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

389 lines
55 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><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 class="current"><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_m"></a>- m -</h3><ul>
Updated doxygen
2018-09-20 05:02:08 +08:00
<li>m()
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_1gemm_1_1BatchedGemmCoord.html#ac78745410bc514e7978bcafb22fd0843">cutlass::gemm::BatchedGemmCoord</a>
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#a8199f5e336a20c31e54d68b11e9fa3d3">cutlass::gemm::GemmCoord</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>main_loop
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="unioncutlass_1_1gemm_1_1kernel_1_1Gemm_1_1SharedStorage.html#a25ca6f379b42d97b73de07473e2fdf02">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::SharedStorage</a>
, <a class="el" href="unioncutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1SharedStorage.html#ae9f9d72c08cbf031961d95dca828d573">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::SharedStorage</a>
, <a class="el" href="unioncutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1SharedStorage.html#a9a4c11bb49a8cadae9cb2c17c92bf6db">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::SharedStorage</a>
</li>
<li>mantissa()
: <a class="el" href="structcutlass_1_1half__t.html#ad21f512b39dff2e51e6fc2bf0ae2ebcf">cutlass::half_t</a>
</li>
<li>Mask()
: <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1InterleavedPredicatedTileIterator_1_1Mask.html#a1e08665b2b5cb30736b03e69ec215298">cutlass::epilogue::threadblock::InterleavedPredicatedTileIterator&lt; ThreadMap_, Element_, InterleavedK &gt;::Mask</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1PredicatedTileIterator_1_1Mask.html#aefa401d42f20dd3740d90a410bcedc83">cutlass::epilogue::threadblock::PredicatedTileIterator&lt; ThreadMap_, Element_ &gt;::Mask</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__da632779aba661c0f4cfaaa78126b771.html#a5e9a837e34af696d41f235341f01d0a2">cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__1790abaa54a01f277d75766d5882fec8.html#a3158eacb57d29742d9119710f5152a0a">cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__7327fa15996bcb8502cdfcc192350fe1.html#ac8b5b1def7e2de1da7faf02fbf04e2ea">cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen89c687c583745a73cb485041911a4c4e.html#a33876a55f6ecc68c03f13f03cf97c1fc">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemenab63a1e105bf37f6371516cb9e2c5a7a.html#a98b107b995cb3e29be1f8b2851be6b03">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen784a0e9da3f55064c47e5613791f51f7.html#a4e672808cc7e6c9addb4eb1f4f252fbf">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen9838736ad62fae54213fbaf722a989ab.html#ab4c0b4e67d67cf817b56fe95a5693a66">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen809793e785fb4211888c6b4e5dcfcb39.html#acda8e8e61036e923301a908b1a985dc2">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_0165b39a630d10785a3558406f9adb99b9.html#a8275a7f49a107c4122a4d3278aba94ce">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#a20a0f1a208b301ecd6f29cced4f7cbb7">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#aef7b4321bef75c2ad969fedfb7e1aeee">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#a9e22fa23c9927277741d6bea462f7589">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#ac9583e3c6f7046733a0fa5671d8f32be">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#a6b9ef24ac3350b33f76a2c580e31a579">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#a278a0d69e5ffe2602faac7a29f5f9e74">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#aa047aa70ae6f0eabebbafa2620692833">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessSize &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>math_instruction
: <a class="el" href="structcutlass_1_1library_1_1TileDescription.html#aad96c46afa29cb6c02f8e0737eeaf963">cutlass::library::TileDescription</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>MathInstructionDescription()
: <a class="el" href="structcutlass_1_1library_1_1MathInstructionDescription.html#a2a14df3a23c8b961ec50a949377e9e8b">cutlass::library::MathInstructionDescription</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>Matrix()
: <a class="el" href="classcutlass_1_1thread_1_1Matrix.html#a96d351d1686a79eececdc4ee2e6b886e">cutlass::thread::Matrix&lt; Element, Rows, Columns, Layout &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>MatrixCoord()
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_1MatrixCoord.html#a64bddbf8238dc937a01a140722f7f39c">cutlass::MatrixCoord</a>
</li>
<li>max
: <a class="el" href="structcutlass_1_1Distribution.html#a824641fd3addfa360999614970adfac0">cutlass::Distribution</a>
, <a class="el" href="structstd_1_1numeric__limits_3_01cutlass_1_1half__t_01_4.html#a39a5774583daedbb5ac4aaaaa8034883">std::numeric_limits&lt; cutlass::half_t &gt;</a>
</li>
<li>max_dim_index()
: <a class="el" href="structcutlass_1_1Coord.html#abe58b7c8f153a6029c2adc173f340fe0">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
</li>
<li>max_size()
: <a class="el" href="structcutlass_1_1AlignedBuffer.html#a673d7413585d44f0c025840c9b84b6b3">cutlass::AlignedBuffer&lt; T, N, Align &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4.html#a8f982c95366ce4fda90e35281adfe63c">cutlass::Array&lt; T, N, false &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#a3391b79db2b9f3bac9576c9bc7af0402">cutlass::Array&lt; T, N, true &gt;</a>
</li>
<li>maximum_compute_capability
: <a class="el" href="structcutlass_1_1library_1_1TileDescription.html#a37f5925a2f04995091c56c034f4f2572">cutlass::library::TileDescription</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<li>maxInReg
: <a class="el" href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#af11a3284195a24e580d2f379f179f05a">cutlass::reduction::BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt;</a>
</li>
<li>maxOutReg
: <a class="el" href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#ac28e31791c5888bbe7b04abe6376a422">cutlass::reduction::BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt;</a>
</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>mean
: <a class="el" href="structcutlass_1_1Distribution.html#a776df53c7ad1b7de983c9f9d17d7438c">cutlass::Distribution</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomGaussianFunc_1_1Params.html#a07d12eba25e8e2e9da03ce735b1b8113">cutlass::reference::device::detail::RandomGaussianFunc&lt; Element &gt;::Params</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomGaussianFunc.html#a5da7c7b953aa178bb356ba8578245d89">cutlass::reference::host::detail::RandomGaussianFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomGaussianFunc_3_01complex_3_01Element_01_4_01_4.html#ad8c60b0630a2867fd80a0d09a3cf63cd">cutlass::reference::host::detail::RandomGaussianFunc&lt; complex&lt; Element &gt; &gt;</a>
</li>
<li>min
: <a class="el" href="structcutlass_1_1Distribution.html#a846430e3a21ed25c779fc6e714bc1bcc">cutlass::Distribution</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomUniformFunc_1_1Params.html#a5f49e5175fa2fc3c3c7fb495fe6958db">cutlass::reference::device::detail::RandomUniformFunc&lt; Element &gt;::Params</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomUniformFunc.html#ab3de08baeb9246f5472a1c8ec19c68fa">cutlass::reference::host::detail::RandomUniformFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomUniformFunc_3_01complex_3_01Element_01_4_01_4.html#a005299ae9f0e4533f6847a7f6ff9e6ad">cutlass::reference::host::detail::RandomUniformFunc&lt; complex&lt; Element &gt; &gt;</a>
, <a class="el" href="structstd_1_1numeric__limits_3_01cutlass_1_1half__t_01_4.html#ad9175b4d7b32fe18cf9c07e4f559b32c">std::numeric_limits&lt; cutlass::half_t &gt;</a>
</li>
<li>min_dim_index()
: <a class="el" href="structcutlass_1_1Coord.html#ae89e8a9fa3f07308f8938052ef1aa1fb">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
</li>
<li>minimum_compute_capability
: <a class="el" href="structcutlass_1_1library_1_1TileDescription.html#a25557ce7220e284c095ef7f691b16fa0">cutlass::library::TileDescription</a>
</li>
<li>mk()
: <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#a68e79e339f5de2ce79fb90f2ec099233">cutlass::gemm::GemmCoord</a>
</li>
<li>Mma
: <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1DefaultGemm_3_01ElementA_00_01layout_1_1ColumnMajorInterleave661fe54d13cc2c9153dcdf31e4beaa30.html#a3ff721fd6df7783d2481ff466ef9041b">cutlass::gemm::kernel::DefaultGemm&lt; ElementA, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, kAlignmentA, ElementB, layout::RowMajorInterleaved&lt; InterleavedK &gt;, kAlignmentB, ElementC, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, int32_t, arch::OpClassTensorOp, arch::Sm75, ThreadblockShape, WarpShape, InstructionShape, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator, IsBetaZero &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1DefaultGemm_3_01ElementA_00_01LayoutA_00_01kAlignmentA_00_01Edd80343e6570718ed237122e4ebf7fb5.html#a2582e9547846a602ce69e9b9ba445393">cutlass::gemm::kernel::DefaultGemm&lt; ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementC, layout::RowMajor, ElementAccumulator, arch::OpClassSimt, ArchTag, ThreadblockShape, WarpShape, GemmShape&lt; 1, 1, 1 &gt;, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1DefaultGemm_3_01ElementA_00_01LayoutA_00_01kAlignmentA_00_01E044b039b2fe402f29b04a9f5feee5342.html#aa24fd6fd5e9ab4de4e0edec3226e044b">cutlass::gemm::kernel::DefaultGemm&lt; ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementC, layout::RowMajor, ElementAccumulator, arch::OpClassTensorOp, arch::Sm70, ThreadblockShape, WarpShape, GemmShape&lt; 8, 8, 4 &gt;, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1DefaultGemm_3_01ElementA_00_01LayoutA_00_01kAlignmentA_00_01E5d78d37a9ae2ec08d7d477d571df036e.html#a04323cdfbd5109ceabcb5bf31d435798">cutlass::gemm::kernel::DefaultGemm&lt; ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementC, layout::RowMajor, ElementAccumulator, arch::OpClassTensorOp, arch::Sm75, ThreadblockShape, WarpShape, InstructionShape, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1DefaultGemm_3_01int8__t_00_01LayoutA_00_01kAlignmentA_00_01inf48440732c1c5f42ddbfaba179861815.html#a9c029aa1ac8275a18aeaefa8a4469dfb">cutlass::gemm::kernel::DefaultGemm&lt; int8_t, LayoutA, kAlignmentA, int8_t, LayoutB, kAlignmentB, ElementC, LayoutC, ElementAccumulator, arch::OpClassSimt, ArchTag, ThreadblockShape, WarpShape, GemmShape&lt; 1, 1, 4 &gt;, EpilogueOutputOp, ThreadblockSwizzle, 2, SplitKSerial, Operator, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1DefaultGemmSplitKParallel.html#a4365444d154a4fa7e5b7eb5f2ffa6e77">cutlass::gemm::kernel::DefaultGemmSplitKParallel&lt; ElementA_, LayoutA_, kAlignmentA, ElementB_, LayoutB_, kAlignmentB, ElementC_, LayoutC_, ElementAccumulator, OperatorClass, ArchTag, ThreadblockShape, WarpShape, InstructionShape, EpilogueOutputOp, ThreadblockSwizzle, Stages, Operator &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm.html#a950fcca6c690f22061706faccef9877a">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched.html#a4365b372cc097b012c0375c612c67f59">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel.html#af603f7f29cfb208585dfececcbed9b02">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;</a>
</li>
<li>mma
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOp.html#a2eddba2b1644d2df5ade4de535b95ed7">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#aa4837cb1fd3eac3713e99d0625f14842">cutlass::gemm::warp::MmaVoltaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
</li>
<li>MmaBase()
: <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaBase.html#a3bc85aaa0a10e2b1846b7f9d4269e498">cutlass::gemm::threadblock::MmaBase&lt; Shape_, Policy_, Stages, Enable &gt;</a>
Updating Doxygen docs
2018-10-27 05:54:58 +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>MmaComplexTensorOp()
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaComplexTensorOp_3_01Shape___00_01complex_3_01RealElementA_01_146441010dad1f40eb51b6dae3ded216.html#a028ca8153d786b8d0855db3c9076bf2d">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>
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>MmaCore
: <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMma_3_01ElementA_00_01LayoutA_00_01kAlignmentA_0010764e1fd5a3251a57eddafbd83eab8e.html#afb9ea2c8962b31211307d744a6543451">cutlass::gemm::threadblock::DefaultMma&lt; ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementAccumulator, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, OperatorClass, ArchTag, ThreadblockShape, WarpShape, InstructionShape, 2, Operator, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMma_3_01ElementA_00_01LayoutA_00_01kAlignmentA_00c67c16f9881e4f2fda76d8ed83ebabd6.html#a234366375e417a362efcc6ace66d59b6">cutlass::gemm::threadblock::DefaultMma&lt; ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementAccumulator, layout::RowMajor, arch::OpClassSimt, ArchTag, ThreadblockShape, WarpShape, InstructionShape, 2, Operator, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMma_3_01ElementA_00_01LayoutA_00_01kAlignmentA_00ce36642cae579bce6605ff8edde3c6ab.html#a84ba5d319587fe073f92e6e84a6dc9f7">cutlass::gemm::threadblock::DefaultMma&lt; ElementA, LayoutA, kAlignmentA, ElementB, LayoutB, kAlignmentB, ElementAccumulator, layout::RowMajor, arch::OpClassTensorOp, ArchTag, ThreadblockShape, WarpShape, InstructionShape, 2, Operator, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMma_3_01int8__t_00_01LayoutA_00_01kAlignmentA_00_07e7230d4011ada5e22cfcb29103b696.html#a0badaa2f2872010545f4f18e158dd4b2">cutlass::gemm::threadblock::DefaultMma&lt; int8_t, LayoutA, kAlignmentA, int8_t, LayoutB, kAlignmentB, ElementAccumulator, layout::RowMajor, arch::OpClassSimt, ArchTag, ThreadblockShape, WarpShape, GemmShape&lt; 1, 1, 4 &gt;, 2, Operator, false &gt;</a>
</li>
<li>MmaCount
: <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1InterleavedOutputTileThreadMap.html#a8a829dc89f7c7a2ed2c47078d44ebd44">cutlass::epilogue::threadblock::InterleavedOutputTileThreadMap&lt; WarpCount_, MmaCount_, Threads, ElementsPerAccess, ElementSize &gt;</a>
</li>
<li>MmaIterations
: <a class="el" href="structcutlass_1_1epilogue_1_1warp_1_1VoltaTensorOpPolicy_3_01WarpShape___00_01gemm_1_1GemmShape_136ce744d4c1c6e8707f5a9785196194.html#aa1ae4efe90c09565eb128c5899b0bd74">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#a9c9f7ef856b9f44a823bcf054ab945c3">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_1MmaComplexTensorOp_3_01Shape___00_01complex_3_01RealElementA_01_146441010dad1f40eb51b6dae3ded216.html#a9a09561812d929c605312ab1c80d16fa">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="structcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___0d35fa5dc4e4b4f72784c943fd857fc1d.html#a45aaaf52d132111ece0b1a5463da5a0e">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajor, InstructionShape_, OpDelta_ &gt;::Policy</a>
, <a class="el" href="structcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___03822d9be37f3725022005a5434441f22.html#a9aa3c185c4cf51ba6b29f2bdfae2ce6c">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajorInterleaved&lt; InterleavedN &gt;, InstructionShape_, OpDelta_ &gt;::Policy</a>
, <a class="el" href="structcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___093b5d2838ac5a742704ef62b5c8688f0.html#a29f781ce62fd4a067019ae90aa35c075">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ &gt;::Policy</a>
, <a class="el" href="structcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpAccumulatorTileIterator_1_1Policy.html#a097dab943f286373446529301b6c5f19">cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, Layout_, InstructionShape_, OpDelta_ &gt;::Policy</a>
</li>
<li>MmaOp
: <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1MmaGeneric.html#a19ba069f4e07c10243abdef545a32529">cutlass::gemm::thread::MmaGeneric&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Operator_ &gt;</a>
</li>
<li>MmaPipelined()
: <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaPipelined.html#ac8dc63b0f3ced3d8f615b56678e26400">cutlass::gemm::threadblock::MmaPipelined&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, TransformA_, TransformB_, Enable &gt;</a>
</li>
<li>MmaPolicy
: <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShab94a11a77dd0565102710907089acee0.html#a62de2763f572e5a71e9587f49d0905e6">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShafafd5c61db86cbfe90863578ddd11092.html#a05d5677b44fd111bfe6ebcc6fb7d1676">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha46446d1e3871e31d2e728f710d78c8c1.html#acc44f3b1dd43e72319731a5a44c4d0a4">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_, &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha8da7a0cfbbe859b701fdd9f2b8566aa7.html#a19db74df10235f4d5e36518bac33b4c6">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha84e9f8afb6a4ca9f5dcd219b182d16e7.html#af2c71eb39f54a898753b9ff25e6f3072">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha2c0d0b7cdb5c4bcb11e83c058eb65345.html#a520a725ea60dc096d0cdf0e101d0e547">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 4 &gt;, int8_t, layout::ColumnMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha34a52cc7b2942e8c290f0032b6779b52.html#aada86baebc846e6614f3616688477fba">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 4 &gt;, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShaaf312aafe9da92ea9d417bcc12a8e7dc.html#ad8f801b18df3a58f83aa5e8ce325fcf6">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 4 &gt;, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha863d4139ccaa713bc4bde32c425f4067.html#aabe20df5c98890c82bdd20c351fb9709">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 4 &gt;, int8_t, layout::RowMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShaf03a122202ad10acdc96f280106d678b.html#a463db7b6cbc0d7282f40527d4d19b1a7">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 8, 8, 4 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha69bef08ea63dd930f99d9788105873dd.html#a902194945a2d446b1c5b8b8b9f2b7cfe">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 8, 8, 4 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha3adf608332a8c9ee7014fced0da8a9ca.html#ae5c7db9853bf902893d74216bd1e152a">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 8, 8, 4 &gt;, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShab7edfba3cdf43a07e3c4d719d87565a4.html#a2cd92e3d2ee0b2e50bfa0d1aec1a0e97">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 8, 8, 4 &gt;, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instruc803d38bc1e4618c07c47f54c87ae2678.html#aab58d254a612fc24677fdf1d401a0e43">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instrucf60fe02fcdd80d28b7fd419133465dcc.html#a594192a60d21cfc23a84d226c19eda14">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instruc2bf00737f4ad0a9da9a8be6d3e66c152.html#a437e37a3722b84eaaaf850808f80950f">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, ElementB_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instruc24092ddc01fc83dabb7db4c14880fe60.html#ad3d2144895ccb48a231821598bc0c91b">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instruc4fee9f2965b8468bfb42b94a74527d22.html#af40acc821237daa148a5dc1eae8db11f">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
</li>
<li>MmaShape
: <a class="el" href="structcutlass_1_1gemm_1_1warp_1_1MmaSimtPolicy.html#a740b65916634aad45538f7abb4b295a9">cutlass::gemm::warp::MmaSimtPolicy&lt; WarpShape_, LaneLayout_, LaneMmaShape_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1warp_1_1MmaTensorOpPolicy.html#a67047c73033653a03c240e430d553c07">cutlass::gemm::warp::MmaTensorOpPolicy&lt; Operator_, OpDelta_ &gt;</a>
</li>
<li>MmaSimt()
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimt.html#a43d8aee45470b6c6717eadfad68ecd10">cutlass::gemm::warp::MmaSimt&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK, Enable &gt;</a>
</li>
<li>MmaSimtOp
: <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultGemvCore.html#ad8a1d5fe1ab735863d08bb2e477c5b58">cutlass::gemm::threadblock::DefaultGemvCore&lt; Shape_, ThreadShape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_ &gt;</a>
</li>
<li>MmaSimtPolicy
: <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1DefaultThreadMapSimt.html#a9a6af9fcf3e00b00012d0ca488fd6fb8">cutlass::epilogue::threadblock::DefaultThreadMapSimt&lt; ThreadblockShape_, WarpShape_, MmaSimtPolicy_, PartitionsK, Element_, ElementsPerAccess &gt;</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1warp_1_1SimtPolicy_3_01WarpShape___00_01Operator___00_01layout_1_1Rcef1c60e23e997017ae176c92931151d.html#a07ecea6a8c3407343add0be43bb2beba">cutlass::epilogue::warp::SimtPolicy&lt; WarpShape_, Operator_, layout::RowMajor, MmaSimtPolicy_ &gt;</a>
</li>
<li>MmaSimtTileIterator()
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kA_00_01Element_67ca7e11a38e38f2c51b84767654a90f.html#a117441c8b77715fb415d711c15cbba96">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#a264541432a58abbad656747ba97c5834">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#a26705cfbbbb37fde18f9034ca87e7638">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#a12189641e509bb936d0539a6a3c5ef1e">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#a4ea5fd150b537b5b447aaf608f5cc575">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#a4c81d469ef0aff493fb62e04af27fb50">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kC, Element_, layout::RowMajor, Policy_ &gt;</a>
</li>
<li>MmaSingleStage()
: <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaSingleStage.html#a285069a8af925b60ce72945b3ac5171d">cutlass::gemm::threadblock::MmaSingleStage&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
</li>
<li>MmaTensorOp
: <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShaf03a122202ad10acdc96f280106d678b.html#a4d249a541b0bde7825e83388111e3d14">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 8, 8, 4 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha69bef08ea63dd930f99d9788105873dd.html#adf395731ef027df47e8f4012e95bb887">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 8, 8, 4 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha3adf608332a8c9ee7014fced0da8a9ca.html#a8b08686b206f0d15e05a06e4740a5387">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 8, 8, 4 &gt;, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShab7edfba3cdf43a07e3c4d719d87565a4.html#a6ae03edb110147c3338282230d9814c7">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 8, 8, 4 &gt;, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instruc803d38bc1e4618c07c47f54c87ae2678.html#ac18f54c8e22e3683cdcba2e59f6ae5aa">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instrucf60fe02fcdd80d28b7fd419133465dcc.html#a52f33e5e7e2474943d05170e2de4f3a7">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instruc2bf00737f4ad0a9da9a8be6d3e66c152.html#a7bff8f08d3d0b836752ec48df837d0c4">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, ElementB_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_, AccumulatorsInRowMajor &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instruc24092ddc01fc83dabb7db4c14880fe60.html#ae3fc451c287ee81283d2a4c38b8c030d">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01Instruc4fee9f2965b8468bfb42b94a74527d22.html#aa59f4707244fea0ae23b03ee301c5ca3">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, InstructionShape_, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassTensorOp, 2, Operator_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOp.html#ae344dbb59b9234254768238bee17f2d7">cutlass::gemm::warp::MmaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK_, AccumulatorsInRowMajor, PartitionsN_, Enable &gt;</a>
</li>
<li>MmaTensorOpAccumulatorTileIterator()
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___008f607b871a2b3d854eb4def64712c042.html#a5b36b747c47da28a51a5ec94fa51edef">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#aaad8afd436ee02d232e37aebdd0e2a8e">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#a9b33490455b2ec60311e97712a5ae7da">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ &gt;</a>
</li>
<li>MmaTensorOpMultiplicandTileIterator()
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0b84f53cd44b339eccc12067c9f86e11c.html#ab01cd5a33f54ab40ffea40556a042329">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#a3ce87fe4f5ea64612fe20bc97dcab01d">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#a459d61e80b72846e8a096b5c93e3b7a8">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#ad9ac79132faaf9d6913e13915f1d7cde">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#aed04d00e3dd336bdfc9eff39e1123861">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#a82e4782707be8ef551bb2649178efd67">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>
</li>
<li>MmaVoltaTensorOp()
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOp.html#a7ba75978dcba9b1806da57ebb568a57d">cutlass::gemm::warp::MmaVoltaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
</li>
<li>MmaVoltaTensorOpAccumulatorTileIterator()
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpAccumulatorTileIterator.html#a9a5997c666e1e5f0dd19e238c4f2307b">cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, Layout_, InstructionShape_, OpDelta_ &gt;</a>
</li>
<li>MmaVoltaTensorOpMultiplicandTileIterator()
: <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan0d3248553e52cd61ed8a2b3b12a20343.html#aba989c98c040d2b54287cfcbd629535f">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#a499d86e246b80a13ce5e8c3d4ec07980">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#a3e386fc8b9fa4715b3ed437f85695b4a">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#a468ea714e860b51a6735245b7be2cae5">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#a7a6f4590fa188e7353c8e742b030f9e5">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#ac86a65d3aec0b83fc689c14e3fb7e647">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#ac1ee4a04df22d2d24d1a4eaf2e8cca89">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>
</li>
<li>MmaWarpSimt
: <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShab94a11a77dd0565102710907089acee0.html#a5da02e8f8e757c09c7e59ad2750a7833">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShafafd5c61db86cbfe90863578ddd11092.html#accc92ce0cfc6c8e19420557d8b888edb">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::ColumnMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha8da7a0cfbbe859b701fdd9f2b8566aa7.html#ad76d08483cf17fb0bbd2c5670a1a1613">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::RowMajor, ElementB_, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha84e9f8afb6a4ca9f5dcd219b182d16e7.html#a72f4f59aa197a618e56037ff006482d2">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 1 &gt;, ElementA_, layout::RowMajor, ElementB_, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha2c0d0b7cdb5c4bcb11e83c058eb65345.html#a99137700de654a22fea6d9cdc8f09866">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 4 &gt;, int8_t, layout::ColumnMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha34a52cc7b2942e8c290f0032b6779b52.html#ad2299d968635843d88000386bd2ca20d">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 4 &gt;, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShaaf312aafe9da92ea9d417bcc12a8e7dc.html#acba9ead4e2ec4c4b99d7d22f492acf74">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 4 &gt;, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmSha863d4139ccaa713bc4bde32c425f4067.html#a4c28590242538f752a18f7fc9d7959da">cutlass::gemm::threadblock::DefaultMmaCore&lt; Shape_, WarpShape_, GemmShape&lt; 1, 1, 4 &gt;, int8_t, layout::RowMajor, int8_t, layout::RowMajor, ElementC_, LayoutC_, arch::OpClassSimt, 2, Operator_ &gt;</a>
</li>
<li>mn()
: <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#ad8b9f6a9a69546f7a245e0d9a9296137">cutlass::gemm::GemmCoord</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>mnk()
: <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#aca44ec13cf0d22f3074b920401b4a089">cutlass::gemm::BatchedGemmCoord</a>
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#ae86d164050023469df1a5cd86e055c6f">cutlass::gemm::GemmCoord</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>mnkb()
: <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#a881825fbd6a51c3300a939220eb66443">cutlass::gemm::BatchedGemmCoord</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>msg
: <a class="el" href="classcutlass_1_1cuda__exception.html#af748a69a87ad9863985f6a77260ba77a">cutlass::cuda_exception</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>multiply_add()
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_1reference_1_1device_1_1thread_1_1Gemm.html#a6f0aa8fc056eaba23b030efea31c518e">cutlass::reference::device::thread::Gemm&lt; TensorRefA, TensorRefB, TensorRefC, ScalarType, AccumulatorType, OutputTile, InnerProductOp, ConvertOp &gt;</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