squall/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
7a458f00a6
cutlass/docs/tensor__view_8h_source.html

157 lines
67 KiB
HTML
Raw Normal View History

CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<!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"/>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<meta name="generator" content="Doxygen 1.8.11"/>
<title>CUTLASS: tensor_view.h Source File</title>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<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>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<script type="text/javascript">
$(document).ready(function() { init_search(); });
</script>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
extensions: ["tex2jax.js"],
jax: ["input/TeX","output/HTML-CSS"],
});
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
</script><script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js"></script>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<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;">
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<td id="projectlogo"><img alt="Logo" src="cutlass-logo-small.png"/></td>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<td id="projectalign" style="padding-left: 0.5em;">
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<div id="projectname">CUTLASS
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</div>
<div id="projectbrief">CUDA Templates for Linear Algebra Subroutines and Solvers</div>
</td>
</tr>
</tbody>
</table>
</div>
<!-- end header part -->
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<!-- Generated by Doxygen 1.8.11 -->
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<script type="text/javascript">
var searchBox = new SearchBox("searchBox", "search",false,'Search');
</script>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<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><a href="namespaces.html"><span>Namespaces</span></a></li>
<li><a href="annotated.html"><span>Classes</span></a></li>
<li class="current"><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="files.html"><span>File&#160;List</span></a></li>
<li><a href="globals.html"><span>File&#160;Members</span></a></li>
</ul>
</div>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<!-- 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>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<li class="navelem"><a class="el" href="dir_d44c64559bbebec7f509842c48db8b23.html">include</a></li><li class="navelem"><a class="el" href="dir_6baf2bb612a2f0daa69af3101ede80a1.html">cutlass</a></li> </ul>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</div>
</div><!-- top -->
<div class="header">
<div class="headertitle">
<div class="title">tensor_view.h</div> </div>
</div><!--header-->
<div class="contents">
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<a href="tensor__view_8h.html">Go to the documentation of this file.</a><div class="fragment"><div class="line"><a name="l00001"></a><span class="lineno"> 1</span>&#160;<span class="comment">/***************************************************************************************************</span></div><div class="line"><a name="l00002"></a><span class="lineno"> 2</span>&#160;<span class="comment"> * Copyright (c) 2017-2019, NVIDIA CORPORATION. All rights reserved.</span></div><div class="line"><a name="l00003"></a><span class="lineno"> 3</span>&#160;<span class="comment"> *</span></div><div class="line"><a name="l00004"></a><span class="lineno"> 4</span>&#160;<span class="comment"> * Redistribution and use in source and binary forms, with or without modification, are permitted</span></div><div class="line"><a name="l00005"></a><span class="lineno"> 5</span>&#160;<span class="comment"> * provided that the following conditions are met:</span></div><div class="line"><a name="l00006"></a><span class="lineno"> 6</span>&#160;<span class="comment"> * * Redistributions of source code must retain the above copyright notice, this list of</span></div><div class="line"><a name="l00007"></a><span class="lineno"> 7</span>&#160;<span class="comment"> * conditions and the following disclaimer.</span></div><div class="line"><a name="l00008"></a><span class="lineno"> 8</span>&#160;<span class="comment"> * * Redistributions in binary form must reproduce the above copyright notice, this list of</span></div><div class="line"><a name="l00009"></a><span class="lineno"> 9</span>&#160;<span class="comment"> * conditions and the following disclaimer in the documentation and/or other materials</span></div><div class="line"><a name="l00010"></a><span class="lineno"> 10</span>&#160;<span class="comment"> * provided with the distribution.</span></div><div class="line"><a name="l00011"></a><span class="lineno"> 11</span>&#160;<span class="comment"> * * Neither the name of the NVIDIA CORPORATION nor the names of its contributors may be used</span></div><div class="line"><a name="l00012"></a><span class="lineno"> 12</span>&#160;<span class="comment"> * to endorse or promote products derived from this software without specific prior written</span></div><div class="line"><a name="l00013"></a><span class="lineno"> 13</span>&#160;<span class="comment"> * permission.</span></div><div class="line"><a name="l00014"></a><span class="lineno"> 14</span>&#160;<span class="comment"> *</span></div><div class="line"><a name="l00015"></a><span class="lineno"> 15</span>&#160;<span class="comment"> * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS &quot;AS IS&quot; AND ANY EXPRESS OR</span></div><div class="line"><a name="l00016"></a><span class="lineno"> 16</span>&#160;<span class="comment"> * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND</span></div><div class="line"><a name="l00017"></a><span class="lineno"> 17</span>&#160;<span class="comment"> * FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NVIDIA CORPORATION BE LIABLE</span></div><div class="line"><a name="l00018"></a><span class="lineno"> 18</span>&#160;<span class="comment"> * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,</span></div><div class="line"><a name="l00019"></a><span class="lineno"> 19</span>&#160;<span class="comment"> * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;</span></div><div class="line"><a name="l00020"></a><span class="lineno"> 20</span>&#160;<span class="comment"> * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,</span></div><div class="line"><a name="l00021"></a><span class="lineno"> 21</span>&#160;<span class="comment"> * STRICT LIABILITY, OR TOR (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE</span></div><div class="line"><a name="l00022"
<div class="ttc" id="namespacecutlass_html"><div class="ttname"><a href="namespacecutlass.html">cutlass</a></div><div class="ttdef"><b>Definition:</b> aligned_buffer.h:35</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_ad7c287afe582ff3681a8cd6e53d6a4f5"><div class="ttname"><a href="classcutlass_1_1TensorView.html#ad7c287afe582ff3681a8cd6e53d6a4f5">cutlass::TensorView&lt; Element, Layout &gt;::capacity</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE size_t capacity() const</div><div class="ttdoc">Returns the number of scalar elements needed to store tensor. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:215</div></div>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<div class="ttc" id="tensor__ref_8h_html"><div class="ttname"><a href="tensor__ref_8h.html">tensor_ref.h</a></div><div class="ttdoc">Defines a structure containing strides, bounds, and a pointer to tensor data. </div></div>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<div class="ttc" id="structcutlass_1_1platform_1_1remove__const_html_ac3662947fa50251daf58240a9c798085"><div class="ttname"><a href="structcutlass_1_1platform_1_1remove__const.html#ac3662947fa50251daf58240a9c798085">cutlass::platform::remove_const::type</a></div><div class="ttdeci">T type</div><div class="ttdef"><b>Definition:</b> platform.h:351</div></div>
<div class="ttc" id="classcutlass_1_1TensorRef_html_ac7db3ca62ab1dfe0d3ea08bcadbc9352"><div class="ttname"><a href="classcutlass_1_1TensorRef.html#ac7db3ca62ab1dfe0d3ea08bcadbc9352">cutlass::TensorRef::data</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE Element * data() const </div><div class="ttdoc">Returns the pointer to referenced data. </div><div class="ttdef"><b>Definition:</b> tensor_ref.h:254</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a7d3914dd5042c9c40be9e21a7b4e9ece"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a7d3914dd5042c9c40be9e21a7b4e9ece">cutlass::TensorView::extent</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorCoord const &amp; extent() const </div><div class="ttdoc">Returns the extent of the view (the size along each logical dimension). </div><div class="ttdef"><b>Definition:</b> tensor_view.h:167</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_ab50def50420ed64afefe864d108f3c58"><div class="ttname"><a href="classcutlass_1_1TensorView.html#ab50def50420ed64afefe864d108f3c58">cutlass::TensorView::kRank</a></div><div class="ttdeci">static int const kRank</div><div class="ttdoc">Logical rank of tensor index space. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:78</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a3c3417b9b777f103ec86607a7751e668"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a3c3417b9b777f103ec86607a7751e668">cutlass::TensorView::resize</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE void resize(TensorCoord extent)</div><div class="ttdoc">Changes the size of the view without affecting pointer or layout. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:161</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a481a22e9efb27945c1bccae2381c03fb"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a481a22e9efb27945c1bccae2381c03fb">cutlass::TensorView&lt; Element, Layout &gt;::operator+</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView operator+(TensorCoord const &amp;b) const</div><div class="ttdoc">Returns a TensorView offset by a given amount. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:221</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a1e1533d9b7942be33df1d122667d5ec6"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a1e1533d9b7942be33df1d122667d5ec6">cutlass::TensorView&lt; Element, Layout &gt;::operator-=</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView &amp; operator-=(TensorCoord const &amp;b)</div><div class="ttdoc">Returns a TensorRef offset by a given amount. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:253</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a1d57ad0c37d33a544a07e71925605a1d"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a1d57ad0c37d33a544a07e71925605a1d">cutlass::TensorView&lt; Element, Layout &gt;::operator-</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView operator-(TensorCoord const &amp;b) const</div><div class="ttdoc">Returns a TensorRef offset by a given amount. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:242</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a0baf266b25b2dd5adbc1fa262b510990"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a0baf266b25b2dd5adbc1fa262b510990">cutlass::TensorView::TensorRef</a></div><div class="ttdeci">Base TensorRef</div><div class="ttdoc">Underlying TensorRef type. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:69</div></div>
<div class="ttc" id="classcutlass_1_1TensorRef_html_ad3c5c9466713f62a5191e720827f34da"><div class="ttname"><a href="classcutlass_1_1TensorRef.html#ad3c5c9466713f62a5191e720827f34da">cutlass::TensorRef::ConstTensorRef</a></div><div class="ttdeci">TensorRef&lt; typename platform::remove_const&lt; Element &gt;::type const, Layout &gt; ConstTensorRef</div><div class="ttdoc">TensorRef to constant data. </div><div class="ttdef"><b>Definition:</b> tensor_ref.h:179</div></div>
<div class="ttc" id="classcutlass_1_1TensorRef_html_a4bed879c428963070de8ffbdc5d6e4f9"><div class="ttname"><a href="classcutlass_1_1TensorRef.html#a4bed879c428963070de8ffbdc5d6e4f9">cutlass::TensorRef::add_coord_offset</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorRef &amp; add_coord_offset(TensorCoord const &amp;coord)</div><div class="ttdoc">Adds an offset to each pointer. </div><div class="ttdef"><b>Definition:</b> tensor_ref.h:326</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_afe228764eb67b664fb5ca320c092903b"><div class="ttname"><a href="classcutlass_1_1TensorView.html#afe228764eb67b664fb5ca320c092903b">cutlass::TensorView&lt; Element, Layout &gt;&lt; Element, Layout &gt;::Element</a></div><div class="ttdeci">Element Element</div><div class="ttdoc">Data type of individual access. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:72</div></div>
<div class="ttc" id="cutlass_8h_html_a4b1c9f25ab6eaa25e1f2258dd63e6ce4"><div class="ttname"><a href="cutlass_8h.html#a4b1c9f25ab6eaa25e1f2258dd63e6ce4">CUTLASS_PRAGMA_UNROLL</a></div><div class="ttdeci">#define CUTLASS_PRAGMA_UNROLL</div><div class="ttdef"><b>Definition:</b> cutlass.h:110</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a71def6d54ed28dfe3b17fde3e6461578"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a71def6d54ed28dfe3b17fde3e6461578">cutlass::TensorView::ConstTensorView</a></div><div class="ttdeci">TensorView&lt; typename platform::remove_const&lt; Element &gt;::type const, Layout &gt; ConstTensorView</div><div class="ttdoc">TensorView pointing to constant memory. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:95</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html"><div class="ttname"><a href="classcutlass_1_1TensorView.html">cutlass::TensorView</a></div><div class="ttdef"><b>Definition:</b> tensor_view.h:56</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_add0de4f0548c957df00cae0ee0b9257c"><div class="ttname"><a href="classcutlass_1_1TensorView.html#add0de4f0548c957df00cae0ee0b9257c">cutlass::TensorView&lt; Element, Layout &gt;::reset</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE void reset(Element *ptr, Layout const &amp;layout, TensorCoord size)</div><div class="ttdoc">Updates the pointer and layout object. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:154</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a893017197cb29d46773feea6cdbb25db"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a893017197cb29d46773feea6cdbb25db">cutlass::TensorView&lt; Element, Layout &gt;&lt; Element, Layout &gt;::TensorCoord</a></div><div class="ttdeci">typename Layout::TensorCoord TensorCoord</div><div class="ttdoc">Coordinate in logical tensor space. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:87</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a4d8af7fd842866a218722e7686d6bc3c"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a4d8af7fd842866a218722e7686d6bc3c">cutlass::TensorView&lt; Element, Layout &gt;&lt; Element, Layout &gt;::Reference</a></div><div class="ttdeci">Element &amp; Reference</div><div class="ttdoc">Reference type to an element. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:75</div></div>
<div class="ttc" id="classcutlass_1_1TensorRef_html_a9c2149162016bc19c7735b824d57eb9e"><div class="ttname"><a href="classcutlass_1_1TensorRef.html#a9c2149162016bc19c7735b824d57eb9e">cutlass::TensorRef::reset</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE void reset(Element *ptr=nullptr)</div><div class="ttdoc">Updates only the pointer. </div><div class="ttdef"><b>Definition:</b> tensor_ref.h:235</div></div>
<div class="ttc" id="classcutlass_1_1TensorRef_html"><div class="ttname"><a href="classcutlass_1_1TensorRef.html">cutlass::TensorRef</a></div><div class="ttdef"><b>Definition:</b> tensor_ref.h:146</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a33385840adab818ef25076edab42175d"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a33385840adab818ef25076edab42175d">cutlass::TensorView::ref</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorRef ref() const </div><div class="ttdoc">Returns a TensorRef pointing to the first element of the tensor. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:187</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a6b1bff47f56b66a6da2b2ea35afe583f"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a6b1bff47f56b66a6da2b2ea35afe583f">cutlass::TensorView&lt; Element, Layout &gt;&lt; Element, Layout &gt;::Stride</a></div><div class="ttdeci">typename Layout::Stride Stride</div><div class="ttdoc">Coordinate in storage n-D array. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:90</div></div>
<div class="ttc" id="cutlass_8h_html_a28c2443a142676d3d71effdae1a986b1"><div class="ttname"><a href="cutlass_8h.html#a28c2443a142676d3d71effdae1a986b1">CUTLASS_HOST_DEVICE</a></div><div class="ttdeci">#define CUTLASS_HOST_DEVICE</div><div class="ttdef"><b>Definition:</b> cutlass.h:89</div></div>
<div class="ttc" id="classcutlass_1_1TensorRef_html_a4166ac2a0754574ac21d5d57d74f34e5"><div class="ttname"><a href="classcutlass_1_1TensorRef.html#a4166ac2a0754574ac21d5d57d74f34e5">cutlass::TensorRef::offset</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE LongIndex offset(TensorCoord const &amp;coord) const </div><div class="ttdoc">Computes the offset of an index from the origin of the tensor. </div><div class="ttdef"><b>Definition:</b> tensor_ref.h:301</div></div>
<div class="ttc" id="platform_8h_html_adde4c9ea91b753491851361a4198c009"><div class="ttname"><a href="platform_8h.html#adde4c9ea91b753491851361a4198c009">static_assert</a></div><div class="ttdeci">#define static_assert(__e, __m)</div><div class="ttdef"><b>Definition:</b> platform.h:153</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a1bc72b40e616cbb86a306b8ef0be1575"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a1bc72b40e616cbb86a306b8ef0be1575">cutlass::TensorView&lt; Element, Layout &gt;::extent</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE Index extent(int dim) const</div><div class="ttdoc">Returns the extent along a particular logical dimension. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:171</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a554ebbcae400782746b7cf728ad48093"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a554ebbcae400782746b7cf728ad48093">cutlass::TensorView&lt; Element, Layout &gt;::subview</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView subview(TensorCoord extent, TensorCoord const &amp;location=TensorCoord()) const</div><div class="ttdoc">Returns a Tensor_view given location and size quantities. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:205</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_af9ed97090d3e4df3e32bb177c6745217"><div class="ttname"><a href="classcutlass_1_1TensorView.html#af9ed97090d3e4df3e32bb177c6745217">cutlass::TensorView&lt; Element, Layout &gt;::TensorView</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView(TensorRef const &amp;ref, TensorCoord const &amp;extent)</div><div class="ttdoc">Constructs a TensorView object. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:137</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a0326d4b92019837828d3b737be4f99af"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a0326d4b92019837828d3b737be4f99af">cutlass::TensorView&lt; Element, Layout &gt;::TensorView</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView(NonConstTensorView const &amp;view)</div><div class="ttdoc">Converting constructor from TensorRef to non-constant data. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:147</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_acf1126519c4d26c85ddd20dc66a13e14"><div class="ttname"><a href="classcutlass_1_1TensorView.html#acf1126519c4d26c85ddd20dc66a13e14">cutlass::TensorView&lt; Element, Layout &gt;&lt; Element, Layout &gt;::Index</a></div><div class="ttdeci">typename Layout::Index Index</div><div class="ttdoc">Index type. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:81</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a3c2ec2c816648b7c95d9b9e4b24311ae"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a3c2ec2c816648b7c95d9b9e4b24311ae">cutlass::TensorView::Base</a></div><div class="ttdeci">cutlass::TensorRef&lt; Element_, Layout_ &gt; Base</div><div class="ttdoc">Base tensor reference. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:60</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_ab7794e21b87340f8b73aeb4dca2cb80c"><div class="ttname"><a href="classcutlass_1_1TensorView.html#ab7794e21b87340f8b73aeb4dca2cb80c">cutlass::TensorView&lt; Element, Layout &gt;::const_view</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE ConstTensorView const_view() const</div><div class="ttdoc">Returns a TensorView to const data. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:199</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a48c934ddac84fa964fb9b1364ec44164"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a48c934ddac84fa964fb9b1364ec44164">cutlass::TensorView&lt; Element, Layout &gt;&lt; Element, Layout &gt;::ConstTensorRef</a></div><div class="ttdeci">typename Base::ConstTensorRef ConstTensorRef</div><div class="ttdoc">TensorRef pointing to constant memory. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:66</div></div>
<div class="ttc" id="classcutlass_1_1TensorRef_html_a7771fcd932e36c51c15572305f5b5520"><div class="ttname"><a href="classcutlass_1_1TensorRef.html#a7771fcd932e36c51c15572305f5b5520">cutlass::TensorRef::layout</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE Layout &amp; layout()</div><div class="ttdoc">Returns the layout object. </div><div class="ttdef"><b>Definition:</b> tensor_ref.h:265</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_aad4fd32dda17f8a4fba57cc04ad33004"><div class="ttname"><a href="classcutlass_1_1TensorView.html#aad4fd32dda17f8a4fba57cc04ad33004">cutlass::TensorView&lt; Element, Layout &gt;::TensorView</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView(Element *ptr, Layout const &amp;layout, TensorCoord const &amp;extent)</div><div class="ttdoc">Constructs a TensorView object. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:126</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_ad554523c77f7166cf0a86f44c359bc32"><div class="ttname"><a href="classcutlass_1_1TensorView.html#ad554523c77f7166cf0a86f44c359bc32">cutlass::TensorView&lt; Element, Layout &gt;::TensorView</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView(TensorCoord const &amp;extent=TensorCoord())</div><div class="ttdoc">Constructs a TensorView object. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:120</div></div>
<div class="ttc" id="namespacecutlass_html_a29c8084a6b077ef2d0acb6a4e80f11c8"><div class="ttname"><a href="namespacecutlass.html#a29c8084a6b077ef2d0acb6a4e80f11c8">cutlass::make_TensorView</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorView&lt; Element, Layout &gt; make_TensorView(Element *ptr, Layout const &amp;layout, typename Layout::TensorCoord const &amp;extent)</div><div class="ttdoc">Constructs a TensorRef, deducing types from arguments. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:269</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a6a6a1f99d06abd8fb3f5a8e4e0fea25e"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a6a6a1f99d06abd8fb3f5a8e4e0fea25e">cutlass::TensorView&lt; Element, Layout &gt;::contains</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE bool contains(TensorCoord const &amp;coord) const</div><div class="ttdoc">Determines whether a location is within a tensor. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:175</div></div>
<div class="ttc" id="classcutlass_1_1TensorRef_html_a6bbcd0e512915565cabfeccdb1b6417d"><div class="ttname"><a href="classcutlass_1_1TensorRef.html#a6bbcd0e512915565cabfeccdb1b6417d">cutlass::TensorRef::add_pointer_offset</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE TensorRef &amp; add_pointer_offset(LongIndex offset_)</div><div class="ttdoc">Adds an offset to each pointer. </div><div class="ttdef"><b>Definition:</b> tensor_ref.h:319</div></div>
<div class="ttc" id="cutlass_8h_html"><div class="ttname"><a href="cutlass_8h.html">cutlass.h</a></div><div class="ttdoc">Basic include for CUTLASS. </div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_aef27ab5348a53539286057a0da8720fc"><div class="ttname"><a href="classcutlass_1_1TensorView.html#aef27ab5348a53539286057a0da8720fc">cutlass::TensorView&lt; Element, Layout &gt;::const_ref</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE ConstTensorRef const_ref() const</div><div class="ttdoc">Returns a TensorRef pointing to the first element of the tensor. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:193</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_a59dc3d93fa2f5132659f768c9d254627"><div class="ttname"><a href="classcutlass_1_1TensorView.html#a59dc3d93fa2f5132659f768c9d254627">cutlass::TensorView&lt; Element, Layout &gt;&lt; Element, Layout &gt;::LongIndex</a></div><div class="ttdeci">typename Layout::LongIndex LongIndex</div><div class="ttdoc">Long index used for pointer offsets. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:84</div></div>
<div class="ttc" id="classcutlass_1_1TensorView_html_ae5168523626a13d6e7391a64bca8caf0"><div class="ttname"><a href="classcutlass_1_1TensorView.html#ae5168523626a13d6e7391a64bca8caf0">cutlass::TensorView&lt; Element, Layout &gt;&lt; Element, Layout &gt;::Layout</a></div><div class="ttdeci">Layout Layout</div><div class="ttdoc">Mapping function from logical coordinate to internal n-D array. </div><div class="ttdef"><b>Definition:</b> tensor_view.h:63</div></div>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</div><!-- fragment --></div><!-- contents -->
<!-- start footer part -->
<hr class="footer"/><address class="footer"><small>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
Generated by &#160;<a href="http://www.doxygen.org/index.html">
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<img class="footer" src="doxygen.png" alt="doxygen"/>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
</a> 1.8.11
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</small></address>
</body>
</html>
Reference in New Issue Copy Permalink