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

513 lines
78 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><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 class="current"><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_p"></a>- p -</h3><ul>
<li>packed()
: <a class="el" href="classcutlass_1_1layout_1_1ColumnMajor.html#afd7374b993fd41d55a8c02691b02ead7">cutlass::layout::ColumnMajor</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorBlockLinear.html#a17c4a1e21d92742ca725379a963a3fc7">cutlass::layout::ColumnMajorBlockLinear&lt; BlockRows, BlockColumns &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorInterleaved.html#ab690e548a771a3e35357f454f3c924a5">cutlass::layout::ColumnMajorInterleaved&lt; Interleave &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorTensorOpMultiplicandCongruous.html#ae068c7ff740e38ea4c06a35a57416595">cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorTensorOpMultiplicandCrosswise.html#afd79156b28c636812e72d0fcccd455e7">cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandBCongruous.html#a61d5ba91f20217170458cb0affbe3e02">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandCongruous.html#af4efe777f1d9ec8c62d870e3ced85114">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandCrosswise.html#ac0c8aea5f38628ce078dd1562e44ff06">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; ElementSize, KBlock &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ContiguousMatrix.html#a4c5f7cb47727b7ca7fdce09984c8669d">cutlass::layout::ContiguousMatrix</a>
, <a class="el" href="structcutlass_1_1layout_1_1GeneralMatrix.html#a41dac2461e2f7b652fd06df7631df259">cutlass::layout::GeneralMatrix</a>
, <a class="el" href="classcutlass_1_1layout_1_1PackedVectorLayout.html#aaa31d06cab1db8ff0c303515396b5b09">cutlass::layout::PackedVectorLayout</a>
, <a class="el" href="classcutlass_1_1layout_1_1PitchLinear.html#a3c76f195505f55b26d17c136a13c5f32">cutlass::layout::PitchLinear</a>
, <a class="el" href="classcutlass_1_1layout_1_1RowMajor.html#a770edcc93145fc3dfa4dfdf37a7c515e">cutlass::layout::RowMajor</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorBlockLinear.html#a21b674f37ee92e3ef1413ece1e2f6d49">cutlass::layout::RowMajorBlockLinear&lt; BlockRows, BlockColumns &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorInterleaved.html#a4c012aaca002750d6dda744388f6cd30">cutlass::layout::RowMajorInterleaved&lt; Interleave &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCongruous.html#a443b469af1a3602e3173b132ae3dd40b">cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCrosswise.html#a5b3ff99ea22d5d6c5bce9aa9ab228b46">cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandBCongruous.html#a860c3a44cff442eada17a765b5744049">cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandCongruous.html#a0930c8e2ba61cdfd454817c841ef5135">cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandCrosswise.html#a6ccf0dc730b1b68b5f842c10d4fb710c">cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; ElementSize, KBlock &gt;</a>
, <a class="el" href="classcutlass_1_1layout_1_1TensorCxRSKx.html#a48194ca13de771a824c37d48bf3885d7">cutlass::layout::TensorCxRSKx&lt; Interleave &gt;</a>
, <a class="el" href="classcutlass_1_1layout_1_1TensorNCHW.html#a3096dbd7c7243eaa54c139d137b7e2b5">cutlass::layout::TensorNCHW</a>
, <a class="el" href="classcutlass_1_1layout_1_1TensorNCxHWx.html#af6a8593026ce08ec962f43db36c84496">cutlass::layout::TensorNCxHWx&lt; Interleave &gt;</a>
, <a class="el" href="classcutlass_1_1layout_1_1TensorNHWC.html#ab38ea68b38cf635ea661360967edb8d2">cutlass::layout::TensorNHWC</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicand.html#aae09584d5e60d401d73484d579722007">cutlass::layout::TensorOpMultiplicand&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandColumnMajorInterleaved.html#a559487c37a4387d88726a51cd08b17b7">cutlass::layout::TensorOpMultiplicandColumnMajorInterleaved&lt; ElementSize, InterleavedK &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCongruous.html#a9307437b4b93a3dd450244bb54aa50d4">cutlass::layout::TensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCongruous_3_0132_00_01Crosswise_01_4.html#aafa9812d3c82b30b2d416738c586bfe0">cutlass::layout::TensorOpMultiplicandCongruous&lt; 32, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCrosswise.html#a1a69654a56446c0271037c5df7fbfc86">cutlass::layout::TensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandRowMajorInterleaved.html#acb62895338493de3e37649e46b1c8c05">cutlass::layout::TensorOpMultiplicandRowMajorInterleaved&lt; ElementSize, InterleavedK &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandBCongruous.html#a9e54d47e637060ba318d04416d35f4f4">cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandCongruous.html#a0f6dfb913a9f6d1dd6a6e4d83e3c87ec">cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandCrosswise.html#aa27ee17b10d393e8578afdbdedaa190e">cutlass::layout::VoltaTensorOpMultiplicandCrosswise&lt; ElementSize, KBlock &gt;</a>
</li>
<li>PackedVectorLayout()
: <a class="el" href="classcutlass_1_1layout_1_1PackedVectorLayout.html#a8a6ead8c5b4b2a9e22d1e6ae779ff038">cutlass::layout::PackedVectorLayout</a>
</li>
<li>Padding
: <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1DefaultEpilogueComplexTensorOp.html#a6685913f5820773e1b26ea2eb76866cd">cutlass::epilogue::threadblock::DefaultEpilogueComplexTensorOp&lt; Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess &gt;</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1DefaultEpilogueSimt.html#afa279c7d787f83ecb8c0c04de6b07bc6">cutlass::epilogue::threadblock::DefaultEpilogueSimt&lt; Shape_, WarpMmaSimt_, OutputOp_, ElementsPerAccess &gt;</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1DefaultEpilogueTensorOp.html#a101568b1a0371b99f0f0b02d7f203812">cutlass::epilogue::threadblock::DefaultEpilogueTensorOp&lt; Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess &gt;</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1DefaultEpilogueVoltaTensorOp.html#a70fe4440f09f6428bc57a4da0017f454">cutlass::epilogue::threadblock::DefaultEpilogueVoltaTensorOp&lt; Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess &gt;</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1DefaultEpilogueWmmaTensorOp.html#a97250de320f3d20e1eff7ee6c633a28d">cutlass::epilogue::threadblock::DefaultEpilogueWmmaTensorOp&lt; Shape_, WarpMmaTensorOp_, PartitionsK, OutputOp_, ElementsPerAccess &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1Epilogue.html#a36970da339d478df9807c01bd26fb87a">cutlass::epilogue::threadblock::Epilogue&lt; Shape_, WarpMmaOperator_, PartitionsK, OutputTileIterator_, AccumulatorFragmentIterator_, WarpTileIterator_, SharedLoadIterator_, OutputOp_, Padding_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1EpilogueBase.html#a18d185fd1a896120f0ceb22c83758635">cutlass::epilogue::threadblock::EpilogueBase&lt; Shape_, WarpMmaOperator_, PartitionsK, AccumulatorFragmentIterator_, WarpTileIterator_, Padding_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1TileIteratorSimt_3_01WarpShape___00_01Operator___00_01Elemenf2bd262ed3e202b25d5802d83965bf3b.html#aa56eb3d1c6b3aea627b8ee024be0e451">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#a96dd094804882c2103e1a457632cf182">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#a3f3dc7225c2cb2f44e5257ad4b3d8b31">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#a516c1e4268bc2629c3539a995963ffe4">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#aab81005bb67d46819c7d9c2571295876">cutlass::epilogue::warp::TileIteratorWmmaTensorOp&lt; WarpShape_, OperatorShape_, OperatorFragment_, layout::RowMajor &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>Params()
: <a class="el" href="structcutlass_1_1epilogue_1_1EpilogueWorkspace_1_1Params.html#a1d05c8ac7337fa0437a7870e024b58e3">cutlass::epilogue::EpilogueWorkspace&lt; Shape_, WarpCount, FragmentC_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1Convert_1_1Params.html#a5a6c0959f8481505164e6f4748a77dbe">cutlass::epilogue::thread::Convert&lt; ElementOutput_, Count, ElementAccumulator_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombination_1_1Params.html#ad7ac223e67928ea2aa44b64602089e81">cutlass::epilogue::thread::LinearCombination&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationClamp_1_1Params.html#ab19de8ec29f12815e5d558ad3a32ba66">cutlass::epilogue::thread::LinearCombinationClamp&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_1_1Params.html#aab0a1bf972d12ad08acec38f10702bf0">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_3_01ElementOutput___00_01Count_00_00274a94522c46cd041d0b10d484e2ef3.html#ab4301431271a1a57a51904b8c00584b2">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, int, float, Round &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1DirectEpilogueTensorOp_1_1Params.html#a7890b7910f0b680ff5b9230d91efc2ef">cutlass::epilogue::threadblock::DirectEpilogueTensorOp&lt; Shape_, Operator_, PartitionsK, Element_, OutputOp_, ConvertOp_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1InterleavedPredicatedTileIterator_1_1Params.html#af64fa3173c9790da33060e7fe7574d7b">cutlass::epilogue::threadblock::InterleavedPredicatedTileIterator&lt; ThreadMap_, Element_, InterleavedK &gt;::Params</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1PredicatedTileIterator_1_1Params.html#af884dcf4ef98ad19a5e9e5af9dfa3e40">cutlass::epilogue::threadblock::PredicatedTileIterator&lt; ThreadMap_, Element_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#af09f4fcf7702d3a6bd4904a379d77e8c">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#a83a5e9c4325affc7e04175d3df448977">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1Params.html#a3699384e74a290c6418ef40350c43655">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<li>params
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#a9c9d8378c735597f39927c6ab32519ef">cutlass::reduction::BatchedReduction&lt; BatchedReductionTraits_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>Params
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#a213c6812458c008435ed3ea710fe2454">cutlass::reduction::BatchedReduction&lt; BatchedReductionTraits_ &gt;</a>
, <a class="el" href="structcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK_1_1Params.html#a7613c14f567f1179108896db24f61901">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;::Params</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>params
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_1thread_1_1ReduceAdd.html#ac3ba3575c91e948c1f622d068a181428">cutlass::reduction::thread::ReduceAdd&lt; ElementAccumulator_, Element_, Count &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomGaussianFunc.html#acd40b7369356ac0ad4e83db8742677a5">cutlass::reference::device::detail::RandomGaussianFunc&lt; Element &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>Params()
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomGaussianFunc_1_1Params.html#a18c7d6191e9334ecc64abeccc418fa42">cutlass::reference::device::detail::RandomGaussianFunc&lt; Element &gt;::Params</a>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
</li>
<li>params
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_1detail_1_1RandomUniformFunc.html#a0306f6102c710d11428e5fdbbc2d3fc6">cutlass::reference::device::detail::RandomUniformFunc&lt; Element &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>Params()
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_1detail_1_1RandomUniformFunc_1_1Params.html#a903b050c7855e8b1dbd70ab5b201db46">cutlass::reference::device::detail::RandomUniformFunc&lt; Element &gt;::Params</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>params
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_1detail_1_1TensorCopyDiagonalInFunc.html#a74b866ebefe84dd33f31977f189adebe">cutlass::reference::device::detail::TensorCopyDiagonalInFunc&lt; Element, Layout &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>Params()
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_1detail_1_1TensorCopyDiagonalInFunc_1_1Params.html#a24b17f2db455bfb0d86f6534c6850766">cutlass::reference::device::detail::TensorCopyDiagonalInFunc&lt; Element, Layout &gt;::Params</a>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
</li>
<li>params
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_1detail_1_1TensorCopyDiagonalOutFunc.html#a68d828562ead4350eed8ad3901ba1237">cutlass::reference::device::detail::TensorCopyDiagonalOutFunc&lt; Element, Layout &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>Params()
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorCopyDiagonalOutFunc_1_1Params.html#a32775ff6e9303eac6dd60b3ef8bedcde">cutlass::reference::device::detail::TensorCopyDiagonalOutFunc&lt; Element, Layout &gt;::Params</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>params
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_1detail_1_1TensorFillDiagonalFunc.html#a8e9cb15084811d890b00124378ee2660">cutlass::reference::device::detail::TensorFillDiagonalFunc&lt; Element, Layout &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>Params()
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_1detail_1_1TensorFillDiagonalFunc_1_1Params.html#a9e5512d91acbfdcf4bc74d029b7a93e7">cutlass::reference::device::detail::TensorFillDiagonalFunc&lt; Element, Layout &gt;::Params</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>params
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_1detail_1_1TensorFillLinearFunc.html#a062cc0662f2c4f00715889679141143f">cutlass::reference::device::detail::TensorFillLinearFunc&lt; Element, Layout &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>Params()
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_1detail_1_1TensorFillLinearFunc_1_1Params.html#a2c6db4c2b5fc5d61b1568a4a1ea60915">cutlass::reference::device::detail::TensorFillLinearFunc&lt; Element, Layout &gt;::Params</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>params
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomGaussianFunc.html#a7310921bd7e3f168f2d89ad5a459a95a">cutlass::reference::device::detail::TensorFillRandomGaussianFunc&lt; Element, Layout &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>Params()
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomGaussianFunc_1_1Params.html#a3021a3a4f0da196005d8f17397c3d7c8">cutlass::reference::device::detail::TensorFillRandomGaussianFunc&lt; Element, Layout &gt;::Params</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>params
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomUniformFunc.html#a54451c0609b552e9775c5ad2680d89c3">cutlass::reference::device::detail::TensorFillRandomUniformFunc&lt; Element, Layout &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>Params()
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomUniformFunc_1_1Params.html#a57398f088e1f1d96c731d4778497d608">cutlass::reference::device::detail::TensorFillRandomUniformFunc&lt; Element, Layout &gt;::Params</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>params
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorUpdateDiagonalFunc.html#a529415be152f110de60f66ce52c2709d">cutlass::reference::device::detail::TensorUpdateDiagonalFunc&lt; Element, Layout &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>Params()
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorUpdateDiagonalFunc_1_1Params.html#a6edd7198bf1d58e6e34cc40d4c4f184d">cutlass::reference::device::detail::TensorUpdateDiagonalFunc&lt; Element, Layout &gt;::Params</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>params
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorUpdateOffDiagonalFunc.html#a0ad8679159037d6cd2f665af29e33d37">cutlass::reference::device::detail::TensorUpdateOffDiagonalFunc&lt; Element, Layout &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>Params()
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorUpdateOffDiagonalFunc_1_1Params.html#a54230167d62dee10a2fffda235a01fe1">cutlass::reference::device::detail::TensorUpdateOffDiagonalFunc&lt; Element, Layout &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__18e9cf25bb3b8edfaad595241a6dc2d7.html#a045183b00ed8b72bd6b43e6f66915475">cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__8ccc62d47a092afc8bee32ffe9d1e4ba.html#a65b4c9b38fb709f3b86944e17517f194">cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__a56cbccec33ee916292ad9d068474609.html#ac05e8dea32490f8850b507e3e1b6b081">cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemenc07b5ec72f83e782121ac629288d61fe.html#ae8306adee1dd43c641abc176f49ba22f">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemena9b06926a275b569ee9f7f142604b997.html#abfe6c715cc340cefc036f2071895780d">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen41e459f664d17473570cf22fb616845f.html#a134195552cfb4a327133e7f0e53f0d9a">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen44ce348364e78f5a56fa0c2cef6af930.html#aeecf6872696baab51189c956825e8b4f">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen058417e2cdd86f3cd6ad5458581571c8.html#a4dacbac0f3e525860d753031e3801a59">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_01e11ed7192af5d7ad1bce5641fa13112e.html#af4d9c8e63b88ca4a42c87a86d1cb9abc">cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Transpose_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_0145ef045e8f7d57dc718098adcb00cf3d.html#ad7de484ae6fd7e295771003ff2dc155f">cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Transpose_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_0102e766863c6ac9ec2063a02c4803eecb.html#ac9762c9d5012de584dfac064d9928d00">cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Transpose_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___00a6b756b1bcfbb35fe4a3e68ff074e380.html#a234151421c93148ed80209c99415bf7f">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, AccessSize &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___00ebd1a63351e1085d0b718582ec7b06c8.html#ae36baf85b0cc499418c5c0c6622e2b8a">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::ColumnMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessSize &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___006a5f2f7a8271031e6cdc5daa5441f2af.html#a45e7ac8084883fdf4d84d71afddf45d6">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessSize &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___004d0f9b5e19c29acc17bcdc360dafebbd.html#a40aa9cc4c64cc49e926ce791326f70bb">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessSize &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___009fd89f6dad84238fd7d63df0a0c0364f.html#a565c30dda08950f30fa66b6ca1d0bae1">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessSize &gt;::Params</a>
</li>
<li>params_A
: <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#a93f15acc09f27c23dc5a213d63359b5c">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#a5db93e5cd101892903cf66c7e373e4a0">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1Params.html#a22986b223a6aac9b0d742d359b12e36a">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
</li>
<li>params_B
: <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#a910309fbd40055ab81150b055407f5cc">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#ae95fc48d5d43cb3061e21b697a5a01b4">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1Params.html#ab2ad41c17aee55f453946a81f2d87fa3">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
</li>
<li>params_C
: <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#a78dd936eb07a5415c93d1841a0fc7ff3">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#abd68830b012c0cb7de5f883d9fbc316c">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
</li>
<li>params_D
: <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#a8eb01bbf1b150e2779ecc05de9155f38">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#a0a12100dd4dc325a550a50c4b8ec92f5">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1Params.html#aee1eb37b66934e28f6e2c633389bc71c">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
</li>
<li>parsed_argc()
: <a class="el" href="structcutlass_1_1CommandLine.html#a228e1a273d223eec4b2f6d73135d3c1e">cutlass::CommandLine</a>
</li>
<li>partition_stride
: <a class="el" href="structcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK_1_1Params.html#a10fb9f2ac4dc43b02aeb0714ab4ba889">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;::Params</a>
</li>
<li>PartitionCount
: <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorTensorOpMultiplicandCongruous.html#abd26edad15a0ce1c3a24d6a9c96f66a3">cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorTensorOpMultiplicandCrosswise.html#afe8f9e93641b51a0172a0a724e6cfb9c">cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandBCongruous.html#a50beb822a56d11fe7d7511b2514d9eeb">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandCongruous.html#aa9095d999b45d7e37dbeaa102112696a">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCongruous.html#a6db9483dd8c793e687445514fd00124f">cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCrosswise.html#ac0f9fe7e728edb0eff201fd1c356db4a">cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandBCongruous.html#ac8c8479bd68f8ff26281ccf20257c416">cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandCongruous.html#aab81368102d96f090a143aabf4f72595">cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicand.html#a439376ce6d4333d65d971c0012674931">cutlass::layout::TensorOpMultiplicand&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCongruous.html#a9c818862df1951d9e8ebb31165b61fb5">cutlass::layout::TensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCongruous_3_0132_00_01Crosswise_01_4.html#a10feb79f61f6dec862da9541baa37425">cutlass::layout::TensorOpMultiplicandCongruous&lt; 32, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCrosswise.html#a3ff5e9dcf1e98e074b40a2e06fa56df0">cutlass::layout::TensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandBCongruous.html#a0e5c5d56b6dc1daf9268b32ec18c44b0">cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandCongruous.html#a13f44ba4804056f56d5166a2f0403377">cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
</li>
<li>partitions
: <a class="el" href="structcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK_1_1Params.html#a355a2740ee735de6616705c523b68fdd">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;::Params</a>
</li>
<li>PartitionShape
: <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorTensorOpMultiplicandCongruous.html#ade26abdd740cb37ed4ef0383a0096c01">cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorTensorOpMultiplicandCrosswise.html#a8de62d289c4d9e3123736e42b4c3b33d">cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandBCongruous.html#a4606210cdd1342bd1dc17ab37c4ef133">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandCongruous.html#acc538c9d418a299ce3cffb8f914af15e">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCongruous.html#aef568e91d5233b70fffe5315633bcc0d">cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCrosswise.html#abdbf2109359d6c902676eb3519484ab2">cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandBCongruous.html#ac2ea9ce6186f5940fbfbb9d8528c450d">cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandCongruous.html#af3c9f1bd2d159857671da73de894d6c9">cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicand.html#a46f64537deddaaedfaca7b5ae3cc3e6e">cutlass::layout::TensorOpMultiplicand&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCongruous.html#a33b3e232203ea774f1bfd568f710f36d">cutlass::layout::TensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCongruous_3_0132_00_01Crosswise_01_4.html#ae1dfd7c1955567ade7e136b779fda2e7">cutlass::layout::TensorOpMultiplicandCongruous&lt; 32, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCrosswise.html#a8ef78b952d245778ef8df144df21de2e">cutlass::layout::TensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandBCongruous.html#ae54ee072f8405ef10574d8898489b543">cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandCongruous.html#ab823a7af93e0b440aaaef19f40d11500">cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
</li>
<li>PartitionsK
: <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShab94a11a77dd0565102710907089acee0.html#aa33ac47916bd5e9458989915426c8b3b">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#ab949799d5ae5e367142e0c4370241fc6">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#ae2344ae8e3350817d9a2275936685165">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#a3f408abe089b8476d19f809c976992d7">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#a31f2cf101f49b119db8e04da688144c1">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#a521119adcd3eddf50876448c506a5e38">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#ae22e247d96d9d8e2e49ae69fce1002c3">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#ab81183f42ddd0474c45bd4ecfd2f91cb">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>pi
: <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomGaussianFunc.html#a6b96ec27d0f7df4c4a1bd451b29a69fe">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#a61f603633246ab86c8f46e6cbe0f257c">cutlass::reference::host::detail::RandomGaussianFunc&lt; complex&lt; Element &gt; &gt;</a>
</li>
<li>PitchLinear()
: <a class="el" href="classcutlass_1_1layout_1_1PitchLinear.html#a2c8c3651e1cf3282e623451a8d60f42b">cutlass::layout::PitchLinear</a>
</li>
<li>PitchLinearCoord()
: <a class="el" href="structcutlass_1_1layout_1_1PitchLinearCoord.html#ada179709a13a11ca3e60b21b2fc69ed9">cutlass::layout::PitchLinearCoord</a>
</li>
<li>PitchLinearThreadMap
: <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Eleb60d066756d1c18f05fceee6a27bdb8a.html#a48683c4e40689d85b21cb1ee6caafe2d">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::ColumnMajorInterleaved&lt; 4 &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>pointer
: <a class="el" href="structcutlass_1_1AlignedBuffer.html#a1c0b77fef16a9f3d7007817a9fc32bf1">cutlass::AlignedBuffer&lt; T, N, Align &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4.html#a2a77712281a0ddbf880a4f6fb9aa2ea3">cutlass::Array&lt; T, N, false &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#a949beb7b21ad69d3a3bc394235dd8ec0">cutlass::Array&lt; T, N, true &gt;</a>
, <a class="el" href="classcutlass_1_1platform_1_1unique__ptr.html#ab6ce60d03d11b269c1e151dfa7c696f9">cutlass::platform::unique_ptr&lt; T, Deleter &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>Pointer
: <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__da632779aba661c0f4cfaaa78126b771.html#a050471ccc537ccb537d272c94623172e">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#ac3f24a8e7b61140ddb822186d9617986">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#ad6092b73dd2ef54f61c2c1baa911f9c5">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#a11fa9c51781866db17b2028864c2cdc8">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#aff02d8d269168d0bcc6ab6c984afd42b">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#a72f33398ccf4d0d5579ff9db34845adb">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#a48f202a1c8357b78073828b4c1101ad3">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#a9940278e9c3fa63485be0ca911972fbf">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#abbc6f561480d48c05001038ab7a0bebb">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#aeae69750c7c478c23653af6e785892af">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#a9ecc09cd70be9f7e56f309465d878f79">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#a77ec03286075f26cb5b3a020e3156745">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#a5761c878c096fe769e76b5dbb951d1f5">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#af79b4a0ff3191a3dbdd01afd164fbc68">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#a0f1b733f3d1da13bd967b6eb76f04ae6">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#a838b764b815761c22bb879e4ceebccf9">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessSize &gt;</a>
</li>
<li>pointer_mode
: <a class="el" href="structcutlass_1_1library_1_1GemmArguments.html#ab20c7493f073047a254c5e14996067db">cutlass::library::GemmArguments</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmArrayArguments.html#ac11dab7bcbce01b1f9a79aeff1175763">cutlass::library::GemmArrayArguments</a>
</li>
<li>Policy
: <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorComplexTensorOp_3_01WarpShape___00_01Operato8cf03c624cf3210c71b7cbd580b080f8.html#a9230b73de63fcf673cc1e9125113d9c4">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#a8e8487fee1e71fe537e5927143b92ebc">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#af682ef0a9152d8f1c34e0d14ee084e0d">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#a008247244142e3fadbd130bb9c268e24">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#ae776e5f6feac3733565fb81c52725f2c">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#a64367042cbb3b202b98f8b02a07b2de2">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#af16566eb4e56d790c875c842f592fd4b">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#ad70c4e067ea653db52e72cd07a918cc7">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#aa43845436513f3eec39906a41a275953">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#ad011e908ce60c548082a985a2c896f39">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#ac25c73990d9bb57ccaaa536b5569490f">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#ae1270cd0109c82ba4f315e3b3de6d1c4">cutlass::epilogue::warp::TileIteratorWmmaTensorOp&lt; WarpShape_, OperatorShape_, OperatorFragment_, layout::RowMajor &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1threadblock_1_1DefaultMmaCore_3_01Shape___00_01WarpShape___00_01GemmShab94a11a77dd0565102710907089acee0.html#a5c0553c4e049d6ebcd87ba251080b77d">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#a5e6932a753e9283f543e87b6a4f7d40b">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#a83650df8cb2ab42b64c369ec37b92871">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#a0eebd627f3649fab55d71ea147f6ec82">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#a346489e42e33dc95e7568793d43c5c64">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#a6d1fb04576af9af3d989c39545d810d4">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#a5b4fe0717b80227cb8af9e705a058c05">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#a76f447f264ee662e71140dab01124398">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#a830a1897d9dfd27e43bb71b4f1bedc35">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#acb4f53137e01145a137fd4eaebb71f67">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#a562f35290b06d22790ce63d44cb79be0">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#a4b74f2c5f0d0a94ca74c16166659f47d">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="classcutlass_1_1gemm_1_1threadblock_1_1MmaBase.html#ad3e8e34907fb1611c876f204195ea85f">cutlass::gemm::threadblock::MmaBase&lt; Shape_, Policy_, Stages, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaPipelined.html#a09e5e5bfa35c1398346c452eb3deeb97">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#aec6d96bf7381ba99a6b13275de1d3fb0">cutlass::gemm::threadblock::MmaSingleStage&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1warp_1_1DefaultMmaTensorOp.html#a9e8b45cc517276a2e5d7f8d2c0839295">cutlass::gemm::warp::DefaultMmaTensorOp&lt; WarpShape_, InstructionShape_, ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, Operator_, PartitionsK, AccumulatorsInRowMajor, PartitionsN &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaComplexTensorOp_3_01Shape___00_01complex_3_01RealElementA_01_146441010dad1f40eb51b6dae3ded216.html#ad3452123ebc1b384b4eef4b6823a10d8">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#a26f5eb70804a9bcaee23d216d0166740">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_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kA_00_01Element_67ca7e11a38e38f2c51b84767654a90f.html#a13de4e93620cabff6b648c40d72fede9">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#a01a2b04fe36321906945e3e245096e13">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#a36852208ae3044a879c7ee002a57f557">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#a6f249c42aff57628bf0b5a1377a736c8">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#ab364c2018c4c79222b58d49a0ca737dd">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#a72832b7b24f40aafd9e21573bf86f8e7">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kC, Element_, layout::RowMajor, Policy_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOp.html#aca53ae2c2f98f6b1e735b332904aa9aa">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#ac7dbce8ff671e1b4d101aa386f379100">cutlass::gemm::warp::MmaVoltaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
</li>
<li>PredicatedTileAccessIterator()
: <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen89c687c583745a73cb485041911a4c4e.html#af0d24b5f67377109966b009a95c1f9fb">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#a28ef2a11ff8aa22e6e3091cf47fdfc04">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_01Elemen41e459f664d17473570cf22fb616845f.html#a5afebae9ece3156030fd38901e730184">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator_3_01Shape___00_01Elemen784a0e9da3f55064c47e5613791f51f7.html#a01973ed26ee44cc778e3913b8b64df1b">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#a89ea0762a8628807ffd46b18d9ea4f91">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#a8af4e1b2338f22dea60c857e13377e6c">cutlass::transform::threadblock::PredicatedTileAccessIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
</li>
<li>PredicatedTileAccessIterator2dThreadTile()
: <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__da632779aba661c0f4cfaaa78126b771.html#af1b744fd89da7f5c305ad26fc042c17b">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__8ccc62d47a092afc8bee32ffe9d1e4ba.html#a350269fda2d7e4e8025bfebdc83921de">cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessType_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__1790abaa54a01f277d75766d5882fec8.html#a10500645804b2cb6a0e537f03b57f384">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#ae2cf64f40d954803ad851128c785dc07">cutlass::transform::threadblock::PredicatedTileAccessIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, AccessType_ &gt;</a>
</li>
<li>PredicatedTileIterator()
: <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1PredicatedTileIterator.html#a0b73a5e03549ccf747bf5bf3c07f6f27">cutlass::epilogue::threadblock::PredicatedTileIterator&lt; ThreadMap_, Element_ &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___0068b3e874b5d93d11f0fa902c7f1d11d9.html#a79a6680479608730781b18d8f9ef86fb">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#a9bf65422579611ecaf4275516b6d9e1f">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___006a5f2f7a8271031e6cdc5daa5441f2af.html#ae0f61aca69c0d05ec21684cd08d040fc">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, AccessSize &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator_3_01Shape___00_01Element___00e7c2c404e7aedfe60ad56bb5571306a1.html#a562259e56e4981326c16c96b00b67987">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#ab77e9c9d9d7337dea6c27ee24f09ea86">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#adc98a00c097c5b93755b182ea493d2c6">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessSize &gt;</a>
</li>
<li>PredicatedTileIterator2dThreadTile()
: <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_0165b39a630d10785a3558406f9adb99b9.html#a26211dc2b48ab669363e9af8738f1e13">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_0145ef045e8f7d57dc718098adcb00cf3d.html#a3f0c8d2081d8d23cd6cb81627a77390a">cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Transpose_ &gt;::Params</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileIterator2dThreadTile_3_01Shape___00_017a517f3c73efd795ab05059cc9b111e1.html#a344627cd2525bf07e4c631d2cc3ca1d9">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#a206d1ee28d98522f4e6577c41cce5ddb">cutlass::transform::threadblock::PredicatedTileIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Transpose_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>predicates
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_1epilogue_1_1threadblock_1_1InterleavedPredicatedTileIterator_1_1Mask.html#a7491a28ffa24251ca9b1999392c443d2">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#a1a03001c853066670b8533eba6866b52">cutlass::epilogue::threadblock::PredicatedTileIterator&lt; ThreadMap_, Element_ &gt;::Mask</a>
</li>
<li>PredicateVector()
: <a class="el" href="structcutlass_1_1PredicateVector.html#aec1201df19c0ed0516810a3f19353c21">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>problem_size
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_1device_1_1Gemm_1_1Arguments.html#ad68cd06dadc163a13f5ed29e07d6535b">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_3_01ElementA___00_01LayoutA___00_01ElementB___00_01Layou1b211cc9c97c022d8fe10f2dd32c8709.html#acd02e86dfff866eade08415e0043ccc3">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_1_1Arguments.html#aab4fc258e38ebcf9b430a5dee6daba5e">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_3_01ElementA___00_01LayoutA___00_01ElementB___00_213d78696663f4231cd52c6a277c60e5.html#ad0469cc3e961d21e212d026bccf6fe1a">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_1_1Arguments.html#a0323a91544c8b67fe9c4c2c0c40d75cd">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_3_01ElementA___00_01LayoutA___00_01ElementB___00_a3923967cafb5cb9774c320dc24baa77.html#a29159f430d4a733ec3fac550d0458e18">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_1_1Arguments.html#a24cbcd47c87175248ed1c55a9a0e5426">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_3_01ElementA___00_01LayoutA___00_01Elementafcb1aeaf2035a7ac769d7acc233423b.html#adee4f1a66aa6b6cb0400f6159ec52eb9">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#a8ee835b21f77e387ea0ebff58f9b0135">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#a05909ba49e633c7eeb0707166c72a4ee">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1Params.html#a1594128a193160fef9a01185360d54bd">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmArrayConfiguration.html#a823157e106610cd0255f034c999fa202">cutlass::library::GemmArrayConfiguration</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmBatchedConfiguration.html#ab9de6786f3fecf882048f461dc793d40">cutlass::library::GemmBatchedConfiguration</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmConfiguration.html#a831da43ec2dfbce23420411aeec0cad5">cutlass::library::GemmConfiguration</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmPlanarComplexBatchedConfiguration.html#a27594f0e372c6ada5076cbfa75fb5176">cutlass::library::GemmPlanarComplexBatchedConfiguration</a>
, <a class="el" href="structcutlass_1_1library_1_1GemmPlanarComplexConfiguration.html#ab40101e20f043ae5bc1eb6fc392169fc">cutlass::library::GemmPlanarComplexConfiguration</a>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
, <a class="el" href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#adbd0cf20c4f7033016d1b8fdaca6aeae">cutlass::reduction::BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt;::Params</a>
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_1kernel_1_1ReduceSplitK_1_1Params.html#a4e71c9fe8dab59b795bd7a4a2d33cf0c">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;::Params</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>product()
: <a class="el" href="structcutlass_1_1Coord.html#ad5a2fb5b6b57e0726624c2b6e7c6545c">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &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>ptr
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorCopyDiagonalInFunc_1_1Params.html#a088d18e084a3bd3c60ef12069b70b03c">cutlass::reference::device::detail::TensorCopyDiagonalInFunc&lt; Element, Layout &gt;::Params</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorCopyDiagonalOutFunc_1_1Params.html#aec7f77c57f4eaa7afc539b92f1016646">cutlass::reference::device::detail::TensorCopyDiagonalOutFunc&lt; Element, Layout &gt;::Params</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>ptr_C
: <a class="el" href="structcutlass_1_1epilogue_1_1EpilogueWorkspace_1_1Params.html#a11183b5f5352f7c38442be03152f0210">cutlass::epilogue::EpilogueWorkspace&lt; Shape_, WarpCount, FragmentC_ &gt;::Params</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</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