squall/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
a8f2c80db0
cutlass/docs/functions_func_o.html

778 lines
146 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 - Functions</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><a href="functions.html"><span>All</span></a></li>
<li class="current"><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_func.html#index_a"><span>a</span></a></li>
<li><a href="functions_func_b.html#index_b"><span>b</span></a></li>
<li><a href="functions_func_c.html#index_c"><span>c</span></a></li>
<li><a href="functions_func_d.html#index_d"><span>d</span></a></li>
<li><a href="functions_func_e.html#index_e"><span>e</span></a></li>
<li><a href="functions_func_f.html#index_f"><span>f</span></a></li>
<li><a href="functions_func_g.html#index_g"><span>g</span></a></li>
<li><a href="functions_func_h.html#index_h"><span>h</span></a></li>
<li><a href="functions_func_i.html#index_i"><span>i</span></a></li>
<li><a href="functions_func_k.html#index_k"><span>k</span></a></li>
<li><a href="functions_func_l.html#index_l"><span>l</span></a></li>
<li><a href="functions_func_m.html#index_m"><span>m</span></a></li>
<li><a href="functions_func_n.html#index_n"><span>n</span></a></li>
<li class="current"><a href="functions_func_o.html#index_o"><span>o</span></a></li>
<li><a href="functions_func_p.html#index_p"><span>p</span></a></li>
<li><a href="functions_func_q.html#index_q"><span>q</span></a></li>
<li><a href="functions_func_r.html#index_r"><span>r</span></a></li>
<li><a href="functions_func_s.html#index_s"><span>s</span></a></li>
<li><a href="functions_func_t.html#index_t"><span>t</span></a></li>
<li><a href="functions_func_u.html#index_u"><span>u</span></a></li>
<li><a href="functions_func_v.html#index_v"><span>v</span></a></li>
<li><a href="functions_func_w.html#index_w"><span>w</span></a></li>
<li><a href="functions_func_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">
&#160;
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_o"></a>- o -</h3><ul>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<li>offset()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1HostTensor.html#ac3e08629bab5304f999fb5115453a714">cutlass::HostTensor&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorRef.html#a4166ac2a0754574ac21d5d57d74f34e5">cutlass::TensorRef&lt; Element_, Layout_ &gt;</a>
</li>
<li>operand_A_ref()
: <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaBase_1_1SharedStorage.html#adf0d8d1da3ff1fc83e45ba437d68bf8e">cutlass::gemm::threadblock::MmaBase&lt; Shape_, Policy_, Stages, Enable &gt;::SharedStorage</a>
</li>
<li>operand_B_ref()
: <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaBase_1_1SharedStorage.html#a7b0fc26d9dd92732884cde342afcc131">cutlass::gemm::threadblock::MmaBase&lt; Shape_, Policy_, Stages, Enable &gt;::SharedStorage</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>OperationDescription()
: <a class="el" href="structcutlass_1_1library_1_1OperationDescription.html#addc07f7fc02839f928a071840d3c5835">cutlass::library::OperationDescription</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>operations()
: <a class="el" href="classcutlass_1_1library_1_1Manifest.html#a379f07f92b5ed10fa299852e830e9209">cutlass::library::Manifest</a>
</li>
<li>operator B *()
: <a class="el" href="structcutlass_1_1platform_1_1is__base__of__helper_1_1dummy.html#a8185b54df44dd6bed772eb4678f4e246">cutlass::platform::is_base_of_helper&lt; BaseT, DerivedT &gt;::dummy&lt; B, D &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator bool()
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_1Coord.html#a88096d051dd05111cf265a011a89f7f6">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1half__t.html#aa24f262b2f41c17f2a118c890beb39f5">cutlass::half_t</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="classcutlass_1_1platform_1_1unique__ptr.html#a5791650488ae864f10ad04bec4a31005">cutlass::platform::unique_ptr&lt; T, Deleter &gt;</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_1reference_1_1host_1_1detail_1_1TensorContainsFunc.html#a7766766e9565c6ece73c89baa9ae1d70">cutlass::reference::host::detail::TensorContainsFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorEqualsFunc.html#a3e00265cbfaf8471fdb1adb2ba590572">cutlass::reference::host::detail::TensorEqualsFunc&lt; Element, Layout &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>operator cuDoubleComplex()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#a4b9b799c37101651ce7b369af96a24b8">cutlass::complex&lt; T &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>operator cuFloatComplex()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#a20853f692a65e361e0f06b8af3a4a394">cutlass::complex&lt; T &gt;</a>
</li>
<li>operator D *()
: <a class="el" href="structcutlass_1_1platform_1_1is__base__of__helper_1_1dummy.html#a53c3a19beaa23e025b7e1e98e83c224a">cutlass::platform::is_base_of_helper&lt; BaseT, DerivedT &gt;::dummy&lt; B, D &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>operator double()
: <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#a5e735e155826e5c2a0e33b0a456cc5ab">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1half__t.html#af5c8d2a225114cc0bc877464ab934754">cutlass::half_t</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#a7b2d1ddc31b0e9a8f128c775e99b1ba9">cutlass::SubbyteReference&lt; Element_, Storage_ &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>operator Element()
: <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#a908a341f883db06f529536fae196a7c9">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#a3989d8d0402257fa1fa723bc730a8219">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
</li>
<li>operator float()
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1const__reference.html#afa022bf34a7086c43b5bd45b40c2b25f">cutlass::Array&lt; T, N, false &gt;::const_reference</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1reference.html#ac800b24861e676e21c7d6201338175bc">cutlass::Array&lt; T, N, false &gt;::reference</a>
, <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#ae794d96fa19581472489bece95fcf344">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1half__t.html#a46ed3fd0ec19acfd5a5019128dd83557">cutlass::half_t</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#a31c36b7beb7e70079e9264be19c74544">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
</li>
<li>operator int()
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1const__reference.html#a61648afeb4e15881fb001611c37df1ec">cutlass::Array&lt; T, N, false &gt;::const_reference</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1reference.html#ad50d57afc8e33ce406f31dc15564cb3a">cutlass::Array&lt; T, N, false &gt;::reference</a>
, <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#abe87d46c924861d6a9e2b06d2d4d69cc">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1half__t.html#a9d051d2e20f89d66a9ed4f9315bf51bd">cutlass::half_t</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#aa7cf6337b97cf7d07ee2f7efef1c6c82">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
</li>
<li>operator int64_t()
: <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#ad47d3da6dca44d5a8c821f63ca37ded7">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#aac7fd7d470166f68da0a882c3f965e5c">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
</li>
<li>operator RowMajor()
: <a class="el" href="classcutlass_1_1layout_1_1TensorNHWC.html#ae700570f695693a3f5e344366cf1ae65">cutlass::layout::TensorNHWC</a>
</li>
<li>operator T()
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1const__reference.html#a7c5f7d59a22d89a7dd5c923d9bcebd97">cutlass::Array&lt; T, N, false &gt;::const_reference</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1reference.html#ac51b14ff76b80e7f7bf4142a1af01d82">cutlass::Array&lt; T, N, false &gt;::reference</a>
, <a class="el" href="structcutlass_1_1integer__subbyte.html#a69962ee56f9192d2dd7d30c34efebb01">cutlass::integer_subbyte&lt; Bits, Signed &gt;</a>
</li>
<li>operator uint64_t()
: <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#abd60b998269ea5f771d2f51ed736d3c1">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#a31b1157e88da26c11944ea8aa0d33a14">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</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>operator value_type()
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_1platform_1_1integral__constant.html#adfcb0c650152fe171f00c7b817c69ae1">cutlass::platform::integral_constant&lt; value_t, V &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
Updated doxygen
2018-09-20 05:02:08 +08:00
<li>operator!()
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_1Coord.html#aa733c6fae0da553053530cba2dddcaa0">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</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>operator!=()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1const__iterator.html#ad8a6c87e370a53e7ff783ee4ad3d1198">cutlass::Array&lt; T, N, false &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1iterator.html#ae72e0d7919ac6d40e1d4f8ce5458af1e">cutlass::Array&lt; T, N, false &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__iterator.html#a6a6e0f4caab421bbe90a94d199df0281">cutlass::Array&lt; T, N, true &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__reverse__iterator.html#aec971b0beb331c02f2b6739bd78037dd">cutlass::Array&lt; T, N, true &gt;::const_reverse_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1iterator.html#a498750b21a735e4ed7cd3c5ff4512497">cutlass::Array&lt; T, N, true &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1reverse__iterator.html#aa3d086a10cf92a20c213c4bedbaa60ff">cutlass::Array&lt; T, N, true &gt;::reverse_iterator</a>
, <a class="el" href="classcutlass_1_1complex.html#adc03f7ab9c6a1c3b61073be3a4384c28">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#a8183b9203a213d4b6381ad7dc120deea">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1integer__subbyte.html#a47732b2b9eee34b308e41c6298058917">cutlass::integer_subbyte&lt; Bits, Signed &gt;</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1ConstIterator.html#a0ed3c88f8229d50b812f3cf151778ad9">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#a95d40d20eb0749e2916d03088d49f680">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</a>
</li>
<li>operator&amp;=()
: <a class="el" href="structcutlass_1_1PredicateVector.html#afe47d16e877fc7e02f2a693b5bd4b6d0">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator()()
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_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01complex_30fa42e1ad201df010637cd22fc070a1.html#af37b41df0067e4c878baafc8d71574d9">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, complex&lt; double &gt;, LayoutA, complex&lt; double &gt;, LayoutB, complex&lt; double &gt;, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01complex_48b3a43bc03fff93a111ac01abe7e40d.html#a8f8180dcd03dad7b45d73f09e87c060b">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, complex&lt; double &gt;, LayoutA, double, LayoutB, complex&lt; double &gt;, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01complex_76f9d24016e1b4167b16f4d7628c9546.html#ae914107ff2d7b1524cc9f8c70237a8f9">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, complex&lt; float &gt;, LayoutA, complex&lt; float &gt;, LayoutB, complex&lt; float &gt;, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01complex_f1c9d2ee842455cd0c5b71d56108d468.html#a3b69fa99a09158d9be274651f6b74980">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, complex&lt; float &gt;, LayoutA, float, LayoutB, complex&lt; float &gt;, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01double_070b94670e040ed5855e5b42d5ca8a443.html#a672562083112f8463a9791113482c4e9">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, double, LayoutA, complex&lt; double &gt;, LayoutB, complex&lt; double &gt;, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01double_0aa57e6a2e6b5da37d10688bf99419a23.html#acc4d8ede06a490f3fbfec8e3dd75b0b1">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, double, LayoutA, double, LayoutB, double, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01ElementAb6e65b2cf5ede7f41cb070a767158dee.html#af8276da36c235c482881864f4ee93c58">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, Operator &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01float_00e3e12e263df6506b8cf06c3f4d478b8e.html#a175c9e89f95837838e533687f4c4078d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, float, LayoutA, complex&lt; float &gt;, LayoutB, complex&lt; float &gt;, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01float_004bb3fd76ca2af7b3210676fa9644d95b.html#a47e6d9cb8a24ae8246272985741f7f0d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, float, LayoutA, float, LayoutB, float, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01half__t_4f30ee91f7bb3844ff7579c68d078818.html#aa1e0584011f1b74dc6a541693d6a2dc2">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, half_t, LayoutA, half_t, LayoutB, float, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_011_01_4_00_011_00_01int_00_00b2dff9ce8caad9aff5bc6a355539161.html#a147ecaa2af6851b26a17eb8f8d95f9d0">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 1 &gt;, 1, int, LayoutA, int, LayoutB, int, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_012_01_4_00_011_00_01int16__t8c4bac365710598317a69c489f7239db.html#acb6436ae8878e9c988517485d02ffc92">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 2 &gt;, 1, int16_t, layout::RowMajor, int16_t, layout::ColumnMajor, int, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_011_00_014_01_4_00_011_00_01int8__t_a1ef6624fc8c10126f17f4ee88283d72.html#a35db4d9db337ee70832b44e5aa76fe23">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 1, 4 &gt;, 1, int8_t, LayoutA, int8_t, LayoutB, int, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_011_00_012_00_011_01_4_00_011_00_01half__t_f3dc2e59f857ada163d1e0781ea8f391.html#a58bfd183b9425b6eebe825192569fd3f">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 1, 2, 1 &gt;, 1, half_t, LayoutA, half_t, LayoutB, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_0116_00_018_00_018_01_4_00_0132_00_01half__02a3f19a78995f97d793a668e0e4d4f0.html#a35ad9fdeb81d431f127606e643d5ffad">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 16, 8, 8 &gt;, 32, half_t, layout::RowMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_0116_00_018_00_018_01_4_00_0132_00_01half__96363097c47b056f0ca1911afd7f8b7a.html#afdbfd1e7162fca08934cfd9d03971b29">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 16, 8, 8 &gt;, 32, half_t, layout::RowMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_012_00_011_00_011_01_4_00_011_00_01half__t_8cf78649807b93684f3d431bfa34ee28.html#a602ba0c72022da14a4def190581ca0de">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 2, 1, 1 &gt;, 1, half_t, LayoutA, half_t, LayoutB, half_t, LayoutC, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_012_00_012_00_011_01_4_00_011_00_01half__t_ccde11d1bbbdab3702772ce44eb9729a.html#a6e7f4b55f118e81ee493bcbf51519e77">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 2, 2, 1 &gt;, 1, half_t, layout::ColumnMajor, half_t, layout::RowMajor, half_t, layout::ColumnMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_012_00_012_00_011_01_4_00_011_00_01half__t_c07cc6439298fa5486a719e577be2538.html#a37ea44d16dbc07ca46888812caf5773c">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 2, 2, 1 &gt;, 1, half_t, layout::ColumnMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_01128_01_4_00_0132_00_01uint15918972b95027764b3a849b03075ed2b.html#ab23a775098428cdf19a6a9a4f03e2506">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 128 &gt;, 32, uint1b_t, layout::RowMajor, uint1b_t, layout::ColumnMajor, int, layout::RowMajor, OpXorPopc &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__927179f46017ea5f58f859f1196c4829.html#a4c2fe667d51ad6740d458ba46b9b4b5d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__8ebae0cbdf333fddfe5c24d35ebe8e02.html#aa4adee271cdf27a7e7c3b1f545e2831d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__5299c9c90c8f2f521be0c8cec1c3eb08.html#a4c58b7ab353251872a0e3c082ee87f15">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01int8__f083347e265b1e9eea5572d86ddb6bf9.html#ad4035659907d79a371ab13da41d306de">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, int8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_a62aa63a212985df306fb27e8a50aeae.html#a659eca51b1e715b849aac3e9e086a096">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_ab741d81fdc991345cb9e43c29fca573.html#aa5df706b43e6be8ef9c1484f004df26a">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_5221708cec5828d35db1d1c47cb4964e.html#aef68bd9a000449c3f84763a706f0614c">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0116_01_4_00_0132_00_01uint8_bef0c048bc0f8ba2d875cb7ab26d363b.html#ab0fd65eba5dfc4927f6034c5b5eda8b6">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 16 &gt;, 32, uint8_t, layout::RowMajor, uint8_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_6e513ccbc44ae7909a60d93b9b5435b3.html#a8137013b1551deb37b4015e2cfd19b97">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_0ee08a4520882d24ba9026879265e892.html#a8ba5f06bdafd276c6db6fe3e1a0d147e">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_4746fc55e614df0016c518d3fda2677e.html#a0d76a6b029d10c796d2923f000eeea29">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01int4b_546e9ec6de6a5970b326da6f6280f1d4.html#ab4e2440c4628b14dffc417970364c88d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, int4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b03e3b50dbcb30d0d1ac062f3a9d5abef.html#a71e4dbe9d9592afdb55b5526db40d93a">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b6d968039dde5c9f062ab15f90a8049fe.html#a3834efe6e24befed0df2868a376af894">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, int4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4bc4b6ba004e25c44bfd9266c61f937dfb.html#a5ee758f7bb9880c5dc69161654b6fd2a">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_0132_01_4_00_0132_00_01uint4b451d5cf5d7e8cbbe476afe3dab5c09b2.html#abab710c2ed9ee3fa031a8f4ad2bac3c2">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 32 &gt;, 32, uint4b_t, layout::RowMajor, uint4b_t, layout::ColumnMajor, int, layout::RowMajor, OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_b0242d7a01097510effbc4718040d3e5.html#af2246459842af9b8c696b06c940ff026">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_c7f88bfd32a544fba8111d2dcadeab11.html#a38a710562c29787bbe64676077668310">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_44a3b2a8df88a2b067f1284515cb5371.html#ae2789576da3917acd47de1cea88f054d">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_4b7308177b308a272c1889fbe9670275.html#aeb017b2c1a20a6e0044d1648a9979919">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::ColumnMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_5a9888862cebd333ecaf11f7262f77d4.html#a4424e01778af2b9f622d207c18f917e3">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::ColumnMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_31defda8ea2b7d855642ffd77da1a411.html#ad07499e28e4e7fae9412be6e29d28965">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::ColumnMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_839a7c8bb938d1661f4611e68f85d8cb.html#a445b51d63c2301ee34f1ac98b9036908">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::RowMajor, float, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1arch_1_1Mma_3_01gemm_1_1GemmShape_3_018_00_018_00_014_01_4_00_018_00_01half__t_73d9802d6b944a5299bc255887db6bbc.html#a8039db821c629b57e638a6f14eddc48b">cutlass::arch::Mma&lt; gemm::GemmShape&lt; 8, 8, 4 &gt;, 8, half_t, layout::RowMajor, half_t, layout::RowMajor, half_t, layout::RowMajor, OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1device__memory_1_1allocation_1_1deleter.html#a21c95062a05aba32c3aa8551458ab4e2">cutlass::device_memory::allocation&lt; T &gt;::deleter</a>
, <a class="el" href="structcutlass_1_1divides.html#a45c044899c466720712fa3a55dab57f1">cutlass::divides&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1divides_3_01Array_3_01half__t_00_01N_01_4_01_4.html#a79459449652aa53d523b043e7098604c">cutlass::divides&lt; Array&lt; half_t, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1divides_3_01Array_3_01T_00_01N_01_4_01_4.html#abe16d29dc2547723268d6a0a204550f4">cutlass::divides&lt; Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1EpilogueWorkspace.html#a40eda72233c29dd344513b21ae94db0e">cutlass::epilogue::EpilogueWorkspace&lt; Shape_, WarpCount, FragmentC_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1Convert.html#afb63b7c6166990357cfe2686ed528909">cutlass::epilogue::thread::Convert&lt; ElementOutput_, Count, ElementAccumulator_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombination.html#a68ce9da2c3697fb84e6c7ecc73e9b6cc">cutlass::epilogue::thread::LinearCombination&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationClamp.html#a22d4448a6fc26e7a028316157292af9e">cutlass::epilogue::thread::LinearCombinationClamp&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu.html#a93c794d586c7d0a6054a24db70490539">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, ElementAccumulator_, ElementCompute_, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1LinearCombinationRelu_3_01ElementOutput___00_01Count_00_01int_00_01float_00_01Round_01_4.html#abb3916aa0408eb67e72809ccd80e03a5">cutlass::epilogue::thread::LinearCombinationRelu&lt; ElementOutput_, Count, int, float, Round &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1ReductionOpPlus.html#a5237e6034d1a3d08cccfddf3a4cd556d">cutlass::epilogue::thread::ReductionOpPlus&lt; Element_, Count &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1DirectEpilogueTensorOp.html#a8a8dd685344c17952aed17bcd9d6b831">cutlass::epilogue::threadblock::DirectEpilogueTensorOp&lt; Shape_, Operator_, PartitionsK, Element_, OutputOp_, ConvertOp_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1Epilogue.html#ac7819d17866af3356d4f7d4b8d4c0c2c">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_1InterleavedEpilogue.html#a152930367cc10492305daeee170f8f45">cutlass::epilogue::threadblock::InterleavedEpilogue&lt; Shape_, WarpMmaOperator_, PartitionsK, OutputTileIterator_, AccumulatorFragmentIterator_, OutputOp_, InterleavedK, IsBetaZero &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1device_1_1Gemm.html#a114b9c14e102d333ef3dcad7865a4efb">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;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1device_1_1Gemm_3_01ElementA___00_01LayoutA___00_01ElementB___00_01Layout4d0960ae6b1d1bf19e6239dbd002249c.html#a384db4125183e504fafc5a946b7ba757">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;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1device_1_1GemmBatched.html#afcdd646be7e79a60bac8dede563c56fa">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1device_1_1GemmBatched_3_01ElementA___00_01LayoutA___00_01ElementB___00_0c9bb6f4463ab6085e6008b5d5ad6abfd.html#a00805989734182945f982cab23a5dca8">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;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1device_1_1GemmComplex.html#a8aed4fea755d53435c829b2ee118b9d9">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1device_1_1GemmComplex_3_01ElementA___00_01LayoutA___00_01ElementB___00_07c56401b4df75709ae636675d9980a9a.html#a375220f643161478c1fb5bcd24f8b5cd">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;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel.html#a5d934b66a5ead9df63bea323abf8f83f">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;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_3_01ElementA___00_01LayoutA___00_01ElementBbe7c1f7154ad5b5bf9d4d28301e2b457.html#a72f5de19ad97e08241157d5106f2f66a">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;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1detail_1_1GemvBatchedStridedEpilogueScaling.html#a822364cf966d0eb32949ebb7f17817d7">cutlass::gemm::kernel::detail::GemvBatchedStridedEpilogueScaling&lt; ElementAlphaBeta, BetaIsZero &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm.html#afc8edf524286b2b3720336f22674a012">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched.html#a9545b33639695004a139b977240ad925">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel.html#a5490540efabcbcc110a6294005c7cf5e">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_72621f7ab9ae4a4ba4fe9725cf8e89c1.html#ae13e7521a412dd15a6438a565f0db467">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::ColumnMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_94c813e3bbfb6f9857c155166f772687.html#a1dcec61fd63296bec70da397ab6f9fc7">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::ColumnMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_17070298bc4cced0a1b98aee2bb6b455.html#a6b501c0685b9b181fa48dba331d5a6ec">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::RowMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1ColumnMajor_00_bf6d29bb09a025e7b96942809743e28a.html#ae321ab37ce6b6db0a3546c1626c746e7">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::ColumnMajor, layout::RowMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l26a133b13650c1d058273e3649f60f04.html#a2349d9640c696096eba961539c79bb3f">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::ColumnMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01lbba3a796be96a0276693ef6b259ecc4a.html#a620a8cd5007a41463672e2dcfdfe4b2e">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::ColumnMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l2aa4d2fd2e940e0d0cf7c47bc8f6017c.html#a1d4b72705fca11cc4f01e477fb6aefe5">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::RowMajor, layout::ColumnMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01layout_1_1RowMajor_00_01l086c058a15d6c79558e4f3d9ff1dc148.html#a882d7a806108fbec527cce945c976e9f">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, layout::RowMajor, layout::RowMajor, layout::RowMajor, true &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01LayoutA_00_01LayoutB_00_0e1104c65871c539155bd3a0c7631928b.html#a82f72b2e8729b28e3f8786ea1ededd28">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, LayoutA, LayoutB, layout::ColumnMajor, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1detail_1_1Mma__HFMA2_3_01Shape_00_01LayoutA_00_01LayoutB_00_07ac147cb320ee0d28ff8e78eb4cd330e.html#ac79f4b4fe20aa32e58ffd1d6daba7557">cutlass::gemm::thread::detail::Mma_HFMA2&lt; Shape, LayoutA, LayoutB, layout::RowMajor, false &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01ElementA___00_01LayoutA___00_01ElementB_e41c1cd6078b6d1347fac239b0639d56.html#ae98fd835ed4750d4f22d7e4e50b5e59f">cutlass::gemm::thread::Mma&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, arch::OpMultiplyAdd, bool &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01half__t_00_01LayoutA_00_01half__t_00_01L066c9d2371712cdf0cac099ca9bcc578.html#a7eb69f25c0b516fda203957a230df3ee">cutlass::gemm::thread::Mma&lt; Shape_, half_t, LayoutA, half_t, LayoutB, half_t, LayoutC, arch::OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01half__t_00_01LayoutA___00_01half__t_00_088f0e99e501b6012297eb30b4e89bcea.html#a72fad6edd8b029407aad12fb22937358">cutlass::gemm::thread::Mma&lt; Shape_, half_t, LayoutA_, half_t, LayoutB_, half_t, layout::RowMajor, arch::OpMultiplyAdd, typename platform::enable_if&lt; detail::EnableMma_Crow_SM60&lt; LayoutA_, LayoutB_ &gt;::value &gt;::type &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01int8__t_00_01layout_1_1ColumnMajor_00_013f3785e722edc6e9aab6f866309b8623.html#ae518a9644f8a0842921d78216b5ac952">cutlass::gemm::thread::Mma&lt; Shape_, int8_t, layout::ColumnMajor, int8_t, layout::RowMajor, int32_t, LayoutC_, arch::OpMultiplyAdd, int8_t &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1Mma_3_01Shape___00_01int8__t_00_01layout_1_1RowMajor_00_01int89c659e7faf47264972bdba6cd80f42b.html#a827c554a04cd668a357b87a20dc8abbe">cutlass::gemm::thread::Mma&lt; Shape_, int8_t, layout::RowMajor, int8_t, layout::ColumnMajor, int32_t, LayoutC_, arch::OpMultiplyAdd, bool &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1thread_1_1MmaGeneric.html#a1131288ac68988bb611e39d6523026f4">cutlass::gemm::thread::MmaGeneric&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Operator_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1Gemv.html#a76265ca41d7fb75089d3839f8fbd509f">cutlass::gemm::threadblock::Gemv&lt; Core_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1threadblock_1_1MmaPipelined.html#a2e45872db16c8ebe01289d7cb74fe3d7">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#a96a18e6ed18ffa66215fabebb5e7a8c6">cutlass::gemm::threadblock::MmaSingleStage&lt; Shape_, IteratorA_, SmemIteratorA_, IteratorB_, SmemIteratorB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaComplexTensorOp_3_01Shape___00_01complex_3_01RealElementA_01_146441010dad1f40eb51b6dae3ded216.html#ab44fc2721d13e50c3cc4396d6adbc03f">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#ad92ee7c01214791d98a473ad0b215b1e">cutlass::gemm::warp::MmaSimt&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, PartitionsK, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOp.html#ac7b9462528c81aa23995760c8c104f47">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#ab06b4c9da74987b1c2322039a2ddf765">cutlass::gemm::warp::MmaVoltaTensorOp&lt; Shape_, ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, Policy_, Enable &gt;</a>
, <a class="el" href="classcutlass_1_1IdentityTensorLayout.html#a471e797514da29a5cad6c61fffe9eb5c">cutlass::IdentityTensorLayout&lt; Rank &gt;</a>
, <a class="el" href="classcutlass_1_1layout_1_1ColumnMajor.html#a28531f0da7b36cfa1f42e84c04b8da04">cutlass::layout::ColumnMajor</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorBlockLinear.html#a849bbda685abc0b855352c429a8f7e08">cutlass::layout::ColumnMajorBlockLinear&lt; BlockRows, BlockColumns &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorInterleaved.html#a3917dbbf6f698bb3267a17f307664e21">cutlass::layout::ColumnMajorInterleaved&lt; Interleave &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorTensorOpMultiplicandCongruous.html#a4707586d7a899943c62fa9aaee2c612e">cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorTensorOpMultiplicandCrosswise.html#a8adbc25064e0edb48c429a59008f6d6a">cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandBCongruous.html#a2f828abd034bcb9d47106bcac109d989">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandCongruous.html#ae7bcec1e4b5e2f8960c6b4ba475df7a0">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ColumnMajorVoltaTensorOpMultiplicandCrosswise.html#a28756bc06eb1cf2567c51ee78f161911">cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; ElementSize, KBlock &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1ContiguousMatrix.html#a1d93f4fd30ebb534de71c0b51bb8a0ef">cutlass::layout::ContiguousMatrix</a>
, <a class="el" href="structcutlass_1_1layout_1_1GeneralMatrix.html#a3d01a8ce1802bd6c6d4edc5095b1fcb8">cutlass::layout::GeneralMatrix</a>
, <a class="el" href="classcutlass_1_1layout_1_1PackedVectorLayout.html#a39cc27d637aa946f1515df04972b1885">cutlass::layout::PackedVectorLayout</a>
, <a class="el" href="classcutlass_1_1layout_1_1PitchLinear.html#a76f5a01eaffaf7e880468efa47de36eb">cutlass::layout::PitchLinear</a>
, <a class="el" href="classcutlass_1_1layout_1_1RowMajor.html#abb932e4b3cb345451a8c6b682d4e7e2d">cutlass::layout::RowMajor</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorBlockLinear.html#a9faa57c538087f23bc1fe4abf0f6822b">cutlass::layout::RowMajorBlockLinear&lt; BlockRows, BlockColumns &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorInterleaved.html#a96a8d6609ee434b21e4a284874582672">cutlass::layout::RowMajorInterleaved&lt; Interleave &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCongruous.html#a7e55b41a76b968ade75d77330471e10e">cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCrosswise.html#a7f702111a44362cf73b81c11a1907b3a">cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandBCongruous.html#abd1a744d04f46d408e2bdce12660ccf2">cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandCongruous.html#af6003c4aee50471c4c09ef74f3996560">cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandCrosswise.html#a28d2430f1d6fc98eb1cccd36fe985099">cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; ElementSize, KBlock &gt;</a>
, <a class="el" href="classcutlass_1_1layout_1_1TensorCxRSKx.html#a1f7f0c8accc79e69057fc4ac487142b4">cutlass::layout::TensorCxRSKx&lt; Interleave &gt;</a>
, <a class="el" href="classcutlass_1_1layout_1_1TensorNCHW.html#adb289aec0ed0a3e5db329210ec55c9be">cutlass::layout::TensorNCHW</a>
, <a class="el" href="classcutlass_1_1layout_1_1TensorNCxHWx.html#aee9b7569e5e31b15d35ca64cfa0bc0d6">cutlass::layout::TensorNCxHWx&lt; Interleave &gt;</a>
, <a class="el" href="classcutlass_1_1layout_1_1TensorNHWC.html#a681b2a1c1a1df6b68b6fb68a18a8392a">cutlass::layout::TensorNHWC</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicand.html#a13c4bd4700ca704d527a4f83f0e58365">cutlass::layout::TensorOpMultiplicand&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandColumnMajorInterleaved.html#ad20058caa6ef3c9f540a33fea56ac0d3">cutlass::layout::TensorOpMultiplicandColumnMajorInterleaved&lt; ElementSize, InterleavedK &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCongruous.html#a3b9ea3abdcd5cc5e4c5b14f7f329bb4f">cutlass::layout::TensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCongruous_3_0132_00_01Crosswise_01_4.html#a3881a944fbeeda82884adfd5853f7577">cutlass::layout::TensorOpMultiplicandCongruous&lt; 32, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandCrosswise.html#ab78912683077ccfdd5c664069de713f1">cutlass::layout::TensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1TensorOpMultiplicandRowMajorInterleaved.html#ac7b0202ef4a01f7fdff37ac678a60d34">cutlass::layout::TensorOpMultiplicandRowMajorInterleaved&lt; ElementSize, InterleavedK &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandBCongruous.html#a89ab20e8b1809e094ffa09ded1042c69">cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandCongruous.html#a7054268bd8313c4e3b55b0b27d662f8f">cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1VoltaTensorOpMultiplicandCrosswise.html#abf01787fadeda14d212bf99efe40e32e">cutlass::layout::VoltaTensorOpMultiplicandCrosswise&lt; ElementSize, KBlock &gt;</a>
, <a class="el" href="structcutlass_1_1maximum.html#a5425d2fc8a8028241332554378e980d7">cutlass::maximum&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1maximum_3_01Array_3_01T_00_01N_01_4_01_4.html#af02b086d2775d769b9675de752a5351c">cutlass::maximum&lt; Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1maximum_3_01float_01_4.html#adf9daaeef5c378bb6036c62c3eed9b3e">cutlass::maximum&lt; float &gt;</a>
, <a class="el" href="structcutlass_1_1minimum.html#a15eb1ef96e29e5230ee59e89beadfa6b">cutlass::minimum&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1minimum_3_01Array_3_01T_00_01N_01_4_01_4.html#a4a45f7d2dad800165887d6bec93b16f1">cutlass::minimum&lt; Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1minimum_3_01float_01_4.html#a13bac6699851e89e63b7b852983d2212">cutlass::minimum&lt; float &gt;</a>
, <a class="el" href="structcutlass_1_1minus.html#aa0db9bb7741aa930c9a52e795d12f227">cutlass::minus&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1minus_3_01Array_3_01half__t_00_01N_01_4_01_4.html#a76637faf0a196f09e6e70318ac574a4b">cutlass::minus&lt; Array&lt; half_t, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1minus_3_01Array_3_01T_00_01N_01_4_01_4.html#ac11e165f6f34a5bc47cdecf6b85db1ab">cutlass::minus&lt; Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1multiplies.html#adff981528415c4c40b251a3401e642ad">cutlass::multiplies&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1multiplies_3_01Array_3_01half__t_00_01N_01_4_01_4.html#a7de36b2383e45e0a4e7349e55321a887">cutlass::multiplies&lt; Array&lt; half_t, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1multiplies_3_01Array_3_01T_00_01N_01_4_01_4.html#a85beeba0468104af0df3c04bebe690d3">cutlass::multiplies&lt; Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1multiply__add.html#a842a468a8d77bda4b789fb9248926655">cutlass::multiply_add&lt; A, B, C &gt;</a>
, <a class="el" href="structcutlass_1_1multiply__add_3_01Array_3_01half__t_00_01N_01_4_00_01Array_3_01half__t_00_01N_01adaeadb27c0e4439444709c0eb30963.html#abc9dd51cad4f2997dae521fad8f5b486">cutlass::multiply_add&lt; Array&lt; half_t, N &gt;, Array&lt; half_t, N &gt;, Array&lt; half_t, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1multiply__add_3_01Array_3_01T_00_01N_01_4_00_01Array_3_01T_00_01N_01_4_00_01Array_3_01T_00_01N_01_4_01_4.html#a070cba31672b435b1edf16e6ded8c955">cutlass::multiply_add&lt; Array&lt; T, N &gt;, Array&lt; T, N &gt;, Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1multiply__add_3_01complex_3_01T_01_4_00_01complex_3_01T_01_4_00_01complex_3_01T_01_4_01_4.html#af46b43d672441fccd96c1aa3b89886fa">cutlass::multiply_add&lt; complex&lt; T &gt;, complex&lt; T &gt;, complex&lt; T &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1multiply__add_3_01complex_3_01T_01_4_00_01T_00_01complex_3_01T_01_4_01_4.html#a28aeeb385a3a47fb72a3d6f421c35294">cutlass::multiply_add&lt; complex&lt; T &gt;, T, complex&lt; T &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1multiply__add_3_01T_00_01complex_3_01T_01_4_00_01complex_3_01T_01_4_01_4.html#a24b5049dfd834f3a84537b3fceba4e65">cutlass::multiply_add&lt; T, complex&lt; T &gt;, complex&lt; T &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1negate.html#a63ff0d5bd31a54bcee8e1512182c8253">cutlass::negate&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1negate_3_01Array_3_01half__t_00_01N_01_4_01_4.html#aecf43e7a3abad362be771be2b98c7f71">cutlass::negate&lt; Array&lt; half_t, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1negate_3_01Array_3_01T_00_01N_01_4_01_4.html#a96bbedf72574b7657c21661ad866639e">cutlass::negate&lt; Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1NumericArrayConverter.html#aec6947576fd0a9330859e62b8afc040a">cutlass::NumericArrayConverter&lt; T, S, N, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericArrayConverter_3_01float_00_01half__t_00_012_00_01Round_01_4.html#a1f9de10dc27776b96dc17c80de91de8f">cutlass::NumericArrayConverter&lt; float, half_t, 2, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericArrayConverter_3_01float_00_01half__t_00_01N_00_01Round_01_4.html#a350690c1f46c4d1f8bcba3d54792ea5d">cutlass::NumericArrayConverter&lt; float, half_t, N, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericArrayConverter_3_01half__t_00_01float_00_012_00_01FloatRoundStyle_1_1round__to__nearest_01_4.html#a177cc3209f728a6629754cf2d685a37a">cutlass::NumericArrayConverter&lt; half_t, float, 2, FloatRoundStyle::round_to_nearest &gt;</a>
, <a class="el" href="structcutlass_1_1NumericArrayConverter_3_01half__t_00_01float_00_01N_00_01Round_01_4.html#aaceb23262eba4b9457e5020396e7b0f9">cutlass::NumericArrayConverter&lt; half_t, float, N, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverter.html#aeb946a1caf2882aafc57fae06f1bb1f8">cutlass::NumericConverter&lt; T, S, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverter_3_01float_00_01half__t_00_01Round_01_4.html#a69d1086aef51891da27dcbebddadac6e">cutlass::NumericConverter&lt; float, half_t, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverter_3_01half__t_00_01float_00_01FloatRoundStyle_1_1round__to__nearest_01_4.html#a619f44798aa208650afaf2c584454d4b">cutlass::NumericConverter&lt; half_t, float, FloatRoundStyle::round_to_nearest &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverter_3_01half__t_00_01float_00_01FloatRoundStyle_1_1round__toward__zero_01_4.html#a9ee10f5cf5ff71842a31e305d3a83947">cutlass::NumericConverter&lt; half_t, float, FloatRoundStyle::round_toward_zero &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverter_3_01int8__t_00_01float_00_01Round_01_4.html#a260bca91e4314cd401ad25f33c1b03cd">cutlass::NumericConverter&lt; int8_t, float, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverter_3_01T_00_01T_00_01Round_01_4.html#a30f44d01e40b5daf0fb683c608f34904">cutlass::NumericConverter&lt; T, T, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverterClamp.html#a38f56632e80a03a52f4b0197c2c8a769">cutlass::NumericConverterClamp&lt; T, S &gt;</a>
, <a class="el" href="structcutlass_1_1platform_1_1default__delete.html#a134480d0848e3c213e5a0bee120c90e0">cutlass::platform::default_delete&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1platform_1_1default__delete_3_01T[]_4.html#a5db6aa72c16a39437c9de615fa163f06">cutlass::platform::default_delete&lt; T[]&gt;</a>
, <a class="el" href="structcutlass_1_1platform_1_1integral__constant.html#a56a164b98a93b1430c26cdfde6f3720d">cutlass::platform::integral_constant&lt; value_t, V &gt;</a>
, <a class="el" href="structcutlass_1_1plus.html#a95661210dd2176c4b8ab86ffc3d5eb50">cutlass::plus&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1plus_3_01Array_3_01half__t_00_01N_01_4_01_4.html#a7eb77069757fffa01b7fba384c1e4bd2">cutlass::plus&lt; Array&lt; half_t, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1plus_3_01Array_3_01T_00_01N_01_4_01_4.html#a93dda0b911aabecf4a115ec7692ff5bd">cutlass::plus&lt; Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="classcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK.html#a8cd3ee6c0e54206393bf5931dc060fc9">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;</a>
, <a class="el" href="structcutlass_1_1reduction_1_1thread_1_1Reduce_3_01plus_3_01half__t_01_4_00_01AlignedArray_3_01half__t_00_01N_01_4_01_4.html#a37155e3dcc591e896df2d80c8daff4d4">cutlass::reduction::thread::Reduce&lt; plus&lt; half_t &gt;, AlignedArray&lt; half_t, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1reduction_1_1thread_1_1Reduce_3_01plus_3_01half__t_01_4_00_01Array_3_01half__t_00_01N_01_4_01_4.html#afe8938d7e9d806511480831668ef1563">cutlass::reduction::thread::Reduce&lt; plus&lt; half_t &gt;, Array&lt; half_t, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1reduction_1_1thread_1_1Reduce_3_01plus_3_01T_01_4_00_01Array_3_01T_00_01N_01_4_01_4.html#abb3c25f1d38931ffe40a24307fbf49b1">cutlass::reduction::thread::Reduce&lt; plus&lt; T &gt;, Array&lt; T, N &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1reduction_1_1thread_1_1Reduce_3_01plus_3_01T_01_4_00_01T_01_4.html#adc27a687496822f8477a427bf6ecd8b5">cutlass::reduction::thread::Reduce&lt; plus&lt; T &gt;, T &gt;</a>
, <a class="el" href="structcutlass_1_1reduction_1_1thread_1_1ReduceAdd.html#ac2fd8699c5a40b3c633fe8a59f02d8da">cutlass::reduction::thread::ReduceAdd&lt; ElementAccumulator_, Element_, Count &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomGaussianFunc.html#a8926a03c72ad2d9720dd1e4f39e0496e">cutlass::reference::device::detail::RandomGaussianFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomUniformFunc.html#a4f38cd67a3805c355f6820b0d5d376c6">cutlass::reference::device::detail::RandomUniformFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorCopyDiagonalInFunc.html#af700e9ac6ece02af0ce80fb8ef792084">cutlass::reference::device::detail::TensorCopyDiagonalInFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorCopyDiagonalOutFunc.html#a1f305e80ba94fcb123d4ce12d986322d">cutlass::reference::device::detail::TensorCopyDiagonalOutFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillDiagonalFunc.html#ae183dc13d4fc39856aa6b4bffe90a86e">cutlass::reference::device::detail::TensorFillDiagonalFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillLinearFunc.html#a0a70ddd37f599b4158f249d0f6cd896f">cutlass::reference::device::detail::TensorFillLinearFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomGaussianFunc.html#aa9e6a27d0a27d7769519181471ea7227">cutlass::reference::device::detail::TensorFillRandomGaussianFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomUniformFunc.html#a9d81b32b6e16509537540d7fb5085e9d">cutlass::reference::device::detail::TensorFillRandomUniformFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorUpdateDiagonalFunc.html#ac164c91a3b6236098c3ec87e70a8d66e">cutlass::reference::device::detail::TensorUpdateDiagonalFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorUpdateOffDiagonalFunc.html#ae1f12f1efd80ced9b4976698515bac41">cutlass::reference::device::detail::TensorUpdateOffDiagonalFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout4e016ab7cfc644acd7cb4ae770339773.html#aaa583b83e88d991c2e34e46e2858bbf8">cutlass::reference::device::Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, AccumulatorType, arch::OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout30b72addd464a2ca4a26785cbfd77a8e.html#a505e5a44cacbf991e68638dcd6eba466">cutlass::reference::device::Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, AccumulatorType, arch::OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01Layout660562b232f408218828ca5915b7e73a.html#a00dfb8cae5e94bed01f5a09bbd515410">cutlass::reference::device::Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, AccumulatorType, arch::OpXorPopc &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomGaussianFunc.html#a0208d33ceb87f0fc273dc17a66a5e990">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#a04ad19f3f63c0cba77a76a2ce243b727">cutlass::reference::host::detail::RandomGaussianFunc&lt; complex&lt; Element &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomUniformFunc.html#a09db614d063fec17a6a2fbe23389f10a">cutlass::reference::host::detail::RandomUniformFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomUniformFunc_3_01complex_3_01Element_01_4_01_4.html#a6ef7020f1108432fe51853dffb7e727c">cutlass::reference::host::detail::RandomUniformFunc&lt; complex&lt; Element &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorContainsFunc.html#a6e4363ba937eebda8642074662fb985b">cutlass::reference::host::detail::TensorContainsFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorCopyIf.html#ac2df07db0906c5cbed9f4eea92718e0e">cutlass::reference::host::detail::TensorCopyIf&lt; DstElement, DstLayout, SrcElement, SrcLayout, F &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorEqualsFunc.html#a8c3f73d72f88355b5810e81199ca3a77">cutlass::reference::host::detail::TensorEqualsFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFillDiagonalFunc.html#aa4b7ce383219da51d8021a4c27b3a57b">cutlass::reference::host::detail::TensorFillDiagonalFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFillFunc.html#a5fd28f9610e15afd33e1e8f56e1b9021">cutlass::reference::host::detail::TensorFillFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFillGaussianFunc.html#a4e447a80bd94cde69fa66f9e9d882b28">cutlass::reference::host::detail::TensorFillGaussianFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFillLinearFunc.html#a660184169f203d2499799692b136d589">cutlass::reference::host::detail::TensorFillLinearFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFillRandomUniformFunc.html#a9841b5a0391cd0dc747f5a30660ea7d9">cutlass::reference::host::detail::TensorFillRandomUniformFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFuncBinaryOp.html#a22f913a5999084fde6565bbb1c90a73b">cutlass::reference::host::detail::TensorFuncBinaryOp&lt; ElementA, LayoutA, ElementB, LayoutB, ElementD, LayoutD, BinaryFunc &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorUpdateOffDiagonalFunc.html#a4b610b1ca5801a747afb4a2c5ad0fadd">cutlass::reference::host::detail::TensorUpdateOffDiagonalFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TrivialConvert.html#aaaf16d4fc00a6c2a507cb964b41b3f84">cutlass::reference::host::detail::TrivialConvert&lt; DstElement, SrcElement &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01LayoutB_193dd3a37f00deff1e5dcd7c310afb1f.html#ab1279c5fa79550cdd993ce5241eaac24">cutlass::reference::host::Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, ComputeType, arch::OpMultiplyAdd &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01LayoutB_55729eac7dbd6bf311ea36f680e83e93.html#ac41fc498b787cd86d4433608121caffc">cutlass::reference::host::Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, ComputeType, arch::OpMultiplyAddSaturate &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1Gemm_3_01ElementA_00_01LayoutA_00_01ElementB_00_01LayoutB_4f3f32c4b336238abfd741e87bfced46.html#acb564b7ad68fa082a6c785e919a9de6a">cutlass::reference::host::Gemm&lt; ElementA, LayoutA, ElementB, LayoutB, ElementC, LayoutC, ScalarType, ComputeType, arch::OpXorPopc &gt;</a>
, <a class="el" href="structcutlass_1_1xor__add.html#a2bda9d781aee02f9aa90c75071a73f98">cutlass::xor_add&lt; T &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator*()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1const__iterator.html#a36aa6aa70a9536a7d2750d83d53f39f3">cutlass::Array&lt; T, N, false &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1iterator.html#ac43741ba9bcacd11dfb91fe02c57bef5">cutlass::Array&lt; T, N, false &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__iterator.html#af34c15c6d1d13db36ffe4b112bc75d47">cutlass::Array&lt; T, N, true &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__reverse__iterator.html#a49ed1c872c1640dca056fd4f8ced2261">cutlass::Array&lt; T, N, true &gt;::const_reverse_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1iterator.html#a2e7e50d7d99fda71be7ba03e0ce80417">cutlass::Array&lt; T, N, true &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1reverse__iterator.html#a1c8dca20abc1818ddab6973280394888">cutlass::Array&lt; T, N, true &gt;::reverse_iterator</a>
, <a class="el" href="classcutlass_1_1complex.html#ad87393e41d5f5d9df5a34759aadb84e9">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#ac6b5fd8d0e5cb856d363fbff9a5b89dd">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#a769289de7df1e4c24ba4d06ef9f0e8ba">cutlass::gemm::BatchedGemmCoord</a>
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#ae5fad509fbf1354e4717bf2bca593a25">cutlass::gemm::GemmCoord</a>
, <a class="el" href="structcutlass_1_1layout_1_1PitchLinearCoord.html#a4a9890d8c90801b38d11562e7622a8b5">cutlass::layout::PitchLinearCoord</a>
, <a class="el" href="structcutlass_1_1MatrixCoord.html#a2d0b67b654f39527b392505b1f1c77a5">cutlass::MatrixCoord</a>
, <a class="el" href="classcutlass_1_1platform_1_1unique__ptr.html#a0bd8240db28f3964714da5113cd9e354">cutlass::platform::unique_ptr&lt; T, Deleter &gt;</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1ConstIterator.html#a31fad0a6c23f15b0e7add2e130f5fa0b">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#a57dd1dfd84701160273fff79789f1137">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</a>
, <a class="el" href="structcutlass_1_1PredicateVector_1_1TrivialIterator.html#aa6774c80e75fdd469bf1e42d98f8959a">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::TrivialIterator</a>
, <a class="el" href="structcutlass_1_1Tensor4DCoord.html#aa9fd53334c3a6fdcce9b83896355d429">cutlass::Tensor4DCoord</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator*=()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#a5a2b48db6f225da9f7463408bbc1bd16">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#a00e618bc944d355badf67c0edd791412">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#a28f4434c4a25348b9ac0f99766ec262d">cutlass::gemm::BatchedGemmCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#a69fb0bb5e73f35d3c8df71a0174d6520">cutlass::gemm::GemmCoord</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_1layout_1_1PitchLinearCoord.html#a91764cdd9964cb7ad518538a6a230ab3">cutlass::layout::PitchLinearCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1MatrixCoord.html#a5fd3c3b58af1147a5c73657c05a16f5b">cutlass::MatrixCoord</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_1Tensor4DCoord.html#ab56a0b2352264f7a3753b621d1d850d6">cutlass::Tensor4DCoord</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator+()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#a09919ee4b99816bffed389b53f28dbf0">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#a0afaee4126a794f9db58ed4bd079b792">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#aec4c529a728118c0df6a3f527daba746">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#a510fc64da099c2a1e5c0a93c020776f5">cutlass::gemm::BatchedGemmCoord</a>
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#a20618962d649b187cb5d5613572946db">cutlass::gemm::GemmCoord</a>
, <a class="el" href="structcutlass_1_1layout_1_1PitchLinearCoord.html#a435d32b48c5526d7df9a8ab1a952ed09">cutlass::layout::PitchLinearCoord</a>
, <a class="el" href="structcutlass_1_1MatrixCoord.html#afc89138032ae6b0bc29edb932959eed4">cutlass::MatrixCoord</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1ConstIterator.html#a8c7811b73f6e0d80fca977b412d466c8">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#a831f15d9f01c7896dab70b94dfad660f">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#a3b58e357232b7c5222628fe871f23efe">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1Tensor4DCoord.html#a28448ff7ebd10f76954d012e7ae9bcd8">cutlass::Tensor4DCoord</a>
, <a class="el" href="classcutlass_1_1TensorRef.html#ad4e1a9d4912a18547cd5391d63e8e7ac">cutlass::TensorRef&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorView.html#a481a22e9efb27945c1bccae2381c03fb">cutlass::TensorView&lt; Element_, Layout_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator++()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1const__iterator.html#adcdcdf49b5d8e3ed801e2555c4f02b99">cutlass::Array&lt; T, N, false &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1iterator.html#a3fa26abda72f9714e39af23bcb5f97df">cutlass::Array&lt; T, N, false &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__iterator.html#ae148a1e543b22c1c4ec20374bc8929b3">cutlass::Array&lt; T, N, true &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__reverse__iterator.html#a76905b3d4bcb451ce2f6bdc73fa958ca">cutlass::Array&lt; T, N, true &gt;::const_reverse_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1iterator.html#a719a46d965659e78b9b4b72ea22aa705">cutlass::Array&lt; T, N, true &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1reverse__iterator.html#a54132c8586c34347e9c607dbf8bf3c39">cutlass::Array&lt; T, N, true &gt;::reverse_iterator</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1InterleavedPredicatedTileIterator.html#aaf23d79b0bdc047bdab776370c916065">cutlass::epilogue::threadblock::InterleavedPredicatedTileIterator&lt; ThreadMap_, Element_, InterleavedK &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1threadblock_1_1PredicatedTileIterator.html#a89c4e151beaab913b025276276fa7b3c">cutlass::epilogue::threadblock::PredicatedTileIterator&lt; ThreadMap_, Element_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorComplexTensorOp_3_01WarpShape___00_01Operato8cf03c624cf3210c71b7cbd580b080f8.html#ac18b19d15654fbbae90af8de4f5c4540">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#a71335e9f7fd399900035c397f1d5cfb1">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#af51b1e23c36ba486a00eeb614589967e">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#ac638be718341c995d2fb0948eab54d34">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#ae3b3c2db69782cfa3a1720e999901ae9">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#a69c832f3eb6db78a28910a8f8fba2383">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#ae56192089322c323496ad7bfb3f5a35f">cutlass::epilogue::warp::FragmentIteratorWmmaTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kA_00_01Element_67ca7e11a38e38f2c51b84767654a90f.html#a43622d4d221982dc3899e86da6ddf464">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#a786b7b06e620fac11f4cef8390d575c5">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#a6ebe8cd19a201e8310ed9925e3e764f7">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#ae5ad0892bae1856b535949cf9aba1015">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#a3d27aa97735386b44d163b1faa6a9e19">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#aa01726492e6ff5d624f2192a5585f76a">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kC, Element_, layout::RowMajor, Policy_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___008f607b871a2b3d854eb4def64712c042.html#a085b222207880b579b8ffa91c905e071">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___00027dabdc144edd6276f664ca74088510.html#a5143b72d5470049e8b879d2fab598b9b">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajorInterleaved&lt; InterleavedN &gt;, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___006c39f57875e0aa9d0ad82c8043ed8b98.html#ada65ee7584604de9a688d88838399809">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0b84f53cd44b339eccc12067c9f86e11c.html#aacfc1f7f194fad9371391134b2acf801">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0e52ad425e1ee3e68544873f66733237b.html#aa0317bb349e4b16873cfe51dfa748519">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___039819fb3ccd43786d556c2c9669508ef.html#acce80cd5cb5ab6c23795a8748c427610">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0352e0dcab42bc8360606874e00173556.html#aaad33f64f25dbcca8a34f577ad247595">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0ed7daaeba1c095e77f68533d4d2c475c.html#a9e8bb224bd9d307c1e8c7b0b898c3912">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, 64 &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0c7d419c589d601ce4eb603be566fea21.html#ae4af2d212abc11a8cf1a97d47b230f00">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpAccumulatorTileIterator.html#a1f4fce885f0c011704b89375ab18fec8">cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, Layout_, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan0d3248553e52cd61ed8a2b3b12a20343.html#adb4c78fa4cedb201d10334748aa1fdc9">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan34be8e21a40af3ebd2dc3dff460dca72.html#a21c2fa0fc65c51bb16fa9f6ddad1eda4">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand734577b7e54a074d143aba59828c2f2.html#a4cbfc44c1eac4ae4fd6571455b7b503d">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan16c56cdc2dda5eeb996af8ec0242d501.html#a8b5d7a05b12dd435a6acb4e598643cd1">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan5a221944f4a0e16ccab77ba684856942.html#a670d4132ca27fcd64f3d7e9eef192b7c">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operandcc9821c435540895138bc9af495f321.html#a0d18d3d63940b906fdd1ab48c4b2799a">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operana2f40b28f0d2286b84d86f7238d67b52.html#a75ea2d47c7a8fbec0a359736777b4a46">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1ConstIterator.html#a10ee4bb2f206432aa5ee1a83cb046b70">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#a6c7333ad14d545cafc707e78752bf1e3">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1PredicateVector_1_1TrivialIterator.html#ad24e9b451064e99fb19955f772c30e6a">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::TrivialIterator</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="classcutlass_1_1transform_1_1threadblock_1_1PredicatedTileAccessIterator2dThreadTile_3_01Shape__da632779aba661c0f4cfaaa78126b771.html#aebcde97662a51598dd410fa4684019ca">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#ad61f39942634bdb06917e989c768b1fd">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#aa8ea4f4cf66de36f0d72717540d229cc">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#aeec66a4a787a4f24ff006a2b5cee50a6">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#a262e1b3148cdac596af9f9d12571181b">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#a007063b233b0f4fe961c23ce2cfe9665">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#a31cb06df2a40b9741a6ddb550b001118">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#a727cf2ef7716c9ee28180a35e7a44ed9">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#a24603f7fa0795c885eaefd097579ca5e">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#a928b6720a862da0315fb5fef5da5921f">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#a5e72ab609350946f7d0773c347d3f4e4">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#ac34969c60025d6206b848433319da4fa">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#ac87a36086f64417f60da797760236aaa">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#a139515c9d839b495b661d1f7a5a58b48">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#a03832ac3ba57eb81b9cbd3eb63f4af70">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#a2465bdfb439468c269dd32c441b54a86">cutlass::transform::threadblock::PredicatedTileIterator&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; InterleavedK &gt;, AdvanceRank, ThreadMap_, AccessSize &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__eb7d20f8b9d69e0ae5e7ef51dc480867.html#a3e9cc8a43ac32bb97829adbe08f9942a">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__2c1476eaf582bfe972793e17babfe985.html#a3c9014675466aec05102de18ef6eb40a">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__a3c11cf1f00ef7a1efb8389ac6e4c6e0.html#a98e43fc279be18523a6839bd309695ab">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__0855e9d9ab619202d2397180c1e4c4a5.html#a8c3d182c37328fc568fcd025f73fdd98">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__f04332958a49a47d6fb2b25201764630.html#a3b6779627df8bc30fce0133eff084ec6">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__6baada077236f1a368c61c5e11b45b72.html#ad3153ccc12d036260b80a0a6ab4081d5">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::RowMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__0184b7188941788a96624510a4b2f876.html#a76f3aed087b22e12afaa00fbbba21462">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::RowMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__ebf4714349612673e8b6609b763eeb6f.html#a3d07f09c89750c04f1056ae2b2285296">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::TensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__e9a9e0f4286f652f55eb9b863b21effe.html#a352c7685cb9a37ba214609fe9ccb12e2">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Eleb60d066756d1c18f05fceee6a27bdb8a.html#a420dad81a508231290458ea637cb9036">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::ColumnMajorInterleaved&lt; 4 &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Ele76ed82829532ae1c17f4c78158f036c7.html#ad9bf859dd599ed0377e9cadcd86158b7">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Ele654c8f6161ae5340f040397a4e2e045c.html#aacef8150dc24810fee9ea4f7feee2c64">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; 4 &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_011d3637dbd8bc58bcb020b51bf57fbfc0.html#a215a832690fd3f0b770cc64eabc3eb2f">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_017982f81d4ef592e19c8427de2ea933a3.html#ac7cbbaf82a9a4a2dd785ec518bb5bef1">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_010889a732373c350de9b9a9f6c13cd761.html#ac0b1a5dd492d6f10f30401404d1aa114">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01187f8574e1fe9d7d5e8fbf09bd834bf0.html#a9147d2161f9fb8a5676495369e303c7a">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01793f74bfd8f116a827948ab01a37349a.html#a712e63fc221198ed9c2e3a3790f235de">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01bd31b3810c1fedf2e7e5959ff92b5d3d.html#a2f670f194227664f39fdcdac863a1a21">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Shape_::kRow &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0184a89653916f5d51ab59d1b386989a17.html#a3874988a2d312ab538f242e16f6a3b83">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0149454d361ea5885cf5166a920b5145df.html#a4d9b37a6c2bed1209a61fdcbf225f691">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01c20d35180520077a5a09b1e33543c1a5.html#a12d739452a931da5ca64f477b5a3b9b2">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01a31b454d9c930525c1e9ca406a514f40.html#a8ef7a6eb5d7c1059c0d9124902c86a3f">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0104ad31bd559a88cc418ae1cab7492ed5.html#a1b07f5faace323f365200325d4527173">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01f6f6511b5033cad31083644ac69c54d8.html#a6f6036167ff8e431d14155b12f5863ba">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01b3fa5720e807697de61b9f937b269cd0.html#abc2ea2a8705fa16f36e819ada74e2ae5">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Shape_::kColumn &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01efd5013a2503d6567e2bf6b40c97360c.html#a00da62dde4f9727452f281420cdaa7c8">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::TensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0197fef2242a3454a7d1cebe61aee28b43.html#a2c521b37a48b6b5cff910c1b137c464a">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01a75d2cd74e722d6ad6a3b41aabfd432d.html#ac921a2ca39e59829154ea0d99969b094">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::VoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01f96bbeb63e6d4ce4a2551279de3a9f0e.html#a75c3a995d2c15b28417a8f085defecdc">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::VoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_01dbd6b8468d5bd787308d2f615a24d123.html#a43f7bbb7a5ee381a159015d789bb1d4a">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::VoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Shape_::kContiguous &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator+=()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#acc9f522dfb539d55ac28e3629ce94da4">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#ad7e5cf02325b590fffa2fc5bfcb9da09">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#acb799faf60a17b708d0802f9e23c812f">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1TileIteratorSimt_3_01WarpShape___00_01Operator___00_01Elemenf2bd262ed3e202b25d5802d83965bf3b.html#a1a76d3b0fd2419e78d66640650ac511c">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#a46b0395432768c15516edf5d4ce5af73">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#ac043e1542408af1268bef2b8a09ed7d9">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#a10e475f4cadd7ec7371c5b74742a6081">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#a10fd20879beda76f48772465a1b491f4">cutlass::epilogue::warp::TileIteratorWmmaTensorOp&lt; WarpShape_, OperatorShape_, OperatorFragment_, layout::RowMajor &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#ab7192255fd439e763138f2d9167581d3">cutlass::gemm::BatchedGemmCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#a5465f7308778eac5d14d8020179a65e0">cutlass::gemm::GemmCoord</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="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___008f607b871a2b3d854eb4def64712c042.html#af8d0bb15e6c967e41c175d7989c728b2">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___00027dabdc144edd6276f664ca74088510.html#afefe082ee98be3a0a7dfdadba702d58f">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajorInterleaved&lt; InterleavedN &gt;, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___006c39f57875e0aa9d0ad82c8043ed8b98.html#ad48f385d6822485184461d82bd70d4de">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0b84f53cd44b339eccc12067c9f86e11c.html#aae859de5fea22c7ee1653d9eabe0468d">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0e52ad425e1ee3e68544873f66733237b.html#a2b8a09d931c351b3c252ee1a72147208">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___039819fb3ccd43786d556c2c9669508ef.html#ad657ba5ffa2eb7c80c4328a47783e67a">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0352e0dcab42bc8360606874e00173556.html#ab44ddbcd150470e1d234b2138ee1e8ef">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0ed7daaeba1c095e77f68533d4d2c475c.html#aa711fab987df637da35e00d655dcaf65">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, 64 &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0c7d419c589d601ce4eb603be566fea21.html#a7407a4b2ac07e96c39c1e53aeb76a594">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpAccumulatorTileIterator.html#a9c6b5b280940abc435d27d67d89f9238">cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, Layout_, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan0d3248553e52cd61ed8a2b3b12a20343.html#a226402c19b486c0424b8ac6d8ef21409">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan34be8e21a40af3ebd2dc3dff460dca72.html#a528adebab8aa3fa3233554782652047e">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand734577b7e54a074d143aba59828c2f2.html#a646f6d40a4b8d446aedc0c709a04f9b4">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan16c56cdc2dda5eeb996af8ec0242d501.html#ab3049e27aa39203d23e891527e7ae36d">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan5a221944f4a0e16ccab77ba684856942.html#ab25abf8b419fe5aa8d2ff0993859cd6c">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operandcc9821c435540895138bc9af495f321.html#ac1e657b6f0e2e671d2c5c2d66a95e904">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operana2f40b28f0d2286b84d86f7238d67b52.html#aa4c1c2d490a62ea8f742b88d281d7630">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1PitchLinearCoord.html#a0bda872098a003cd045c67dd215bc1d1">cutlass::layout::PitchLinearCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1MatrixCoord.html#ad105615dbf7ede75caa0e778c873bd06">cutlass::MatrixCoord</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="classcutlass_1_1PredicateVector_1_1ConstIterator.html#a1ab6e127b815ec870abf80ecfa94963c">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#a5dcf36597690ce9ad7ef95d82b50654e">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#ac580da95d9109736c3a091ee3b0340f9">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1Tensor4DCoord.html#af5f312484425767c77f0192cc89eef3d">cutlass::Tensor4DCoord</a>
, <a class="el" href="classcutlass_1_1TensorRef.html#ae82b8ba3bbf7ebb28fe063932db3dc6b">cutlass::TensorRef&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorView.html#a5ea9e8f707291de961532f083554c2b6">cutlass::TensorView&lt; Element_, Layout_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator-()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#a5fd9515e7f802221d6eb7c36f83586cd">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#aeb37aceee94bbef99217d011b28d89f8">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#a2e1a706629eae28128230a0fa34b84a0">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#a3cbb255d020bd6023325f6be3e781b65">cutlass::gemm::BatchedGemmCoord</a>
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#aff600813c52c6f0a8e08b4a9a2925c02">cutlass::gemm::GemmCoord</a>
, <a class="el" href="structcutlass_1_1layout_1_1PitchLinearCoord.html#a5f721cafc24e18805b4f844397b7465e">cutlass::layout::PitchLinearCoord</a>
, <a class="el" href="structcutlass_1_1MatrixCoord.html#af7da194fa0200966f24a06dda344c6df">cutlass::MatrixCoord</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1ConstIterator.html#a4d9ebd49895cd9d218118edfdc3350d7">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#aca8c844b4e4444fa6dc9779761356a75">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#a7b548b80615296b7298d3911a5521032">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1Tensor4DCoord.html#a9f850c2e2a7b4cda1e58a04884d8be47">cutlass::Tensor4DCoord</a>
, <a class="el" href="classcutlass_1_1TensorRef.html#aa275fd660406665884f25c6f693a8892">cutlass::TensorRef&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorView.html#a1d57ad0c37d33a544a07e71925605a1d">cutlass::TensorView&lt; Element_, Layout_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator--()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1const__iterator.html#aa2c9f9bb9601208bd784bdc821b62f3a">cutlass::Array&lt; T, N, false &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1iterator.html#ae0fd752d82e67eb74ace86cfdaa69020">cutlass::Array&lt; T, N, false &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__iterator.html#a39085604c1b7a0dee3f5a0b96776d297">cutlass::Array&lt; T, N, true &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__reverse__iterator.html#a9bbf80876fe3e72dbfe39ee880ec2f59">cutlass::Array&lt; T, N, true &gt;::const_reverse_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1iterator.html#a1bd27b5796ef75d377514d91ac8ab03c">cutlass::Array&lt; T, N, true &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1reverse__iterator.html#a45521aa859c43140ac0dd95f5b161036">cutlass::Array&lt; T, N, true &gt;::reverse_iterator</a>
, <a class="el" href="classcutlass_1_1epilogue_1_1warp_1_1FragmentIteratorComplexTensorOp_3_01WarpShape___00_01Operato8cf03c624cf3210c71b7cbd580b080f8.html#a5ec88193a070b08cfe7b0b9bfed042f8">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#a21f202bb39729599ab91d2e852c9bc7b">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#a6faf5b99575aac462873322e194dd2d4">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#a91c59ef117cc8d16241b58d5f0c18399">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#a331cd6e1757dacbcc8461fcd2d0e6139">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#a9c9998058fa65fbf0f4ff4b73658e3d8">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#a6afb744ed20be944b1f8ea0b0dd506f6">cutlass::epilogue::warp::FragmentIteratorWmmaTensorOp&lt; WarpShape_, OperatorShape_, OperatorElementC_, OperatorFragmentC_, layout::RowMajor &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaSimtTileIterator_3_01Shape___00_01Operand_1_1kA_00_01Element_67ca7e11a38e38f2c51b84767654a90f.html#a0687c1a81ea61fa23073d8eee66e5dc2">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#a8cbbbe47e7d5c19e227f62daefa36a45">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#a11e9cfb5508eb7d5aeaee3a842be011d">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#ae7affdc67286b181c282f00a6da6ff0d">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#a1fb00b7856386aab1b784e6079346a46">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#a76b31f3bcc2fb27c9168712d455d97f8">cutlass::gemm::warp::MmaSimtTileIterator&lt; Shape_, Operand::kC, Element_, layout::RowMajor, Policy_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___008f607b871a2b3d854eb4def64712c042.html#ac6dc64ff8af8291455978f9fd168c6fc">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___00027dabdc144edd6276f664ca74088510.html#aa6757114de97d47339a97caa1ff5d863">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajorInterleaved&lt; InterleavedN &gt;, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___006c39f57875e0aa9d0ad82c8043ed8b98.html#abba62827ecc9247b59e1d03898c5feb2">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0b84f53cd44b339eccc12067c9f86e11c.html#adb4fa15e29e8f14136d9f6a9826bbfd2">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0e52ad425e1ee3e68544873f66733237b.html#ab319010cfbb0e20723c0fbdefac6a10e">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___039819fb3ccd43786d556c2c9669508ef.html#a9ce64fa398716ceb18123a34cb8c7eef">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0352e0dcab42bc8360606874e00173556.html#ac698bc8cd44572e1f3ee52771fb2cefb">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0ed7daaeba1c095e77f68533d4d2c475c.html#ab273a94817606b8bdbc714a7964bd95b">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, 64 &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0c7d419c589d601ce4eb603be566fea21.html#af02b9089644245ee0d7cd4fab17b2026">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpAccumulatorTileIterator.html#abf77a8981beee6433afcae3c820719e2">cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, Layout_, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan0d3248553e52cd61ed8a2b3b12a20343.html#ac920b9ce72a6e473dc7da0c55592d515">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan34be8e21a40af3ebd2dc3dff460dca72.html#ad262c03181b674266af1daa43162ef8e">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand734577b7e54a074d143aba59828c2f2.html#ad70d6d18a516d29a9a2234823ed0379e">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan16c56cdc2dda5eeb996af8ec0242d501.html#a170dc47832a0e4e2a2d4c95ee1408931">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan5a221944f4a0e16ccab77ba684856942.html#ab1c2a3d187a87f69f218be46c0d91e48">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operandcc9821c435540895138bc9af495f321.html#aab514462e6a1799ccdb7fffaa0d10601">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operana2f40b28f0d2286b84d86f7238d67b52.html#a79ebd8c200848da932f1de7abd7adb3c">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1ConstIterator.html#a2763012a9284e97650b14e20c5668286">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#aad709a11f43b84c88e3ce3a0394f8e8a">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</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="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Eleb60d066756d1c18f05fceee6a27bdb8a.html#a5939b39c5781c86a5ab8b72c443ebd90">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::ColumnMajorInterleaved&lt; 4 &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Ele76ed82829532ae1c17f4c78158f036c7.html#ae596f8af2a0482bddeaf589417247a69">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Ele654c8f6161ae5340f040397a4e2e045c.html#ab056380a296291b2f8470ef86a9bda2d">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; 4 &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_011d3637dbd8bc58bcb020b51bf57fbfc0.html#a3b0cbdbbec0e63e759f86131cbcad32e">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::ColumnMajor, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0184a89653916f5d51ab59d1b386989a17.html#ae59497fa13e9b0d4c695ea7030cd2fca">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::PitchLinear, AdvanceRank, ThreadMap_, Alignment &gt;</a>
, <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_0149454d361ea5885cf5166a920b5145df.html#a8b62546dd2f7373ec5a45b0b309bbd64">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::RowMajor, AdvanceRank, ThreadMap_, Alignment &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator-=()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#ac955777830f1a5a487a57852528ea9ef">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="classcutlass_1_1ConstSubbyteReference.html#abc062e38c9eede21a770f22ac957dec0">cutlass::ConstSubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#a15ac170c861b34d418432aeb62ea86e0">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#a7d2794db4b9a1874d26d6caddc466003">cutlass::gemm::BatchedGemmCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#a12ba250be3d5474b7c6fc4eddd4f58d5">cutlass::gemm::GemmCoord</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="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___008f607b871a2b3d854eb4def64712c042.html#a0e3aca80cfc279e9e207a3044e925dc9">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___00027dabdc144edd6276f664ca74088510.html#aa5e85b5ad8fcadd76dfc664c626875dc">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::ColumnMajorInterleaved&lt; InterleavedN &gt;, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpAccumulatorTileIterator_3_01Shape___00_01Element___006c39f57875e0aa9d0ad82c8043ed8b98.html#affd5c6e12f2fc9c00ec08d43e8f8d2f6">cutlass::gemm::warp::MmaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, cutlass::layout::RowMajor, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0b84f53cd44b339eccc12067c9f86e11c.html#a8379144007ec28f9b99ff85f255bed49">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0e52ad425e1ee3e68544873f66733237b.html#adfcfe12e980652302eddabcf06911208">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___039819fb3ccd43786d556c2c9669508ef.html#ad0e1edd1608df393c451b9ac6bd079de">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, int(128/sizeof(Element_))&gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0352e0dcab42bc8360606874e00173556.html#a14abb245ffc13842b281db8f72b71361">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0ed7daaeba1c095e77f68533d4d2c475c.html#a8da3cde971d4dd1beb123645b59e90db">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value, 64 &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand___0c7d419c589d601ce4eb603be566fea21.html#a97d09b2ad300996739ff4d1935804aaa">cutlass::gemm::warp::MmaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, InstructionShape_, OpDelta_, 32, PartitionsK_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpAccumulatorTileIterator.html#a2e77e28efc5efcb411258d4e55c63a4e">cutlass::gemm::warp::MmaVoltaTensorOpAccumulatorTileIterator&lt; Shape_, Element_, Layout_, InstructionShape_, OpDelta_ &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan0d3248553e52cd61ed8a2b3b12a20343.html#a2eae25791f7973945d760fcae0f35cf7">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan34be8e21a40af3ebd2dc3dff460dca72.html#aae34807a2f795d57a71a0661182cdb6a">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kA, Element_, cutlass::layout::VoltaTensorOpMultiplicandCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operand734577b7e54a074d143aba59828c2f2.html#a561d838825c39ba946afc2125dcc9fe0">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan16c56cdc2dda5eeb996af8ec0242d501.html#ad3244c12457461bd2f21035b202016d3">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand::kB, Element_, cutlass::layout::VoltaTensorOpMultiplicandBCongruous&lt; sizeof_bits&lt; Element_ &gt;::value &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operan5a221944f4a0e16ccab77ba684856942.html#a59083f63ea63934c727af2d2afc5ea98">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::ColumnMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operandcc9821c435540895138bc9af495f321.html#a73efe1a38ac07d8c1bd72a72fed00b57">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="classcutlass_1_1gemm_1_1warp_1_1MmaVoltaTensorOpMultiplicandTileIterator_3_01Shape___00_01Operana2f40b28f0d2286b84d86f7238d67b52.html#ac52e7b70d38ab7f0d071eee0730b21ab">cutlass::gemm::warp::MmaVoltaTensorOpMultiplicandTileIterator&lt; Shape_, Operand_, Element_, cutlass::layout::VoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, KBlock &gt;, InstructionShape_, OpDelta_, 32 &gt;</a>
, <a class="el" href="structcutlass_1_1layout_1_1PitchLinearCoord.html#afc72732cfb36f432ecc369fa21b05db8">cutlass::layout::PitchLinearCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1MatrixCoord.html#a6feef48cf24733d22ca53a27cbc33ac0">cutlass::MatrixCoord</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="classcutlass_1_1PredicateVector_1_1ConstIterator.html#a6c178f795233f0212e89f5b522097d7b">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#a13a19c8d85bf415fe5f4c17fce9cf6ca">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#a1f3a56c15363b7287665c2bd93a66ffd">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
, <a class="el" href="structcutlass_1_1Tensor4DCoord.html#a05798b41f0fbaa92f766902bac286609">cutlass::Tensor4DCoord</a>
, <a class="el" href="classcutlass_1_1TensorRef.html#ac5f24cda8149ddfa8f19467ff2629f49">cutlass::TensorRef&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorView.html#a1e1533d9b7942be33df1d122667d5ec6">cutlass::TensorView&lt; Element_, Layout_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator-&gt;()
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_1device__memory_1_1allocation.html#a793e663af04f7fb81f7df710147c8fc8">cutlass::device_memory::allocation&lt; T &gt;</a>
, <a class="el" href="classcutlass_1_1platform_1_1unique__ptr.html#afa52edcaef23461ce1f9c1dac349c24b">cutlass::platform::unique_ptr&lt; T, Deleter &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator/()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#af3ecad26ad7c54e461392b1c808f6ee2">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#a6ae9f189d1a7a5ce7bb5e4416559c79f">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#a509d8d73926232df2d05b309a01f1f94">cutlass::gemm::BatchedGemmCoord</a>
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#a2d03dab79e54b8a17c0e7d91b9e0afd7">cutlass::gemm::GemmCoord</a>
, <a class="el" href="structcutlass_1_1layout_1_1PitchLinearCoord.html#add84bf353eda9ec51c3d8a6b240d50ee">cutlass::layout::PitchLinearCoord</a>
, <a class="el" href="structcutlass_1_1MatrixCoord.html#a2542660b34b184b5b8f9d0ad3dedc40a">cutlass::MatrixCoord</a>
, <a class="el" href="structcutlass_1_1Tensor4DCoord.html#ab2fd81b93d9130ff969b783dbaab54b2">cutlass::Tensor4DCoord</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator/=()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1complex.html#a41f71af804002f802c0a6fbc4d8938a0">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#af515e669363986dbbd60951ea6b69e14">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1gemm_1_1BatchedGemmCoord.html#a72577cab5f8f329c22d622a1b9d817c5">cutlass::gemm::BatchedGemmCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1gemm_1_1GemmCoord.html#aac1e1bad751237fd76d32a1ea10f6c40">cutlass::gemm::GemmCoord</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_1layout_1_1PitchLinearCoord.html#a2e07dd084a7194ea6ad847fce9fa23a5">cutlass::layout::PitchLinearCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
, <a class="el" href="structcutlass_1_1MatrixCoord.html#aab345c8ddb8048bfe3d667bc7ce6522f">cutlass::MatrixCoord</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_1Tensor4DCoord.html#ab591c052af780e65a77ea3e0f33d46aa">cutlass::Tensor4DCoord</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>operator&lt;()
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_1Coord.html#a47ad37153eb8d291266a51b39ead5948">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1integer__subbyte.html#a3579d7cc4925f055e0b4069cec222bbe">cutlass::integer_subbyte&lt; Bits, Signed &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>operator&lt;=()
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_1Coord.html#a0e6405e081936a4fb23f15160e94ad08">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1integer__subbyte.html#ab7d9011bc706eeccf5d81b7308c93992">cutlass::integer_subbyte&lt; Bits, Signed &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator=()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1reference.html#a629bfbf64481de5252896b45721254ad">cutlass::Array&lt; T, N, false &gt;::reference</a>
, <a class="el" href="classcutlass_1_1complex.html#aedddbb22051f5b3a5b45a6b0377cfa9c">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1device__memory_1_1allocation.html#a3d6109a5a85d407b2e43d34766af1d85">cutlass::device_memory::allocation&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1half__t.html#a60416b7575abcb835de85be6ae56c847">cutlass::half_t</a>
, <a class="el" href="classcutlass_1_1SubbyteReference.html#a6eb29d35b4536cfa7e0d351bd49cf04c">cutlass::SubbyteReference&lt; Element_, Storage_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator==()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1const__iterator.html#a42dd93a0f0df4ec86de4880fa9cc5843">cutlass::Array&lt; T, N, false &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1iterator.html#ac6173c654b3cc22cb357e5ad847dffc9">cutlass::Array&lt; T, N, false &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__iterator.html#a25c770f60b9e8f8d7eb2e58efcb7c3e1">cutlass::Array&lt; T, N, true &gt;::const_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1const__reverse__iterator.html#af5f68e8bd235243c74e7dc256fa37408">cutlass::Array&lt; T, N, true &gt;::const_reverse_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1iterator.html#aeab63d32f97e5046f55473d652a69bd7">cutlass::Array&lt; T, N, true &gt;::iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1reverse__iterator.html#a6837b73dd030c8d327041639fa8e8a27">cutlass::Array&lt; T, N, true &gt;::reverse_iterator</a>
, <a class="el" href="classcutlass_1_1complex.html#acea055c11a695f0c14c4da2280b311f1">cutlass::complex&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#a8424ccd74e7e0ff1cf358ef571779cba">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1integer__subbyte.html#a1097f17193b9d3ab0a6d7e96a1148142">cutlass::integer_subbyte&lt; Bits, Signed &gt;</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1ConstIterator.html#acce1de47d03b30131d36ca02327b77f6">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::ConstIterator</a>
, <a class="el" href="classcutlass_1_1PredicateVector_1_1Iterator.html#a349edcbf86ca0a6827462a07e8754320">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;::Iterator</a>
</li>
<li>operator&gt;()
: <a class="el" href="structcutlass_1_1Coord.html#a08507ff13f518a93a7d16ea0018f8a53">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1integer__subbyte.html#aabc7863a44172ac29a43447c23e28767">cutlass::integer_subbyte&lt; Bits, Signed &gt;</a>
</li>
<li>operator&gt;=()
: <a class="el" href="structcutlass_1_1Coord.html#ae37243e9f51b2b92b5fd09de69392657">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="structcutlass_1_1integer__subbyte.html#a7cf206db2da7612ce1708e8869570a59">cutlass::integer_subbyte&lt; Bits, Signed &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator[]()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4.html#a35db1c6ac0d42a486eb3a0a0eee95c80">cutlass::Array&lt; T, N, false &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#a9812d796007116dbd8b20117976deb48">cutlass::Array&lt; T, N, true &gt;</a>
, <a class="el" href="structcutlass_1_1Coord.html#a6eee93e5fdbe147f751ec108b28275a1">cutlass::Coord&lt; Rank_, Index_, LongIndex_ &gt;</a>
, <a class="el" href="classcutlass_1_1platform_1_1unique__ptr.html#a353b89f40cb21e0cfd2af2094d417043">cutlass::platform::unique_ptr&lt; T, Deleter &gt;</a>
, <a class="el" href="structcutlass_1_1PredicateVector.html#a5a759fda07cb29bfa9041f45e1725a07">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorRef.html#a10e36681cae687d9a45ec01bad2626b7">cutlass::TensorRef&lt; Element_, Layout_ &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>operator|=()
: <a class="el" href="structcutlass_1_1PredicateVector.html#aab9de134132c62de1c062ca57582cdbc">cutlass::PredicateVector&lt; kPredicates_, kPredicatesPerByte_, kPredicateStart_ &gt;</a>
</li>
</ul>
</div><!-- contents -->
<!-- start footer part -->
<hr class="footer"/><address class="footer"><small>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
Generated by &#160;<a href="http://www.doxygen.org/index.html">
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<img class="footer" src="doxygen.png" alt="doxygen"/>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
</a> 1.8.11
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</small></address>
</body>
</html>
Reference in New Issue Copy Permalink