mirror of
https://github.com/facebookresearch/faiss.git
synced 2025-06-03 21:54:02 +08:00
0df3f646fc
various bugfixes from github issues kmean with some frozen centroids GPU better tiling for large flat datasets default AVX for vector ops
262 lines
29 KiB
HTML
262 lines
29 KiB
HTML
<!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"/>
|
|
<meta name="generator" content="Doxygen 1.8.5"/>
|
|
<title>Faiss: /data/users/matthijs/github_faiss/faiss/gpu/utils/Transpose.cuh Source File</title>
|
|
<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/search.js"></script>
|
|
<script type="text/javascript">
|
|
$(document).ready(function() { searchBox.OnSelectItem(0); });
|
|
</script>
|
|
<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;">
|
|
<td style="padding-left: 0.5em;">
|
|
<div id="projectname">Faiss
|
|
</div>
|
|
</td>
|
|
</tr>
|
|
</tbody>
|
|
</table>
|
|
</div>
|
|
<!-- end header part -->
|
|
<!-- Generated by Doxygen 1.8.5 -->
|
|
<script type="text/javascript">
|
|
var searchBox = new SearchBox("searchBox", "search",false,'Search');
|
|
</script>
|
|
<div id="navrow1" class="tabs">
|
|
<ul class="tablist">
|
|
<li><a href="index.html"><span>Main Page</span></a></li>
|
|
<li><a href="namespaces.html"><span>Namespaces</span></a></li>
|
|
<li><a href="annotated.html"><span>Classes</span></a></li>
|
|
<li class="current"><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="files.html"><span>File List</span></a></li>
|
|
</ul>
|
|
</div>
|
|
<!-- window showing the filter options -->
|
|
<div id="MSearchSelectWindow"
|
|
onmouseover="return searchBox.OnSearchSelectShow()"
|
|
onmouseout="return searchBox.OnSearchSelectHide()"
|
|
onkeydown="return searchBox.OnSearchSelectKey(event)">
|
|
<a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(0)"><span class="SelectionMark"> </span>All</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(1)"><span class="SelectionMark"> </span>Classes</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(2)"><span class="SelectionMark"> </span>Namespaces</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(3)"><span class="SelectionMark"> </span>Functions</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(4)"><span class="SelectionMark"> </span>Variables</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(5)"><span class="SelectionMark"> </span>Typedefs</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(6)"><span class="SelectionMark"> </span>Enumerations</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(7)"><span class="SelectionMark"> </span>Enumerator</a><a class="SelectItem" href="javascript:void(0)" onclick="searchBox.OnSelectItem(8)"><span class="SelectionMark"> </span>Friends</a></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 id="nav-path" class="navpath">
|
|
<ul>
|
|
<li class="navelem"><a class="el" href="dir_6b3ae6988449b0834e9596fad5d75199.html">gpu</a></li><li class="navelem"><a class="el" href="dir_498271007b03b2a0521055e88776887b.html">utils</a></li> </ul>
|
|
</div>
|
|
</div><!-- top -->
|
|
<div class="header">
|
|
<div class="headertitle">
|
|
<div class="title">Transpose.cuh</div> </div>
|
|
</div><!--header-->
|
|
<div class="contents">
|
|
<div class="fragment"><div class="line"><a name="l00001"></a><span class="lineno"> 1</span> <span class="comment">/**</span></div>
|
|
<div class="line"><a name="l00002"></a><span class="lineno"> 2</span> <span class="comment"> * Copyright (c) 2015-present, Facebook, Inc.</span></div>
|
|
<div class="line"><a name="l00003"></a><span class="lineno"> 3</span> <span class="comment"> * All rights reserved.</span></div>
|
|
<div class="line"><a name="l00004"></a><span class="lineno"> 4</span> <span class="comment"> *</span></div>
|
|
<div class="line"><a name="l00005"></a><span class="lineno"> 5</span> <span class="comment"> * This source code is licensed under the BSD+Patents license found in the</span></div>
|
|
<div class="line"><a name="l00006"></a><span class="lineno"> 6</span> <span class="comment"> * LICENSE file in the root directory of this source tree.</span></div>
|
|
<div class="line"><a name="l00007"></a><span class="lineno"> 7</span> <span class="comment"> */</span></div>
|
|
<div class="line"><a name="l00008"></a><span class="lineno"> 8</span> </div>
|
|
<div class="line"><a name="l00009"></a><span class="lineno"> 9</span> <span class="comment">// Copyright 2004-present Facebook. All Rights Reserved.</span></div>
|
|
<div class="line"><a name="l00010"></a><span class="lineno"> 10</span> </div>
|
|
<div class="line"><a name="l00011"></a><span class="lineno"> 11</span> <span class="preprocessor">#pragma once</span></div>
|
|
<div class="line"><a name="l00012"></a><span class="lineno"> 12</span> <span class="preprocessor"></span></div>
|
|
<div class="line"><a name="l00013"></a><span class="lineno"> 13</span> <span class="preprocessor">#include "../../FaissAssert.h"</span></div>
|
|
<div class="line"><a name="l00014"></a><span class="lineno"> 14</span> <span class="preprocessor">#include "Tensor.cuh"</span></div>
|
|
<div class="line"><a name="l00015"></a><span class="lineno"> 15</span> <span class="preprocessor">#include "DeviceUtils.h"</span></div>
|
|
<div class="line"><a name="l00016"></a><span class="lineno"> 16</span> <span class="preprocessor">#include <cuda.h></span></div>
|
|
<div class="line"><a name="l00017"></a><span class="lineno"> 17</span> </div>
|
|
<div class="line"><a name="l00018"></a><span class="lineno"> 18</span> <span class="preprocessor">#include <stdio.h></span></div>
|
|
<div class="line"><a name="l00019"></a><span class="lineno"> 19</span> </div>
|
|
<div class="line"><a name="l00020"></a><span class="lineno"> 20</span> <span class="keyword">namespace </span>faiss { <span class="keyword">namespace </span>gpu {</div>
|
|
<div class="line"><a name="l00021"></a><span class="lineno"> 21</span> </div>
|
|
<div class="line"><a name="l00022"></a><span class="lineno"> 22</span> <span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> IndexT></div>
|
|
<div class="line"><a name="l00023"></a><span class="lineno"><a class="line" href="structfaiss_1_1gpu_1_1TensorInfo.html"> 23</a></span> <span class="keyword">struct </span><a class="code" href="structfaiss_1_1gpu_1_1TensorInfo.html">TensorInfo</a> {</div>
|
|
<div class="line"><a name="l00024"></a><span class="lineno"> 24</span>  <span class="keyword">static</span> constexpr <span class="keywordtype">int</span> kMaxDims = 8;</div>
|
|
<div class="line"><a name="l00025"></a><span class="lineno"> 25</span> </div>
|
|
<div class="line"><a name="l00026"></a><span class="lineno"> 26</span>  T* data;</div>
|
|
<div class="line"><a name="l00027"></a><span class="lineno"> 27</span>  IndexT sizes[kMaxDims];</div>
|
|
<div class="line"><a name="l00028"></a><span class="lineno"> 28</span>  IndexT strides[kMaxDims];</div>
|
|
<div class="line"><a name="l00029"></a><span class="lineno"> 29</span>  <span class="keywordtype">int</span> dims;</div>
|
|
<div class="line"><a name="l00030"></a><span class="lineno"> 30</span> };</div>
|
|
<div class="line"><a name="l00031"></a><span class="lineno"> 31</span> </div>
|
|
<div class="line"><a name="l00032"></a><span class="lineno"> 32</span> <span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> IndexT, <span class="keywordtype">int</span> Dim></div>
|
|
<div class="line"><a name="l00033"></a><span class="lineno"><a class="line" href="structfaiss_1_1gpu_1_1TensorInfoOffset.html"> 33</a></span> <span class="keyword">struct </span><a class="code" href="structfaiss_1_1gpu_1_1TensorInfoOffset.html">TensorInfoOffset</a> {</div>
|
|
<div class="line"><a name="l00034"></a><span class="lineno"> 34</span>  __device__ <span class="keyword">inline</span> <span class="keyword">static</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">get</span>(<span class="keyword">const</span> <a class="code" href="structfaiss_1_1gpu_1_1TensorInfo.html">TensorInfo<T, IndexT></a>& info,</div>
|
|
<div class="line"><a name="l00035"></a><span class="lineno"> 35</span>  IndexT linearId) {</div>
|
|
<div class="line"><a name="l00036"></a><span class="lineno"> 36</span>  IndexT offset = 0;</div>
|
|
<div class="line"><a name="l00037"></a><span class="lineno"> 37</span> </div>
|
|
<div class="line"><a name="l00038"></a><span class="lineno"> 38</span> <span class="preprocessor">#pragma unroll</span></div>
|
|
<div class="line"><a name="l00039"></a><span class="lineno"> 39</span> <span class="preprocessor"></span> <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = Dim - 1; i >= 0; --i) {</div>
|
|
<div class="line"><a name="l00040"></a><span class="lineno"> 40</span>  IndexT curDimIndex = linearId % info.sizes[i];</div>
|
|
<div class="line"><a name="l00041"></a><span class="lineno"> 41</span>  IndexT curDimOffset = curDimIndex * info.strides[i];</div>
|
|
<div class="line"><a name="l00042"></a><span class="lineno"> 42</span> </div>
|
|
<div class="line"><a name="l00043"></a><span class="lineno"> 43</span>  offset += curDimOffset;</div>
|
|
<div class="line"><a name="l00044"></a><span class="lineno"> 44</span> </div>
|
|
<div class="line"><a name="l00045"></a><span class="lineno"> 45</span>  <span class="keywordflow">if</span> (i > 0) {</div>
|
|
<div class="line"><a name="l00046"></a><span class="lineno"> 46</span>  linearId /= info.sizes[i];</div>
|
|
<div class="line"><a name="l00047"></a><span class="lineno"> 47</span>  }</div>
|
|
<div class="line"><a name="l00048"></a><span class="lineno"> 48</span>  }</div>
|
|
<div class="line"><a name="l00049"></a><span class="lineno"> 49</span> </div>
|
|
<div class="line"><a name="l00050"></a><span class="lineno"> 50</span>  <span class="keywordflow">return</span> offset;</div>
|
|
<div class="line"><a name="l00051"></a><span class="lineno"> 51</span>  }</div>
|
|
<div class="line"><a name="l00052"></a><span class="lineno"> 52</span> };</div>
|
|
<div class="line"><a name="l00053"></a><span class="lineno"> 53</span> </div>
|
|
<div class="line"><a name="l00054"></a><span class="lineno"> 54</span> <span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> IndexT></div>
|
|
<div class="line"><a name="l00055"></a><span class="lineno"><a class="line" href="structfaiss_1_1gpu_1_1TensorInfoOffset_3_01T_00_01IndexT_00-1_01_4.html"> 55</a></span> <span class="keyword">struct </span><a class="code" href="structfaiss_1_1gpu_1_1TensorInfoOffset.html">TensorInfoOffset</a><T, IndexT, -1> {</div>
|
|
<div class="line"><a name="l00056"></a><span class="lineno"> 56</span>  __device__ <span class="keyword">inline</span> <span class="keyword">static</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <span class="keyword">get</span>(<span class="keyword">const</span> <a class="code" href="structfaiss_1_1gpu_1_1TensorInfo.html">TensorInfo<T, IndexT></a>& info,</div>
|
|
<div class="line"><a name="l00057"></a><span class="lineno"> 57</span>  IndexT linearId) {</div>
|
|
<div class="line"><a name="l00058"></a><span class="lineno"> 58</span>  <span class="keywordflow">return</span> linearId;</div>
|
|
<div class="line"><a name="l00059"></a><span class="lineno"> 59</span>  }</div>
|
|
<div class="line"><a name="l00060"></a><span class="lineno"> 60</span> };</div>
|
|
<div class="line"><a name="l00061"></a><span class="lineno"> 61</span> </div>
|
|
<div class="line"><a name="l00062"></a><span class="lineno"> 62</span> <span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> IndexT, <span class="keywordtype">int</span> Dim></div>
|
|
<div class="line"><a name="l00063"></a><span class="lineno"> 63</span> <a class="code" href="structfaiss_1_1gpu_1_1TensorInfo.html">TensorInfo<T, IndexT></a> getTensorInfo(<span class="keyword">const</span> <a class="code" href="classfaiss_1_1gpu_1_1Tensor.html">Tensor<T, Dim, true></a>& t) {</div>
|
|
<div class="line"><a name="l00064"></a><span class="lineno"> 64</span>  <a class="code" href="structfaiss_1_1gpu_1_1TensorInfo.html">TensorInfo<T, IndexT></a> info;</div>
|
|
<div class="line"><a name="l00065"></a><span class="lineno"> 65</span> </div>
|
|
<div class="line"><a name="l00066"></a><span class="lineno"> 66</span>  <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0; i < Dim; ++i) {</div>
|
|
<div class="line"><a name="l00067"></a><span class="lineno"> 67</span>  info.sizes[i] = (IndexT) t.<a class="code" href="classfaiss_1_1gpu_1_1Tensor.html#a6699c311648457f257afa340c61f417c">getSize</a>(i);</div>
|
|
<div class="line"><a name="l00068"></a><span class="lineno"> 68</span>  info.strides[i] = (IndexT) t.<a class="code" href="classfaiss_1_1gpu_1_1Tensor.html#a0b8bba630f7a1fa217f90b20d298420a">getStride</a>(i);</div>
|
|
<div class="line"><a name="l00069"></a><span class="lineno"> 69</span>  }</div>
|
|
<div class="line"><a name="l00070"></a><span class="lineno"> 70</span> </div>
|
|
<div class="line"><a name="l00071"></a><span class="lineno"> 71</span>  info.data = t.<a class="code" href="classfaiss_1_1gpu_1_1Tensor.html#a50411ce4d0fa32ef715e3321b6e33212">data</a>();</div>
|
|
<div class="line"><a name="l00072"></a><span class="lineno"> 72</span>  info.dims = Dim;</div>
|
|
<div class="line"><a name="l00073"></a><span class="lineno"> 73</span> </div>
|
|
<div class="line"><a name="l00074"></a><span class="lineno"> 74</span>  <span class="keywordflow">return</span> info;</div>
|
|
<div class="line"><a name="l00075"></a><span class="lineno"> 75</span> }</div>
|
|
<div class="line"><a name="l00076"></a><span class="lineno"> 76</span> </div>
|
|
<div class="line"><a name="l00077"></a><span class="lineno"> 77</span> <span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keyword">typename</span> IndexT, <span class="keywordtype">int</span> DimInput, <span class="keywordtype">int</span> DimOutput></div>
|
|
<div class="line"><a name="l00078"></a><span class="lineno"> 78</span> __global__ <span class="keywordtype">void</span> transposeAny(TensorInfo<T, IndexT> input,</div>
|
|
<div class="line"><a name="l00079"></a><span class="lineno"> 79</span>  TensorInfo<T, IndexT> output,</div>
|
|
<div class="line"><a name="l00080"></a><span class="lineno"> 80</span>  IndexT totalSize) {</div>
|
|
<div class="line"><a name="l00081"></a><span class="lineno"> 81</span>  <span class="keywordflow">for</span> (IndexT i = blockIdx.x * blockDim.x + threadIdx.x;</div>
|
|
<div class="line"><a name="l00082"></a><span class="lineno"> 82</span>  i < totalSize;</div>
|
|
<div class="line"><a name="l00083"></a><span class="lineno"> 83</span>  i += gridDim.x + blockDim.x) {</div>
|
|
<div class="line"><a name="l00084"></a><span class="lineno"> 84</span>  <span class="keyword">auto</span> inputOffset = TensorInfoOffset<T, IndexT, DimInput>::get(input, i);</div>
|
|
<div class="line"><a name="l00085"></a><span class="lineno"> 85</span>  <span class="keyword">auto</span> outputOffset = TensorInfoOffset<T, IndexT, DimOutput>::get(output, i);</div>
|
|
<div class="line"><a name="l00086"></a><span class="lineno"> 86</span> </div>
|
|
<div class="line"><a name="l00087"></a><span class="lineno"> 87</span> <span class="preprocessor">#if __CUDA_ARCH__ >= 350</span></div>
|
|
<div class="line"><a name="l00088"></a><span class="lineno"> 88</span> <span class="preprocessor"></span> output.data[outputOffset] = __ldg(&input.data[inputOffset]);</div>
|
|
<div class="line"><a name="l00089"></a><span class="lineno"> 89</span> <span class="preprocessor">#else</span></div>
|
|
<div class="line"><a name="l00090"></a><span class="lineno"> 90</span> <span class="preprocessor"></span> output.data[outputOffset] = input.data[inputOffset];</div>
|
|
<div class="line"><a name="l00091"></a><span class="lineno"> 91</span> <span class="preprocessor">#endif</span></div>
|
|
<div class="line"><a name="l00092"></a><span class="lineno"> 92</span> <span class="preprocessor"></span> }</div>
|
|
<div class="line"><a name="l00093"></a><span class="lineno"> 93</span> }</div>
|
|
<div class="line"><a name="l00094"></a><span class="lineno"> 94</span> <span class="comment"></span></div>
|
|
<div class="line"><a name="l00095"></a><span class="lineno"> 95</span> <span class="comment">/// Performs an out-of-place transposition between any two dimensions.</span></div>
|
|
<div class="line"><a name="l00096"></a><span class="lineno"> 96</span> <span class="comment">/// Best performance is if the transposed dimensions are not</span></div>
|
|
<div class="line"><a name="l00097"></a><span class="lineno"> 97</span> <span class="comment">/// innermost, since the reads and writes will be coalesced.</span></div>
|
|
<div class="line"><a name="l00098"></a><span class="lineno"> 98</span> <span class="comment">/// Could include a shared memory transposition if the dimensions</span></div>
|
|
<div class="line"><a name="l00099"></a><span class="lineno"> 99</span> <span class="comment">/// being transposed are innermost, but would require support for</span></div>
|
|
<div class="line"><a name="l00100"></a><span class="lineno"> 100</span> <span class="comment">/// arbitrary rectangular matrices.</span></div>
|
|
<div class="line"><a name="l00101"></a><span class="lineno"> 101</span> <span class="comment">/// This linearized implementation seems to perform well enough,</span></div>
|
|
<div class="line"><a name="l00102"></a><span class="lineno"> 102</span> <span class="comment">/// especially for cases that we care about (outer dimension</span></div>
|
|
<div class="line"><a name="l00103"></a><span class="lineno"> 103</span> <span class="comment">/// transpositions).</span></div>
|
|
<div class="line"><a name="l00104"></a><span class="lineno"> 104</span> <span class="comment"></span><span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keywordtype">int</span> Dim></div>
|
|
<div class="line"><a name="l00105"></a><span class="lineno"> 105</span> <span class="keywordtype">void</span> runTransposeAny(Tensor<T, Dim, true>& in,</div>
|
|
<div class="line"><a name="l00106"></a><span class="lineno"> 106</span>  <span class="keywordtype">int</span> dim1, <span class="keywordtype">int</span> dim2,</div>
|
|
<div class="line"><a name="l00107"></a><span class="lineno"> 107</span>  Tensor<T, Dim, true>& out,</div>
|
|
<div class="line"><a name="l00108"></a><span class="lineno"> 108</span>  cudaStream_t stream) {</div>
|
|
<div class="line"><a name="l00109"></a><span class="lineno"> 109</span>  static_assert(Dim <= TensorInfo<T, unsigned int>::kMaxDims,</div>
|
|
<div class="line"><a name="l00110"></a><span class="lineno"> 110</span>  <span class="stringliteral">"too many dimensions"</span>);</div>
|
|
<div class="line"><a name="l00111"></a><span class="lineno"> 111</span> </div>
|
|
<div class="line"><a name="l00112"></a><span class="lineno"> 112</span>  FAISS_ASSERT(dim1 != dim2);</div>
|
|
<div class="line"><a name="l00113"></a><span class="lineno"> 113</span>  FAISS_ASSERT(dim1 < Dim && dim2 < Dim);</div>
|
|
<div class="line"><a name="l00114"></a><span class="lineno"> 114</span> </div>
|
|
<div class="line"><a name="l00115"></a><span class="lineno"> 115</span>  <span class="keywordtype">int</span> outSize[Dim];</div>
|
|
<div class="line"><a name="l00116"></a><span class="lineno"> 116</span> </div>
|
|
<div class="line"><a name="l00117"></a><span class="lineno"> 117</span>  <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0; i < Dim; ++i) {</div>
|
|
<div class="line"><a name="l00118"></a><span class="lineno"> 118</span>  outSize[i] = in.getSize(i);</div>
|
|
<div class="line"><a name="l00119"></a><span class="lineno"> 119</span>  }</div>
|
|
<div class="line"><a name="l00120"></a><span class="lineno"> 120</span> </div>
|
|
<div class="line"><a name="l00121"></a><span class="lineno"> 121</span>  std::swap(outSize[dim1], outSize[dim2]);</div>
|
|
<div class="line"><a name="l00122"></a><span class="lineno"> 122</span> </div>
|
|
<div class="line"><a name="l00123"></a><span class="lineno"> 123</span>  <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0; i < Dim; ++i) {</div>
|
|
<div class="line"><a name="l00124"></a><span class="lineno"> 124</span>  FAISS_ASSERT(out.getSize(i) == outSize[i]);</div>
|
|
<div class="line"><a name="l00125"></a><span class="lineno"> 125</span>  }</div>
|
|
<div class="line"><a name="l00126"></a><span class="lineno"> 126</span> </div>
|
|
<div class="line"><a name="l00127"></a><span class="lineno"> 127</span>  <span class="keywordtype">size_t</span> totalSize = in.numElements();</div>
|
|
<div class="line"><a name="l00128"></a><span class="lineno"> 128</span>  <span class="keywordtype">size_t</span> block = std::min((<span class="keywordtype">size_t</span>) getMaxThreadsCurrentDevice(), totalSize);</div>
|
|
<div class="line"><a name="l00129"></a><span class="lineno"> 129</span> </div>
|
|
<div class="line"><a name="l00130"></a><span class="lineno"> 130</span>  <span class="keywordflow">if</span> (totalSize <= (<span class="keywordtype">size_t</span>) std::numeric_limits<int>::max()) {</div>
|
|
<div class="line"><a name="l00131"></a><span class="lineno"> 131</span>  <span class="comment">// div/mod seems faster with unsigned types</span></div>
|
|
<div class="line"><a name="l00132"></a><span class="lineno"> 132</span>  <span class="keyword">auto</span> inInfo = getTensorInfo<T, unsigned int, Dim>(in);</div>
|
|
<div class="line"><a name="l00133"></a><span class="lineno"> 133</span>  <span class="keyword">auto</span> outInfo = getTensorInfo<T, unsigned int, Dim>(out);</div>
|
|
<div class="line"><a name="l00134"></a><span class="lineno"> 134</span> </div>
|
|
<div class="line"><a name="l00135"></a><span class="lineno"> 135</span>  std::swap(inInfo.sizes[dim1], inInfo.sizes[dim2]);</div>
|
|
<div class="line"><a name="l00136"></a><span class="lineno"> 136</span>  std::swap(inInfo.strides[dim1], inInfo.strides[dim2]);</div>
|
|
<div class="line"><a name="l00137"></a><span class="lineno"> 137</span> </div>
|
|
<div class="line"><a name="l00138"></a><span class="lineno"> 138</span>  <span class="keyword">auto</span> grid = std::min(utils::divUp(totalSize, block), (<span class="keywordtype">size_t</span>) 4096);</div>
|
|
<div class="line"><a name="l00139"></a><span class="lineno"> 139</span> </div>
|
|
<div class="line"><a name="l00140"></a><span class="lineno"> 140</span>  transposeAny<T, <span class="keywordtype">unsigned</span> int, Dim, -1></div>
|
|
<div class="line"><a name="l00141"></a><span class="lineno"> 141</span>  <<<grid, block, 0, stream>>>(inInfo, outInfo, totalSize);</div>
|
|
<div class="line"><a name="l00142"></a><span class="lineno"> 142</span>  } <span class="keywordflow">else</span> {</div>
|
|
<div class="line"><a name="l00143"></a><span class="lineno"> 143</span>  <span class="keyword">auto</span> inInfo = getTensorInfo<T, unsigned long, Dim>(in);</div>
|
|
<div class="line"><a name="l00144"></a><span class="lineno"> 144</span>  <span class="keyword">auto</span> outInfo = getTensorInfo<T, unsigned long, Dim>(out);</div>
|
|
<div class="line"><a name="l00145"></a><span class="lineno"> 145</span> </div>
|
|
<div class="line"><a name="l00146"></a><span class="lineno"> 146</span>  std::swap(inInfo.sizes[dim1], inInfo.sizes[dim2]);</div>
|
|
<div class="line"><a name="l00147"></a><span class="lineno"> 147</span>  std::swap(inInfo.strides[dim1], inInfo.strides[dim2]);</div>
|
|
<div class="line"><a name="l00148"></a><span class="lineno"> 148</span> </div>
|
|
<div class="line"><a name="l00149"></a><span class="lineno"> 149</span>  <span class="keyword">auto</span> grid = std::min(utils::divUp(totalSize, block), (<span class="keywordtype">size_t</span>) 4096);</div>
|
|
<div class="line"><a name="l00150"></a><span class="lineno"> 150</span> </div>
|
|
<div class="line"><a name="l00151"></a><span class="lineno"> 151</span>  transposeAny<T, <span class="keywordtype">unsigned</span> long, Dim, -1></div>
|
|
<div class="line"><a name="l00152"></a><span class="lineno"> 152</span>  <<<grid, block, 0, stream>>>(inInfo, outInfo, totalSize);</div>
|
|
<div class="line"><a name="l00153"></a><span class="lineno"> 153</span>  }</div>
|
|
<div class="line"><a name="l00154"></a><span class="lineno"> 154</span>  CUDA_TEST_ERROR();</div>
|
|
<div class="line"><a name="l00155"></a><span class="lineno"> 155</span> }</div>
|
|
<div class="line"><a name="l00156"></a><span class="lineno"> 156</span> </div>
|
|
<div class="line"><a name="l00157"></a><span class="lineno"> 157</span> } } <span class="comment">// namespace</span></div>
|
|
<div class="ttc" id="structfaiss_1_1gpu_1_1TensorInfo_html"><div class="ttname"><a href="structfaiss_1_1gpu_1_1TensorInfo.html">faiss::gpu::TensorInfo</a></div><div class="ttdef"><b>Definition:</b> <a href="Transpose_8cuh_source.html#l00023">Transpose.cuh:23</a></div></div>
|
|
<div class="ttc" id="classfaiss_1_1gpu_1_1Tensor_html_a6699c311648457f257afa340c61f417c"><div class="ttname"><a href="classfaiss_1_1gpu_1_1Tensor.html#a6699c311648457f257afa340c61f417c">faiss::gpu::Tensor::getSize</a></div><div class="ttdeci">__host__ __device__ IndexT getSize(int i) const </div><div class="ttdef"><b>Definition:</b> <a href="Tensor_8cuh_source.html#l00226">Tensor.cuh:226</a></div></div>
|
|
<div class="ttc" id="classfaiss_1_1gpu_1_1Tensor_html_a50411ce4d0fa32ef715e3321b6e33212"><div class="ttname"><a href="classfaiss_1_1gpu_1_1Tensor.html#a50411ce4d0fa32ef715e3321b6e33212">faiss::gpu::Tensor::data</a></div><div class="ttdeci">__host__ __device__ DataPtrType data()</div><div class="ttdoc">Returns a raw pointer to the start of our data. </div><div class="ttdef"><b>Definition:</b> <a href="Tensor_8cuh_source.html#l00178">Tensor.cuh:178</a></div></div>
|
|
<div class="ttc" id="classfaiss_1_1gpu_1_1Tensor_html"><div class="ttname"><a href="classfaiss_1_1gpu_1_1Tensor.html">faiss::gpu::Tensor</a></div><div class="ttdoc">Our tensor type. </div><div class="ttdef"><b>Definition:</b> <a href="Tensor_8cuh_source.html#l00030">Tensor.cuh:30</a></div></div>
|
|
<div class="ttc" id="classfaiss_1_1gpu_1_1Tensor_html_a0b8bba630f7a1fa217f90b20d298420a"><div class="ttname"><a href="classfaiss_1_1gpu_1_1Tensor.html#a0b8bba630f7a1fa217f90b20d298420a">faiss::gpu::Tensor::getStride</a></div><div class="ttdeci">__host__ __device__ IndexT getStride(int i) const </div><div class="ttdef"><b>Definition:</b> <a href="Tensor_8cuh_source.html#l00232">Tensor.cuh:232</a></div></div>
|
|
<div class="ttc" id="structfaiss_1_1gpu_1_1TensorInfoOffset_html"><div class="ttname"><a href="structfaiss_1_1gpu_1_1TensorInfoOffset.html">faiss::gpu::TensorInfoOffset</a></div><div class="ttdef"><b>Definition:</b> <a href="Transpose_8cuh_source.html#l00033">Transpose.cuh:33</a></div></div>
|
|
</div><!-- fragment --></div><!-- contents -->
|
|
<!-- start footer part -->
|
|
<hr class="footer"/><address class="footer"><small>
|
|
Generated by  <a href="http://www.doxygen.org/index.html">
|
|
<img class="footer" src="doxygen.png" alt="doxygen"/>
|
|
</a> 1.8.5
|
|
</small></address>
|
|
</body>
|
|
</html>
|