squall/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
7a458f00a6
cutlass/docs/batched__reduction__traits_8h_source.html

151 lines
51 KiB
HTML
Raw Normal View History

Updating Doxygen docs
2018-10-27 05:54:58 +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: batched_reduction_traits.h Source File</title>
Updating Doxygen docs
2018-10-27 05:54:58 +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>
Updating Doxygen docs
2018-10-27 05:54:58 +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>
Updating Doxygen docs
2018-10-27 05:54:58 +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>
Updating Doxygen docs
2018-10-27 05:54:58 +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
Updating Doxygen docs
2018-10-27 05:54:58 +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 -->
Updating Doxygen docs
2018-10-27 05:54:58 +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>
Updating Doxygen docs
2018-10-27 05:54:58 +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><li class="navelem"><a class="el" href="dir_ac488927e63b76ba9cb3ad9c317bbde9.html">reduction</a></li> </ul>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
</div>
</div><!-- top -->
<div class="header">
<div class="headertitle">
<div class="title">batched_reduction_traits.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="batched__reduction__traits_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"></a><
<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>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html">cutlass::reduction::BatchedReductionTraits::Params</a></div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:138</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_ae0c016bcbe687063774d8abd554939b6"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ae0c016bcbe687063774d8abd554939b6">cutlass::reduction::BatchedReductionTraits::BlockSwizzle</a></div><div class="ttdeci">BlockSwizzle_ BlockSwizzle</div><div class="ttdoc">The thread block swizzle. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:113</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_acfeed4cb4e5eec9d8e1ae2b787cc88e2"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#acfeed4cb4e5eec9d8e1ae2b787cc88e2">cutlass::reduction::BatchedReductionTraits::This_</a></div><div class="ttdeci">BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt; This_</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:91</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_ab35edeae5cd8767bd376fad5f6680e25"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ab35edeae5cd8767bd376fad5f6680e25">cutlass::reduction::BatchedReductionTraits::kThreads</a></div><div class="ttdeci">static int const kThreads</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:122</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="namespacecutlass_html_a7419519fa453a121dfa5f26bf87318d9"><div class="ttname"><a href="namespacecutlass.html#a7419519fa453a121dfa5f26bf87318d9">cutlass::make_Coord</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE Coord&lt; 1 &gt; make_Coord(int _0)</div><div class="ttdoc">Helper to make a 2-element coordinate. </div><div class="ttdef"><b>Definition:</b> coord.h:387</div></div>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_ae7e468b1d372b4b807e2e1089af885ec"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ae7e468b1d372b4b807e2e1089af885ec">cutlass::reduction::BatchedReductionTraits::ScalarAccum</a></div><div class="ttdeci">ScalarAccum_ ScalarAccum</div><div class="ttdoc">The type for accumulation. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:109</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_ab14fa69759f1d600a28df4fabaf59c28"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#ab14fa69759f1d600a28df4fabaf59c28">cutlass::reduction::BatchedReductionTraits::Params::lda</a></div><div class="ttdeci">Index lda</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:150</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_a805f78cae27c3305c988f251207d85f7"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#a805f78cae27c3305c988f251207d85f7">cutlass::reduction::BatchedReductionTraits::Params::beta</a></div><div class="ttdeci">ScalarAlphaBeta beta</div><div class="ttdoc">The beta. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:144</div></div>
<div class="ttc" id="reduction_2threadblock__swizzle_8h_html"><div class="ttname"><a href="reduction_2threadblock__swizzle_8h.html">threadblock_swizzle.h</a></div><div class="ttdoc">Defies functors for mapping blockIdx to partitions of the batched reduction computation. </div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_a29e8adec50bc7e5c0a91f975dff4a7f3"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#a29e8adec50bc7e5c0a91f975dff4a7f3">cutlass::reduction::BatchedReductionTraits::ThreadShape</a></div><div class="ttdeci">ThreadShape_ ThreadShape</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:99</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_ad24096385e51009866b5aafd90b42ad2"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#ad24096385e51009866b5aafd90b42ad2">cutlass::reduction::BatchedReductionTraits::Params::ldd</a></div><div class="ttdeci">Index ldd</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:158</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_abb54e3addfee4097b37deb5cb30fb582"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#abb54e3addfee4097b37deb5cb30fb582">cutlass::reduction::BatchedReductionTraits::ScalarD</a></div><div class="ttdeci">ScalarD_ ScalarD</div><div class="ttdoc">The output pointer type. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:105</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_a5d1463d473d4226b0d19c581b16ed3b2"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#a5d1463d473d4226b0d19c581b16ed3b2">cutlass::reduction::BatchedReductionTraits::Params::reduction_stride</a></div><div class="ttdeci">long long int reduction_stride</div><div class="ttdoc">stride between two element that will be sumed </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:146</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_ad9d4dcb20a69869ebda639747a16e647"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ad9d4dcb20a69869ebda639747a16e647">cutlass::reduction::BatchedReductionTraits::SubTile</a></div><div class="ttdeci">SubTile_ SubTile</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:97</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_1reduction_1_1BatchedReductionTraits_1_1Params_html_ac79830eaf080ea0ffddd2100db6cf3e1"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#ac79830eaf080ea0ffddd2100db6cf3e1">cutlass::reduction::BatchedReductionTraits::Params::d_c</a></div><div class="ttdeci">ScalarC const * d_c</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:152</div></div>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_a3b0e98519b7e551ef6adac69e7406d8f"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#a3b0e98519b7e551ef6adac69e7406d8f">cutlass::reduction::BatchedReductionTraits::OutputTile</a></div><div class="ttdeci">OutputTile_ OutputTile</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:95</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_a6bdb4dfce17da648b55c7aa68ef5b191"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#a6bdb4dfce17da648b55c7aa68ef5b191">cutlass::reduction::BatchedReductionTraits::Params::ldc</a></div><div class="ttdeci">Index ldc</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:154</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_ab89c35cfce0017a47341a1e3b2894e0f"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ab89c35cfce0017a47341a1e3b2894e0f">cutlass::reduction::BatchedReductionTraits::ScalarAlphaBeta</a></div><div class="ttdeci">ScalarAlphaBeta_ ScalarAlphaBeta</div><div class="ttdoc">The alpha beta type. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:107</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_adb39cf54a839bdb2e38fbc8a0bf304a8"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#adb39cf54a839bdb2e38fbc8a0bf304a8">cutlass::reduction::BatchedReductionTraits::ScalarC</a></div><div class="ttdeci">ScalarC_ ScalarC</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:103</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_1reduction_1_1BatchedReductionTraits_1_1Params_html_af1b12ba220602692e84616b420b00f1c"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#af1b12ba220602692e84616b420b00f1c">cutlass::reduction::BatchedReductionTraits::Params::d_a</a></div><div class="ttdeci">ScalarA const * d_a</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:148</div></div>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReduction_html"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReduction.html">cutlass::reduction::BatchedReduction</a></div><div class="ttdef"><b>Definition:</b> batched_reduction.h:52</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_a00c71c9a18aaad84f4a48023dbbb454e"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#a00c71c9a18aaad84f4a48023dbbb454e">cutlass::reduction::BatchedReductionTraits::ReductionSize</a></div><div class="ttdeci">static const int ReductionSize</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:115</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_a1c6cd0a76b2b2fc9cd021b016f30f459"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#a1c6cd0a76b2b2fc9cd021b016f30f459">cutlass::reduction::BatchedReductionTraits::ScalarA</a></div><div class="ttdeci">ScalarA_ ScalarA</div><div class="ttdoc">The input pointer type. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:101</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="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>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<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="structcutlass_1_1Coord_html"><div class="ttname"><a href="structcutlass_1_1Coord.html">cutlass::Coord&lt; 3 &gt;</a></div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_afada1cbad87636228fb58d8577bb8470"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#afada1cbad87636228fb58d8577bb8470">cutlass::reduction::BatchedReductionTraits::Params::alpha</a></div><div class="ttdeci">ScalarAlphaBeta alpha</div><div class="ttdoc">The alpha. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:142</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_af62fd99fae22a4e4d93cfa0560f1dcc5"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#af62fd99fae22a4e4d93cfa0560f1dcc5">cutlass::reduction::BatchedReductionTraits::Functor</a></div><div class="ttdeci">Functor_ Functor</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:119</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_af11a3284195a24e580d2f379f179f05a"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#af11a3284195a24e580d2f379f179f05a">cutlass::reduction::BatchedReductionTraits::maxInReg</a></div><div class="ttdeci">static int const maxInReg</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:124</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_abf9744373a72f3819a616b5a5b3bff22"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#abf9744373a72f3819a616b5a5b3bff22">cutlass::reduction::BatchedReductionTraits::Params::d_d</a></div><div class="ttdeci">ScalarD * d_d</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:156</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_a11e54597165b66e3054c6b9f43210760"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#a11e54597165b66e3054c6b9f43210760">cutlass::reduction::BatchedReductionTraits::Params::functorParams</a></div><div class="ttdeci">Functor::Params functorParams</div><div class="ttdoc">The functor params. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:160</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_ab4f5f457dbfa6bd250a4c34e1d573a85"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ab4f5f457dbfa6bd250a4c34e1d573a85">cutlass::reduction::BatchedReductionTraits::ThreadShapeMultiple2</a></div><div class="ttdeci">static const bool ThreadShapeMultiple2</div><div class="ttdoc">check if threadShape is multiple of 2. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:117</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_ab0c9e548f3ee62746e727127e387a8f4"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ab0c9e548f3ee62746e727127e387a8f4">cutlass::reduction::BatchedReductionTraits::Index</a></div><div class="ttdeci">Index_ Index</div><div class="ttdoc">The index. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:111</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_ac28e31791c5888bbe7b04abe6376a422"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ac28e31791c5888bbe7b04abe6376a422">cutlass::reduction::BatchedReductionTraits::maxOutReg</a></div><div class="ttdeci">static int const maxOutReg</div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:126</div></div>
<div class="ttc" id="batched__reduction_8h_html"><div class="ttname"><a href="batched__reduction_8h.html">batched_reduction.h</a></div><div class="ttdoc">Implements a software-pipelined efficient batched reduction. D = alpha * Reduction(A) + beta * C...</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="cutlass_8h_html"><div class="ttname"><a href="cutlass_8h.html">cutlass.h</a></div><div class="ttdoc">Basic include for CUTLASS. </div></div>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html">cutlass::reduction::BatchedReductionTraits</a></div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:76</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params_html_ac27f42beb3625c5183b76b26677c0cb0"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#ac27f42beb3625c5183b76b26677c0cb0">cutlass::reduction::BatchedReductionTraits::Params::initialize</a></div><div class="ttdeci">CUTLASS_HOST_DEVICE int initialize(Index m_, Index n_, ScalarAlphaBeta alpha_, ScalarAlphaBeta beta_, long long int reduction_stride_, ScalarA const *d_a_, Index lda_, ScalarC const *d_c_, Index ldc_, ScalarD *d_d_, Index ldd_)</div><div class="ttdoc">Initialize the parameters for 2D output tensor. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:162</div></div>
<div class="ttc" id="structcutlass_1_1reduction_1_1BatchedReductionTraits_html_a085c72d54426f5eb60f5bffa9c383229"><div class="ttname"><a href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#a085c72d54426f5eb60f5bffa9c383229">cutlass::reduction::BatchedReductionTraits::KernelClass</a></div><div class="ttdeci">cutlass::reduction::BatchedReduction&lt; This_ &gt; KernelClass</div><div class="ttdoc">The struct that consumes this Traits. </div><div class="ttdef"><b>Definition:</b> batched_reduction_traits.h:93</div></div>
</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">
Updating Doxygen docs
2018-10-27 05:54:58 +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
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
</small></address>
</body>
</html>
Reference in New Issue Copy Permalink