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.
This commit is contained in:
Andrew Kerr
2019-11-19 16:55:34 -08:00
committed by GitHub
parent b5cab177a9
commit fb335f6a5f
5434 changed files with 599799 additions and 250176 deletions
+82 -91
View File
@@ -3,21 +3,23 @@
<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 1.8.14"/>
<meta name="viewport" content="width=device-width, initial-scale=1"/>
<title>Cutlass: cutlass.h File Reference</title>
<meta name="generator" content="Doxygen 1.8.11"/>
<title>CUTLASS: cutlass.h File 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" async src="http://cdn.mathjax.org/mathjax/latest/MathJax.js"></script>
</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>
@@ -26,8 +28,9 @@
<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 id="projectname">CUTLASS
</div>
<div id="projectbrief">CUDA Templates for Linear Algebra Subroutines and Solvers</div>
</td>
@@ -36,22 +39,41 @@
</table>
</div>
<!-- end header part -->
<!-- Generated by Doxygen 1.8.14 -->
<!-- Generated by Doxygen 1.8.11 -->
<script type="text/javascript">
/* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&amp;dn=gpl-2.0.txt GPL-v2 */
var searchBox = new SearchBox("searchBox", "search",false,'Search');
/* @license-end */
</script>
<script type="text/javascript" src="menudata.js"></script>
<script type="text/javascript" src="menu.js"></script>
<script type="text/javascript">
/* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&amp;dn=gpl-2.0.txt GPL-v2 */
$(function() {
initMenu('',true,false,'search.php','Search');
$(document).ready(function() { init_search(); });
});
/* @license-end */</script>
<div id="main-nav"></div>
<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>
<!-- window showing the filter options -->
<div id="MSearchSelectWindow"
onmouseover="return searchBox.OnSearchSelectShow()"
@@ -68,33 +90,24 @@ $(function() {
<div id="nav-path" class="navpath">
<ul>
<li class="navelem"><a class="el" href="dir_1417ee5ebebc309c36b7962f26a92c39.html">cutlass</a></li> </ul>
<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>
</div>
</div><!-- top -->
<div class="header">
<div class="summary">
<a href="#nested-classes">Classes</a> &#124;
<a href="#namespaces">Namespaces</a> &#124;
<a href="#define-members">Macros</a> &#124;
<a href="#func-members">Functions</a> </div>
<a href="#enum-members">Enumerations</a> </div>
<div class="headertitle">
<div class="title">cutlass.h File Reference</div> </div>
</div><!--header-->
<div class="contents">
<p>Basic include for CUTLASS macros.
<p>Basic include for CUTLASS.
<a href="#details">More...</a></p>
<div class="textblock"><code>#include &quot;<a class="el" href="performance__tuning_8h_source.html">cutlass/util/performance_tuning.h</a>&quot;</code><br />
</div>
<p><a href="cutlass_8h_source.html">Go to the source code of this file.</a></p>
<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="structDebugType.html">DebugType&lt; T &gt;</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="structDebugValue.html">DebugValue&lt; Value &gt;</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="namespaces"></a>
Namespaces</h2></td></tr>
<tr class="memitem:namespacecutlass"><td class="memItemLeft" align="right" valign="top"> &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="namespacecutlass.html">cutlass</a></td></tr>
@@ -102,29 +115,38 @@ Namespaces</h2></td></tr>
</table><table class="memberdecls">
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="define-members"></a>
Macros</h2></td></tr>
<tr class="memitem:a8ff3cda9323810c1c504793a0206d4b8"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#a8ff3cda9323810c1c504793a0206d4b8">CUTLASS_MAJOR</a>&#160;&#160;&#160;1</td></tr>
<tr class="separator:a8ff3cda9323810c1c504793a0206d4b8"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:ad114a1ab01f73833ea00020ffb7bcea7"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#ad114a1ab01f73833ea00020ffb7bcea7">CUTLASS_MINOR</a>&#160;&#160;&#160;2</td></tr>
<tr class="separator:ad114a1ab01f73833ea00020ffb7bcea7"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a1d4e5818a594bbfc472e54978955cb8b"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#a1d4e5818a594bbfc472e54978955cb8b">CUTLASS_PATCH</a>&#160;&#160;&#160;0</td></tr>
<tr class="separator:a1d4e5818a594bbfc472e54978955cb8b"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:aa3040eddf073214969f9445bfa925039"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#aa3040eddf073214969f9445bfa925039">CUTLASS_VERSION</a>&#160;&#160;&#160;((<a class="el" href="cutlass_8h.html#a8ff3cda9323810c1c504793a0206d4b8">CUTLASS_MAJOR</a>)*100 + (<a class="el" href="cutlass_8h.html#ad114a1ab01f73833ea00020ffb7bcea7">CUTLASS_MINOR</a>)*10 + <a class="el" href="cutlass_8h.html#a1d4e5818a594bbfc472e54978955cb8b">CUTLASS_PATCH</a>)</td></tr>
<tr class="separator:aa3040eddf073214969f9445bfa925039"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a28c2443a142676d3d71effdae1a986b1"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#a28c2443a142676d3d71effdae1a986b1">CUTLASS_HOST_DEVICE</a></td></tr>
<tr class="memitem:a60467e248cf2879c0531d122a3e2a32d"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#a60467e248cf2879c0531d122a3e2a32d">CUTLASS_ENABLE_TENSOR_CORE_MMA</a>&#160;&#160;&#160;0</td></tr>
<tr class="separator:a60467e248cf2879c0531d122a3e2a32d"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a28c2443a142676d3d71effdae1a986b1"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#a28c2443a142676d3d71effdae1a986b1">CUTLASS_HOST_DEVICE</a>&#160;&#160;&#160;inline</td></tr>
<tr class="separator:a28c2443a142676d3d71effdae1a986b1"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a0159b8e4cd578881a1ccfd0921516af7"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#a0159b8e4cd578881a1ccfd0921516af7">CUTLASS_ASSERT</a>(x)&#160;&#160;&#160;assert(x)</td></tr>
<tr class="separator:a0159b8e4cd578881a1ccfd0921516af7"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a4b1c9f25ab6eaa25e1f2258dd63e6ce4"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#a4b1c9f25ab6eaa25e1f2258dd63e6ce4">CUTLASS_PRAGMA_UNROLL</a></td></tr>
<tr class="separator:a4b1c9f25ab6eaa25e1f2258dd63e6ce4"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:adb3bc73d74b4a4bf13099d5696db3352"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#adb3bc73d74b4a4bf13099d5696db3352">CUTLASS_PRAGMA_NO_UNROLL</a></td></tr>
<tr class="separator:adb3bc73d74b4a4bf13099d5696db3352"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a8e18ced39c05ab3304bb4fcdc0cc9f71"><td class="memItemLeft" align="right" valign="top">#define&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#a8e18ced39c05ab3304bb4fcdc0cc9f71">CUTLASS_GEMM_LOOP</a></td></tr>
<tr class="separator:a8e18ced39c05ab3304bb4fcdc0cc9f71"><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:ab7e23b523490567225b20e2c72649f20"><td class="memTemplParams" colspan="2">template&lt;typename T &gt; </td></tr>
<tr class="memitem:ab7e23b523490567225b20e2c72649f20"><td class="memTemplItemLeft" align="right" valign="top">void&#160;</td><td class="memTemplItemRight" valign="bottom"><a class="el" href="cutlass_8h.html#ab7e23b523490567225b20e2c72649f20">DebugTypeFunc</a> (T const &amp;t)</td></tr>
<tr class="separator:ab7e23b523490567225b20e2c72649f20"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="enum-members"></a>
Enumerations</h2></td></tr>
<tr class="memitem:ac5a88c5840a28a9e0206b9cc7812a18d"><td class="memItemLeft" align="right" valign="top">enum &#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18d">cutlass::Status</a> { <br />
&#160;&#160;<a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18da8c632159fa131f09d04f94e3cbcd8782">cutlass::Status::kSuccess</a>,
<a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18daa4867e1466f5d067dbec566abfe5a67a">cutlass::Status::kErrorMisalignedOperand</a>,
<a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18da7fa8ed49e2d3f7e9af8ad025cdca548c">cutlass::Status::kErrorInvalidLayout</a>,
<a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18da58c16acfe39125423ff22c907487ba2c">cutlass::Status::kErrorInvalidProblem</a>,
<br />
&#160;&#160;<a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18da3af1f6b1c6965aa925910b89e4152ab8">cutlass::Status::kErrorNotSupported</a>,
<a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18da6467125354474c5cff2f3c920abb7d3f">cutlass::Status::kErrorWorkspaceNull</a>,
<a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18da4d3b5847a0b17037c6b69faf8b1d4d71">cutlass::Status::kErrorInternal</a>,
<a class="el" href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18dab10913c938482a8aa4ba85b7a1116cb4">cutlass::Status::kInvalid</a>
<br />
}<tr class="memdesc:ac5a88c5840a28a9e0206b9cc7812a18d"><td class="mdescLeft">&#160;</td><td class="mdescRight">Status code returned by CUTLASS operations. <a href="namespacecutlass.html#ac5a88c5840a28a9e0206b9cc7812a18d">More...</a><br /></td></tr>
</td></tr>
<tr class="separator:ac5a88c5840a28a9e0206b9cc7812a18d"><td class="memSeparator" colspan="2">&#160;</td></tr>
</table>
<h2 class="groupheader">Macro Definition Documentation</h2>
<a id="a0159b8e4cd578881a1ccfd0921516af7"></a>
<h2 class="memtitle"><span class="permalink"><a href="#a0159b8e4cd578881a1ccfd0921516af7">&#9670;&nbsp;</a></span>CUTLASS_ASSERT</h2>
<a class="anchor" id="a0159b8e4cd578881a1ccfd0921516af7"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
@@ -140,91 +162,60 @@ Functions</h2></td></tr>
</div>
</div>
<a id="a28c2443a142676d3d71effdae1a986b1"></a>
<h2 class="memtitle"><span class="permalink"><a href="#a28c2443a142676d3d71effdae1a986b1">&#9670;&nbsp;</a></span>CUTLASS_HOST_DEVICE</h2>
<a class="anchor" id="a60467e248cf2879c0531d122a3e2a32d"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">#define CUTLASS_HOST_DEVICE</td>
<td class="memname">#define CUTLASS_ENABLE_TENSOR_CORE_MMA&#160;&#160;&#160;0</td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a id="a8ff3cda9323810c1c504793a0206d4b8"></a>
<h2 class="memtitle"><span class="permalink"><a href="#a8ff3cda9323810c1c504793a0206d4b8">&#9670;&nbsp;</a></span>CUTLASS_MAJOR</h2>
<a class="anchor" id="a8e18ced39c05ab3304bb4fcdc0cc9f71"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">#define CUTLASS_MAJOR&#160;&#160;&#160;1</td>
<td class="memname">#define CUTLASS_GEMM_LOOP</td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a id="ad114a1ab01f73833ea00020ffb7bcea7"></a>
<h2 class="memtitle"><span class="permalink"><a href="#ad114a1ab01f73833ea00020ffb7bcea7">&#9670;&nbsp;</a></span>CUTLASS_MINOR</h2>
<a class="anchor" id="a28c2443a142676d3d71effdae1a986b1"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">#define CUTLASS_MINOR&#160;&#160;&#160;2</td>
<td class="memname">#define CUTLASS_HOST_DEVICE&#160;&#160;&#160;inline</td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a id="a1d4e5818a594bbfc472e54978955cb8b"></a>
<h2 class="memtitle"><span class="permalink"><a href="#a1d4e5818a594bbfc472e54978955cb8b">&#9670;&nbsp;</a></span>CUTLASS_PATCH</h2>
<a class="anchor" id="adb3bc73d74b4a4bf13099d5696db3352"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">#define CUTLASS_PATCH&#160;&#160;&#160;0</td>
<td class="memname">#define CUTLASS_PRAGMA_NO_UNROLL</td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<a id="aa3040eddf073214969f9445bfa925039"></a>
<h2 class="memtitle"><span class="permalink"><a href="#aa3040eddf073214969f9445bfa925039">&#9670;&nbsp;</a></span>CUTLASS_VERSION</h2>
<a class="anchor" id="a4b1c9f25ab6eaa25e1f2258dd63e6ce4"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">#define CUTLASS_VERSION&#160;&#160;&#160;((<a class="el" href="cutlass_8h.html#a8ff3cda9323810c1c504793a0206d4b8">CUTLASS_MAJOR</a>)*100 + (<a class="el" href="cutlass_8h.html#ad114a1ab01f73833ea00020ffb7bcea7">CUTLASS_MINOR</a>)*10 + <a class="el" href="cutlass_8h.html#a1d4e5818a594bbfc472e54978955cb8b">CUTLASS_PATCH</a>)</td>
</tr>
</table>
</div><div class="memdoc">
</div>
</div>
<h2 class="groupheader">Function Documentation</h2>
<a id="ab7e23b523490567225b20e2c72649f20"></a>
<h2 class="memtitle"><span class="permalink"><a href="#ab7e23b523490567225b20e2c72649f20">&#9670;&nbsp;</a></span>DebugTypeFunc()</h2>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
template&lt;typename T &gt; </div>
<table class="memname">
<tr>
<td class="memname">void DebugTypeFunc </td>
<td>(</td>
<td class="paramtype">T const &amp;&#160;</td>
<td class="paramname"><em>t</em></td><td>)</td>
<td></td>
<td class="memname">#define CUTLASS_PRAGMA_UNROLL</td>
</tr>
</table>
</div><div class="memdoc">
@@ -234,9 +225,9 @@ template&lt;typename T &gt; </div>
</div><!-- contents -->
<!-- start footer part -->
<hr class="footer"/><address class="footer"><small>
Generated on Fri Oct 26 2018 14:53:34 for Cutlass by &#160;<a href="http://www.doxygen.org/index.html">
Generated by &#160;<a href="http://www.doxygen.org/index.html">
<img class="footer" src="doxygen.png" alt="doxygen"/>
</a> 1.8.14
</a> 1.8.11
</small></address>
</body>
</html>