squall/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
06eb90cc0d
cutlass/docs/functions_f.html

456 lines
91 KiB
HTML
Raw Normal View History

CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/>
<meta http-equiv="X-UA-Compatible" content="IE=9"/>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<meta name="generator" content="Doxygen 1.8.11"/>
<title>CUTLASS: Class Members</title>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<link href="tabs.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="jquery.js"></script>
<script type="text/javascript" src="dynsections.js"></script>
<link href="search/search.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="search/searchdata.js"></script>
<script type="text/javascript" src="search/search.js"></script>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<script type="text/javascript">
$(document).ready(function() { init_search(); });
</script>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
extensions: ["tex2jax.js"],
jax: ["input/TeX","output/HTML-CSS"],
});
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
</script><script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js"></script>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<link href="doxygen.css" rel="stylesheet" type="text/css" />
</head>
<body>
<div id="top"><!-- do not remove this div, it is closed by doxygen! -->
<div id="titlearea">
<table cellspacing="0" cellpadding="0">
<tbody>
<tr style="height: 56px;">
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<td id="projectlogo"><img alt="Logo" src="cutlass-logo-small.png"/></td>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<td id="projectalign" style="padding-left: 0.5em;">
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<div id="projectname">CUTLASS
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</div>
<div id="projectbrief">CUDA Templates for Linear Algebra Subroutines and Solvers</div>
</td>
</tr>
</tbody>
</table>
</div>
<!-- end header part -->
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<!-- Generated by Doxygen 1.8.11 -->
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<script type="text/javascript">
var searchBox = new SearchBox("searchBox", "search",false,'Search');
</script>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<div id="navrow1" class="tabs">
<ul class="tablist">
<li><a href="index.html"><span>Main&#160;Page</span></a></li>
<li><a href="modules.html"><span>Modules</span></a></li>
<li><a href="namespaces.html"><span>Namespaces</span></a></li>
<li 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 class="current"><a href="functions.html"><span>All</span></a></li>
<li><a href="functions_func.html"><span>Functions</span></a></li>
<li><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.html#index_a"><span>a</span></a></li>
<li><a href="functions_b.html#index_b"><span>b</span></a></li>
<li><a href="functions_c.html#index_c"><span>c</span></a></li>
<li><a href="functions_d.html#index_d"><span>d</span></a></li>
<li><a href="functions_e.html#index_e"><span>e</span></a></li>
<li class="current"><a href="functions_f.html#index_f"><span>f</span></a></li>
<li><a href="functions_g.html#index_g"><span>g</span></a></li>
<li><a href="functions_h.html#index_h"><span>h</span></a></li>
<li><a href="functions_i.html#index_i"><span>i</span></a></li>
<li><a href="functions_k.html#index_k"><span>k</span></a></li>
<li><a href="functions_l.html#index_l"><span>l</span></a></li>
<li><a href="functions_m.html#index_m"><span>m</span></a></li>
<li><a href="functions_n.html#index_n"><span>n</span></a></li>
<li><a href="functions_o.html#index_o"><span>o</span></a></li>
<li><a href="functions_p.html#index_p"><span>p</span></a></li>
<li><a href="functions_q.html#index_q"><span>q</span></a></li>
<li><a href="functions_r.html#index_r"><span>r</span></a></li>
<li><a href="functions_s.html#index_s"><span>s</span></a></li>
<li><a href="functions_t.html#index_t"><span>t</span></a></li>
<li><a href="functions_u.html#index_u"><span>u</span></a></li>
<li><a href="functions_v.html#index_v"><span>v</span></a></li>
<li><a href="functions_w.html#index_w"><span>w</span></a></li>
<li><a href="functions_y.html#index_y"><span>y</span></a></li>
<li><a href="functions_0x7e.html#index_0x7e"><span>~</span></a></li>
</ul>
</div>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</div><!-- top -->
<!-- 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 class="contents">
<div class="textblock">Here is a list of all class members with links to the classes they belong to:</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
<h3><a class="anchor" id="index_f"></a>- f -</h3><ul>
<li>fetch()
: <a class="el" href="classcutlass_1_1Semaphore.html#af7e78f85e6106c1c82c10bee0b76d454">cutlass::Semaphore</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>fill()
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 class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4.html#a1c31d3673a48b2ed275bd56714fbcfbe">cutlass::Array&lt; T, N, false &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#a0b3f29a6d79dd9cd55de367c96ecfc5c">cutlass::Array&lt; T, N, true &gt;</a>
, <a class="el" href="structcutlass_1_1PredicateVector.html#a236bd1a822479750a809452fd58dd917">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
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>FloatType
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomGaussianFunc.html#ac7cf68adaae0b16b1633a4e3f5d79aa5">cutlass::reference::device::detail::RandomGaussianFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomUniformFunc.html#a7654e2230e68c04b081060496cc188fc">cutlass::reference::device::detail::RandomUniformFunc&lt; Element &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>Fragment
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 class="el" href="classcutlass_1_1epilogue_1_1thread_1_1ReductionOpPlus.html#a1a67e7632e92fed03b0d2d66d1a5f124">cutlass::epilogue::thread::ReductionOpPlus&lt; Element_, Count &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1InterleavedPredicatedTileIterator.html#a185d87d8c1aecb00eb09059068ccf1c6">cutlass::epilogue::threadblock::InterleavedPredicatedTileIterator&lt; ThreadMap_, Element_, InterleavedK &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1PredicatedTileIterator.html#a69151e6217b907fc20888e93f95cc333">cutlass::epilogue::threadblock::PredicatedTileIterator&lt; ThreadMap_, Element_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1SharedLoadIterator.html#afb37ed9a0f3873600b9e743b2dcb805e">cutlass::epilogue::threadblock::SharedLoadIterator&lt; ThreadMap_, Element_, MaxAlignment &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorComplexTensorOp_3_01WarpShape___00_01Operato8cf03c624cf3210c71b7cbd580b080f8.html#ac611843b0222b7f7538ebe8f06f48b6a">cutlass::epilogue::warp::FragmentIteratorComplexTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorSimt_3_01WarpShape___00_01Operator___00_01la3f2abc523201c1b0228df99119ab88e1.html#afaf52901287cfcadc0fe059e71b6842a">cutlass::epilogue::warp::FragmentIteratorSimt&lt; WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorTensorOp_3_01WarpShape___00_01OperatorShape_e459aab140a2ce78336e584f95886726.html#a53fe88441ba1f7bbe5cb30ae60eb2be6">cutlass::epilogue::warp::FragmentIteratorTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::ColumnMajorInterleaved&lt; InterleavedK &gt; &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorTensorOp_3_01WarpShape___00_01OperatorShape_5e78dabe303f20d76b00c600aab61eda.html#ac1cbc7d77f60cedb3e7aa88c24ea04b1">cutlass::epilogue::warp::FragmentIteratorTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorVoltaTensorOp_3_01WarpShape___00_01gemm_1_1Gdb805a2dc5571ac3b66e0fe6ffdcede2.html#a67a41ee64e7d20e92f5841926fabf9a1">cutlass::epilogue::warp::FragmentIteratorVoltaTensorOp&lt; WarpShape_, gemm::GemmShape&lt; 32, 32, 4 &gt;, float, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorVoltaTensorOp_3_01WarpShape___00_01gemm_1_1G16e08718cffa0989cce3fe8dbc4b075b.html#a99fbac47851323828ec2a9617014e6b3">cutlass::epilogue::warp::FragmentIteratorVoltaTensorOp&lt; WarpShape_, gemm::GemmShape&lt; 32, 32, 4 &gt;, half_t, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorWmmaTensorOp_3_01WarpShape___00_01OperatorShfdb1f120c6797383663f9fd11d0fc599.html#a533ce5f9784cd84f7196ce3b16a8441a">cutlass::epilogue::warp::FragmentIteratorWmmaTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1TileIteratorSimt_3_01WarpShape___00_01Operator___00_01Elemenf2bd262ed3e202b25d5802d83965bf3b.html#aa1caeb6928f62ceabba7817c754b08b7">cutlass::epilogue::warp::TileIteratorSimt&lt; WarpShape_, Operator_, Element_, layout::RowMajor, MmaSimtPolicy_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1TileIteratorTensorOp_3_01WarpShape___00_01OperatorShape___003cbb32beb84b4984cb7853662096d289.html#a9156203bccdaf3b36b87286153c63147">cutlass::epilogue::warp::TileIteratorTensorOp&lt; WarpShape_, OperatorShape_, Element_, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1TileIteratorVoltaTensorOp_3_01WarpShape___00_01gemm_1_1GemmS2fe0c60b727c738c622c18fc3dd76644.html#afc5b8faf449e4bd922e017b7c5e47fc4">cutlass::epilogue::warp::TileIteratorVoltaTensorOp&lt; WarpShape_, gemm::GemmShape&lt; 32, 32, 4 &gt;, float, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1TileIteratorVoltaTensorOp_3_01WarpShape___00_01gemm_1_1GemmSa0ceeeddc22575876eb977da7f5416a8.html#a77d5d1e5e8af463d1ec30bfee41f1161">cutlass::epilogue::warp::TileIteratorVoltaTensorOp&lt; WarpShape_, gemm::GemmShape&lt; 32, 32, 4 &gt;, half_t, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1TileIteratorWmmaTensorOp_3_01WarpShape___00_01OperatorShape_fd6a91cd8bbd07ecd1344326b830e3a4.html#aa91a85829f23d19fcd9c406aa45683e5">cutlass::epilogue::warp::TileIteratorWmmaTensorOp&lt; WarpShape_, OperatorShape_, OperatorFragment_, layout::RowMajor &gt;</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1warp_1_1VoltaTensorOpPolicy_3_01WarpShape___00_01gemm_1_1GemmShape_136ce744d4c1c6e8707f5a9785196194.html#ae38ddb7de2a44dc71842e289a8416ca7">cutlass::epilogue::warp::VoltaTensorOpPolicy&lt; WarpShape_, gemm::GemmShape&lt; 32, 32, 4 &gt;, float, layout::RowMajor &gt;</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1warp_1_1VoltaTensorOpPolicy_3_01WarpShape___00_01gemm_1_1GemmShape_1d48185f49e4d066f8e9327bf0856b7f.html#abf178c47cbefcedf0a692ae3327ec5d2">cutlass::epilogue::warp::VoltaTensorOpPolicy&lt; WarpShape_, gemm::GemmShape&lt; 32, 32, 4 &gt;, half_t, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kA_00_01Element_67ca7e11a38e38f2c51b84767654a90f.html#a105714123859727aa4988e7c3b8e733b">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kA, Element_, layout::ColumnMajor, Policy_, PartitionsK, PartitionGroupSize &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kA_00_01Element_f0ce904a9294556f15e1cc9cf7c99a93.html#a37ffe0ab9966006a704d025f38b80e36">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kA, Element_, layout::ColumnMajorInterleaved&lt; 4 &gt;, Policy_, PartitionsK, PartitionGroupSize &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kB_00_01Element_ea0a4e7ce3cd5d25cabf79383efdf4d9.html#a9d1fd03c1dd693d70ac10cdbcd315237">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kB, Element_, layout::RowMajor, Policy_, PartitionsK, PartitionGroupSize &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kB_00_01Element_ada156b62fcbdce47009c5bf1321c92c.html#a4b2f6656d93ed50775f4d1107dbc54df">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kB, Element_, layout::RowMajorInterleaved&lt; 4 &gt;, Policy_, PartitionsK, PartitionGroupSize &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kC_00_01Element_4ccafbc821b3a55cd532602442a74031.html#a20818adbde7ee320d679df8c5124679d">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kC, Element_, layout::ColumnMajor, Policy_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kC_00_01Element_8f92ea79e85febb67169c4b2d94b1b20.html#a2db72e523433c65da79c4d5ad4108b6f">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kC, Element_, layout::RowMajor, Policy_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___008f607b871a2b3d854eb4def64712c042.html#a081c62bfb91fdf105e97427c58b1bca1">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___00027dabdc144edd6276f664ca74088510.html#a81f5416db0a2add284cbfb6b0c422e94">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajorInterleaved&lt; InterleavedN &gt;, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___006c39f57875e0aa9d0ad82c8043ed8b98.html#ac0806ad1ce257502ba7a2c0b580b117b">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0b84f53cd44b339eccc12067c9f86e11c.html#a7f05dbc9436d93dc51401881e9d3b252">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0e52ad425e1ee3e68544873f66733237b.html#ad5154742e24c958288b0f52544aaaa68">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___039819fb3ccd43786d556c2c9669508ef.html#abb80a20c60f1c8702a217cc56fa0b348">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0352e0dcab42bc8360606874e00173556.html#abe2b74533d91f56799cab0365af27fe5">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0ed7daaeba1c095e77f68533d4d2c475c.html#a258a2278561191525ebfab2e6ff93d8f">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, 64 &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0c7d419c589d601ce4eb603be566fea21.html#abed4efcbfe0634ccaeae64eef124cceb">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpAccumulatorTileIterator.html#acd54606d757015b77519e3a4b8491b37">cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, Layout_, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan0d3248553e52cd61ed8a2b3b12a20343.html#add7e4e5c9dbfb389f64bdecdf0889e74">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan34be8e21a40af3ebd2dc3dff460dca72.html#a41ea626da30ce2007fe3bfeb299460ca">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand734577b7e54a074d143aba59828c2f2.html#a348b537e5cdfe105400ad1a66f20e6b4">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan16c56cdc2dda5eeb996af8ec0242d501.html#a68d2f239ff185b0245a3fe219776341c">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan5a221944f4a0e16ccab77ba684856942.html#af99286febf2c6c64d267dfc8fab282dc">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operandcc9821c435540895138bc9af495f321.html#ad3629b3b4b0293d9eca05397077d8626">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operana2f40b28f0d2286b84d86f7238d67b52.html#a3a2ffe3c15142cc6345e02445d2e5256">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="structcutlass_1_1transform_1_1thread_1_1Transpose_3_01ElementCount___00_01layout_1_1PitchLinearS99f8e05faf0bb5ed48a0154afe740d81.html#ab368c16371c6c8e5384fcbc9f4412ab6">cutlass::transform::thread::Transpose&lt; ElementCount_, layout::PitchLinearShape&lt; 4, 4 &gt;, int8_t &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_0165b39a630d10785a3558406f9adb99b9.html#a9be82cb5e5c9a215b5862dba28dc57f4">cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Transpose_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_017a517f3c73efd795ab05059cc9b111e1.html#a0c7351825135690fa82c4fba5f9447c7">cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Transpose_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_013671177d6219bfeb0e1b4dc4c1b5bf11.html#ae342607106778417432c4694b92736b3">cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Transpose_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___0068b3e874b5d93d11f0fa902c7f1d11d9.html#aefea1fc075a0f02ef2d6069b379a48ba">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessSize &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___00f6b3a9dfab5e7c72d5233f7e5e6e3b9b.html#a7712f023dcf819a124a9cb569eda0483">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessSize &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___00e7c2c404e7aedfe60ad56bb5571306a1.html#a8e3ff6897ff18b510d459052e6aa5648">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessSize &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___0041ea81994f8af0d4d071fdb9e66b5ff0.html#ad1ade679ed38f572252668ecbc3b6df4">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessSize &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___00d670f969180a8d182dffb356ebcc957e.html#a799b0d6d28334da4d5f0c16e95813338">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessSize &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Eleb60d066756d1c18f05fceee6a27bdb8a.html#ad1151fe7941db2f895bd9c53015a9ced">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::ColumnMajorInterleaved&lt; 4 &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Ele76ed82829532ae1c17f4c78158f036c7.html#abce03cf9e3e95b31696dc4f34f22d044">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Ele654c8f6161ae5340f040397a4e2e045c.html#a3f4cd0a7a07b3ce2bf755b01107b5ee6">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; 4 &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_011d3637dbd8bc58bcb020b51bf57fbfc0.html#a5e1532d6ba9ceb740f5cdbafb9bad6c3">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_017982f81d4ef592e19c8427de2ea933a3.html#accac975dddce8f98f5cf7c0fd759f668">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_010889a732373c350de9b9a9f6c13cd761.html#a107ef5ee79fef84367e2e1e76696cdc9">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01187f8574e1fe9d7d5e8fbf09bd834bf0.html#acdcedb26b30985f59a77ddb49b1abe27">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01793f74bfd8f116a827948ab01a37349a.html#a20bbd62f67f815c5b06d2310b3a97083">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01bd31b3810c1fedf2e7e5959ff92b5d3d.html#ac8dc6cbe7b5d728373c719d6245194c8">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Shape_::kRow &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0184a89653916f5d51ab59d1b386989a17.html#a23214640848bd844f182c8276d495156">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0149454d361ea5885cf5166a920b5145df.html#a0f8e337d13ada6404909a4a10572b8af">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01c20d35180520077a5a09b1e33543c1a5.html#ac368f1f8ca17a35202185104fc9b9561">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01a31b454d9c930525c1e9ca406a514f40.html#ac92fd47b222851341837f342795af48b">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0104ad31bd559a88cc418ae1cab7492ed5.html#aae78eda22cfd92be06f8397e72abfaf4">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01f6f6511b5033cad31083644ac69c54d8.html#ad89c4d02d6e30efa055e7b4eb92c12be">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01b3fa5720e807697de61b9f937b269cd0.html#a530b747642baaa77452ed811dd7fbae0">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Shape_::kColumn &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01efd5013a2503d6567e2bf6b40c97360c.html#ad819cfb30d80acbf10a40c6bcee257e4">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::TensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0197fef2242a3454a7d1cebe61aee28b43.html#ac020f463c891f79be706588944bcbf81">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01a75d2cd74e722d6ad6a3b41aabfd432d.html#a2e23610a2482c75bc33b4a13bbfaf6eb">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::VoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01f96bbeb63e6d4ce4a2551279de3a9f0e.html#a9e1c6e07838c14d564324a231b0dbe83">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::VoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01dbd6b8468d5bd787308d2f615a24d123.html#a6bb1b49203f861013b8719866d4eefdb">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::VoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Shape_::kContiguous &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
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>FragmentA
: <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_0116_00_018_00_018_01_4_00_0132_00_01half__02a3f19a78995f97d793a668e0e4d4f0.html#a757380f3fa67ee277c719fc74a50faf6">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 16, 8, 8 &gt;, 32, half_t, layout::RowMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_0116_00_018_00_018_01_4_00_0132_00_01half__96363097c47b056f0ca1911afd7f8b7a.html#a86b5a71d720cef78b1d84b667cc566bc">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 16, 8, 8 &gt;, 32, half_t, layout::RowMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_01128_01_4_00_0132_00_01uint15918972b95027764b3a849b03075ed2b.html#ae7d755f4766981293b04cf235c55405e">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 128 &gt;, 32, uint1b_t, layout::RowMajor, uint1b_t, layout::ColumnMajor, int, layout::RowMajor, OpXorPopc &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__927179f46017ea5f58f859f1196c4829.html#a259f7c4fd0d7a74636f741836cd5d56c">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__8ebae0cbdf333fddfe5c24d35ebe8e02.html#a115ad2a707ac2b877d1c77ee988b06e2">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__5299c9c90c8f2f521be0c8cec1c3eb08.html#a39da233858a5241a3f5f7080d6257be9">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__f083347e265b1e9eea5572d86ddb6bf9.html#a0da851ba86ef7e405e487d4fe454a9e5">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_a62aa63a212985df306fb27e8a50aeae.html#a6a8c572e8f24b6b4a0b5d78bb9b09716">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_ab741d81fdc991345cb9e43c29fca573.html#a2ea09a81db7947aba39e8642b8b58a26">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_5221708cec5828d35db1d1c47cb4964e.html#adb1f4ff4e767591dc640319e00e3933d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_bef0c048bc0f8ba2d875cb7ab26d363b.html#aa20f3a3fe79611964509c36a1e627ef8">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_6e513ccbc44ae7909a60d93b9b5435b3.html#a39fb4bf1e9411e0ce49991235fee7ad6">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_0ee08a4520882d24ba9026879265e892.html#a208e6301ec92794bd44ba5e6c500c832">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_4746fc55e614df0016c518d3fda2677e.html#acd536cf5cc0d0da6289a4dd3cf980d53">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_546e9ec6de6a5970b326da6f6280f1d4.html#a053afbc90a78e8eb49e649c0c970ad85">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b03e3b50dbcb30d0d1ac062f3a9d5abef.html#a95b674bbc803a62a462b6bed8b84a6e1">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b6d968039dde5c9f062ab15f90a8049fe.html#a0bdde7ba18ea3de857ae580d602ca372">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4bc4b6ba004e25c44bfd9266c61f937dfb.html#a8e3b15fcf4acdebc2f9c46e494e16042">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b451d5cf5d7e8cbbe476afe3dab5c09b2.html#aa9875fd6612f9d2d18254b39de44bd85">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_b0242d7a01097510effbc4718040d3e5.html#aaa211fadb536084d0afc6ee61b3e843b">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_c7f88bfd32a544fba8111d2dcadeab11.html#a8538ac5fcf576f2643b6f4f62b85683d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_44a3b2a8df88a2b067f1284515cb5371.html#a453c8e5e65ef2e596b6fb650b957554d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_4b7308177b308a272c1889fbe9670275.html#a982b03905212819df6abce5f5fe99d0e">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_5a9888862cebd333ecaf11f7262f77d4.html#a7a5559162324faae78a4aac270c9c633">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_31defda8ea2b7d855642ffd77da1a411.html#a851d5dc5baae4d6443acac2669131252">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_839a7c8bb938d1661f4611e68f85d8cb.html#a500bfb11926a2e4717808976db66fca4">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::RowMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_73d9802d6b944a5299bc255887db6bbc.html#a3d2b9865caf773aadb0ae8d3cf22eb6a">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_72621f7ab9ae4a4ba4fe9725cf8e89c1.html#a5e4e2eb95afb820aad6a4609bd910089">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::ColumnMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_94c813e3bbfb6f9857c155166f772687.html#a511aed0cec5670783f6dc3d2b26a64a5">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::ColumnMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_17070298bc4cced0a1b98aee2bb6b455.html#abd3217abf12bdf7810408cd0af10e684">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::RowMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_bf6d29bb09a025e7b96942809743e28a.html#a32394ddf5b5bb59cc08aaa37e247fb12">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::RowMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l26a133b13650c1d058273e3649f60f04.html#afc64bd91665ee573465a07e4eb6dc563">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::ColumnMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01lbba3a796be96a0276693ef6b259ecc4a.html#a75103e99de4ae19d14678d8ae5b697bd">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::ColumnMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l2aa4d2fd2e940e0d0cf7c47bc8f6017c.html#a125ca8c88a24f1e23716be8f25d6735d">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::RowMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l086c058a15d6c79558e4f3d9ff1dc148.html#abb2a511d7b8b60cac068df2a9caa52c3">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::RowMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01LayoutA_00_01LayoutB_00_0e1104c65871c539155bd3a0c7631928b.html#acf8a91ee6716f8e2c3b9e8f7e805c065">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, LayoutA, LayoutB, layout::ColumnMajor, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01LayoutA_00_01LayoutB_00_07ac147cb320ee0d28ff8e78eb4cd330e.html#aa6c00f1efb2ee4e5d5dd71f83cc3cedd">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, LayoutA, LayoutB, layout::RowMajor, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01ElementA___00_01LayoutA___00_01ElementB_e41c1cd6078b6d1347fac239b0639d56.html#a66abc782808b6b3e68518aff43a0b200">cutlass::gemm::thread::Mma&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, arch::OpMultiplyAdd, bool &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01half__t_00_01LayoutA_00_01half__t_00_01L066c9d2371712cdf0cac099ca9bcc578.html#a64b2cf33786247c4acd872fb8856abd5">cutlass::gemm::thread::Mma&lt; Shape_, half_t, LayoutA, half_t, LayoutB, half_t, LayoutC, arch::OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01half__t_00_01LayoutA___00_01half__t_00_088f0e99e501b6012297eb30b4e89bcea.html#adbd6a51a9e477d917f5739230a023524">cutlass::gemm::thread::Mma&lt; Shape_, half_t, LayoutA_, half_t, LayoutB_, half_t, layout::RowMajor, arch::OpMultiplyAdd, typename platform::enable_if&lt; detail::EnableMma_Crow_SM60&lt; LayoutA_, LayoutB_ &gt;::value &gt;::type &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01int8__t_00_01layout_1_1ColumnMajor_00_013f3785e722edc6e9aab6f866309b8623.html#a0b6a3262da428537ba826a5abf16b894">cutlass::gemm::thread::Mma&lt; Shape_, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, int32_t, LayoutC_, arch::OpMultiplyAdd, int8_t &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01int8__t_00_01layout_1_1RowMajor_00_01int89c659e7faf47264972bdba6cd80f42b.html#af9811ca34314708c1371e37faf3f6cf2">cutlass::gemm::thread::Mma&lt; Shape_, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int32_t, LayoutC_, arch::OpMultiplyAdd, bool &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1MmaGeneric.html#aa52815a830293f0a5b2cfb58689fb4a5">cutlass::gemm::thread::MmaGeneric&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Operator_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1Gemv.html#a37681d1f6e0e41401e183988f041ec09">cutlass::gemm::threadblock::Gemv&lt; Core_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaPipelined.html#a8ab2e07ec7f4cf23ef158974fe3b8dbb">cutlass::gemm::threadblock::MmaPipelined&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, TransformA_, TransformB_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaSingleStage.html#a48dadbc4536a9ba88599ad61a27c9e18">cutlass::gemm::threadblock::MmaSingleStage&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaComplexTensorOp_3_01Shape___00_01complex_3_01RealElementA_01_146441010dad1f40eb51b6dae3ded216.html#a29b3db60b4c7fa4d269e76dc2a6c6732">cutlass::gemm::warp::MmaComplexTensorOp&lt; Shape_, complex&lt; RealElementA &gt;, LayoutA_, complex&lt; RealElementB &gt;, LayoutB_, complex&lt; RealElementC &gt;, LayoutC_, Policy_, TransformA, TransformB, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimt.html#a832af9ae3fa6bb5606c7477efe576738">cutlass::gemm::warp::MmaSimt&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOp.html#ae7958be3c77c61ee500688bba2195b1a">cutlass::gemm::warp::MmaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK_, AccumulatorsInRowMajor, PartitionsN_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOp.html#a3d96fc74476d2c062cea6e45499bbb45">cutlass::gemm::warp::MmaVoltaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
</li>
<li>FragmentAccumulator
: <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1Convert.html#ae914ab18bfbc7835442d948facd649b5">cutlass::epilogue::thread::Convert&lt; ElementOutput_, Count, ElementAccumulator_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombination.html#a01c42c8de4d1092d2e3c2a09eed6c81f">cutlass::epilogue::thread::LinearCombination&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationClamp.html#ae610d34d4c9d712222dbff75455589b9">cutlass::epilogue::thread::LinearCombinationClamp&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu.html#a37cb96ec3a7e16d88dc7402cd5246ce1">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_3_01ElementOutput___00_01Count_00_01int_00_01float_00_01Round_01_4.html#a1a98738c911a0fdc60933b43075e87ed">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, int, float, Round &gt;</a>
, <a class="el" href="classcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK.html#a2482d1e5fb741c0f9c31f09db15c00c2">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;</a>
, <a class="el" href="structcutlass_1_1reduction_1_1thread_1_1ReduceAdd.html#ab8b0d369480bff1792f6439c2bb0ba18">cutlass::reduction::thread::ReduceAdd&lt; ElementAccumulator_, Element_, Count &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
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>FragmentB
: <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_0116_00_018_00_018_01_4_00_0132_00_01half__02a3f19a78995f97d793a668e0e4d4f0.html#a80ac7bebbcb1df44fa791e8e7c27105c">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 16, 8, 8 &gt;, 32, half_t, layout::RowMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_0116_00_018_00_018_01_4_00_0132_00_01half__96363097c47b056f0ca1911afd7f8b7a.html#a39a754967d9ef1a7aa6065f189eb2119">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 16, 8, 8 &gt;, 32, half_t, layout::RowMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_01128_01_4_00_0132_00_01uint15918972b95027764b3a849b03075ed2b.html#af82ee2181ddcd4551eac17dfb758bbc7">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 128 &gt;, 32, uint1b_t, layout::RowMajor, uint1b_t, layout::ColumnMajor, int, layout::RowMajor, OpXorPopc &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__927179f46017ea5f58f859f1196c4829.html#accc5232c75d86ceec3cf158ccb9d485a">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__8ebae0cbdf333fddfe5c24d35ebe8e02.html#a62eabe2936fdb0c0f5c6c3b81ff5ee70">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__5299c9c90c8f2f521be0c8cec1c3eb08.html#a759fc3621a1dac0bf9195aa5247984bc">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__f083347e265b1e9eea5572d86ddb6bf9.html#af3fe248524914cefe81de5f1643a4c04">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_a62aa63a212985df306fb27e8a50aeae.html#a50327c0a9b9a902a252982cb939adbec">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_ab741d81fdc991345cb9e43c29fca573.html#a86b6a9e4a7fd58a19a35b4708fa016c6">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_5221708cec5828d35db1d1c47cb4964e.html#adb56df98af59f51c2c2dc8d24572aba8">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_bef0c048bc0f8ba2d875cb7ab26d363b.html#a57b7d267e7c5abb3ea965928724b35d3">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_6e513ccbc44ae7909a60d93b9b5435b3.html#ac90c60556e063b75b9a308e3a6c0d149">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_0ee08a4520882d24ba9026879265e892.html#a49711a84ef23d8785a80770293fc6c05">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_4746fc55e614df0016c518d3fda2677e.html#a22c4bc1ed799179c65b7b5a3794579ab">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_546e9ec6de6a5970b326da6f6280f1d4.html#a6960707673dacc0bc64a2f1445ed6535">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b03e3b50dbcb30d0d1ac062f3a9d5abef.html#acfa8570d8f7b4d07c0cf42d700b009eb">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b6d968039dde5c9f062ab15f90a8049fe.html#a7cd305897489b398cfeca7487d51fa9b">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4bc4b6ba004e25c44bfd9266c61f937dfb.html#ac4f170abc54f9728b47d2717f4a54530">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b451d5cf5d7e8cbbe476afe3dab5c09b2.html#a5839e113e4f93b3c0f3e80ff3b33f584">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_b0242d7a01097510effbc4718040d3e5.html#a655d3faefcc0bc8c2169a6f77bc13960">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_c7f88bfd32a544fba8111d2dcadeab11.html#ab94f059e28c6b7144028bb9d34ee512e">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_44a3b2a8df88a2b067f1284515cb5371.html#adcf33eba0bd8d1cd85b8e42ce1d59bd2">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_4b7308177b308a272c1889fbe9670275.html#a287ea2ae06df041c682b4afdbfe584ab">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_5a9888862cebd333ecaf11f7262f77d4.html#a05637ac60683b458b5b4945f64a272eb">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_31defda8ea2b7d855642ffd77da1a411.html#af26862b19dc18d3734f27c5ddbba7c66">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_839a7c8bb938d1661f4611e68f85d8cb.html#a03b54f0c36a186f38972856ed10ce960">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::RowMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_73d9802d6b944a5299bc255887db6bbc.html#af5cad7e63eb81a9949b4d0e2816a35d5">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_72621f7ab9ae4a4ba4fe9725cf8e89c1.html#aead4062ad5f251281eeb098351414936">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::ColumnMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_94c813e3bbfb6f9857c155166f772687.html#a38afec0d6f6ff8c7189b0766ecf80425">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::ColumnMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_17070298bc4cced0a1b98aee2bb6b455.html#a99594cf2b14d51d930aeffe5f2a2eedd">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::RowMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_bf6d29bb09a025e7b96942809743e28a.html#a6d82503a204f8ce40c25cd3de379a49a">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::RowMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l26a133b13650c1d058273e3649f60f04.html#a4fa52e77c752fe0977b68b7494043c32">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::ColumnMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01lbba3a796be96a0276693ef6b259ecc4a.html#ac4b19bd9753d7b09acbc087a6fe693fe">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::ColumnMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l2aa4d2fd2e940e0d0cf7c47bc8f6017c.html#ae9681b4ff3427a57ec0f67fb8ae88fe6">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::RowMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l086c058a15d6c79558e4f3d9ff1dc148.html#adbb7b7b310e01a955d0567039b09d55e">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::RowMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01LayoutA_00_01LayoutB_00_0e1104c65871c539155bd3a0c7631928b.html#af3b9e18d19534c5326766bdb87cebb22">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, LayoutA, LayoutB, layout::ColumnMajor, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01LayoutA_00_01LayoutB_00_07ac147cb320ee0d28ff8e78eb4cd330e.html#a3506cc0bc2411d9af2d2f2a019f73111">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, LayoutA, LayoutB, layout::RowMajor, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01ElementA___00_01LayoutA___00_01ElementB_e41c1cd6078b6d1347fac239b0639d56.html#a2deaf8959c027ab4aca92630b85f5211">cutlass::gemm::thread::Mma&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, arch::OpMultiplyAdd, bool &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01half__t_00_01LayoutA_00_01half__t_00_01L066c9d2371712cdf0cac099ca9bcc578.html#a5a00c6305fd345f12f9469b790e99f12">cutlass::gemm::thread::Mma&lt; Shape_, half_t, LayoutA, half_t, LayoutB, half_t, LayoutC, arch::OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01half__t_00_01LayoutA___00_01half__t_00_088f0e99e501b6012297eb30b4e89bcea.html#a687f0bd7056ea8ff518bfed26f027e4f">cutlass::gemm::thread::Mma&lt; Shape_, half_t, LayoutA_, half_t, LayoutB_, half_t, layout::RowMajor, arch::OpMultiplyAdd, typename platform::enable_if&lt; detail::EnableMma_Crow_SM60&lt; LayoutA_, LayoutB_ &gt;::value &gt;::type &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01int8__t_00_01layout_1_1ColumnMajor_00_013f3785e722edc6e9aab6f866309b8623.html#a34af38037b5cd7328f8bd9e463bfd4b7">cutlass::gemm::thread::Mma&lt; Shape_, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, int32_t, LayoutC_, arch::OpMultiplyAdd, int8_t &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01int8__t_00_01layout_1_1RowMajor_00_01int89c659e7faf47264972bdba6cd80f42b.html#acc0894e5afdf193553e9205e7ea56ada">cutlass::gemm::thread::Mma&lt; Shape_, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int32_t, LayoutC_, arch::OpMultiplyAdd, bool &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1MmaGeneric.html#a170523641b3dcccf8d2a65c42d57fe8a">cutlass::gemm::thread::MmaGeneric&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Operator_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1Gemv.html#acbbd9f710b49dab0582d04b3384a33d1">cutlass::gemm::threadblock::Gemv&lt; Core_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaPipelined.html#a8a6b28dec3f9687f46180d59ac9187fb">cutlass::gemm::threadblock::MmaPipelined&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, TransformA_, TransformB_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaSingleStage.html#aa9a1e7b8bad63a676e50f3ff105edbc0">cutlass::gemm::threadblock::MmaSingleStage&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaComplexTensorOp_3_01Shape___00_01complex_3_01RealElementA_01_146441010dad1f40eb51b6dae3ded216.html#a71e50c4aa4dad62ba4cea49c92d377c3">cutlass::gemm::warp::MmaComplexTensorOp&lt; Shape_, complex&lt; RealElementA &gt;, LayoutA_, complex&lt; RealElementB &gt;, LayoutB_, complex&lt; RealElementC &gt;, LayoutC_, Policy_, TransformA, TransformB, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimt.html#ad7c4d427278c0fc5ec09ac0e1710f71b">cutlass::gemm::warp::MmaSimt&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOp.html#a268f27ae31f2e1a4e8a2e734ec1c1816">cutlass::gemm::warp::MmaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK_, AccumulatorsInRowMajor, PartitionsN_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOp.html#abd446811fcf7ef8f4106c4e7ebabcb84">cutlass::gemm::warp::MmaVoltaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
</li>
<li>FragmentC
: <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_0116_00_018_00_018_01_4_00_0132_00_01half__02a3f19a78995f97d793a668e0e4d4f0.html#a6891104cfabe1a08032740b90f6d7e88">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 16, 8, 8 &gt;, 32, half_t, layout::RowMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_0116_00_018_00_018_01_4_00_0132_00_01half__96363097c47b056f0ca1911afd7f8b7a.html#af6d9872ace0a79851cd328872dc94fb2">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 16, 8, 8 &gt;, 32, half_t, layout::RowMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_01128_01_4_00_0132_00_01uint15918972b95027764b3a849b03075ed2b.html#aae7a4f4758e8d71abdf1b5f83c083cfe">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 128 &gt;, 32, uint1b_t, layout::RowMajor, uint1b_t, layout::ColumnMajor, int, layout::RowMajor, OpXorPopc &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__927179f46017ea5f58f859f1196c4829.html#a20875a18df3c97c9dc3444bff7adb137">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__8ebae0cbdf333fddfe5c24d35ebe8e02.html#a7b9bc96139177e113193192e29f2ecfb">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__5299c9c90c8f2f521be0c8cec1c3eb08.html#af47670426067e078d84b7d6195ce6ab8">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__f083347e265b1e9eea5572d86ddb6bf9.html#afe9e13aa24236342941625a29f0bb6cf">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_a62aa63a212985df306fb27e8a50aeae.html#a723d3450071f399e625b5de93ab2f3cc">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_ab741d81fdc991345cb9e43c29fca573.html#a5d474e270e15c468ea4c1ca6f8b9bb05">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_5221708cec5828d35db1d1c47cb4964e.html#a7b929328375ecf4ae3cb67d595ff0487">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_bef0c048bc0f8ba2d875cb7ab26d363b.html#a2cfe3fef5186f3522e1d3e836e57b9af">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_6e513ccbc44ae7909a60d93b9b5435b3.html#aa4e15207db0471ab1ec0b2eee6ff9e62">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_0ee08a4520882d24ba9026879265e892.html#a3e2f438ecd703ea6ae456e2ef6915808">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_4746fc55e614df0016c518d3fda2677e.html#a8f965468f743c33c92de38d80f4dfc25">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_546e9ec6de6a5970b326da6f6280f1d4.html#a6a0342ea8a650dcba6fa710bcfb24278">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b03e3b50dbcb30d0d1ac062f3a9d5abef.html#a01fdded685b706cb128a39eb02b317d8">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b6d968039dde5c9f062ab15f90a8049fe.html#af460b9dc8922a1b84f159129f2c2384f">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4bc4b6ba004e25c44bfd9266c61f937dfb.html#a854f1a6c5d445bed0f4b5cac168af12d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b451d5cf5d7e8cbbe476afe3dab5c09b2.html#ac8240f7916bb7dbafca9c21c407d3fa0">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_b0242d7a01097510effbc4718040d3e5.html#a99b8f43fbf4eeda847d0673358924b76">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_c7f88bfd32a544fba8111d2dcadeab11.html#a21b8496bc263aa017e87c5e43879a94f">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_44a3b2a8df88a2b067f1284515cb5371.html#a2526ed72140753f7d0b96e54f490b063">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_4b7308177b308a272c1889fbe9670275.html#aae5a79ebe8d5033dcc8fa966a956bd18">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_5a9888862cebd333ecaf11f7262f77d4.html#a6e4cab8bcd32ff69c881965656af8995">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_31defda8ea2b7d855642ffd77da1a411.html#ad0e1290950b365bca2517c673c9fe14a">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_839a7c8bb938d1661f4611e68f85d8cb.html#a2b0d9811747225dfef1c4d1a3c994ef4">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::RowMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_73d9802d6b944a5299bc255887db6bbc.html#a2cbf6317b3d9d38e40f8f0bcd8f04557">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1EpilogueWorkspace.html#a2a9e0715486050e4c52985f72392ba97">cutlass::epilogue::EpilogueWorkspace&lt; Shape_, WarpCount, FragmentC_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1DirectEpilogueTensorOp.html#a4e03d7b83604c014d530353a67ba3356">cutlass::epilogue::threadblock::DirectEpilogueTensorOp&lt; Shape_, Operator_, PartitionsK, Element_, OutputOp_, ConvertOp_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_72621f7ab9ae4a4ba4fe9725cf8e89c1.html#a4105c83f7b881c8609c56a55914fcea6">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::ColumnMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_94c813e3bbfb6f9857c155166f772687.html#ab020c25ee9d01c16ec03efc64b6dafe2">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::ColumnMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_17070298bc4cced0a1b98aee2bb6b455.html#a58f282b08f9bdf33863e87838edcff79">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::RowMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_bf6d29bb09a025e7b96942809743e28a.html#a375e71b2b3e9b800b6c5a1f8a8e81881">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::RowMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l26a133b13650c1d058273e3649f60f04.html#ad39651f6394d74322ae9f8ce220ccb6f">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::ColumnMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01lbba3a796be96a0276693ef6b259ecc4a.html#aee67b84a9eca972334278ad84988ffa7">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::ColumnMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l2aa4d2fd2e940e0d0cf7c47bc8f6017c.html#a0d50d8b7290e39ebb5e0cbb7523bd320">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::RowMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l086c058a15d6c79558e4f3d9ff1dc148.html#a8c02fc1a052ea30219a6a31c2dfb92d5">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::RowMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01LayoutA_00_01LayoutB_00_0e1104c65871c539155bd3a0c7631928b.html#a0f5b61c1425281945b40db7038be01d7">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, LayoutA, LayoutB, layout::ColumnMajor, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01LayoutA_00_01LayoutB_00_07ac147cb320ee0d28ff8e78eb4cd330e.html#ae9604312ef68941313db07cd4ea86836">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, LayoutA, LayoutB, layout::RowMajor, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01ElementA___00_01LayoutA___00_01ElementB_e41c1cd6078b6d1347fac239b0639d56.html#a6109558276e8c66a4d3e9ad53fb046d8">cutlass::gemm::thread::Mma&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, arch::OpMultiplyAdd, bool &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01half__t_00_01LayoutA_00_01half__t_00_01L066c9d2371712cdf0cac099ca9bcc578.html#ac5f3a1ad86714fdce4af1e8a4738a4f7">cutlass::gemm::thread::Mma&lt; Shape_, half_t, LayoutA, half_t, LayoutB, half_t, LayoutC, arch::OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01half__t_00_01LayoutA___00_01half__t_00_088f0e99e501b6012297eb30b4e89bcea.html#a0382879142faec5a4b6190869fc6187f">cutlass::gemm::thread::Mma&lt; Shape_, half_t, LayoutA_, half_t, LayoutB_, half_t, layout::RowMajor, arch::OpMultiplyAdd, typename platform::enable_if&lt; detail::EnableMma_Crow_SM60&lt; LayoutA_, LayoutB_ &gt;::value &gt;::type &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01int8__t_00_01layout_1_1ColumnMajor_00_013f3785e722edc6e9aab6f866309b8623.html#ac70df9d0d37586189c6e3796d2b4bfb8">cutlass::gemm::thread::Mma&lt; Shape_, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, int32_t, LayoutC_, arch::OpMultiplyAdd, int8_t &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01int8__t_00_01layout_1_1RowMajor_00_01int89c659e7faf47264972bdba6cd80f42b.html#a9fea49fb8cd291c925ed0a52bccff1e7">cutlass::gemm::thread::Mma&lt; Shape_, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int32_t, LayoutC_, arch::OpMultiplyAdd, bool &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1MmaGeneric.html#a810a6e485277560581a10cf0330fd7bd">cutlass::gemm::thread::MmaGeneric&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Operator_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1Gemv.html#a1de641dbf472a7221873c8c8928be084">cutlass::gemm::threadblock::Gemv&lt; Core_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaPipelined.html#a2993390a8feebe3cf4bfb77a23c5ef33">cutlass::gemm::threadblock::MmaPipelined&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, TransformA_, TransformB_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaSingleStage.html#ae0503fd9e4983496a6168e5a068d3017">cutlass::gemm::threadblock::MmaSingleStage&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaComplexTensorOp_3_01Shape___00_01complex_3_01RealElementA_01_146441010dad1f40eb51b6dae3ded216.html#a39f05934bf2f0201344b464449248f4c">cutlass::gemm::warp::MmaComplexTensorOp&lt; Shape_, complex&lt; RealElementA &gt;, LayoutA_, complex&lt; RealElementB &gt;, LayoutB_, complex&lt; RealElementC &gt;, LayoutC_, Policy_, TransformA, TransformB, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimt.html#aeadf4d2944267ebfb36dbbd9751bac56">cutlass::gemm::warp::MmaSimt&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOp.html#ae03fe3d58f3d96e875055e7721548bab">cutlass::gemm::warp::MmaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK_, AccumulatorsInRowMajor, PartitionsN_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOp.html#a316d26dfb4b537fa0ccdb3542260406b">cutlass::gemm::warp::MmaVoltaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
</li>
<li>FragmentCD
: <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1DefaultGemv.html#a413056056e1b3a4e60cbca72f47ac9d5">cutlass::gemm::kernel::DefaultGemv&lt; ThreadBlockShape_, ThreadShape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementCD_, LayoutCD_, ElementAccumulator_ &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
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>FragmentElement
: <a class="el" href="structcutlass_1_1reduction_1_1thread_1_1ReduceAdd.html#a9cbf7165c8ca9da43b4038d7ea459a73">cutlass::reduction::thread::ReduceAdd&lt; ElementAccumulator_, Element_, Count &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
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>FragmentIteratorComplexTensorOp()
: <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorComplexTensorOp_3_01WarpShape___00_01Operato8cf03c624cf3210c71b7cbd580b080f8.html#acad4794bc84fbe582f58b4328c0d84a4">cutlass::epilogue::warp::FragmentIteratorComplexTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
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>FragmentIteratorSimt()
: <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorSimt_3_01WarpShape___00_01Operator___00_01la3f2abc523201c1b0228df99119ab88e1.html#a3dee783224164a08c24654aba39ddbdb">cutlass::epilogue::warp::FragmentIteratorSimt&lt; WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
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>FragmentIteratorTensorOp()
: <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorTensorOp_3_01WarpShape___00_01OperatorShape_e459aab140a2ce78336e584f95886726.html#ad690dd7f54217a7d1e2033e557306e8d">cutlass::epilogue::warp::FragmentIteratorTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::ColumnMajorInterleaved&lt; InterleavedK &gt; &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorTensorOp_3_01WarpShape___00_01OperatorShape_5e78dabe303f20d76b00c600aab61eda.html#a2fcc41a59fce4e01d3ea7917a73c5ec8">cutlass::epilogue::warp::FragmentIteratorTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
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>FragmentIteratorVoltaTensorOp()
: <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorVoltaTensorOp_3_01WarpShape___00_01gemm_1_1Gdb805a2dc5571ac3b66e0fe6ffdcede2.html#ae7d7fe96fc26493283a5b3b35c9f1c56">cutlass::epilogue::warp::FragmentIteratorVoltaTensorOp&lt; WarpShape_, gemm::GemmShape&lt; 32, 32, 4 &gt;, float, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorVoltaTensorOp_3_01WarpShape___00_01gemm_1_1G16e08718cffa0989cce3fe8dbc4b075b.html#a11fda4be46e143b204c0a6e2325d8490">cutlass::epilogue::warp::FragmentIteratorVoltaTensorOp&lt; WarpShape_, gemm::GemmShape&lt; 32, 32, 4 &gt;, half_t, layout::RowMajor &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
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>FragmentIteratorWmmaTensorOp()
: <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorWmmaTensorOp_3_01WarpShape___00_01OperatorShfdb1f120c6797383663f9fd11d0fc599.html#aaa5e511cd6d134528901955222e0bfe6">cutlass::epilogue::warp::FragmentIteratorWmmaTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
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>FragmentOutput
: <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1Convert.html#a9944ba7d971ea33e7b005d6fba1ef92b">cutlass::epilogue::thread::Convert&lt; ElementOutput_, Count, ElementAccumulator_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombination.html#aecbeb4d1a8e75feb7ca8079b2a2f862c">cutlass::epilogue::thread::LinearCombination&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationClamp.html#a3380b95c93bcfdca8df7e7c80a6e4153">cutlass::epilogue::thread::LinearCombinationClamp&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu.html#a14bfd5ef175cf1d0242cc3a3c28e130d">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_3_01ElementOutput___00_01Count_00_01int_00_01float_00_01Round_01_4.html#a8b815a6cf10f37160ce763030a81a2e0">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, int, float, Round &gt;</a>
, <a class="el" href="classcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK.html#a8571b4b913e4fe4fb44ea79e7f139abb">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
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>FragmentWorkspace
: <a class="el" href="classcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK.html#ad16928ac661803077fe524afc0d21b0b">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
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>front()
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4.html#ab7ebd33505e48ab3beb6b551e8b762e5">cutlass::Array&lt; T, N, false &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#a0a692495c5f7a7d098e60b9292a07e4f">cutlass::Array&lt; T, N, true &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
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>func
: <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFillGaussianFunc.html#aa6cc9355b27c1d33e351a31bcf4aece7">cutlass::reference::host::detail::TensorFillGaussianFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFillRandomUniformFunc.html#a9272276afdeec7284dfcc6605f153c97">cutlass::reference::host::detail::TensorFillRandomUniformFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFuncBinaryOp.html#a2d2bbd2c0a8b8f0292e6820e5d133411">cutlass::reference::host::detail::TensorFuncBinaryOp&lt; ElementA, LayoutA, ElementB, LayoutB, ElementD, LayoutD, BinaryFunc &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>Functor
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 class="el" href="structcutlass_1_1reduction_1_1BatchedReduction.html#a30605f35d51e4364fe80edb80eac5e80">cutlass::reduction::BatchedReduction&lt; BatchedReductionTraits_ &gt;</a>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
</li>
<li>functor
: <a class="el" href="structcutlass_1_1reduction_1_1BatchedReduction.html#a0f6b38c1b3a5800e6f29d9a2c6c1928d">cutlass::reduction::BatchedReduction&lt; BatchedReductionTraits_ &gt;</a>
</li>
<li>Functor
: <a class="el" href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#af62fd99fae22a4e4d93cfa0560f1dcc5">cutlass::reduction::BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt;</a>
</li>
<li>functor_caller()
: <a class="el" href="structcutlass_1_1reduction_1_1BatchedReduction.html#a7c1d173cbe3abd93bd7bd4c4bf0e0d26">cutlass::reduction::BatchedReduction&lt; BatchedReductionTraits_ &gt;</a>
</li>
<li>functorParams
: <a class="el" href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#a11e54597165b66e3054c6b9f43210760">cutlass::reduction::BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt;::Params</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
</ul>
</div><!-- contents -->
<!-- start footer part -->
<hr class="footer"/><address class="footer"><small>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
Generated by &#160;<a href="http://www.doxygen.org/index.html">
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<img class="footer" src="doxygen.png" alt="doxygen"/>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
</a> 1.8.11
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</small></address>
</body>
</html>
Reference in New Issue Copy Permalink