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
+121 -35
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: Class Members - Variables</title>
<meta name="generator" content="Doxygen 1.8.11"/>
<title>CUTLASS: Class Members - Variables</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 +39,78 @@
</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 class="current"><a href="functions.html"><span>Class&#160;Members</span></a></li>
</ul>
</div>
<div id="navrow3" class="tabs2">
<ul class="tablist">
<li><a href="functions.html"><span>All</span></a></li>
<li><a href="functions_func.html"><span>Functions</span></a></li>
<li class="current"><a href="functions_vars.html"><span>Variables</span></a></li>
<li><a href="functions_type.html"><span>Typedefs</span></a></li>
<li><a href="functions_enum.html"><span>Enumerations</span></a></li>
<li><a href="functions_eval.html"><span>Enumerator</span></a></li>
</ul>
</div>
<div id="navrow4" class="tabs3">
<ul class="tablist">
<li><a href="functions_vars.html#index_a"><span>a</span></a></li>
<li class="current"><a href="functions_vars_b.html#index_b"><span>b</span></a></li>
<li><a href="functions_vars_c.html#index_c"><span>c</span></a></li>
<li><a href="functions_vars_d.html#index_d"><span>d</span></a></li>
<li><a href="functions_vars_e.html#index_e"><span>e</span></a></li>
<li><a href="functions_vars_f.html#index_f"><span>f</span></a></li>
<li><a href="functions_vars_g.html#index_g"><span>g</span></a></li>
<li><a href="functions_vars_h.html#index_h"><span>h</span></a></li>
<li><a href="functions_vars_i.html#index_i"><span>i</span></a></li>
<li><a href="functions_vars_k.html#index_k"><span>k</span></a></li>
<li><a href="functions_vars_l.html#index_l"><span>l</span></a></li>
<li><a href="functions_vars_m.html#index_m"><span>m</span></a></li>
<li><a href="functions_vars_n.html#index_n"><span>n</span></a></li>
<li><a href="functions_vars_o.html#index_o"><span>o</span></a></li>
<li><a href="functions_vars_p.html#index_p"><span>p</span></a></li>
<li><a href="functions_vars_r.html#index_r"><span>r</span></a></li>
<li><a href="functions_vars_s.html#index_s"><span>s</span></a></li>
<li><a href="functions_vars_t.html#index_t"><span>t</span></a></li>
<li><a href="functions_vars_u.html#index_u"><span>u</span></a></li>
<li><a href="functions_vars_v.html#index_v"><span>v</span></a></li>
<li><a href="functions_vars_w.html#index_w"><span>w</span></a></li>
</ul>
</div>
</div><!-- top -->
<!-- window showing the filter options -->
<div id="MSearchSelectWindow"
@@ -70,38 +129,65 @@ $(function() {
<div class="contents">
&#160;
<h3><a id="index_b"></a>- b -</h3><ul>
<h3><a class="anchor" id="index_b"></a>- b -</h3><ul>
<li>B
: <a class="el" href="structcutlass_1_1gemm_1_1GemmDesc.html#af0ac89b161f9cad96307f1ff3c80a774">cutlass::gemm::GemmDesc&lt; AType_, BType_, CType_, DType_, SType_, Index_ &gt;</a>
: <a class="el" href="structcutlass_1_1library_1_1GemmArguments.html#ae3b1c625d32bf5cbdbd4d2f520145efc">cutlass::library::GemmArguments</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmArrayArguments.html#a0c3d185b52998f836fbf4c0d27c6e497">cutlass::library::GemmArrayArguments</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmDescription.html#ad6117aecf9e4d22862e621114e95cccf">cutlass::library::GemmDescription</a>
</li>
<li>batch_stride
: <a class="el" href="structcutlass_1_1gemm_1_1GlobalLoadStream_1_1Params.html#ab5d9e5de9b1cc3462920ce7a8f52bbec">cutlass::gemm::GlobalLoadStream&lt; Operand, LoadIterator_, StoreIterator_, Transformer_ &gt;::Params</a>
<li>B_tile
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1thread_1_1Gemm.html#a5329ece817a4d471dfee042a4eb6f7bd">cutlass::reference::device::thread::Gemm&lt; TensorRefA, TensorRefB, TensorRefC, ScalarType, AccumulatorType, OutputTile, InnerProductOp, ConvertOp &gt;</a>
</li>
<li>batch_count
: <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_1_1Arguments.html#ac99ca8f9d8a0053e647a6c99b018bda5">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_3_01ElementA___00_01LayoutA___00_01ElementB___00_213d78696663f4231cd52c6a277c60e5.html#adb66f3083f56c15578b139b7935452b5">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#a7ed96fc1c9cba288ec807736a3ed96e7">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmArrayConfiguration.html#a10fe15d9179998530d3fdd86c78d4a15">cutlass::library::GemmArrayConfiguration</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmBatchedConfiguration.html#ae6cc3b877a073bedb8e4d1c91423b0f5">cutlass::library::GemmBatchedConfiguration</a>
</li>
<li>batch_stride_A
: <a class="el" href="structcutlass_1_1gemm_1_1GemmDesc.html#a8b5a86d14eba0d3c5173753212d62599">cutlass::gemm::GemmDesc&lt; AType_, BType_, CType_, DType_, SType_, Index_ &gt;</a>
: <a class="el" href="structcutlass_1_1library_1_1GemmBatchedConfiguration.html#a7bea2035164b174c45a7589d8132f2af">cutlass::library::GemmBatchedConfiguration</a>
</li>
<li>batch_stride_B
: <a class="el" href="structcutlass_1_1gemm_1_1GemmDesc.html#a95e3fe05e4ca0d4019cbef2b1a54419a">cutlass::gemm::GemmDesc&lt; AType_, BType_, CType_, DType_, SType_, Index_ &gt;</a>
: <a class="el" href="structcutlass_1_1library_1_1GemmBatchedConfiguration.html#ac85cb497652f997f8fa3143be70ac77a">cutlass::library::GemmBatchedConfiguration</a>
</li>
<li>batch_stride_C
: <a class="el" href="structcutlass_1_1gemm_1_1GemmDesc.html#aad3590dffa2e1ba82c834efae6b35ad2">cutlass::gemm::GemmDesc&lt; AType_, BType_, CType_, DType_, SType_, Index_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1GemmEpilogueTraits_1_1Params.html#a4aa03c14de3f0c3c1cf46c3d5f52d802">cutlass::gemm::GemmEpilogueTraits&lt; OutputTile_, Accumulators_, GlobalLoadIteratorC_, GlobalTransformerC_, GlobalTransformerD_, GlobalStoreIteratorD_, SharedStoreIteratorD_, SharedStoreTransformerD_, SharedLoadStreamD_, Iterations_, Delta_, Functor_, Index_ &gt;::Params</a>
: <a class="el" href="structcutlass_1_1library_1_1GemmBatchedConfiguration.html#a1225f9ce96f9819d6d0cfde7a664b921">cutlass::library::GemmBatchedConfiguration</a>
</li>
<li>batch_stride_D
: <a class="el" href="structcutlass_1_1gemm_1_1GemmDesc.html#a7b9afcc7e3105da1d002b1baa68d83de">cutlass::gemm::GemmDesc&lt; AType_, BType_, CType_, DType_, SType_, Index_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1GemmEpilogueTraits_1_1Params.html#a5208998277daef204983989d24707008">cutlass::gemm::GemmEpilogueTraits&lt; OutputTile_, Accumulators_, GlobalLoadIteratorC_, GlobalTransformerC_, GlobalTransformerD_, GlobalStoreIteratorD_, SharedStoreIteratorD_, SharedStoreTransformerD_, SharedLoadStreamD_, Iterations_, Delta_, Functor_, Index_ &gt;::Params</a>
: <a class="el" href="structcutlass_1_1library_1_1GemmBatchedConfiguration.html#aa0b8603417007880a9882774f0f5c988">cutlass::library::GemmBatchedConfiguration</a>
</li>
<li>batched_stride_A
: <a class="el" href="structcutlass_1_1library_1_1GemmPlanarComplexBatchedConfiguration.html#a68643eb068634c6f96719d871363bc09">cutlass::library::GemmPlanarComplexBatchedConfiguration</a>
</li>
<li>batched_stride_B
: <a class="el" href="structcutlass_1_1library_1_1GemmPlanarComplexBatchedConfiguration.html#a0472cd678eeb71d5ffd42fdcab5af409">cutlass::library::GemmPlanarComplexBatchedConfiguration</a>
</li>
<li>batched_stride_C
: <a class="el" href="structcutlass_1_1library_1_1GemmPlanarComplexBatchedConfiguration.html#ab3519e652b982d3b3fdf4c788342bda9">cutlass::library::GemmPlanarComplexBatchedConfiguration</a>
</li>
<li>batched_stride_D
: <a class="el" href="structcutlass_1_1library_1_1GemmPlanarComplexBatchedConfiguration.html#a7245398f9aa9754f6501edf95a8a3ab5">cutlass::library::GemmPlanarComplexBatchedConfiguration</a>
</li>
<li>beta
: <a class="el" href="structcutlass_1_1gemm_1_1GemmDesc.html#adac41a0baad9e65aa4a6fe12d249a02b">cutlass::gemm::GemmDesc&lt; AType_, BType_, CType_, DType_, SType_, Index_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1LinearScaling_1_1Params.html#a0e455ad2e4eba67259867f9123ca817b">cutlass::gemm::LinearScaling&lt; Scalar_, FragmentMultiplyAdd_ &gt;::Params</a>
: <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombination_1_1Params.html#a9677cb04e23e9afa9fcdc3f34074bb56">cutlass::epilogue::thread::LinearCombination&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationClamp_1_1Params.html#a11eb2330d28b470366032dd8f549fe33">cutlass::epilogue::thread::LinearCombinationClamp&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_1_1Params.html#a7f3cb135884b5ef89bdef997159a3844">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_3_01ElementOutput___00_01Count_00_00274a94522c46cd041d0b10d484e2ef3.html#af607bd78ed05e98af8dfe0c413e25091">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, int, float, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1detail_1_1GemvBatchedStridedEpilogueScaling.html#abcd99b63173e4330a75558e78a756296">cutlass::gemm::kernel::detail::GemvBatchedStridedEpilogueScaling&lt; ElementAlphaBeta, BetaIsZero &gt;</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmArguments.html#a91f68b30afc142ea697707fa752c9526">cutlass::library::GemmArguments</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmArrayArguments.html#ac84f31989db6018f465019cf81f83978">cutlass::library::GemmArrayArguments</a>
, <a class="el" href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#a805f78cae27c3305c988f251207d85f7">cutlass::reduction::BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt;::Params</a>
</li>
<li>beta_ptr
: <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombination_1_1Params.html#a01f730dac9a4500cb857bf4ca272bc7b">cutlass::epilogue::thread::LinearCombination&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationClamp_1_1Params.html#adae7ef1a432b24d148df0662954b5bd0">cutlass::epilogue::thread::LinearCombinationClamp&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_1_1Params.html#a2dfa5dc3c851915d39d27bf6b4cc68e6">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_3_01ElementOutput___00_01Count_00_00274a94522c46cd041d0b10d484e2ef3.html#a3b26637eb910d79d8ae6a79011ca85e3">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, int, float, Round &gt;::Params</a>
</li>
<li>block
: <a class="el" href="structcutlass_1_1KernelLaunchConfiguration.html#a09535026bf08f94c6940c358d95d1edd">cutlass::KernelLaunchConfiguration</a>
</li>
<li>bounds
: <a class="el" href="structcutlass_1_1RegularTilePredicateFunctor.html#ac2f49374e6f0a27ad2daffcb1f74708a">cutlass::RegularTilePredicateFunctor&lt; Delta_ &gt;</a>
</li>
<li>byte
: <a class="el" href="structcutlass_1_1platform_1_1alignment__of_1_1pad.html#a86f075f91b80918e968951713430f0b4">cutlass::platform::alignment_of&lt; value_t &gt;::pad</a>
</li>
@@ -109,9 +195,9 @@ $(function() {
</div><!-- contents -->
<!-- start footer part -->
<hr class="footer"/><address class="footer"><small>
Generated on Fri Oct 26 2018 14:53:41 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>