squall/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
2049c6c5a2
cutlass/docs/functions_r.html

456 lines
64 KiB
HTML
Raw Normal View History

CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/>
<meta http-equiv="X-UA-Compatible" content="IE=9"/>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<meta name="generator" content="Doxygen 1.8.11"/>
<title>CUTLASS: Class Members</title>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<link href="tabs.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="jquery.js"></script>
<script type="text/javascript" src="dynsections.js"></script>
<link href="search/search.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="search/searchdata.js"></script>
<script type="text/javascript" src="search/search.js"></script>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<script type="text/javascript">
$(document).ready(function() { init_search(); });
</script>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
extensions: ["tex2jax.js"],
jax: ["input/TeX","output/HTML-CSS"],
});
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
</script><script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js"></script>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<link href="doxygen.css" rel="stylesheet" type="text/css" />
</head>
<body>
<div id="top"><!-- do not remove this div, it is closed by doxygen! -->
<div id="titlearea">
<table cellspacing="0" cellpadding="0">
<tbody>
<tr style="height: 56px;">
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<td id="projectlogo"><img alt="Logo" src="cutlass-logo-small.png"/></td>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<td id="projectalign" style="padding-left: 0.5em;">
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<div id="projectname">CUTLASS
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</div>
<div id="projectbrief">CUDA Templates for Linear Algebra Subroutines and Solvers</div>
</td>
</tr>
</tbody>
</table>
</div>
<!-- end header part -->
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<!-- Generated by Doxygen 1.8.11 -->
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<script type="text/javascript">
var searchBox = new SearchBox("searchBox", "search",false,'Search');
</script>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<div id="navrow1" class="tabs">
<ul class="tablist">
<li><a href="index.html"><span>Main&#160;Page</span></a></li>
<li><a href="modules.html"><span>Modules</span></a></li>
<li><a href="namespaces.html"><span>Namespaces</span></a></li>
<li class="current"><a href="annotated.html"><span>Classes</span></a></li>
<li><a href="files.html"><span>Files</span></a></li>
<li>
<div id="MSearchBox" class="MSearchBoxInactive">
<span class="left">
<img id="MSearchSelect" src="search/mag_sel.png"
onmouseover="return searchBox.OnSearchSelectShow()"
onmouseout="return searchBox.OnSearchSelectHide()"
alt=""/>
<input type="text" id="MSearchField" value="Search" accesskey="S"
onfocus="searchBox.OnSearchFieldFocus(true)"
onblur="searchBox.OnSearchFieldFocus(false)"
onkeyup="searchBox.OnSearchFieldChange(event)"/>
</span><span class="right">
<a id="MSearchClose" href="javascript:searchBox.CloseResultsWindow()"><img id="MSearchCloseImg" border="0" src="search/close.png" alt=""/></a>
</span>
</div>
</li>
</ul>
</div>
<div id="navrow2" class="tabs2">
<ul class="tablist">
<li><a href="annotated.html"><span>Class&#160;List</span></a></li>
<li><a href="classes.html"><span>Class&#160;Index</span></a></li>
<li><a href="inherits.html"><span>Class&#160;Hierarchy</span></a></li>
<li class="current"><a href="functions.html"><span>Class&#160;Members</span></a></li>
</ul>
</div>
<div id="navrow3" class="tabs2">
<ul class="tablist">
<li class="current"><a href="functions.html"><span>All</span></a></li>
<li><a href="functions_func.html"><span>Functions</span></a></li>
<li><a href="functions_vars.html"><span>Variables</span></a></li>
<li><a href="functions_type.html"><span>Typedefs</span></a></li>
<li><a href="functions_enum.html"><span>Enumerations</span></a></li>
<li><a href="functions_eval.html"><span>Enumerator</span></a></li>
</ul>
</div>
<div id="navrow4" class="tabs3">
<ul class="tablist">
<li><a href="functions.html#index_a"><span>a</span></a></li>
<li><a href="functions_b.html#index_b"><span>b</span></a></li>
<li><a href="functions_c.html#index_c"><span>c</span></a></li>
<li><a href="functions_d.html#index_d"><span>d</span></a></li>
<li><a href="functions_e.html#index_e"><span>e</span></a></li>
<li><a href="functions_f.html#index_f"><span>f</span></a></li>
<li><a href="functions_g.html#index_g"><span>g</span></a></li>
<li><a href="functions_h.html#index_h"><span>h</span></a></li>
<li><a href="functions_i.html#index_i"><span>i</span></a></li>
<li><a href="functions_k.html#index_k"><span>k</span></a></li>
<li><a href="functions_l.html#index_l"><span>l</span></a></li>
<li><a href="functions_m.html#index_m"><span>m</span></a></li>
<li><a href="functions_n.html#index_n"><span>n</span></a></li>
<li><a href="functions_o.html#index_o"><span>o</span></a></li>
<li><a href="functions_p.html#index_p"><span>p</span></a></li>
<li><a href="functions_q.html#index_q"><span>q</span></a></li>
<li class="current"><a href="functions_r.html#index_r"><span>r</span></a></li>
<li><a href="functions_s.html#index_s"><span>s</span></a></li>
<li><a href="functions_t.html#index_t"><span>t</span></a></li>
<li><a href="functions_u.html#index_u"><span>u</span></a></li>
<li><a href="functions_v.html#index_v"><span>v</span></a></li>
<li><a href="functions_w.html#index_w"><span>w</span></a></li>
<li><a href="functions_y.html#index_y"><span>y</span></a></li>
<li><a href="functions_0x7e.html#index_0x7e"><span>~</span></a></li>
</ul>
</div>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</div><!-- top -->
<!-- window showing the filter options -->
<div id="MSearchSelectWindow"
onmouseover="return searchBox.OnSearchSelectShow()"
onmouseout="return searchBox.OnSearchSelectHide()"
onkeydown="return searchBox.OnSearchSelectKey(event)">
</div>
<!-- iframe showing the search results (closed by default) -->
<div id="MSearchResultsWindow">
<iframe src="javascript:void(0)" frameborder="0"
name="MSearchResults" id="MSearchResults">
</iframe>
</div>
<div class="contents">
<div class="textblock">Here is a list of all class members with links to the classes they belong to:</div>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<h3><a class="anchor" id="index_r"></a>- r -</h3><ul>
<li>random
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomGaussianFunc_1_1Params.html#aa21637b631a16fc4e2860e27e422765b">cutlass::reference::device::detail::TensorFillRandomGaussianFunc&lt; Element, Layout &gt;::Params</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomGaussianFunc.html#ab7a72f16421d8bc596af374af0fae1d1">cutlass::reference::device::detail::TensorFillRandomGaussianFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomUniformFunc_1_1Params.html#aff721d5c0b74fd3a6edefeecca97debe">cutlass::reference::device::detail::TensorFillRandomUniformFunc&lt; Element, Layout &gt;::Params</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomUniformFunc.html#a134e1c6b57395a313718e8ad5590feab">cutlass::reference::device::detail::TensorFillRandomUniformFunc&lt; Element, Layout &gt;</a>
</li>
<li>RandomFunc
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomGaussianFunc.html#a411755eb3303cc9681111ac2fabb2a48">cutlass::reference::device::detail::TensorFillRandomGaussianFunc&lt; Element, Layout &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1TensorFillRandomUniformFunc.html#afafde9750cffecbaf6d0dba8a55fc5d5">cutlass::reference::device::detail::TensorFillRandomUniformFunc&lt; Element, Layout &gt;</a>
</li>
<li>RandomGaussianFunc()
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomGaussianFunc.html#a38bf4f3bfe2df73c264a23f3956a65fd">cutlass::reference::device::detail::RandomGaussianFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomGaussianFunc.html#a45ca53d7fcac978533361d8fc5f15311">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#a72b71664e9ddc5ce392e1db822cfdada">cutlass::reference::host::detail::RandomGaussianFunc&lt; complex&lt; Element &gt; &gt;</a>
</li>
<li>RandomUniformFunc()
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomUniformFunc.html#a50179367fdbaf59e6fb585b9e871730e">cutlass::reference::device::detail::RandomUniformFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomUniformFunc.html#ac39a320d0975f140d51cf2cd7dd53441">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#aa894090fefa7f3cc99ec66f5b6e7b7ac">cutlass::reference::host::detail::RandomUniformFunc&lt; complex&lt; Element &gt; &gt;</a>
</li>
<li>range
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomUniformFunc_1_1Params.html#ace319d38113a83e3cccc7860897154c3">cutlass::reference::device::detail::RandomUniformFunc&lt; Element &gt;::Params</a>
, <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomUniformFunc.html#ab65f9bd7b329d6ce077daf50fb3148f1">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#a17c91db74ff727d9f42442c09d21d0b3">cutlass::reference::host::detail::RandomUniformFunc&lt; complex&lt; Element &gt; &gt;</a>
</li>
<li>raw()
: <a class="el" href="structcutlass_1_1half__t.html#aaee5cf278f88e1de09cbebf8cdd2cbe8">cutlass::half_t</a>
</li>
<li>raw_data()
: <a class="el" href="structcutlass_1_1AlignedBuffer.html#a3a87c3b8f14893d30f374bde2b88052c">cutlass::AlignedBuffer&lt; T, N, Align &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4.html#a66e2465301e46afebf9e56c4060fb3cb">cutlass::Array&lt; T, N, false &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#ae4e76ed2b36a4deda6ef36b00fdda363">cutlass::Array&lt; T, N, true &gt;</a>
</li>
<li>rbegin()
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4.html#a2098c88aed61f9b27bac37a083130336">cutlass::Array&lt; T, N, false &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#ad8ec17a6d004cb6ffd4450c0686cd924">cutlass::Array&lt; T, N, true &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>real()
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#afaeeace8c2987047400b83973e196aeb">cutlass::complex&lt; T &gt;</a>
</li>
<li>Real
: <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1RandomUniformFunc.html#a317b294d017b81224eb1aca2742e8a3c">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#abd7e66b999df7719f6ac77f0a82a0d5d">cutlass::reference::host::detail::RandomUniformFunc&lt; complex&lt; Element &gt; &gt;</a>
</li>
<li>real
: <a class="el" href="structcutlass_1_1TypeTraits_3_01complex_3_01double_01_4_01_4_1_1integer__type.html#ad95c7b0fd5538d193b3581bf4cd7eca1">cutlass::TypeTraits&lt; complex&lt; double &gt; &gt;::integer_type</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01complex_3_01double_01_4_01_4_1_1unsigned__type.html#ab1d0b72c9376509deaf61bb2eed09dc5">cutlass::TypeTraits&lt; complex&lt; double &gt; &gt;::unsigned_type</a>
</li>
<li>ReduceAdd()
: <a class="el" href="structcutlass_1_1reduction_1_1thread_1_1ReduceAdd.html#ac9ccfca110d19133c4b203227577e2ec">cutlass::reduction::thread::ReduceAdd&lt; ElementAccumulator_, Element_, Count &gt;</a>
</li>
<li>reduction
: <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_1_1Arguments.html#af6008a17370eec923078a23564972383">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_3_01ElementA___00_01LayoutA___00_01Elementafcb1aeaf2035a7ac769d7acc233423b.html#a63048fa3419753d96a60eaee28f6cfe4">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK_1_1Params.html#ab59614242d435c963b9607eb7da6f5b5">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;::Params</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<li>reduction_stride
: <a class="el" href="structcutlass_1_1reduction_1_1BatchedReductionTraits_1_1Params.html#a5d1463d473d4226b0d19c581b16ed3b2">cutlass::reduction::BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt;::Params</a>
</li>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<li>ReductionKernel
: <a class="el" href="classcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel.html#aaf83264eb3effceee610d9547ddf32e9">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#aa69c465611c07990cdc79605c16b04ff">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>
</li>
<li>ReductionOp
: <a class="el" href="classcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel.html#ac82ba3da12b03bc91586a3947ce99fc5">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#a2d8d3a504dd8807ed09e25f37a658783">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="classcutlass_1_1reduction_1_1kernel_1_1ReduceSplitK.html#a9fed5689109358e708a27d487db15232">cutlass::reduction::kernel::ReduceSplitK&lt; Shape_, OutputOp_, ReductionOp_, PartitionsPerStage &gt;</a>
</li>
<li>ReductionOpPlus()
: <a class="el" href="classcutlass_1_1epilogue_1_1thread_1_1ReductionOpPlus.html#a06e34382d5f50a7331f723df7b3f709b">cutlass::epilogue::thread::ReductionOpPlus&lt; Element_, Count &gt;</a>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
</li>
<li>ReductionSize
: <a class="el" href="structcutlass_1_1reduction_1_1BatchedReductionTraits.html#a00c71c9a18aaad84f4a48023dbbb454e">cutlass::reduction::BatchedReductionTraits&lt; ScalarA_, ScalarC_, ScalarD_, ScalarAlphaBeta_, ScalarAccum_, ReductionSize_, OutputTile_, SubTile_, ThreadShape_, Index_, BlockSwizzle_, maxInReg_, maxOutReg_, Functor_ &gt;</a>
</li>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<li>ref()
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_1TensorView.html#a33385840adab818ef25076edab42175d">cutlass::TensorView&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1thread_1_1Matrix.html#a21e0504879de0558718e05977a7b8a73">cutlass::thread::Matrix&lt; Element, Rows, Columns, Layout &gt;</a>
</li>
<li>ref_A
: <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_1_1Arguments.html#a390abae1ca0e01a4b6e58f3724b48eed">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_3_01ElementA___00_01LayoutA___00_01ElementB___00_01Layou1b211cc9c97c022d8fe10f2dd32c8709.html#a9bdaf3563983efcca649460be169b334">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_1_1Arguments.html#a55f32be45559dbf84dcc2db26784f625">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_3_01ElementA___00_01LayoutA___00_01ElementB___00_213d78696663f4231cd52c6a277c60e5.html#a1727630fc0525724df28a75ccf2580b9">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_1_1Arguments.html#a654ced578699b96a0f805434afc5074c">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_3_01ElementA___00_01LayoutA___00_01ElementB___00_a3923967cafb5cb9774c320dc24baa77.html#ac8e9298e3786e9391d740faa4d0566f2">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_1_1Arguments.html#a3f4d6497cd54e624b3e410996a8b7d10">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_3_01ElementA___00_01LayoutA___00_01Elementafcb1aeaf2035a7ac769d7acc233423b.html#a8b10e75e5d6cd348dacc085f5264ee95">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#a3c4db6514188c51f63ee88130d9b9b0c">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#ad1867c0875c10e6327c7fae16acd35a3">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1Params.html#a475c24216aef2580f9b3405a1a42bfd3">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
</li>
<li>ref_a
: <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFuncBinaryOp.html#a77a01e6a9da3273bde175cf5df5d62b4">cutlass::reference::host::detail::TensorFuncBinaryOp&lt; ElementA, LayoutA, ElementB, LayoutB, ElementD, LayoutD, BinaryFunc &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<li>ref_B
: <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_1_1Arguments.html#ae712c362f83fbd45679a6e989315d3dc">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_3_01ElementA___00_01LayoutA___00_01ElementB___00_01Layou1b211cc9c97c022d8fe10f2dd32c8709.html#ab77204c1010b17c6643d26a89f41c3d0">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_1_1Arguments.html#a48844293c34b9c44fe57f577370664ea">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_3_01ElementA___00_01LayoutA___00_01ElementB___00_213d78696663f4231cd52c6a277c60e5.html#ad7d2b82b83d7503b9f920ce3bdcdffa5">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_1_1Arguments.html#a1eb0b3a45baf02021c1d0d12ad728e69">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_3_01ElementA___00_01LayoutA___00_01ElementB___00_a3923967cafb5cb9774c320dc24baa77.html#ab706387c660af35ae2b9579165eec85d">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_1_1Arguments.html#ac37c811b2f58c7756ef720fdfc0d072b">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_3_01ElementA___00_01LayoutA___00_01Elementafcb1aeaf2035a7ac769d7acc233423b.html#a9a22df7c4d515a48e03fd6f16e074217">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#ac9e6c1f13f20d925af51c682e2031a81">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#ade55adc311c5561efe76f53ffd56d1f4">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1Params.html#a2b67aca90833b8a5a85605f1902de539">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<li>ref_b
: <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorFuncBinaryOp.html#a416b2983d56932fa8970a7b988f3d7e6">cutlass::reference::host::detail::TensorFuncBinaryOp&lt; ElementA, LayoutA, ElementB, LayoutB, ElementD, LayoutD, BinaryFunc &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<li>ref_C
: <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_1_1Arguments.html#abdba57a68d6982fffbb1cc3db34ef0f9">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_3_01ElementA___00_01LayoutA___00_01ElementB___00_01Layou1b211cc9c97c022d8fe10f2dd32c8709.html#a590b8da88ae9350042838451e3e37a22">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_1_1Arguments.html#ab0955b722ad4ea0217f725e34b3bcfbe">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_3_01ElementA___00_01LayoutA___00_01ElementB___00_213d78696663f4231cd52c6a277c60e5.html#aa9e30e41627595590421d8b53941b2b2">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_1_1Arguments.html#a6d245eb700f43bb8bf0935e9c5ea2587">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_3_01ElementA___00_01LayoutA___00_01ElementB___00_a3923967cafb5cb9774c320dc24baa77.html#a3a59aa793429bc57d796b40fa4fab622">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_1_1Arguments.html#aa111536496495825c5d2ac2ea7360998">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_3_01ElementA___00_01LayoutA___00_01Elementafcb1aeaf2035a7ac769d7acc233423b.html#a2aabb13f196a087b77245c67c8664b7b">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#a37660d1a2a1031c44f0bb0c27d438ba3">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#a08ecd763b6785dfe872a6e517dc731e6">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
</li>
<li>ref_D
: <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_1_1Arguments.html#a66f9983db4a09ac0d90291c0f8723897">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1Gemm_3_01ElementA___00_01LayoutA___00_01ElementB___00_01Layou1b211cc9c97c022d8fe10f2dd32c8709.html#ab1d4d5865786a415f87db1def1b029e7">cutlass::gemm::device::Gemm&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, SplitKSerial, Operator_, IsBetaZero &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_1_1Arguments.html#ae4450f06a6975191d94026865e445578">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmBatched_3_01ElementA___00_01LayoutA___00_01ElementB___00_213d78696663f4231cd52c6a277c60e5.html#a17c4e381e91229a8ef15b18ee5ec073d">cutlass::gemm::device::GemmBatched&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, AlignmentA, AlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_1_1Arguments.html#a19d893bbaaef4122d9a66a74bb7fa21f">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmComplex_3_01ElementA___00_01LayoutA___00_01ElementB___00_a3923967cafb5cb9774c320dc24baa77.html#a2904e3ad7a47b3d85ea60d94eeebe84b">cutlass::gemm::device::GemmComplex&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ThreadblockSwizzle_, Stages, TransformA, TransformB, SplitKSerial &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_1_1Arguments.html#a1550cfad6be8e22c9dc13b05f5601845">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, LayoutC_, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1device_1_1GemmSplitKParallel_3_01ElementA___00_01LayoutA___00_01Elementafcb1aeaf2035a7ac769d7acc233423b.html#a850da307d8741296e515add0f716eaf9">cutlass::gemm::device::GemmSplitKParallel&lt; ElementA_, LayoutA_, ElementB_, LayoutB_, ElementC_, layout::ColumnMajor, ElementAccumulator_, OperatorClass_, ArchTag_, ThreadblockShape_, WarpShape_, InstructionShape_, EpilogueOutputOp_, ConvertScaledOp_, ReductionOp_, ThreadblockSwizzle_, Stages, kAlignmentA, kAlignmentB, Operator_ &gt;::Arguments</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1Gemm_1_1Params.html#a33618c431b2f6a6730c8ab1f1c1a590f">cutlass::gemm::kernel::Gemm&lt; Mma_, Epilogue_, ThreadblockSwizzle_, SplitKSerial &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmBatched_1_1Params.html#a4f18093b18b0b6dd01a5df0a3813cd40">cutlass::gemm::kernel::GemmBatched&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
, <a class="el" href="structcutlass_1_1gemm_1_1kernel_1_1GemmSplitKParallel_1_1Params.html#ab30d65bd2883bcf15a37a404df944aa4">cutlass::gemm::kernel::GemmSplitKParallel&lt; Mma_, Epilogue_, ThreadblockSwizzle_ &gt;::Params</a>
</li>
<li>reference
: <a class="el" href="structcutlass_1_1AlignedBuffer.html#afa029189fb46528b5eb5f50060cbf28e">cutlass::AlignedBuffer&lt; T, N, Align &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1reference.html#a257c25bee7fa54ff1d492bc0697b05cc">cutlass::Array&lt; T, N, false &gt;::reference</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#a5827968c9c3deca639f5981ad895fe67">cutlass::Array&lt; T, N, true &gt;</a>
, <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1EpilogueBase_1_1SharedStorage.html#a99b724aec92bebece9bc187fb636bc11">cutlass::epilogue::threadblock::EpilogueBase&lt; Shape_, WarpMmaOperator_, PartitionsK, AccumulatorFragmentIterator_, WarpTileIterator_, Padding_ &gt;::SharedStorage</a>
</li>
<li>Reference
: <a class="el" href="classcutlass_1_1HostTensor.html#a33698c7aa33255e7a2e7abc298e28f39">cutlass::HostTensor&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorRef.html#a0aba29ebf8715d217817a49356f95d3f">cutlass::TensorRef&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorView.html#a4d8af7fd842866a218722e7686d6bc3c">cutlass::TensorView&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1thread_1_1Matrix.html#a3c9535ffa08c91d0040b09667c81b201">cutlass::thread::Matrix&lt; Element, Rows, Columns, Layout &gt;</a>
</li>
<li>RegularTileAccessIterator()
: <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileAccessIterator_3_01Shape___00_01Element__eb7d20f8b9d69e0ae5e7ef51dc480867.html#a993c3d66e26ad80da41290a77c49134b">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#a633db64682aa6494c2553e11614b5e1d">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#acc48b323a8a9a8e75e0bfdb77c1cbcc5">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#ad71831a7593c13375f067f8acd048f35">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#a64df8a853b101bd8acddc06d4a973da5">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#abf23348f7e8d3b8aaac6f23c116dcf5c">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#a025187783fb6d708885d31af2492bca4">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#ae4927c7f718b52aa80f603d9911b9639">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#a0f1e846cd4287c51e2d6135a8472c1a8">cutlass::transform::threadblock::RegularTileAccessIterator&lt; Shape_, Element_, layout::TensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Crosswise &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
</li>
<li>RegularTileIterator()
: <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator_3_01Shape___00_01Element___00_011d3637dbd8bc58bcb020b51bf57fbfc0.html#a90dedd03e47edabf0f9fd6d51e70272e">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#a87b0a8a59b5e46f5bcb761b1d8fd620e">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#a0295a753f239eab411f239cd75e05a91">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#a25253a14817a446458b519e368d41d11">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#af66b203bacaabdfe217f003dc256ba7a">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#ac2ccb3732d24e95444a186167dfc122d">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#a77af9f2f9a530fbd63703bb83190135b">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#aad803b85a96da22725a335b911975718">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#a8eff0cc46aa62dd1014eeaaccafdaec1">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#a55a9941eebde58406226a94aab1cdc37">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#a0d0e54f204b90ddf9af598655441864b">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#a2b42db145a79615c5eb6a35c92f06048">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#ab620e71608ef128668b2ac5cc4cb9abb">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#a7786e21804d73da02c2fce5be1b36c99">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#a7bed1f96a9586042f4b69771da9c67e4">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#aa72dc3870a62c972b4abdb085724d6a0">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#a134de9f5c72872a443ce7a051df6f489">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#abc4b12b3974ae4de423d20ef902e3bf3">cutlass::transform::threadblock::RegularTileIterator&lt; Shape_, Element_, layout::VoltaTensorOpMultiplicandCrosswise&lt; sizeof_bits&lt; Element_ &gt;::value, Shape_::kContiguous &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
</li>
<li>RegularTileIterator2dThreadTile()
: <a class="el" href="classcutlass_1_1transform_1_1threadblock_1_1RegularTileIterator2dThreadTile_3_01Shape___00_01Eleb60d066756d1c18f05fceee6a27bdb8a.html#aedc469dde216ca1b0471c1008b0662da">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#a5b5259960083255ea957de82f459d044">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#abc31b5fb85bf3f84818ec89fca787362">cutlass::transform::threadblock::RegularTileIterator2dThreadTile&lt; Shape_, Element_, layout::RowMajorInterleaved&lt; 4 &gt;, AdvanceRank, ThreadMap_, Alignment &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
<li>release()
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#a138bc02627aab06ac09c73da67f06f7a">cutlass::device_memory::allocation&lt; T &gt;</a>
, <a class="el" href="classcutlass_1_1library_1_1Manifest.html#a05c62e11e5face82a64f40e7547d1ac5">cutlass::library::Manifest</a>
, <a class="el" href="classcutlass_1_1platform_1_1unique__ptr.html#a7ac06ebe7bc66573d3225891e12d2279">cutlass::platform::unique_ptr&lt; T, Deleter &gt;</a>
, <a class="el" href="classcutlass_1_1Semaphore.html#a04e893ba5a9ddb20e1b3c6475771c0e9">cutlass::Semaphore</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>remove_negative_zero()
: <a class="el" href="structcutlass_1_1TypeTraits.html#acdf0f64b79e6b8c550e85376232723a8">cutlass::TypeTraits&lt; T &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01complex_3_01double_01_4_01_4.html#a116d619c589a5c64c57bbe01d880962a">cutlass::TypeTraits&lt; complex&lt; double &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01complex_3_01float_01_4_01_4.html#a9c5132250bbcfbe3e7747277b875c6bf">cutlass::TypeTraits&lt; complex&lt; float &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01complex_3_01half__t_01_4_01_4.html#ac3cad0e8f771e936bfa2a64df50982dc">cutlass::TypeTraits&lt; complex&lt; half_t &gt; &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01double_01_4.html#afa38cd9c83856851bf305e15e424b057">cutlass::TypeTraits&lt; double &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01float_01_4.html#a7a81197b153f223cf985e3fae743d437">cutlass::TypeTraits&lt; float &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01half__t_01_4.html#afd4eed55e9957fabbf513c9556973e04">cutlass::TypeTraits&lt; half_t &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01int_01_4.html#a4fec4afa5e9b0aa26b7026e4345d1549">cutlass::TypeTraits&lt; int &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01int64__t_01_4.html#a5a3dbcca8e7766783c0bb55a20e3a822">cutlass::TypeTraits&lt; int64_t &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01int8__t_01_4.html#acb1301812bda85b78dde745889aa0d98">cutlass::TypeTraits&lt; int8_t &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01uint64__t_01_4.html#afeda28af3151a5de60ecb0de5b525dd1">cutlass::TypeTraits&lt; uint64_t &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01uint8__t_01_4.html#a11a6931adbdd24743a919c999f44eda9">cutlass::TypeTraits&lt; uint8_t &gt;</a>
, <a class="el" href="structcutlass_1_1TypeTraits_3_01unsigned_01_4.html#ac596d039114e2c2618da81c0076faa7e">cutlass::TypeTraits&lt; unsigned &gt;</a>
</li>
<li>rend()
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4.html#a39c08a75c7cc22fcd296e6c9fefe754e">cutlass::Array&lt; T, N, false &gt;</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4.html#a6081f288dfc7b60da8d00913be8e83db">cutlass::Array&lt; T, N, true &gt;</a>
</li>
<li>reserve()
: <a class="el" href="classcutlass_1_1HostTensor.html#a6405f217fc2a52f8e23180c9f8160899">cutlass::HostTensor&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1library_1_1Manifest.html#a9dc44049b0d4cd3331b07985bd5b527e">cutlass::library::Manifest</a>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
</li>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<li>reset()
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#aeef3a1cb8cfb96a2823e85f7fa52a31a">cutlass::device_memory::allocation&lt; T &gt;</a>
, <a class="el" href="classcutlass_1_1HostTensor.html#a4f23ad5953f9cd5dff8e97d9edbc0cf0">cutlass::HostTensor&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1platform_1_1unique__ptr.html#a6740f71511f5495d6038cf8878862331">cutlass::platform::unique_ptr&lt; T, Deleter &gt;</a>
, <a class="el" href="classcutlass_1_1TensorRef.html#a1c539841446205c482fe18cc0a99d913">cutlass::TensorRef&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorView.html#add0de4f0548c957df00cae0ee0b9257c">cutlass::TensorView&lt; Element_, Layout_ &gt;</a>
</li>
<li>resize()
: <a class="el" href="classcutlass_1_1HostTensor.html#ac61cee5b25585439f4ccc7d7249ee853">cutlass::HostTensor&lt; Element_, Layout_ &gt;</a>
, <a class="el" href="classcutlass_1_1TensorView.html#a3c3417b9b777f103ec86607a7751e668">cutlass::TensorView&lt; Element_, Layout_ &gt;</a>
</li>
<li>result
: <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorEqualsFunc.html#a71114c0e6c154a9cc3f5a1e380dfc454">cutlass::reference::host::detail::TensorEqualsFunc&lt; Element, Layout &gt;</a>
</li>
<li>result_type
: <a class="el" href="structcutlass_1_1NumericArrayConverter.html#a5ed5dd75f4310d887fa527dd177c6f91">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#af14ce6b66b30bd8942a6693e6a06c8f9">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#a644b3f49bf10e99d6a53061b350ac693">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#ad50d5ce8c7047513745c1fab77c3988c">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#a411c3cd15a3f03d360c96c05025cc3d3">cutlass::NumericArrayConverter&lt; half_t, float, N, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverter.html#a46b1b5c0c96c50176578fcd6f915ee8c">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#ac1adcd31b0db52e8682680c9927a05c8">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#a5dc993f38c6eedd917008e6c839c6300">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#aa5bfe0288e538f1df94d74fa52aa1e17">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#acce8af4bfd5837006dc66803fee491ca">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#a10b599dbfd3f4da945dc3ddb93e48ded">cutlass::NumericConverter&lt; T, T, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverterClamp.html#a8e13f626ae7a0ebb228bf1ef59d4ed09">cutlass::NumericConverterClamp&lt; T, S &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>reverse_iterator()
: <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01false_01_4_1_1reverse__iterator.html#a939c336c7c727748d9efcd5efa066a88">cutlass::Array&lt; T, N, false &gt;::reverse_iterator</a>
, <a class="el" href="classcutlass_1_1Array_3_01T_00_01N_00_01true_01_4_1_1reverse__iterator.html#af5056ae2c80aadeb027fffb8ed35f719">cutlass::Array&lt; T, N, true &gt;::reverse_iterator</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>rhs
: <a class="el" href="structcutlass_1_1reference_1_1host_1_1detail_1_1TensorEqualsFunc.html#a15a7b08ee45105dc3d40b86426824c4c">cutlass::reference::host::detail::TensorEqualsFunc&lt; Element, Layout &gt;</a>
</li>
<li>rng_state
: <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomGaussianFunc.html#a52dd271db62c366ac41e84407b9176c3">cutlass::reference::device::detail::RandomGaussianFunc&lt; Element &gt;</a>
, <a class="el" href="structcutlass_1_1reference_1_1device_1_1detail_1_1RandomUniformFunc.html#aa802faaaf5a6b3f7a5725d26a9d45ef2">cutlass::reference::device::detail::RandomUniformFunc&lt; Element &gt;</a>
</li>
<li>round_error()
: <a class="el" href="structstd_1_1numeric__limits_3_01cutlass_1_1half__t_01_4.html#ab0a036db7a1ad11c65e876020c78b1a5">std::numeric_limits&lt; cutlass::half_t &gt;</a>
</li>
<li>round_style
: <a class="el" href="structcutlass_1_1NumericArrayConverter.html#a3bb36be77b3f0464ab7d1c52b16db6f7">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#a07700d779e1dbfc62f29157e06073f79">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#a5ae106211f42197a8f39a2f037ab6bba">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#a3c31373beb0e6a9c649134b21a02125a">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#ac5060dfecda3f5a29600cfbd2c01def5">cutlass::NumericArrayConverter&lt; half_t, float, N, Round &gt;</a>
, <a class="el" href="structcutlass_1_1NumericConverter.html#a6060b5f316853f791fc87ef3bd180b1a">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#a59f158bedaecd4a6f0d50c7f1567538a">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#affd38515f30c26256ff5c06e5a567080">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#a43fde0bc2ddeeebf1f188c6d1ac7fbe0">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#ae66740da44250e5360e4c3c874634310">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#a5b9a558d800872eb73723e5af3e7d4f3">cutlass::NumericConverter&lt; T, T, Round &gt;</a>
, <a class="el" href="structstd_1_1numeric__limits_3_01cutlass_1_1half__t_01_4.html#ab0af85c1d7c83ca03ed0c083fe22262f">std::numeric_limits&lt; cutlass::half_t &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
<li>row()
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
: <a class="el" href="structcutlass_1_1MatrixCoord.html#a0580610f28427e376b24b71f67602d03">cutlass::MatrixCoord</a>
</li>
<li>RowArrangement
: <a class="el" href="structcutlass_1_1epilogue_1_1threadblock_1_1OutputTileOptimalThreadMap_1_1Detail.html#a028dd493fdfc53f5189aa9936e170941">cutlass::epilogue::threadblock::OutputTileOptimalThreadMap&lt; Shape_, Count_, Threads, ElementsPerAccess, ElementSize &gt;::Detail</a>
</li>
<li>RowMajor()
: <a class="el" href="classcutlass_1_1layout_1_1RowMajor.html#a1fe768ae500e1960013c2cb95e58dce8">cutlass::layout::RowMajor</a>
</li>
<li>RowMajorBlockLinear()
: <a class="el" href="structcutlass_1_1layout_1_1RowMajorBlockLinear.html#ac9c5464ff08cbf9a00a9bd6e87b4367e">cutlass::layout::RowMajorBlockLinear&lt; BlockRows, BlockColumns &gt;</a>
</li>
<li>RowMajorInterleaved()
: <a class="el" href="structcutlass_1_1layout_1_1RowMajorInterleaved.html#a52b330ac2b80e6e38dfde427114baeea">cutlass::layout::RowMajorInterleaved&lt; Interleave &gt;</a>
</li>
<li>RowMajorTensorOpMultiplicandCongruous()
: <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCongruous.html#acd7ea2b943914f72f0040b208915aa89">cutlass::layout::RowMajorTensorOpMultiplicandCongruous&lt; ElementSize, Crosswise &gt;</a>
</li>
<li>RowMajorTensorOpMultiplicandCrosswise()
: <a class="el" href="structcutlass_1_1layout_1_1RowMajorTensorOpMultiplicandCrosswise.html#a90513da18bdcebc20f673d83031551f9">cutlass::layout::RowMajorTensorOpMultiplicandCrosswise&lt; ElementSize, Crosswise &gt;</a>
</li>
<li>RowMajorVoltaTensorOpMultiplicandBCongruous()
: <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandBCongruous.html#aad4ca016701d73fee071bf10c35219ab">cutlass::layout::RowMajorVoltaTensorOpMultiplicandBCongruous&lt; ElementSize &gt;</a>
</li>
<li>RowMajorVoltaTensorOpMultiplicandCongruous()
: <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandCongruous.html#acefc319105f8150e6005d2a1a7ce9366">cutlass::layout::RowMajorVoltaTensorOpMultiplicandCongruous&lt; ElementSize &gt;</a>
Updated doxygen
2018-09-20 05:02:08 +08:00
</li>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
<li>RowMajorVoltaTensorOpMultiplicandCrosswise()
: <a class="el" href="structcutlass_1_1layout_1_1RowMajorVoltaTensorOpMultiplicandCrosswise.html#a602a5e3cd52edc8c6e733cc8ea271484">cutlass::layout::RowMajorVoltaTensorOpMultiplicandCrosswise&lt; ElementSize, KBlock &gt;</a>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</li>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
<li>run()
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_1device_1_1Gemm.html#aef8c133e539ef91efc9dba9012118147">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#a5f4f93ca97b358b4410f3d0b1e0a6387">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#ad7e9e393be872e401a5a777ceda529d9">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#abcae3d15f1ec2ee7ae93690c82fbee8a">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#ac6cfa33deb7e829c4122543aad00964d">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#a4111bba1e9d2000fcc9bba2f114ee801">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#aeadc76210d06ec22776aca4a58de9930">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#a1de7cf5d8bad27b3ff6c803dbc572077">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="classcutlass_1_1library_1_1Operation.html#ae4a336419eb6dba60e3ffc04ef7ebdf3">cutlass::library::Operation</a>
, <a class="el" href="structcutlass_1_1reduction_1_1BatchedReduction.html#a4bfcdfd8f2edeb4fc7081443584d599b">cutlass::reduction::BatchedReduction&lt; BatchedReductionTraits_ &gt;</a>
Updating Doxygen docs
2018-10-27 05:54:58 +08:00
</li>
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</ul>
</div><!-- contents -->
<!-- start footer part -->
<hr class="footer"/><address class="footer"><small>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
Generated by &#160;<a href="http://www.doxygen.org/index.html">
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
<img class="footer" src="doxygen.png" alt="doxygen"/>
CUTLASS 2.0 (#62) CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-20 08:55:34 +08:00
</a> 1.8.11
CUTLASS v1.0 release
2018-05-17 02:44:56 +08:00
</small></address>
</body>
</html>
Reference in New Issue Copy Permalink