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
@@ -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::platform::integral_constant&lt; value_t, V &gt; Struct Template Reference</title>
<meta name="generator" content="Doxygen 1.8.11"/>
<title>CUTLASS: cutlass::platform::integral_constant&lt; value_t, V &gt; Struct Template 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 +39,43 @@
</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 class="current"><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="annotated.html"><span>Class&#160;List</span></a></li>
<li><a href="classes.html"><span>Class&#160;Index</span></a></li>
<li><a href="inherits.html"><span>Class&#160;Hierarchy</span></a></li>
<li><a href="functions.html"><span>Class&#160;Members</span></a></li>
</ul>
</div>
<!-- window showing the filter options -->
<div id="MSearchSelectWindow"
onmouseover="return searchBox.OnSearchSelectShow()"
@@ -89,31 +113,17 @@ $(function() {
<div class="dynheader">
Inheritance diagram for cutlass::platform::integral_constant&lt; value_t, V &gt;:</div>
<div class="dyncontent">
<div class="center">
<img src="structcutlass_1_1platform_1_1integral__constant.png" usemap="#cutlass::platform::integral_5Fconstant_3C_20value_5Ft_2C_20V_20_3E_map" alt=""/>
<map id="cutlass::platform::integral_5Fconstant_3C_20value_5Ft_2C_20V_20_3E_map" name="cutlass::platform::integral_5Fconstant_3C_20value_5Ft_2C_20V_20_3E_map">
<area href="structcutlass_1_1platform_1_1is__integral.html" title="std::is_integral " alt="cutlass::platform::is_integral&lt; T &gt;" shape="rect" coords="364,56,718,80"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01char_01_4.html" alt="cutlass::platform::is_integral&lt; char &gt;" shape="rect" coords="364,112,718,136"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01int_01_4.html" alt="cutlass::platform::is_integral&lt; int &gt;" shape="rect" coords="364,168,718,192"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01long_01_4.html" alt="cutlass::platform::is_integral&lt; long &gt;" shape="rect" coords="364,224,718,248"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01long_01long_01_4.html" alt="cutlass::platform::is_integral&lt; long long &gt;" shape="rect" coords="364,280,718,304"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01short_01_4.html" alt="cutlass::platform::is_integral&lt; short &gt;" shape="rect" coords="364,336,718,360"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01signed_01char_01_4.html" alt="cutlass::platform::is_integral&lt; signed char &gt;" shape="rect" coords="364,392,718,416"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01unsigned_01char_01_4.html" alt="cutlass::platform::is_integral&lt; unsigned char &gt;" shape="rect" coords="364,448,718,472"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01unsigned_01int_01_4.html" alt="cutlass::platform::is_integral&lt; unsigned int &gt;" shape="rect" coords="364,504,718,528"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01unsigned_01long_01_4.html" alt="cutlass::platform::is_integral&lt; unsigned long &gt;" shape="rect" coords="364,560,718,584"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01unsigned_01long_01long_01_4.html" alt="cutlass::platform::is_integral&lt; unsigned long long &gt;" shape="rect" coords="364,616,718,640"/>
<area href="structcutlass_1_1platform_1_1is__integral_3_01unsigned_01short_01_4.html" alt="cutlass::platform::is_integral&lt; unsigned short &gt;" shape="rect" coords="364,672,718,696"/>
<area href="structcutlass_1_1platform_1_1is__pointer__helper.html" title="Helper for std::is_pointer (false specialization) " alt="cutlass::platform::is_pointer_helper&lt; T &gt;" shape="rect" coords="364,728,718,752"/>
<area href="structcutlass_1_1platform_1_1is__pointer__helper_3_01T_01_5_01_4.html" title="Helper for std::is_pointer (true specialization) " alt="cutlass::platform::is_pointer_helper&lt; T * &gt;" shape="rect" coords="364,784,718,808"/>
<area href="structcutlass_1_1platform_1_1is__same.html" title="std::is_same (false specialization) " alt="cutlass::platform::is_same&lt; A, B &gt;" shape="rect" coords="364,840,718,864"/>
<area href="structcutlass_1_1platform_1_1is__same_3_01A_00_01A_01_4.html" title="std::is_same (true specialization) " alt="cutlass::platform::is_same&lt; A, A &gt;" shape="rect" coords="364,896,718,920"/>
<area href="structcutlass_1_1platform_1_1is__volatile.html" title="std::is_volatile " alt="cutlass::platform::is_volatile&lt; T &gt;" shape="rect" coords="364,952,718,976"/>
<area href="structcutlass_1_1platform_1_1is__volatile_3_01volatile_01T_01_4.html" alt="cutlass::platform::is_volatile&lt; volatile T &gt;" shape="rect" coords="364,1008,718,1032"/>
<area href="structcutlass_1_1platform_1_1is__pointer__helper.html" alt="cutlass::platform::is_pointer_helper&lt; remove_cv&lt; T &gt;::type &gt;" shape="rect" coords="364,1064,718,1088"/>
<area href="structcutlass_1_1platform_1_1is__same.html" alt="cutlass::platform::is_same&lt; void, remove_cv&lt; T &gt;::type &gt;" shape="rect" coords="364,1120,718,1144"/>
<div class="center"><img src="structcutlass_1_1platform_1_1integral__constant__inherit__graph.png" border="0" usemap="#cutlass_1_1platform_1_1integral__constant_3_01value__t_00_01V_01_4_inherit__map" alt="Inheritance graph"/></div>
<map name="cutlass_1_1platform_1_1integral__constant_3_01value__t_00_01V_01_4_inherit__map" id="cutlass_1_1platform_1_1integral__constant_3_01value__t_00_01V_01_4_inherit__map">
</map>
</div></div>
<center><span class="legend">[<a href="graph_legend.html">legend</a>]</span></center></div>
<div class="dynheader">
Collaboration diagram for cutlass::platform::integral_constant&lt; value_t, V &gt;:</div>
<div class="dyncontent">
<div class="center"><img src="structcutlass_1_1platform_1_1integral__constant__coll__graph.png" border="0" usemap="#cutlass_1_1platform_1_1integral__constant_3_01value__t_00_01V_01_4_coll__map" alt="Collaboration graph"/></div>
<map name="cutlass_1_1platform_1_1integral__constant_3_01value__t_00_01V_01_4_coll__map" id="cutlass_1_1platform_1_1integral__constant_3_01value__t_00_01V_01_4_coll__map">
</map>
<center><span class="legend">[<a href="graph_legend.html">legend</a>]</span></center></div>
<table class="memberdecls">
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="pub-types"></a>
Public Types</h2></td></tr>
@@ -124,10 +134,10 @@ Public Types</h2></td></tr>
</table><table class="memberdecls">
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="pub-methods"></a>
Public Member Functions</h2></td></tr>
<tr class="memitem:a55d25116387f1c6d978462b1d245d675"><td class="memItemLeft" align="right" valign="top"><a class="el" href="cutlass_8h.html#a28c2443a142676d3d71effdae1a986b1">CUTLASS_HOST_DEVICE</a>&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1platform_1_1integral__constant.html#a55d25116387f1c6d978462b1d245d675">operator value_type</a> () const</td></tr>
<tr class="separator:a55d25116387f1c6d978462b1d245d675"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a5271a533526a535ae8b783c736252f18"><td class="memItemLeft" align="right" valign="top"><a class="el" href="cutlass_8h.html#a28c2443a142676d3d71effdae1a986b1">CUTLASS_HOST_DEVICE</a> const <a class="el" href="structcutlass_1_1platform_1_1integral__constant.html#ab2ed0b3506818139f1f96639742e79fd">value_type</a>&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1platform_1_1integral__constant.html#a5271a533526a535ae8b783c736252f18">operator()</a> () const</td></tr>
<tr class="separator:a5271a533526a535ae8b783c736252f18"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:adfcb0c650152fe171f00c7b817c69ae1"><td class="memItemLeft" align="right" valign="top"><a class="el" href="cutlass_8h.html#a28c2443a142676d3d71effdae1a986b1">CUTLASS_HOST_DEVICE</a>&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1platform_1_1integral__constant.html#adfcb0c650152fe171f00c7b817c69ae1">operator value_type</a> () const </td></tr>
<tr class="separator:adfcb0c650152fe171f00c7b817c69ae1"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a56a164b98a93b1430c26cdfde6f3720d"><td class="memItemLeft" align="right" valign="top"><a class="el" href="cutlass_8h.html#a28c2443a142676d3d71effdae1a986b1">CUTLASS_HOST_DEVICE</a> const <a class="el" href="structcutlass_1_1platform_1_1integral__constant.html#ab2ed0b3506818139f1f96639742e79fd">value_type</a>&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="structcutlass_1_1platform_1_1integral__constant.html#a56a164b98a93b1430c26cdfde6f3720d">operator()</a> () const </td></tr>
<tr class="separator:a56a164b98a93b1430c26cdfde6f3720d"><td class="memSeparator" colspan="2">&#160;</td></tr>
</table><table class="memberdecls">
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="pub-static-attribs"></a>
Static Public Attributes</h2></td></tr>
@@ -135,9 +145,7 @@ Static Public Attributes</h2></td></tr>
<tr class="separator:a9bbaca83ae76941edb9b75b2741d3ad9"><td class="memSeparator" colspan="2">&#160;</td></tr>
</table>
<h2 class="groupheader">Member Typedef Documentation</h2>
<a id="af58810ccead8f16ed88cd6a4afdc6e52"></a>
<h2 class="memtitle"><span class="permalink"><a href="#af58810ccead8f16ed88cd6a4afdc6e52">&#9670;&nbsp;</a></span>type</h2>
<a class="anchor" id="af58810ccead8f16ed88cd6a4afdc6e52"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
@@ -151,9 +159,7 @@ template&lt;typename value_t, value_t V&gt; </div>
</div>
</div>
<a id="ab2ed0b3506818139f1f96639742e79fd"></a>
<h2 class="memtitle"><span class="permalink"><a href="#ab2ed0b3506818139f1f96639742e79fd">&#9670;&nbsp;</a></span>value_type</h2>
<a class="anchor" id="ab2ed0b3506818139f1f96639742e79fd"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
@@ -168,9 +174,7 @@ template&lt;typename value_t, value_t V&gt; </div>
</div>
</div>
<h2 class="groupheader">Member Function Documentation</h2>
<a id="a55d25116387f1c6d978462b1d245d675"></a>
<h2 class="memtitle"><span class="permalink"><a href="#a55d25116387f1c6d978462b1d245d675">&#9670;&nbsp;</a></span>operator value_type()</h2>
<a class="anchor" id="adfcb0c650152fe171f00c7b817c69ae1"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
@@ -195,9 +199,7 @@ template&lt;typename value_t, value_t V&gt; </div>
</div>
</div>
<a id="a5271a533526a535ae8b783c736252f18"></a>
<h2 class="memtitle"><span class="permalink"><a href="#a5271a533526a535ae8b783c736252f18">&#9670;&nbsp;</a></span>operator()()</h2>
<a class="anchor" id="a56a164b98a93b1430c26cdfde6f3720d"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
@@ -223,9 +225,7 @@ template&lt;typename value_t, value_t V&gt; </div>
</div>
</div>
<h2 class="groupheader">Member Data Documentation</h2>
<a id="a9bbaca83ae76941edb9b75b2741d3ad9"></a>
<h2 class="memtitle"><span class="permalink"><a href="#a9bbaca83ae76941edb9b75b2741d3ad9">&#9670;&nbsp;</a></span>value</h2>
<a class="anchor" id="a9bbaca83ae76941edb9b75b2741d3ad9"></a>
<div class="memitem">
<div class="memproto">
<div class="memtemplate">
@@ -253,9 +253,9 @@ template&lt;typename value_t, value_t V&gt; </div>
</div><!-- contents -->
<!-- start footer part -->
<hr class="footer"/><address class="footer"><small>
Generated on Fri Oct 26 2018 14:53:40 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>