a8ab743402a5664eb255b08efd0da3481"><td class="mdescLeft">&#160;</td><td class="mdescRight">Writes a uniform value to all elements in the tensor without modifying diagonal elements. <a href="#a8ab743402a5664eb255b08efd0da3481">More...</a><br /></td></tr>
<tr class="separator:a8ab743402a5664eb255b08efd0da3481"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a37816633b87bce34515e31fa5c2709fa"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:a37816633b87bce34515e31fa5c2709fa"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a37816633b87bce34515e31fa5c2709fa">TensorFillLinear</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view, Array&lt; Element, Layout::kRank &gt; const &amp;v, Element s=Element(0))</td></tr>
<tr class="memdesc:a37816633b87bce34515e31fa5c2709fa"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills tensor with a linear combination of its coordinate and another vector. <a href="#a37816633b87bce34515e31fa5c2709fa">More...</a><br /></td></tr>
<tr class="separator:a37816633b87bce34515e31fa5c2709fa"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a2cf3ac0ae77e672e2af80f4820434cbe"><td class="memTemplParams" colspan="2">template&lt;typename Element &gt; </td></tr>
<tr class="memitem:a2cf3ac0ae77e672e2af80f4820434cbe"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a2cf3ac0ae77e672e2af80f4820434cbe">BlockFillSequential</a> (Element *ptr, int64_t capacity, Element v=Element(1), Element s=Element(0))</td></tr>
<tr class="memdesc:a2cf3ac0ae77e672e2af80f4820434cbe"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a block of data with sequential elements. <a href="#a2cf3ac0ae77e672e2af80f4820434cbe">More...</a><br /></td></tr>
<tr class="separator:a2cf3ac0ae77e672e2af80f4820434cbe"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:af6b21c6d90a1bb3f10dffd0a4adb644a"><td class="memTemplParams" colspan="2">template&lt;typename Element &gt; </td></tr>
<tr class="memitem:af6b21c6d90a1bb3f10dffd0a4adb644a"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#af6b21c6d90a1bb3f10dffd0a4adb644a">BlockFillRandom</a> (Element *ptr, size_t capacity, uint64_t seed, <a class="el" href="structcutlass_1_1Distribution.html">Distribution</a> dist)</td></tr>
<tr class="memdesc:af6b21c6d90a1bb3f10dffd0a4adb644a"><td class="mdescLeft">&#160;</td><td class="mdescRight">Fills a block of data with sequential elements. <a href="#af6b21c6d90a1bb3f10dffd0a4adb644a">More...</a><br /></td></tr>
<tr class="separator:af6b21c6d90a1bb3f10dffd0a4adb644a"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a3d11dd00b1bdaa15fdb96345c5ac613a"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:a3d11dd00b1bdaa15fdb96345c5ac613a"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a3d11dd00b1bdaa15fdb96345c5ac613a">TensorCopyDiagonalIn</a> (<a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view, Element const *ptr)</td></tr>
<tr class="memdesc:a3d11dd00b1bdaa15fdb96345c5ac613a"><td class="mdescLeft">&#160;</td><td class="mdescRight">Copies a diagonal in from host memory without modifying off-diagonal elements. <a href="#a3d11dd00b1bdaa15fdb96345c5ac613a">More...</a><br /></td></tr>
<tr class="separator:a3d11dd00b1bdaa15fdb96345c5ac613a"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a299cab22dca6be5ddf6ff62e23566a24"><td class="memTemplParams" colspan="2">template&lt;typename Element , typename Layout &gt; </td></tr>
<tr class="memitem:a299cab22dca6be5ddf6ff62e23566a24"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="namespacecutlass_1_1reference_1_1device.html#a299cab22dca6be5ddf6ff62e23566a24">TensorCopyDiagonalOut</a> (Element *ptr, <a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt; view)</td></tr>
<tr class="memdesc:a299cab22dca6be5ddf6ff62e23566a24"><td class="mdescLeft">&#160;</td><td class="mdescRight">Copies the diagonal of a tensor into a dense buffer in host memory. <a href="#a299cab22dca6be5ddf6ff62e23566a24">More...</a><br /></td></tr>
<tr class="separator:a299cab22dca6be5ddf6ff62e23566a24"><td class="memSeparator" colspan="2">&#160;</td></tr>
</table>
<h2 class="groupheader">Function Documentation</h2>
<a class="anchor" id="aaa524d4e141cc8934eb9a981e1c89fc5"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename TensorRefCollectionA , typename TensorRefCollectionB , typename TensorRefCollectionC , typename ScalarType , typename AccumulatorType , typename InnerProductOp , typename ConvertOp &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::BatchedGemm </td>
<td>(</td>
<td class="paramtype"><a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html">gemm::GemmCoord</a>&#160;</td>
<td class="paramname"><em>problem_size</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>batch_count</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">ScalarType&#160;</td>
<td class="paramname"><em>alpha</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">TensorRefCollectionA const &amp;&#160;</td>
<td class="paramname"><em>tensor_a</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">TensorRefCollectionB const &amp;&#160;</td>
<td class="paramname"><em>tensor_b</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">ScalarType&#160;</td>
<td class="paramname"><em>beta</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">TensorRefCollectionC &amp;&#160;</td>
<td class="paramname"><em>tensor_c</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">AccumulatorType&#160;</td>
<td class="paramname"><em>initial_accum</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a class="anchor" id="abbb24b1a372b793bf35320443c179875"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename TensorRefCollectionA , typename TensorRefCollectionB , typename TensorRefCollectionC , typename ScalarType , typename AccumulatorType &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::BatchedGemm </td>
<td>(</td>
<td class="paramtype"><a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html">gemm::GemmCoord</a>&#160;</td>
<td class="paramname"><em>problem_size</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>batch_count</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">ScalarType&#160;</td>
<td class="paramname"><em>alpha</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">TensorRefCollectionA const &amp;&#160;</td>
<td class="paramname"><em>tensor_a</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">TensorRefCollectionB const &amp;&#160;</td>
<td class="paramname"><em>tensor_b</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">ScalarType&#160;</td>
<td class="paramname"><em>beta</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">TensorRefCollectionC &amp;&#160;</td>
<td class="paramname"><em>tensor_c</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>Computes a general matrix product among matrices (tensors of rank=2) pointed to by <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a> objects. </p>
</div>
</div>
<a class="anchor" id="aad19927d67f15b89e66560cb77f2a813"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element &gt; </div>
<table class="memname">
<tr>
<td class="memname">bool cutlass::reference::device::BlockCompareEqual </td>
<td>(</td>
<td class="paramtype">Element const *&#160;</td>
<td class="paramname"><em>ptr_A</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element const *&#160;</td>
<td class="paramname"><em>ptr_B</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">size_t&#160;</td>
<td class="paramname"><em>capacity</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>grid_size</em> = <code>0</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>block_size</em> = <code>0</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a class="anchor" id="a286d24a9faabc0be18f96e1069dca23e"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element &gt; </div>
<table class="memname">
<tr>
<td class="memname">bool cutlass::reference::device::BlockCompareRelativelyEqual </td>
<td>(</td>
<td class="paramtype">Element const *&#160;</td>
<td class="paramname"><em>ptr_A</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element const *&#160;</td>
<td class="paramname"><em>ptr_B</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">size_t&#160;</td>
<td class="paramname"><em>capacity</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>epsilon</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>nonzero_floor</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>grid_size</em> = <code>0</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>block_size</em> = <code>0</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a class="anchor" id="af6b21c6d90a1bb3f10dffd0a4adb644a"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::BlockFillRandom </td>
<td>(</td>
<td class="paramtype">Element *&#160;</td>
<td class="paramname"><em>ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">size_t&#160;</td>
<td class="paramname"><em>capacity</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint64_t&#160;</td>
<td class="paramname"><em>seed</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="structcutlass_1_1Distribution.html">Distribution</a>&#160;</td>
<td class="paramname"><em>dist</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a class="anchor" id="a478e311bfbe901d167090032b6c28732"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::BlockFillRandomGaussian </td>
<td>(</td>
<td class="paramtype">Element *&#160;</td>
<td class="paramname"><em>ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">size_t&#160;</td>
<td class="paramname"><em>capacity</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint64_t&#160;</td>
<td class="paramname"><em>seed</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>mean</em> = <code>Element(0)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>stddev</em> = <code>Element(1)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>bits</em> = <code>-1</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Element type </p>
<p>&lt; If non-negative, specifies number of fractional bits that are not truncated to zero. Permits reducing precision of data. </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">seed</td><td>seed for RNG </td></tr>
<tr><td class="paramname">mean</td><td>Gaussian distribution's mean </td></tr>
<tr><td class="paramname">stddev</td><td>Gaussian distribution's standard deviation </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="a6f7f618350cf975e261a4ee758650c66"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::BlockFillRandomUniform </td>
<td>(</td>
<td class="paramtype">Element *&#160;</td>
<td class="paramname"><em>ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">size_t&#160;</td>
<td class="paramname"><em>capacity</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint64_t&#160;</td>
<td class="paramname"><em>seed</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>max</em> = <code>Element(1)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>min</em> = <code>Element(0)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>bits</em> = <code>-1</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; If non-negative, specifies number of fractional bits that are not truncated to zero. Permits reducing precision of data. </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">seed</td><td>seed for RNG </td></tr>
<tr><td class="paramname">max</td><td>upper bound of distribution </td></tr>
<tr><td class="paramname">min</td><td>lower bound for distribution </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="a2cf3ac0ae77e672e2af80f4820434cbe"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::BlockFillSequential </td>
<td>(</td>
<td class="paramtype">Element *&#160;</td>
<td class="paramname"><em>ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int64_t&#160;</td>
<td class="paramname"><em>capacity</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>v</em> = <code>Element(1)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>s</em> = <code>Element(0)</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a class="anchor" id="a4b872e5b16985b2cf31530a9090a8423"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename ElementA , typename LayoutA , typename ElementB , typename LayoutB , typename ElementC , typename LayoutC , typename ScalarType , typename AccumulatorType , typename InnerProductOp = multiply_add&lt;AccumulatorType&gt;, typename ConvertOp = NumericConverter&lt;ElementC, ScalarType&gt;&gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::compute_gemm </td>
<td>(</td>
<td class="paramtype"><a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html">gemm::GemmCoord</a>&#160;</td>
<td class="paramname"><em>problem_size</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">ScalarType&#160;</td>
<td class="paramname"><em>alpha</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementA, LayoutA &gt;&#160;</td>
<td class="paramname"><em>tensor_a</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementB, LayoutB &gt;&#160;</td>
<td class="paramname"><em>tensor_b</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">ScalarType&#160;</td>
<td class="paramname"><em>beta</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementC, LayoutC &gt;&#160;</td>
<td class="paramname"><em>tensor_c</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementC, LayoutC &gt;&#160;</td>
<td class="paramname"><em>tensor_d</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">AccumulatorType&#160;</td>
<td class="paramname"><em>initial_accum</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>Computes a general matrix product among matrices (tensors of rank=2) pointed to by <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a> objects.</p>
<p>Explicitly naming types needed by this template can be cumbersome, particularly for the accumulator type, so a function argument 'initial_accum' is exposed. Passing AccumulatorType(0) as the last function argument can be easier than naming all template arguments explicitly. </p>
</div>
</div>
<a class="anchor" id="aa1b04f721cb13fb3f110acf6b29dc53b"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename ElementA , typename LayoutA , typename ElementB , typename LayoutB , typename ElementC , typename LayoutC , typename ScalarType , typename AccumulatorType , typename InnerProductOp = multiply_add&lt;AccumulatorType&gt;, typename ConvertOp = NumericConverter&lt;ElementC, ScalarType&gt;&gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::compute_gemm </td>
<td>(</td>
<td class="paramtype"><a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html">gemm::GemmCoord</a>&#160;</td>
<td class="paramname"><em>problem_size</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">ScalarType&#160;</td>
<td class="paramname"><em>alpha</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementA, LayoutA &gt;&#160;</td>
<td class="paramname"><em>tensor_a</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementB, LayoutB &gt;&#160;</td>
<td class="paramname"><em>tensor_b</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">ScalarType&#160;</td>
<td class="paramname"><em>beta</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a>&lt; ElementC, LayoutC &gt;&#160;</td>
<td class="paramname"><em>tensor_c</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">AccumulatorType&#160;</td>
<td class="paramname"><em>initial_accum</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>Computes a general matrix product among matrices (tensors of rank=2) pointed to by <a class="el" href="classcutlass_1_1TensorRef.html">TensorRef</a> objects.</p>
<p>This assumes the accumulator type is the same type as the scalars. </p>
</div>
</div>
<a class="anchor" id="a3d11dd00b1bdaa15fdb96345c5ac613a"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorCopyDiagonalIn </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element const *&#160;</td>
<td class="paramname"><em>ptr</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<p>&lt; dense buffer of elements </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">view</td><td>destination tensor </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="a299cab22dca6be5ddf6ff62e23566a24"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorCopyDiagonalOut </td>
<td>(</td>
<td class="paramtype">Element *&#160;</td>
<td class="paramname"><em>ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<p>&lt; source tensor </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">ptr</td><td>dense buffer of elements </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="a6e23d479ebb3760d5846ed1b67e450e4"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorFill </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>val</em> = <code>Element(0)</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<p>&lt; value to uniformly fill it with </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">view</td><td>destination tensor </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="aee20536c8ac0a5adcbb162c76eb89c00"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorFillDiagonal </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>diag</em> = <code>Element(1)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>other</em> = <code>Element(0)</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<p>&lt; value to write off the diagonal </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">view</td><td>destination tensor </td></tr>
<tr><td class="paramname">diag</td><td>value to write in the diagonal </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="a6b0f21995c4fd5c33617550e6905c78e"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorFillIdentity </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em></td><td>)</td>
<td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<p>&lt; destination tensor </p>
</div>
</div>
<a class="anchor" id="a37816633b87bce34515e31fa5c2709fa"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorFillLinear </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Array&lt; Element, Layout::kRank &gt; const &amp;&#160;</td>
<td class="paramname"><em>v</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>s</em> = <code>Element(0)</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">view</td><td>destination tensor </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="ad71c8103c1f6a2d46a9ba6877844a69a"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorFillRandomGaussian </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint64_t&#160;</td>
<td class="paramname"><em>seed</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>mean</em> = <code>Element(0)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>stddev</em> = <code>Element(1)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>bits</em> = <code>-1</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<p>&lt; If non-negative, specifies number of fractional bits that are not truncated to zero. Permits reducing precision of data. </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">view</td><td>destination tensor </td></tr>
<tr><td class="paramname">seed</td><td>seed for RNG </td></tr>
<tr><td class="paramname">mean</td><td>Gaussian distribution's mean </td></tr>
<tr><td class="paramname">stddev</td><td>Gaussian distribution's standard deviation </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="a448cf6f610939c95615ab66d7ca18b4c"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorFillRandomUniform </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint64_t&#160;</td>
<td class="paramname"><em>seed</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>max</em> = <code>Element(1)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>min</em> = <code>Element(0)</code>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>bits</em> = <code>-1</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<p>&lt; If non-negative, specifies number of fractional bits that are not truncated to zero. Permits reducing precision of data. </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">view</td><td>destination tensor </td></tr>
<tr><td class="paramname">seed</td><td>seed for RNG </td></tr>
<tr><td class="paramname">max</td><td>upper bound of distribution </td></tr>
<tr><td class="paramname">min</td><td>lower bound for distribution </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="aaff3d7919a2f2dce14eb254c17eead9a"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorUpdateDiagonal </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>diag</em> = <code>Element(1)</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">view</td><td>destination tensor </td></tr>
</table>
</dd>
</dl>
</div>
</div>
<a class="anchor" id="a8ab743402a5664eb255b08efd0da3481"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename Element , typename Layout &gt; </div>
<table class="memname">
<tr>
<td class="memname">void cutlass::reference::device::TensorUpdateOffDiagonal </td>
<td>(</td>
<td class="paramtype"><a class="el" href="classcutlass_1_1TensorView.html">TensorView</a>&lt; Element, Layout &gt;&#160;</td>
<td class="paramname"><em>view</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">Element&#160;</td>
<td class="paramname"><em>other</em> = <code>Element(1)</code>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>&lt; Layout function </p>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramname">view</td><td>destination tensor </td></tr>
</table>
</dd>
</dl>
</div>
</div>
</div><!-- contents -->
<!-- start footer part -->
<hr class="footer"/><address class="footer"><small>
Generated by &#160;<a href="http://www.doxygen.org/index.html">
<img class="footer" src="doxygen.png" alt="doxygen"/>
</a> 1.8.11
</small></address>
</body>
</html>