| <!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 content="text/xhtml;charset=utf-8" http-equiv="Content-Type"/> |
| <meta content="IE=9" http-equiv="X-UA-Compatible"/> |
| <meta content="Doxygen 1.8.11" name="generator"/> |
| <title>mxnet: src/common/cuda_utils.h Source File</title> |
| <link href="tabs.css" rel="stylesheet" type="text/css"/> |
| <script src="jquery.js" type="text/javascript"></script> |
| <script src="dynsections.js" type="text/javascript"></script> |
| <link href="search/search.css" rel="stylesheet" type="text/css"/> |
| <script src="search/searchdata.js" type="text/javascript"></script> |
| <script src="search/search.js" type="text/javascript"></script> |
| <script type="text/javascript"> |
| $(document).ready(function() { init_search(); }); |
| </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 cellpadding="0" cellspacing="0"> |
| <tbody> |
| <tr style="height: 56px;"> |
| <td id="projectalign" style="padding-left: 0.5em;"> |
| <div id="projectname">mxnet |
| </div> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| </div> |
| <!-- end header part --> |
| <!-- Generated by Doxygen 1.8.11 --> |
| <script type="text/javascript"> |
| var searchBox = new SearchBox("searchBox", "search",false,'Search'); |
| </script> |
| <div class="tabs" id="navrow1"> |
| <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 class="MSearchBoxInactive" id="MSearchBox"> |
| <span class="left"> |
| <img alt="" id="MSearchSelect" onmouseout="return searchBox.OnSearchSelectHide()" onmouseover="return searchBox.OnSearchSelectShow()" src="search/mag_sel.png"/> |
| <input accesskey="S" id="MSearchField" onblur="searchBox.OnSearchFieldFocus(false)" onfocus="searchBox.OnSearchFieldFocus(true)" onkeyup="searchBox.OnSearchFieldChange(event)" type="text" value="Search"/> |
| </span><span class="right"> |
| <a href="javascript:searchBox.CloseResultsWindow()" id="MSearchClose"><img alt="" border="0" id="MSearchCloseImg" src="search/close.png"/></a> |
| </span> |
| </div> |
| </li> |
| </ul> |
| </div> |
| <div class="tabs2" id="navrow2"> |
| <ul class="tablist"> |
| <li><a href="files.html"><span>File List</span></a></li> |
| <li><a href="globals.html"><span>File Members</span></a></li> |
| </ul> |
| </div> |
| <!-- window showing the filter options --> |
| <div id="MSearchSelectWindow" onkeydown="return searchBox.OnSearchSelectKey(event)" onmouseout="return searchBox.OnSearchSelectHide()" onmouseover="return searchBox.OnSearchSelectShow()"> |
| </div> |
| <!-- iframe showing the search results (closed by default) --> |
| <div id="MSearchResultsWindow"> |
| <iframe frameborder="0" id="MSearchResults" name="MSearchResults" src="javascript:void(0)"> |
| </iframe> |
| </div> |
| <div class="navpath" id="nav-path"> |
| <ul> |
| <li class="navelem"><a class="el" href="dir_68267d1309a1af8e8297ef4c3efbcdba.html">src</a></li><li class="navelem"><a class="el" href="dir_fdedb0aba14d44ce9d99bc100e026e6a.html">common</a></li> </ul> |
| </div> |
| </div><!-- top --> |
| <div class="header"> |
| <div class="headertitle"> |
| <div class="title">cuda_utils.h</div> </div> |
| </div><!--header--> |
| <div class="contents"> |
| <a href="cuda__utils_8h.html">Go to the documentation of this file.</a><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"> * Licensed to the Apache Software Foundation (ASF) under one</span></div><div class="line"><a name="l00003"></a><span class="lineno"> 3</span> <span class="comment"> * or more contributor license agreements. See the NOTICE file</span></div><div class="line"><a name="l00004"></a><span class="lineno"> 4</span> <span class="comment"> * distributed with this work for additional information</span></div><div class="line"><a name="l00005"></a><span class="lineno"> 5</span> <span class="comment"> * regarding copyright ownership. The ASF licenses this file</span></div><div class="line"><a name="l00006"></a><span class="lineno"> 6</span> <span class="comment"> * to you under the Apache License, Version 2.0 (the</span></div><div class="line"><a name="l00007"></a><span class="lineno"> 7</span> <span class="comment"> * "License"); you may not use this file except in compliance</span></div><div class="line"><a name="l00008"></a><span class="lineno"> 8</span> <span class="comment"> * with the License. You may obtain a copy of the License at</span></div><div class="line"><a name="l00009"></a><span class="lineno"> 9</span> <span class="comment"> *</span></div><div class="line"><a name="l00010"></a><span class="lineno"> 10</span> <span class="comment"> * http://www.apache.org/licenses/LICENSE-2.0</span></div><div class="line"><a name="l00011"></a><span class="lineno"> 11</span> <span class="comment"> *</span></div><div class="line"><a name="l00012"></a><span class="lineno"> 12</span> <span class="comment"> * Unless required by applicable law or agreed to in writing,</span></div><div class="line"><a name="l00013"></a><span class="lineno"> 13</span> <span class="comment"> * software distributed under the License is distributed on an</span></div><div class="line"><a name="l00014"></a><span class="lineno"> 14</span> <span class="comment"> * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY</span></div><div class="line"><a name="l00015"></a><span class="lineno"> 15</span> <span class="comment"> * KIND, either express or implied. See the License for the</span></div><div class="line"><a name="l00016"></a><span class="lineno"> 16</span> <span class="comment"> * specific language governing permissions and limitations</span></div><div class="line"><a name="l00017"></a><span class="lineno"> 17</span> <span class="comment"> * under the License.</span></div><div class="line"><a name="l00018"></a><span class="lineno"> 18</span> <span class="comment"> */</span></div><div class="line"><a name="l00019"></a><span class="lineno"> 19</span> </div><div class="line"><a name="l00025"></a><span class="lineno"> 25</span> <span class="preprocessor">#ifndef MXNET_COMMON_CUDA_UTILS_H_</span></div><div class="line"><a name="l00026"></a><span class="lineno"> 26</span> <span class="preprocessor">#define MXNET_COMMON_CUDA_UTILS_H_</span></div><div class="line"><a name="l00027"></a><span class="lineno"> 27</span> </div><div class="line"><a name="l00028"></a><span class="lineno"> 28</span> <span class="preprocessor">#include <dmlc/logging.h></span></div><div class="line"><a name="l00029"></a><span class="lineno"> 29</span> <span class="preprocessor">#include <dmlc/parameter.h></span></div><div class="line"><a name="l00030"></a><span class="lineno"> 30</span> <span class="preprocessor">#include <dmlc/optional.h></span></div><div class="line"><a name="l00031"></a><span class="lineno"> 31</span> <span class="preprocessor">#include <mshadow/base.h></span></div><div class="line"><a name="l00032"></a><span class="lineno"> 32</span> </div><div class="line"><a name="l00034"></a><span class="lineno"> 34</span> <span class="preprocessor">#ifdef __JETBRAINS_IDE__</span></div><div class="line"><a name="l00035"></a><span class="lineno"> 35</span> <span class="preprocessor">#define __CUDACC__ 1</span></div><div class="line"><a name="l00036"></a><span class="lineno"> 36</span> <span class="preprocessor">#define __host__</span></div><div class="line"><a name="l00037"></a><span class="lineno"> 37</span> <span class="preprocessor">#define __device__</span></div><div class="line"><a name="l00038"></a><span class="lineno"> 38</span> <span class="preprocessor">#define __global__</span></div><div class="line"><a name="l00039"></a><span class="lineno"> 39</span> <span class="preprocessor">#define __forceinline__</span></div><div class="line"><a name="l00040"></a><span class="lineno"> 40</span> <span class="preprocessor">#define __shared__</span></div><div class="line"><a name="l00041"></a><span class="lineno"> 41</span> <span class="keyword">inline</span> <span class="keywordtype">void</span> __syncthreads() {}</div><div class="line"><a name="l00042"></a><span class="lineno"> 42</span> <span class="keyword">inline</span> <span class="keywordtype">void</span> __threadfence_block() {}</div><div class="line"><a name="l00043"></a><span class="lineno"> 43</span> <span class="keyword">template</span><<span class="keyword">class</span> T> <span class="keyword">inline</span> T __clz(<span class="keyword">const</span> T val) { <span class="keywordflow">return</span> val; }</div><div class="line"><a name="l00044"></a><span class="lineno"> 44</span> <span class="keyword">struct </span>__cuda_fake_struct { <span class="keywordtype">int</span> x; <span class="keywordtype">int</span> y; <span class="keywordtype">int</span> z; };</div><div class="line"><a name="l00045"></a><span class="lineno"> 45</span> <span class="keyword">extern</span> __cuda_fake_struct blockDim;</div><div class="line"><a name="l00046"></a><span class="lineno"> 46</span> <span class="keyword">extern</span> __cuda_fake_struct threadIdx;</div><div class="line"><a name="l00047"></a><span class="lineno"> 47</span> <span class="keyword">extern</span> __cuda_fake_struct blockIdx;</div><div class="line"><a name="l00048"></a><span class="lineno"> 48</span> <span class="preprocessor">#endif</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="preprocessor">#if MXNET_USE_CUDA</span></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> <span class="preprocessor">#include <cuda_runtime.h></span></div><div class="line"><a name="l00053"></a><span class="lineno"> 53</span> <span class="preprocessor">#include <cublas_v2.h></span></div><div class="line"><a name="l00054"></a><span class="lineno"> 54</span> <span class="preprocessor">#include <curand.h></span></div><div class="line"><a name="l00055"></a><span class="lineno"> 55</span> </div><div class="line"><a name="l00060"></a><span class="lineno"> 60</span> <span class="preprocessor">#ifdef __CUDACC__</span></div><div class="line"><a name="l00061"></a><span class="lineno"> 61</span> <span class="keyword">inline</span> __device__ <span class="keywordtype">bool</span> __is_supported_cuda_architecture() {</div><div class="line"><a name="l00062"></a><span class="lineno"> 62</span> <span class="preprocessor">#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ < 300</span></div><div class="line"><a name="l00063"></a><span class="lineno"> 63</span> <span class="preprocessor">#error "Fermi and earlier GPU architectures are not supported (architecture versions less than 3.0)"</span></div><div class="line"><a name="l00064"></a><span class="lineno"> 64</span> <span class="keywordflow">return</span> <span class="keyword">false</span>;</div><div class="line"><a name="l00065"></a><span class="lineno"> 65</span> <span class="preprocessor">#else</span></div><div class="line"><a name="l00066"></a><span class="lineno"> 66</span> <span class="keywordflow">return</span> <span class="keyword">true</span>;</div><div class="line"><a name="l00067"></a><span class="lineno"> 67</span> <span class="preprocessor">#endif // __CUDA_ARCH__ < 300</span></div><div class="line"><a name="l00068"></a><span class="lineno"> 68</span> }</div><div class="line"><a name="l00069"></a><span class="lineno"> 69</span> <span class="preprocessor">#endif // __CUDACC__</span></div><div class="line"><a name="l00070"></a><span class="lineno"> 70</span> </div><div class="line"><a name="l00075"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#afc69a418242c5b851993bc2307b1c897"> 75</a></span> <span class="preprocessor">#define CHECK_CUDA_ERROR(msg) \</span></div><div class="line"><a name="l00076"></a><span class="lineno"> 76</span> <span class="preprocessor"> { \</span></div><div class="line"><a name="l00077"></a><span class="lineno"> 77</span> <span class="preprocessor"> cudaError_t e = cudaGetLastError(); \</span></div><div class="line"><a name="l00078"></a><span class="lineno"> 78</span> <span class="preprocessor"> CHECK_EQ(e, cudaSuccess) << (msg) << " CUDA: " << cudaGetErrorString(e); \</span></div><div class="line"><a name="l00079"></a><span class="lineno"> 79</span> <span class="preprocessor"> }</span></div><div class="line"><a name="l00080"></a><span class="lineno"> 80</span> </div><div class="line"><a name="l00087"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a06cc7d24ca66505e69f5ad40009f5e8d"> 87</a></span> <span class="preprocessor">#define CUDA_CALL(func) \</span></div><div class="line"><a name="l00088"></a><span class="lineno"> 88</span> <span class="preprocessor"> { \</span></div><div class="line"><a name="l00089"></a><span class="lineno"> 89</span> <span class="preprocessor"> cudaError_t e = (func); \</span></div><div class="line"><a name="l00090"></a><span class="lineno"> 90</span> <span class="preprocessor"> CHECK(e == cudaSuccess || e == cudaErrorCudartUnloading) \</span></div><div class="line"><a name="l00091"></a><span class="lineno"> 91</span> <span class="preprocessor"> << "CUDA: " << cudaGetErrorString(e); \</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="l00100"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a685d7ca3c9370ff471665abcacdeb381"> 100</a></span> <span class="preprocessor">#define CUBLAS_CALL(func) \</span></div><div class="line"><a name="l00101"></a><span class="lineno"> 101</span> <span class="preprocessor"> { \</span></div><div class="line"><a name="l00102"></a><span class="lineno"> 102</span> <span class="preprocessor"> cublasStatus_t e = (func); \</span></div><div class="line"><a name="l00103"></a><span class="lineno"> 103</span> <span class="preprocessor"> CHECK_EQ(e, CUBLAS_STATUS_SUCCESS) \</span></div><div class="line"><a name="l00104"></a><span class="lineno"> 104</span> <span class="preprocessor"> << "cuBLAS: " << mxnet::common::cuda::CublasGetErrorString(e); \</span></div><div class="line"><a name="l00105"></a><span class="lineno"> 105</span> <span class="preprocessor"> }</span></div><div class="line"><a name="l00106"></a><span class="lineno"> 106</span> </div><div class="line"><a name="l00113"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#ab38940ff6950f84102baa4573675b670"> 113</a></span> <span class="preprocessor">#define CUSOLVER_CALL(func) \</span></div><div class="line"><a name="l00114"></a><span class="lineno"> 114</span> <span class="preprocessor"> { \</span></div><div class="line"><a name="l00115"></a><span class="lineno"> 115</span> <span class="preprocessor"> cusolverStatus_t e = (func); \</span></div><div class="line"><a name="l00116"></a><span class="lineno"> 116</span> <span class="preprocessor"> CHECK_EQ(e, CUSOLVER_STATUS_SUCCESS) \</span></div><div class="line"><a name="l00117"></a><span class="lineno"> 117</span> <span class="preprocessor"> << "cuSolver: " << mxnet::common::cuda::CusolverGetErrorString(e); \</span></div><div class="line"><a name="l00118"></a><span class="lineno"> 118</span> <span class="preprocessor"> }</span></div><div class="line"><a name="l00119"></a><span class="lineno"> 119</span> </div><div class="line"><a name="l00126"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a82d7233550780a8c186e79c24aed8406"> 126</a></span> <span class="preprocessor">#define CURAND_CALL(func) \</span></div><div class="line"><a name="l00127"></a><span class="lineno"> 127</span> <span class="preprocessor"> { \</span></div><div class="line"><a name="l00128"></a><span class="lineno"> 128</span> <span class="preprocessor"> curandStatus_t e = (func); \</span></div><div class="line"><a name="l00129"></a><span class="lineno"> 129</span> <span class="preprocessor"> CHECK_EQ(e, CURAND_STATUS_SUCCESS) \</span></div><div class="line"><a name="l00130"></a><span class="lineno"> 130</span> <span class="preprocessor"> << "cuRAND: " << mxnet::common::cuda::CurandGetErrorString(e); \</span></div><div class="line"><a name="l00131"></a><span class="lineno"> 131</span> <span class="preprocessor"> }</span></div><div class="line"><a name="l00132"></a><span class="lineno"> 132</span> </div><div class="line"><a name="l00139"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a63b6d263b94df9e33474894ad02b792d"> 139</a></span> <span class="preprocessor">#define NVRTC_CALL(x) \</span></div><div class="line"><a name="l00140"></a><span class="lineno"> 140</span> <span class="preprocessor"> { \</span></div><div class="line"><a name="l00141"></a><span class="lineno"> 141</span> <span class="preprocessor"> nvrtcResult result = x; \</span></div><div class="line"><a name="l00142"></a><span class="lineno"> 142</span> <span class="preprocessor"> CHECK_EQ(result, NVRTC_SUCCESS) \</span></div><div class="line"><a name="l00143"></a><span class="lineno"> 143</span> <span class="preprocessor"> << #x " failed with error " \</span></div><div class="line"><a name="l00144"></a><span class="lineno"> 144</span> <span class="preprocessor"> << nvrtcGetErrorString(result); \</span></div><div class="line"><a name="l00145"></a><span class="lineno"> 145</span> <span class="preprocessor"> }</span></div><div class="line"><a name="l00146"></a><span class="lineno"> 146</span> </div><div class="line"><a name="l00153"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a0d9b08b9ef45122c54bf5a121aeab5c3"> 153</a></span> <span class="preprocessor">#define CUDA_DRIVER_CALL(func) \</span></div><div class="line"><a name="l00154"></a><span class="lineno"> 154</span> <span class="preprocessor"> { \</span></div><div class="line"><a name="l00155"></a><span class="lineno"> 155</span> <span class="preprocessor"> CUresult e = (func); \</span></div><div class="line"><a name="l00156"></a><span class="lineno"> 156</span> <span class="preprocessor"> if (e != CUDA_SUCCESS) { \</span></div><div class="line"><a name="l00157"></a><span class="lineno"> 157</span> <span class="preprocessor"> char const * err_msg = nullptr; \</span></div><div class="line"><a name="l00158"></a><span class="lineno"> 158</span> <span class="preprocessor"> if (cuGetErrorString(e, &err_msg) == CUDA_ERROR_INVALID_VALUE) { \</span></div><div class="line"><a name="l00159"></a><span class="lineno"> 159</span> <span class="preprocessor"> LOG(FATAL) << "CUDA Driver: Unknown error " << e; \</span></div><div class="line"><a name="l00160"></a><span class="lineno"> 160</span> <span class="preprocessor"> } else { \</span></div><div class="line"><a name="l00161"></a><span class="lineno"> 161</span> <span class="preprocessor"> LOG(FATAL) << "CUDA Driver: " << err_msg; \</span></div><div class="line"><a name="l00162"></a><span class="lineno"> 162</span> <span class="preprocessor"> } \</span></div><div class="line"><a name="l00163"></a><span class="lineno"> 163</span> <span class="preprocessor"> } \</span></div><div class="line"><a name="l00164"></a><span class="lineno"> 164</span> <span class="preprocessor"> }</span></div><div class="line"><a name="l00165"></a><span class="lineno"> 165</span> </div><div class="line"><a name="l00166"></a><span class="lineno"> 166</span> </div><div class="line"><a name="l00167"></a><span class="lineno"> 167</span> <span class="preprocessor">#if !defined(_MSC_VER)</span></div><div class="line"><a name="l00168"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a685e3713856baaafb1d4edea43725c83"> 168</a></span> <span class="preprocessor">#define CUDA_UNROLL _Pragma("unroll")</span></div><div class="line"><a name="l00169"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#addb314f15d765a2ba72ae37dab23c03b"> 169</a></span> <span class="preprocessor">#define CUDA_NOUNROLL _Pragma("nounroll")</span></div><div class="line"><a name="l00170"></a><span class="lineno"> 170</span> <span class="preprocessor">#else</span></div><div class="line"><a name="l00171"></a><span class="lineno"> 171</span> <span class="preprocessor">#define CUDA_UNROLL</span></div><div class="line"><a name="l00172"></a><span class="lineno"> 172</span> <span class="preprocessor">#define CUDA_NOUNROLL</span></div><div class="line"><a name="l00173"></a><span class="lineno"> 173</span> <span class="preprocessor">#endif</span></div><div class="line"><a name="l00174"></a><span class="lineno"> 174</span> </div><div class="line"><a name="l00175"></a><span class="lineno"> 175</span> <span class="keyword">namespace </span><a class="code" href="namespacemxnet.html">mxnet</a> {</div><div class="line"><a name="l00176"></a><span class="lineno"> 176</span> <span class="keyword">namespace </span>common {</div><div class="line"><a name="l00178"></a><span class="lineno"><a class="line" href="namespacemxnet_1_1common_1_1cuda.html"> 178</a></span> <span class="keyword">namespace </span>cuda {</div><div class="line"><a name="l00184"></a><span class="lineno"><a class="line" href="namespacemxnet_1_1common_1_1cuda.html#a9feee613a4f16a954dd68e55345a72ac"> 184</a></span> <span class="keyword">inline</span> <span class="keyword">const</span> <span class="keywordtype">char</span>* <a class="code" href="namespacemxnet_1_1common_1_1cuda.html#a9feee613a4f16a954dd68e55345a72ac">CublasGetErrorString</a>(cublasStatus_t error) {</div><div class="line"><a name="l00185"></a><span class="lineno"> 185</span> <span class="keywordflow">switch</span> (error) {</div><div class="line"><a name="l00186"></a><span class="lineno"> 186</span> <span class="keywordflow">case</span> CUBLAS_STATUS_SUCCESS:</div><div class="line"><a name="l00187"></a><span class="lineno"> 187</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_SUCCESS"</span>;</div><div class="line"><a name="l00188"></a><span class="lineno"> 188</span> <span class="keywordflow">case</span> CUBLAS_STATUS_NOT_INITIALIZED:</div><div class="line"><a name="l00189"></a><span class="lineno"> 189</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_NOT_INITIALIZED"</span>;</div><div class="line"><a name="l00190"></a><span class="lineno"> 190</span> <span class="keywordflow">case</span> CUBLAS_STATUS_ALLOC_FAILED:</div><div class="line"><a name="l00191"></a><span class="lineno"> 191</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_ALLOC_FAILED"</span>;</div><div class="line"><a name="l00192"></a><span class="lineno"> 192</span> <span class="keywordflow">case</span> CUBLAS_STATUS_INVALID_VALUE:</div><div class="line"><a name="l00193"></a><span class="lineno"> 193</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_INVALID_VALUE"</span>;</div><div class="line"><a name="l00194"></a><span class="lineno"> 194</span> <span class="keywordflow">case</span> CUBLAS_STATUS_ARCH_MISMATCH:</div><div class="line"><a name="l00195"></a><span class="lineno"> 195</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_ARCH_MISMATCH"</span>;</div><div class="line"><a name="l00196"></a><span class="lineno"> 196</span> <span class="keywordflow">case</span> CUBLAS_STATUS_MAPPING_ERROR:</div><div class="line"><a name="l00197"></a><span class="lineno"> 197</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_MAPPING_ERROR"</span>;</div><div class="line"><a name="l00198"></a><span class="lineno"> 198</span> <span class="keywordflow">case</span> CUBLAS_STATUS_EXECUTION_FAILED:</div><div class="line"><a name="l00199"></a><span class="lineno"> 199</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_EXECUTION_FAILED"</span>;</div><div class="line"><a name="l00200"></a><span class="lineno"> 200</span> <span class="keywordflow">case</span> CUBLAS_STATUS_INTERNAL_ERROR:</div><div class="line"><a name="l00201"></a><span class="lineno"> 201</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_INTERNAL_ERROR"</span>;</div><div class="line"><a name="l00202"></a><span class="lineno"> 202</span> <span class="keywordflow">case</span> CUBLAS_STATUS_NOT_SUPPORTED:</div><div class="line"><a name="l00203"></a><span class="lineno"> 203</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUBLAS_STATUS_NOT_SUPPORTED"</span>;</div><div class="line"><a name="l00204"></a><span class="lineno"> 204</span> <span class="keywordflow">default</span>:</div><div class="line"><a name="l00205"></a><span class="lineno"> 205</span> <span class="keywordflow">break</span>;</div><div class="line"><a name="l00206"></a><span class="lineno"> 206</span> }</div><div class="line"><a name="l00207"></a><span class="lineno"> 207</span> <span class="keywordflow">return</span> <span class="stringliteral">"Unknown cuBLAS status"</span>;</div><div class="line"><a name="l00208"></a><span class="lineno"> 208</span> }</div><div class="line"><a name="l00209"></a><span class="lineno"> 209</span> </div><div class="line"><a name="l00215"></a><span class="lineno"><a class="line" href="namespacemxnet_1_1common_1_1cuda.html#abf9bcb4cb696e9ae61b818510dac39c8"> 215</a></span> <span class="keyword">inline</span> <span class="keyword">const</span> <span class="keywordtype">char</span>* <a class="code" href="namespacemxnet_1_1common_1_1cuda.html#abf9bcb4cb696e9ae61b818510dac39c8">CusolverGetErrorString</a>(cusolverStatus_t error) {</div><div class="line"><a name="l00216"></a><span class="lineno"> 216</span> <span class="keywordflow">switch</span> (error) {</div><div class="line"><a name="l00217"></a><span class="lineno"> 217</span> <span class="keywordflow">case</span> CUSOLVER_STATUS_SUCCESS:</div><div class="line"><a name="l00218"></a><span class="lineno"> 218</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUSOLVER_STATUS_SUCCESS"</span>;</div><div class="line"><a name="l00219"></a><span class="lineno"> 219</span> <span class="keywordflow">case</span> CUSOLVER_STATUS_NOT_INITIALIZED:</div><div class="line"><a name="l00220"></a><span class="lineno"> 220</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUSOLVER_STATUS_NOT_INITIALIZED"</span>;</div><div class="line"><a name="l00221"></a><span class="lineno"> 221</span> <span class="keywordflow">case</span> CUSOLVER_STATUS_ALLOC_FAILED:</div><div class="line"><a name="l00222"></a><span class="lineno"> 222</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUSOLVER_STATUS_ALLOC_FAILED"</span>;</div><div class="line"><a name="l00223"></a><span class="lineno"> 223</span> <span class="keywordflow">case</span> CUSOLVER_STATUS_INVALID_VALUE:</div><div class="line"><a name="l00224"></a><span class="lineno"> 224</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUSOLVER_STATUS_INVALID_VALUE"</span>;</div><div class="line"><a name="l00225"></a><span class="lineno"> 225</span> <span class="keywordflow">case</span> CUSOLVER_STATUS_ARCH_MISMATCH:</div><div class="line"><a name="l00226"></a><span class="lineno"> 226</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUSOLVER_STATUS_ARCH_MISMATCH"</span>;</div><div class="line"><a name="l00227"></a><span class="lineno"> 227</span> <span class="keywordflow">case</span> CUSOLVER_STATUS_EXECUTION_FAILED:</div><div class="line"><a name="l00228"></a><span class="lineno"> 228</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUSOLVER_STATUS_EXECUTION_FAILED"</span>;</div><div class="line"><a name="l00229"></a><span class="lineno"> 229</span> <span class="keywordflow">case</span> CUSOLVER_STATUS_INTERNAL_ERROR:</div><div class="line"><a name="l00230"></a><span class="lineno"> 230</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUSOLVER_STATUS_INTERNAL_ERROR"</span>;</div><div class="line"><a name="l00231"></a><span class="lineno"> 231</span> <span class="keywordflow">case</span> CUSOLVER_STATUS_MATRIX_TYPE_NOT_SUPPORTED:</div><div class="line"><a name="l00232"></a><span class="lineno"> 232</span> <span class="keywordflow">return</span> <span class="stringliteral">"CUSOLVER_STATUS_MATRIX_TYPE_NOT_SUPPORTED"</span>;</div><div class="line"><a name="l00233"></a><span class="lineno"> 233</span> <span class="keywordflow">default</span>:</div><div class="line"><a name="l00234"></a><span class="lineno"> 234</span> <span class="keywordflow">break</span>;</div><div class="line"><a name="l00235"></a><span class="lineno"> 235</span> }</div><div class="line"><a name="l00236"></a><span class="lineno"> 236</span> <span class="keywordflow">return</span> <span class="stringliteral">"Unknown cuSOLVER status"</span>;</div><div class="line"><a name="l00237"></a><span class="lineno"> 237</span> }</div><div class="line"><a name="l00238"></a><span class="lineno"> 238</span> </div><div class="line"><a name="l00244"></a><span class="lineno"><a class="line" href="namespacemxnet_1_1common_1_1cuda.html#a97c06b2f4d26445a7386b0f54fae1feb"> 244</a></span> <span class="keyword">inline</span> <span class="keyword">const</span> <span class="keywordtype">char</span>* <a class="code" href="namespacemxnet_1_1common_1_1cuda.html#a97c06b2f4d26445a7386b0f54fae1feb">CurandGetErrorString</a>(curandStatus_t status) {</div><div class="line"><a name="l00245"></a><span class="lineno"> 245</span> <span class="keywordflow">switch</span> (status) {</div><div class="line"><a name="l00246"></a><span class="lineno"> 246</span> <span class="keywordflow">case</span> CURAND_STATUS_SUCCESS:</div><div class="line"><a name="l00247"></a><span class="lineno"> 247</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_SUCCESS"</span>;</div><div class="line"><a name="l00248"></a><span class="lineno"> 248</span> <span class="keywordflow">case</span> CURAND_STATUS_VERSION_MISMATCH:</div><div class="line"><a name="l00249"></a><span class="lineno"> 249</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_VERSION_MISMATCH"</span>;</div><div class="line"><a name="l00250"></a><span class="lineno"> 250</span> <span class="keywordflow">case</span> CURAND_STATUS_NOT_INITIALIZED:</div><div class="line"><a name="l00251"></a><span class="lineno"> 251</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_NOT_INITIALIZED"</span>;</div><div class="line"><a name="l00252"></a><span class="lineno"> 252</span> <span class="keywordflow">case</span> CURAND_STATUS_ALLOCATION_FAILED:</div><div class="line"><a name="l00253"></a><span class="lineno"> 253</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_ALLOCATION_FAILED"</span>;</div><div class="line"><a name="l00254"></a><span class="lineno"> 254</span> <span class="keywordflow">case</span> CURAND_STATUS_TYPE_ERROR:</div><div class="line"><a name="l00255"></a><span class="lineno"> 255</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_TYPE_ERROR"</span>;</div><div class="line"><a name="l00256"></a><span class="lineno"> 256</span> <span class="keywordflow">case</span> CURAND_STATUS_OUT_OF_RANGE:</div><div class="line"><a name="l00257"></a><span class="lineno"> 257</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_OUT_OF_RANGE"</span>;</div><div class="line"><a name="l00258"></a><span class="lineno"> 258</span> <span class="keywordflow">case</span> CURAND_STATUS_LENGTH_NOT_MULTIPLE:</div><div class="line"><a name="l00259"></a><span class="lineno"> 259</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_LENGTH_NOT_MULTIPLE"</span>;</div><div class="line"><a name="l00260"></a><span class="lineno"> 260</span> <span class="keywordflow">case</span> CURAND_STATUS_DOUBLE_PRECISION_REQUIRED:</div><div class="line"><a name="l00261"></a><span class="lineno"> 261</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_DOUBLE_PRECISION_REQUIRED"</span>;</div><div class="line"><a name="l00262"></a><span class="lineno"> 262</span> <span class="keywordflow">case</span> CURAND_STATUS_LAUNCH_FAILURE:</div><div class="line"><a name="l00263"></a><span class="lineno"> 263</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_LAUNCH_FAILURE"</span>;</div><div class="line"><a name="l00264"></a><span class="lineno"> 264</span> <span class="keywordflow">case</span> CURAND_STATUS_PREEXISTING_FAILURE:</div><div class="line"><a name="l00265"></a><span class="lineno"> 265</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_PREEXISTING_FAILURE"</span>;</div><div class="line"><a name="l00266"></a><span class="lineno"> 266</span> <span class="keywordflow">case</span> CURAND_STATUS_INITIALIZATION_FAILED:</div><div class="line"><a name="l00267"></a><span class="lineno"> 267</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_INITIALIZATION_FAILED"</span>;</div><div class="line"><a name="l00268"></a><span class="lineno"> 268</span> <span class="keywordflow">case</span> CURAND_STATUS_ARCH_MISMATCH:</div><div class="line"><a name="l00269"></a><span class="lineno"> 269</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_ARCH_MISMATCH"</span>;</div><div class="line"><a name="l00270"></a><span class="lineno"> 270</span> <span class="keywordflow">case</span> CURAND_STATUS_INTERNAL_ERROR:</div><div class="line"><a name="l00271"></a><span class="lineno"> 271</span> <span class="keywordflow">return</span> <span class="stringliteral">"CURAND_STATUS_INTERNAL_ERROR"</span>;</div><div class="line"><a name="l00272"></a><span class="lineno"> 272</span> }</div><div class="line"><a name="l00273"></a><span class="lineno"> 273</span> <span class="keywordflow">return</span> <span class="stringliteral">"Unknown cuRAND status"</span>;</div><div class="line"><a name="l00274"></a><span class="lineno"> 274</span> }</div><div class="line"><a name="l00275"></a><span class="lineno"> 275</span> </div><div class="line"><a name="l00276"></a><span class="lineno"> 276</span> <span class="keyword">template</span> <<span class="keyword">typename</span> DType></div><div class="line"><a name="l00277"></a><span class="lineno"><a class="line" href="namespacemxnet_1_1common_1_1cuda.html#a6f3ee04eb382c57e10916108db3efd80"> 277</a></span> <span class="keyword">inline</span> DType __device__ <a class="code" href="namespacemxnet_1_1common_1_1cuda.html#a6f3ee04eb382c57e10916108db3efd80">CudaMax</a>(DType a, DType b) {</div><div class="line"><a name="l00278"></a><span class="lineno"> 278</span> <span class="keywordflow">return</span> a > b ? a : b;</div><div class="line"><a name="l00279"></a><span class="lineno"> 279</span> }</div><div class="line"><a name="l00280"></a><span class="lineno"> 280</span> </div><div class="line"><a name="l00281"></a><span class="lineno"> 281</span> <span class="keyword">template</span> <<span class="keyword">typename</span> DType></div><div class="line"><a name="l00282"></a><span class="lineno"><a class="line" href="namespacemxnet_1_1common_1_1cuda.html#a03888f252f813f6d052ae84bf8801498"> 282</a></span> <span class="keyword">inline</span> DType __device__ <a class="code" href="namespacemxnet_1_1common_1_1cuda.html#a03888f252f813f6d052ae84bf8801498">CudaMin</a>(DType a, DType b) {</div><div class="line"><a name="l00283"></a><span class="lineno"> 283</span> <span class="keywordflow">return</span> a < b ? a : b;</div><div class="line"><a name="l00284"></a><span class="lineno"> 284</span> }</div><div class="line"><a name="l00285"></a><span class="lineno"> 285</span> </div><div class="line"><a name="l00286"></a><span class="lineno"><a class="line" href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html"> 286</a></span> <span class="keyword">class </span><a class="code" href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html">DeviceStore</a> {</div><div class="line"><a name="l00287"></a><span class="lineno"> 287</span> <span class="keyword">public</span>:</div><div class="line"><a name="l00289"></a><span class="lineno"><a class="line" href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#ad9878a09a93d4fcaf9d0639b3613d9f7"> 289</a></span> <span class="keyword">explicit</span> <a class="code" href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#ad9878a09a93d4fcaf9d0639b3613d9f7">DeviceStore</a>(<span class="keywordtype">int</span> requested_device = -1, <span class="keywordtype">bool</span> restore = <span class="keyword">true</span>) :</div><div class="line"><a name="l00290"></a><span class="lineno"> 290</span> restore_device_(-1),</div><div class="line"><a name="l00291"></a><span class="lineno"> 291</span> current_device_(requested_device),</div><div class="line"><a name="l00292"></a><span class="lineno"> 292</span> restore_(restore) {</div><div class="line"><a name="l00293"></a><span class="lineno"> 293</span> <span class="keywordflow">if</span> (restore_)</div><div class="line"><a name="l00294"></a><span class="lineno"> 294</span> <a class="code" href="cuda__utils_8h.html#a06cc7d24ca66505e69f5ad40009f5e8d">CUDA_CALL</a>(cudaGetDevice(&restore_device_));</div><div class="line"><a name="l00295"></a><span class="lineno"> 295</span> <span class="keywordflow">if</span> (requested_device != restore_device_) {</div><div class="line"><a name="l00296"></a><span class="lineno"> 296</span> SetDevice(requested_device);</div><div class="line"><a name="l00297"></a><span class="lineno"> 297</span> }</div><div class="line"><a name="l00298"></a><span class="lineno"> 298</span> }</div><div class="line"><a name="l00299"></a><span class="lineno"> 299</span> </div><div class="line"><a name="l00300"></a><span class="lineno"><a class="line" href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#a701d38ae493688ee2136995fe8611aa0"> 300</a></span> <a class="code" href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#a701d38ae493688ee2136995fe8611aa0">~DeviceStore</a>() {</div><div class="line"><a name="l00301"></a><span class="lineno"> 301</span> <span class="keywordflow">if</span> (restore_ &&</div><div class="line"><a name="l00302"></a><span class="lineno"> 302</span> current_device_ != restore_device_ &&</div><div class="line"><a name="l00303"></a><span class="lineno"> 303</span> current_device_ != -1 &&</div><div class="line"><a name="l00304"></a><span class="lineno"> 304</span> restore_device_ != -1)</div><div class="line"><a name="l00305"></a><span class="lineno"> 305</span> <a class="code" href="cuda__utils_8h.html#a06cc7d24ca66505e69f5ad40009f5e8d">CUDA_CALL</a>(cudaSetDevice(restore_device_));</div><div class="line"><a name="l00306"></a><span class="lineno"> 306</span> }</div><div class="line"><a name="l00307"></a><span class="lineno"> 307</span> </div><div class="line"><a name="l00308"></a><span class="lineno"><a class="line" href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#a01163fd4915e74bdd81dd7305917f0e4"> 308</a></span> <span class="keywordtype">void</span> <a class="code" href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#a01163fd4915e74bdd81dd7305917f0e4">SetDevice</a>(<span class="keywordtype">int</span> device) {</div><div class="line"><a name="l00309"></a><span class="lineno"> 309</span> <span class="keywordflow">if</span> (device != -1) {</div><div class="line"><a name="l00310"></a><span class="lineno"> 310</span> <a class="code" href="cuda__utils_8h.html#a06cc7d24ca66505e69f5ad40009f5e8d">CUDA_CALL</a>(cudaSetDevice(device));</div><div class="line"><a name="l00311"></a><span class="lineno"> 311</span> current_device_ = device;</div><div class="line"><a name="l00312"></a><span class="lineno"> 312</span> }</div><div class="line"><a name="l00313"></a><span class="lineno"> 313</span> }</div><div class="line"><a name="l00314"></a><span class="lineno"> 314</span> </div><div class="line"><a name="l00315"></a><span class="lineno"> 315</span> <span class="keyword">private</span>:</div><div class="line"><a name="l00316"></a><span class="lineno"> 316</span> <span class="keywordtype">int</span> restore_device_;</div><div class="line"><a name="l00317"></a><span class="lineno"> 317</span> <span class="keywordtype">int</span> current_device_;</div><div class="line"><a name="l00318"></a><span class="lineno"> 318</span> <span class="keywordtype">bool</span> restore_;</div><div class="line"><a name="l00319"></a><span class="lineno"> 319</span> };</div><div class="line"><a name="l00320"></a><span class="lineno"> 320</span> </div><div class="line"><a name="l00321"></a><span class="lineno"> 321</span> } <span class="comment">// namespace cuda</span></div><div class="line"><a name="l00322"></a><span class="lineno"> 322</span> } <span class="comment">// namespace common</span></div><div class="line"><a name="l00323"></a><span class="lineno"> 323</span> } <span class="comment">// namespace mxnet</span></div><div class="line"><a name="l00324"></a><span class="lineno"> 324</span> </div><div class="line"><a name="l00330"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#aa79f548df23452162de37663f171e99d"> 330</a></span> <span class="keyword">inline</span> <span class="keywordtype">int</span> <a class="code" href="cuda__utils_8h.html#aa79f548df23452162de37663f171e99d">ComputeCapabilityMajor</a>(<span class="keywordtype">int</span> device_id) {</div><div class="line"><a name="l00331"></a><span class="lineno"> 331</span> <span class="keywordtype">int</span> major = 0;</div><div class="line"><a name="l00332"></a><span class="lineno"> 332</span> <a class="code" href="cuda__utils_8h.html#a06cc7d24ca66505e69f5ad40009f5e8d">CUDA_CALL</a>(cudaDeviceGetAttribute(&major,</div><div class="line"><a name="l00333"></a><span class="lineno"> 333</span> cudaDevAttrComputeCapabilityMajor, device_id));</div><div class="line"><a name="l00334"></a><span class="lineno"> 334</span> <span class="keywordflow">return</span> major;</div><div class="line"><a name="l00335"></a><span class="lineno"> 335</span> }</div><div class="line"><a name="l00336"></a><span class="lineno"> 336</span> </div><div class="line"><a name="l00342"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a7c16e8770e4f399cabed1fc231ffd9b6"> 342</a></span> <span class="keyword">inline</span> <span class="keywordtype">int</span> <a class="code" href="cuda__utils_8h.html#a7c16e8770e4f399cabed1fc231ffd9b6">ComputeCapabilityMinor</a>(<span class="keywordtype">int</span> device_id) {</div><div class="line"><a name="l00343"></a><span class="lineno"> 343</span> <span class="keywordtype">int</span> minor = 0;</div><div class="line"><a name="l00344"></a><span class="lineno"> 344</span> <a class="code" href="cuda__utils_8h.html#a06cc7d24ca66505e69f5ad40009f5e8d">CUDA_CALL</a>(cudaDeviceGetAttribute(&minor,</div><div class="line"><a name="l00345"></a><span class="lineno"> 345</span> cudaDevAttrComputeCapabilityMinor, device_id));</div><div class="line"><a name="l00346"></a><span class="lineno"> 346</span> <span class="keywordflow">return</span> minor;</div><div class="line"><a name="l00347"></a><span class="lineno"> 347</span> }</div><div class="line"><a name="l00348"></a><span class="lineno"> 348</span> </div><div class="line"><a name="l00354"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a9779e3ad0efd0faec7fbe431c0db896d"> 354</a></span> <span class="keyword">inline</span> <span class="keywordtype">int</span> <a class="code" href="cuda__utils_8h.html#a9779e3ad0efd0faec7fbe431c0db896d">SMArch</a>(<span class="keywordtype">int</span> device_id) {</div><div class="line"><a name="l00355"></a><span class="lineno"> 355</span> <span class="keyword">auto</span> major = <a class="code" href="cuda__utils_8h.html#aa79f548df23452162de37663f171e99d">ComputeCapabilityMajor</a>(device_id);</div><div class="line"><a name="l00356"></a><span class="lineno"> 356</span> <span class="keyword">auto</span> minor = <a class="code" href="cuda__utils_8h.html#a7c16e8770e4f399cabed1fc231ffd9b6">ComputeCapabilityMinor</a>(device_id);</div><div class="line"><a name="l00357"></a><span class="lineno"> 357</span> <span class="keywordflow">return</span> 10 * major + minor;</div><div class="line"><a name="l00358"></a><span class="lineno"> 358</span> }</div><div class="line"><a name="l00359"></a><span class="lineno"> 359</span> </div><div class="line"><a name="l00366"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#afb4268417c1d8886a39142c85c8f188f"> 366</a></span> <span class="keyword">inline</span> <span class="keywordtype">bool</span> <a class="code" href="cuda__utils_8h.html#afb4268417c1d8886a39142c85c8f188f">SupportsFloat16Compute</a>(<span class="keywordtype">int</span> device_id) {</div><div class="line"><a name="l00367"></a><span class="lineno"> 367</span> <span class="keywordflow">if</span> (device_id < 0) {</div><div class="line"><a name="l00368"></a><span class="lineno"> 368</span> <span class="keywordflow">return</span> <span class="keyword">false</span>;</div><div class="line"><a name="l00369"></a><span class="lineno"> 369</span> } <span class="keywordflow">else</span> {</div><div class="line"><a name="l00370"></a><span class="lineno"> 370</span> <span class="comment">// Kepler and most Maxwell GPUs do not support fp16 compute</span></div><div class="line"><a name="l00371"></a><span class="lineno"> 371</span> <span class="keywordtype">int</span> computeCapabilityMajor = <a class="code" href="cuda__utils_8h.html#aa79f548df23452162de37663f171e99d">ComputeCapabilityMajor</a>(device_id);</div><div class="line"><a name="l00372"></a><span class="lineno"> 372</span> <span class="keywordflow">return</span> (computeCapabilityMajor > 5) ||</div><div class="line"><a name="l00373"></a><span class="lineno"> 373</span> (computeCapabilityMajor == 5 && <a class="code" href="cuda__utils_8h.html#a7c16e8770e4f399cabed1fc231ffd9b6">ComputeCapabilityMinor</a>(device_id) >= 3);</div><div class="line"><a name="l00374"></a><span class="lineno"> 374</span> }</div><div class="line"><a name="l00375"></a><span class="lineno"> 375</span> }</div><div class="line"><a name="l00376"></a><span class="lineno"> 376</span> </div><div class="line"><a name="l00383"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#af7e22ce6d80d61e8ca37df23880ff1a9"> 383</a></span> <span class="keyword">inline</span> <span class="keywordtype">bool</span> <a class="code" href="cuda__utils_8h.html#af7e22ce6d80d61e8ca37df23880ff1a9">SupportsTensorCore</a>(<span class="keywordtype">int</span> device_id) {</div><div class="line"><a name="l00384"></a><span class="lineno"> 384</span> <span class="comment">// Volta (sm_70) supports TensorCore algos</span></div><div class="line"><a name="l00385"></a><span class="lineno"> 385</span> <span class="keywordflow">return</span> device_id >= 0 &&</div><div class="line"><a name="l00386"></a><span class="lineno"> 386</span> <a class="code" href="cuda__utils_8h.html#aa79f548df23452162de37663f171e99d">ComputeCapabilityMajor</a>(device_id) >=7;</div><div class="line"><a name="l00387"></a><span class="lineno"> 387</span> }</div><div class="line"><a name="l00388"></a><span class="lineno"> 388</span> </div><div class="line"><a name="l00389"></a><span class="lineno"> 389</span> <span class="comment">// The policy if the user hasn't set the environment variable MXNET_CUDA_ALLOW_TENSOR_CORE</span></div><div class="line"><a name="l00390"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#aa7ba00b841d6b7ba443b0e58dac9ab88"> 390</a></span> <span class="preprocessor">#define MXNET_CUDA_ALLOW_TENSOR_CORE_DEFAULT true</span></div><div class="line"><a name="l00391"></a><span class="lineno"> 391</span> </div><div class="line"><a name="l00396"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#a464dee13053e3b0b1006c6307069196c"> 396</a></span> <span class="keyword">inline</span> <span class="keywordtype">bool</span> <a class="code" href="cuda__utils_8h.html#a464dee13053e3b0b1006c6307069196c">GetEnvAllowTensorCore</a>() {</div><div class="line"><a name="l00397"></a><span class="lineno"> 397</span> <span class="comment">// Since these statics are in the '.h' file, they will exist and will be set</span></div><div class="line"><a name="l00398"></a><span class="lineno"> 398</span> <span class="comment">// separately in each compilation unit. Not ideal, but cleaner than creating a</span></div><div class="line"><a name="l00399"></a><span class="lineno"> 399</span> <span class="comment">// cuda_utils.cc solely to have a single instance and initialization.</span></div><div class="line"><a name="l00400"></a><span class="lineno"> 400</span> <span class="keyword">static</span> <span class="keywordtype">bool</span> allow_tensor_core = <span class="keyword">false</span>;</div><div class="line"><a name="l00401"></a><span class="lineno"> 401</span> <span class="keyword">static</span> <span class="keywordtype">bool</span> is_set = <span class="keyword">false</span>;</div><div class="line"><a name="l00402"></a><span class="lineno"> 402</span> <span class="keywordflow">if</span> (!is_set) {</div><div class="line"><a name="l00403"></a><span class="lineno"> 403</span> <span class="comment">// Use of optional<bool> here permits: "0", "1", "true" and "false" to all be legal.</span></div><div class="line"><a name="l00404"></a><span class="lineno"> 404</span> <span class="keywordtype">bool</span> default_value = <a class="code" href="cuda__utils_8h.html#aa7ba00b841d6b7ba443b0e58dac9ab88">MXNET_CUDA_ALLOW_TENSOR_CORE_DEFAULT</a>;</div><div class="line"><a name="l00405"></a><span class="lineno"> 405</span> allow_tensor_core = dmlc::GetEnv(<span class="stringliteral">"MXNET_CUDA_ALLOW_TENSOR_CORE"</span>,</div><div class="line"><a name="l00406"></a><span class="lineno"> 406</span> dmlc::optional<bool>(default_value)).value();</div><div class="line"><a name="l00407"></a><span class="lineno"> 407</span> is_set = <span class="keyword">true</span>;</div><div class="line"><a name="l00408"></a><span class="lineno"> 408</span> }</div><div class="line"><a name="l00409"></a><span class="lineno"> 409</span> <span class="keywordflow">return</span> allow_tensor_core;</div><div class="line"><a name="l00410"></a><span class="lineno"> 410</span> }</div><div class="line"><a name="l00411"></a><span class="lineno"> 411</span> </div><div class="line"><a name="l00412"></a><span class="lineno"> 412</span> <span class="comment">// The policy if the user hasn't set the environment variable</span></div><div class="line"><a name="l00413"></a><span class="lineno"> 413</span> <span class="comment">// CUDNN_TENSOR_OP_MATH_ALLOW_CONVERSION</span></div><div class="line"><a name="l00414"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#aa16d34c218441b0d4074baa8c66a5521"> 414</a></span> <span class="preprocessor">#define MXNET_CUDA_TENSOR_OP_MATH_ALLOW_CONVERSION_DEFAULT false</span></div><div class="line"><a name="l00415"></a><span class="lineno"> 415</span> </div><div class="line"><a name="l00419"></a><span class="lineno"><a class="line" href="cuda__utils_8h.html#ad77e70546b7f35ecba0098caa2d07523"> 419</a></span> <span class="keyword">inline</span> <span class="keywordtype">bool</span> <a class="code" href="cuda__utils_8h.html#ad77e70546b7f35ecba0098caa2d07523">GetEnvAllowTensorCoreConversion</a>() {</div><div class="line"><a name="l00420"></a><span class="lineno"> 420</span> <span class="comment">// Use of optional<bool> here permits: "0", "1", "true" and "false" to all be</span></div><div class="line"><a name="l00421"></a><span class="lineno"> 421</span> <span class="comment">// legal.</span></div><div class="line"><a name="l00422"></a><span class="lineno"> 422</span> <span class="keywordtype">bool</span> default_value = <a class="code" href="cuda__utils_8h.html#aa16d34c218441b0d4074baa8c66a5521">MXNET_CUDA_TENSOR_OP_MATH_ALLOW_CONVERSION_DEFAULT</a>;</div><div class="line"><a name="l00423"></a><span class="lineno"> 423</span> <span class="keywordflow">return</span> dmlc::GetEnv(<span class="stringliteral">"MXNET_CUDA_TENSOR_OP_MATH_ALLOW_CONVERSION"</span>,</div><div class="line"><a name="l00424"></a><span class="lineno"> 424</span> dmlc::optional<bool>(default_value))</div><div class="line"><a name="l00425"></a><span class="lineno"> 425</span> .value();</div><div class="line"><a name="l00426"></a><span class="lineno"> 426</span> }</div><div class="line"><a name="l00427"></a><span class="lineno"> 427</span> </div><div class="line"><a name="l00428"></a><span class="lineno"> 428</span> <span class="preprocessor">#if CUDA_VERSION >= 9000</span></div><div class="line"><a name="l00429"></a><span class="lineno"> 429</span> <span class="comment">// Sets the cuBLAS math mode that determines the 'allow TensorCore' policy. Returns previous.</span></div><div class="line"><a name="l00430"></a><span class="lineno"> 430</span> <span class="keyword">inline</span> cublasMath_t SetCublasMathMode(cublasHandle_t blas_handle, cublasMath_t new_math_type) {</div><div class="line"><a name="l00431"></a><span class="lineno"> 431</span> <span class="keyword">auto</span> handle_math_mode = CUBLAS_DEFAULT_MATH;</div><div class="line"><a name="l00432"></a><span class="lineno"> 432</span> <a class="code" href="cuda__utils_8h.html#a685d7ca3c9370ff471665abcacdeb381">CUBLAS_CALL</a>(cublasGetMathMode(blas_handle, &handle_math_mode));</div><div class="line"><a name="l00433"></a><span class="lineno"> 433</span> <a class="code" href="cuda__utils_8h.html#a685d7ca3c9370ff471665abcacdeb381">CUBLAS_CALL</a>(cublasSetMathMode(blas_handle, new_math_type));</div><div class="line"><a name="l00434"></a><span class="lineno"> 434</span> <span class="keywordflow">return</span> handle_math_mode;</div><div class="line"><a name="l00435"></a><span class="lineno"> 435</span> }</div><div class="line"><a name="l00436"></a><span class="lineno"> 436</span> <span class="preprocessor">#endif</span></div><div class="line"><a name="l00437"></a><span class="lineno"> 437</span> </div><div class="line"><a name="l00438"></a><span class="lineno"> 438</span> <span class="preprocessor">#endif // MXNET_USE_CUDA</span></div><div class="line"><a name="l00439"></a><span class="lineno"> 439</span> </div><div class="line"><a name="l00440"></a><span class="lineno"> 440</span> <span class="preprocessor">#if MXNET_USE_CUDNN</span></div><div class="line"><a name="l00441"></a><span class="lineno"> 441</span> </div><div class="line"><a name="l00442"></a><span class="lineno"> 442</span> <span class="preprocessor">#include <cudnn.h></span></div><div class="line"><a name="l00443"></a><span class="lineno"> 443</span> </div><div class="line"><a name="l00444"></a><span class="lineno"> 444</span> <span class="preprocessor">#define CUDNN_CALL(func) \</span></div><div class="line"><a name="l00445"></a><span class="lineno"> 445</span> <span class="preprocessor"> { \</span></div><div class="line"><a name="l00446"></a><span class="lineno"> 446</span> <span class="preprocessor"> cudnnStatus_t e = (func); \</span></div><div class="line"><a name="l00447"></a><span class="lineno"> 447</span> <span class="preprocessor"> CHECK_EQ(e, CUDNN_STATUS_SUCCESS) << "cuDNN: " << cudnnGetErrorString(e); \</span></div><div class="line"><a name="l00448"></a><span class="lineno"> 448</span> <span class="preprocessor"> }</span></div><div class="line"><a name="l00449"></a><span class="lineno"> 449</span> </div><div class="line"><a name="l00457"></a><span class="lineno"> 457</span> <span class="keyword">inline</span> <span class="keywordtype">int</span> MaxForwardAlgos(cudnnHandle_t cudnn_handle) {</div><div class="line"><a name="l00458"></a><span class="lineno"> 458</span> <span class="preprocessor">#if CUDNN_MAJOR >= 7</span></div><div class="line"><a name="l00459"></a><span class="lineno"> 459</span> <span class="keywordtype">int</span> max_algos = 0;</div><div class="line"><a name="l00460"></a><span class="lineno"> 460</span> CUDNN_CALL(cudnnGetConvolutionForwardAlgorithmMaxCount(cudnn_handle, &max_algos));</div><div class="line"><a name="l00461"></a><span class="lineno"> 461</span> <span class="keywordflow">return</span> max_algos;</div><div class="line"><a name="l00462"></a><span class="lineno"> 462</span> <span class="preprocessor">#else</span></div><div class="line"><a name="l00463"></a><span class="lineno"> 463</span> <span class="keywordflow">return</span> 10;</div><div class="line"><a name="l00464"></a><span class="lineno"> 464</span> <span class="preprocessor">#endif</span></div><div class="line"><a name="l00465"></a><span class="lineno"> 465</span> }</div><div class="line"><a name="l00466"></a><span class="lineno"> 466</span> </div><div class="line"><a name="l00474"></a><span class="lineno"> 474</span> <span class="keyword">inline</span> <span class="keywordtype">int</span> MaxBackwardFilterAlgos(cudnnHandle_t cudnn_handle) {</div><div class="line"><a name="l00475"></a><span class="lineno"> 475</span> <span class="preprocessor">#if CUDNN_MAJOR >= 7</span></div><div class="line"><a name="l00476"></a><span class="lineno"> 476</span> <span class="keywordtype">int</span> max_algos = 0;</div><div class="line"><a name="l00477"></a><span class="lineno"> 477</span> CUDNN_CALL(cudnnGetConvolutionBackwardFilterAlgorithmMaxCount(cudnn_handle, &max_algos));</div><div class="line"><a name="l00478"></a><span class="lineno"> 478</span> <span class="keywordflow">return</span> max_algos;</div><div class="line"><a name="l00479"></a><span class="lineno"> 479</span> <span class="preprocessor">#else</span></div><div class="line"><a name="l00480"></a><span class="lineno"> 480</span> <span class="keywordflow">return</span> 10;</div><div class="line"><a name="l00481"></a><span class="lineno"> 481</span> <span class="preprocessor">#endif</span></div><div class="line"><a name="l00482"></a><span class="lineno"> 482</span> }</div><div class="line"><a name="l00483"></a><span class="lineno"> 483</span> </div><div class="line"><a name="l00491"></a><span class="lineno"> 491</span> <span class="keyword">inline</span> <span class="keywordtype">int</span> MaxBackwardDataAlgos(cudnnHandle_t cudnn_handle) {</div><div class="line"><a name="l00492"></a><span class="lineno"> 492</span> <span class="preprocessor">#if CUDNN_MAJOR >= 7</span></div><div class="line"><a name="l00493"></a><span class="lineno"> 493</span> <span class="keywordtype">int</span> max_algos = 0;</div><div class="line"><a name="l00494"></a><span class="lineno"> 494</span> CUDNN_CALL(cudnnGetConvolutionBackwardDataAlgorithmMaxCount(cudnn_handle, &max_algos));</div><div class="line"><a name="l00495"></a><span class="lineno"> 495</span> <span class="keywordflow">return</span> max_algos;</div><div class="line"><a name="l00496"></a><span class="lineno"> 496</span> <span class="preprocessor">#else</span></div><div class="line"><a name="l00497"></a><span class="lineno"> 497</span> <span class="keywordflow">return</span> 10;</div><div class="line"><a name="l00498"></a><span class="lineno"> 498</span> <span class="preprocessor">#endif</span></div><div class="line"><a name="l00499"></a><span class="lineno"> 499</span> }</div><div class="line"><a name="l00500"></a><span class="lineno"> 500</span> </div><div class="line"><a name="l00501"></a><span class="lineno"> 501</span> <span class="preprocessor">#endif // MXNET_USE_CUDNN</span></div><div class="line"><a name="l00502"></a><span class="lineno"> 502</span> </div><div class="line"><a name="l00503"></a><span class="lineno"> 503</span> <span class="comment">// Overload atomicAdd to work for floats on all architectures</span></div><div class="line"><a name="l00504"></a><span class="lineno"> 504</span> <span class="preprocessor">#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ < 600</span></div><div class="line"><a name="l00505"></a><span class="lineno"> 505</span> <span class="comment">// From CUDA Programming Guide</span></div><div class="line"><a name="l00506"></a><span class="lineno"> 506</span> <span class="keyword">static</span> <span class="keyword">inline</span> __device__ <span class="keywordtype">void</span> atomicAdd(<span class="keywordtype">double</span> *address, <span class="keywordtype">double</span> val) {</div><div class="line"><a name="l00507"></a><span class="lineno"> 507</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">long</span> <span class="keywordtype">long</span>* address_as_ull = <span class="comment">// NOLINT(*)</span></div><div class="line"><a name="l00508"></a><span class="lineno"> 508</span> <span class="keyword">reinterpret_cast<</span><span class="keywordtype">unsigned</span> <span class="keywordtype">long</span> <span class="keywordtype">long</span>*<span class="keyword">></span>(address); <span class="comment">// NOLINT(*)</span></div><div class="line"><a name="l00509"></a><span class="lineno"> 509</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">long</span> <span class="keywordtype">long</span> old = *address_as_ull; <span class="comment">// NOLINT(*)</span></div><div class="line"><a name="l00510"></a><span class="lineno"> 510</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">long</span> <span class="keywordtype">long</span> assumed; <span class="comment">// NOLINT(*)</span></div><div class="line"><a name="l00511"></a><span class="lineno"> 511</span> </div><div class="line"><a name="l00512"></a><span class="lineno"> 512</span> <span class="keywordflow">do</span> {</div><div class="line"><a name="l00513"></a><span class="lineno"> 513</span> assumed = old;</div><div class="line"><a name="l00514"></a><span class="lineno"> 514</span> old = atomicCAS(address_as_ull, assumed,</div><div class="line"><a name="l00515"></a><span class="lineno"> 515</span> __double_as_longlong(val +</div><div class="line"><a name="l00516"></a><span class="lineno"> 516</span> __longlong_as_double(assumed)));</div><div class="line"><a name="l00517"></a><span class="lineno"> 517</span> </div><div class="line"><a name="l00518"></a><span class="lineno"> 518</span> <span class="comment">// Note: uses integer comparison to avoid hang in case of NaN (since NaN != NaN)</span></div><div class="line"><a name="l00519"></a><span class="lineno"> 519</span> } <span class="keywordflow">while</span> (assumed != old);</div><div class="line"><a name="l00520"></a><span class="lineno"> 520</span> }</div><div class="line"><a name="l00521"></a><span class="lineno"> 521</span> <span class="preprocessor">#endif</span></div><div class="line"><a name="l00522"></a><span class="lineno"> 522</span> </div><div class="line"><a name="l00523"></a><span class="lineno"> 523</span> <span class="comment">// Overload atomicAdd for half precision</span></div><div class="line"><a name="l00524"></a><span class="lineno"> 524</span> <span class="comment">// Taken from:</span></div><div class="line"><a name="l00525"></a><span class="lineno"> 525</span> <span class="comment">// https://github.com/torch/cutorch/blob/master/lib/THC/THCAtomics.cuh</span></div><div class="line"><a name="l00526"></a><span class="lineno"> 526</span> <span class="preprocessor">#if defined(__CUDA_ARCH__)</span></div><div class="line"><a name="l00527"></a><span class="lineno"> 527</span> <span class="keyword">static</span> <span class="keyword">inline</span> __device__ <span class="keywordtype">void</span> atomicAdd(mshadow::half::half_t *address,</div><div class="line"><a name="l00528"></a><span class="lineno"> 528</span> mshadow::half::half_t val) {</div><div class="line"><a name="l00529"></a><span class="lineno"> 529</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *address_as_ui =</div><div class="line"><a name="l00530"></a><span class="lineno"> 530</span> <span class="keyword">reinterpret_cast<</span><span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *<span class="keyword">></span>(<span class="keyword">reinterpret_cast<</span><span class="keywordtype">char</span> *<span class="keyword">></span>(address) -</div><div class="line"><a name="l00531"></a><span class="lineno"> 531</span> (reinterpret_cast<size_t>(address) & 2));</div><div class="line"><a name="l00532"></a><span class="lineno"> 532</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> old = *address_as_ui;</div><div class="line"><a name="l00533"></a><span class="lineno"> 533</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> assumed;</div><div class="line"><a name="l00534"></a><span class="lineno"> 534</span> </div><div class="line"><a name="l00535"></a><span class="lineno"> 535</span> <span class="keywordflow">do</span> {</div><div class="line"><a name="l00536"></a><span class="lineno"> 536</span> assumed = old;</div><div class="line"><a name="l00537"></a><span class="lineno"> 537</span> mshadow::half::half_t hsum;</div><div class="line"><a name="l00538"></a><span class="lineno"> 538</span> hsum.half_ =</div><div class="line"><a name="l00539"></a><span class="lineno"> 539</span> <span class="keyword">reinterpret_cast<</span><span class="keywordtype">size_t</span><span class="keyword">></span>(address) & 2 ? (old >> 16) : (old & 0xffff);</div><div class="line"><a name="l00540"></a><span class="lineno"> 540</span> hsum += val;</div><div class="line"><a name="l00541"></a><span class="lineno"> 541</span> old = <span class="keyword">reinterpret_cast<</span><span class="keywordtype">size_t</span><span class="keyword">></span>(address) & 2</div><div class="line"><a name="l00542"></a><span class="lineno"> 542</span> ? (old & 0xffff) | (hsum.half_ << 16)</div><div class="line"><a name="l00543"></a><span class="lineno"> 543</span> : (old & 0xffff0000) | hsum.half_;</div><div class="line"><a name="l00544"></a><span class="lineno"> 544</span> old = atomicCAS(address_as_ui, assumed, old);</div><div class="line"><a name="l00545"></a><span class="lineno"> 545</span> } <span class="keywordflow">while</span> (assumed != old);</div><div class="line"><a name="l00546"></a><span class="lineno"> 546</span> }</div><div class="line"><a name="l00547"></a><span class="lineno"> 547</span> </div><div class="line"><a name="l00548"></a><span class="lineno"> 548</span> <span class="keyword">static</span> <span class="keyword">inline</span> __device__ <span class="keywordtype">void</span> atomicAdd(uint8_t *address, uint8_t val) {</div><div class="line"><a name="l00549"></a><span class="lineno"> 549</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> * address_as_ui = (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *) (address - ((<span class="keywordtype">size_t</span>)address & 0x3));</div><div class="line"><a name="l00550"></a><span class="lineno"> 550</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> old = *address_as_ui;</div><div class="line"><a name="l00551"></a><span class="lineno"> 551</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> shift = (((size_t)address & 0x3) << 3);</div><div class="line"><a name="l00552"></a><span class="lineno"> 552</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespacemxnet_1_1cpp.html#acb4fccaa546283e233c15b46f6465443">sum</a>;</div><div class="line"><a name="l00553"></a><span class="lineno"> 553</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> assumed;</div><div class="line"><a name="l00554"></a><span class="lineno"> 554</span> </div><div class="line"><a name="l00555"></a><span class="lineno"> 555</span> <span class="keywordflow">do</span> {</div><div class="line"><a name="l00556"></a><span class="lineno"> 556</span> assumed = old;</div><div class="line"><a name="l00557"></a><span class="lineno"> 557</span> sum = val + <span class="keyword">static_cast<</span>uint8_t<span class="keyword">></span>((old >> shift) & 0xff);</div><div class="line"><a name="l00558"></a><span class="lineno"> 558</span> old = (old & ~(0x000000ff << shift)) | (sum << shift);</div><div class="line"><a name="l00559"></a><span class="lineno"> 559</span> old = atomicCAS(address_as_ui, assumed, old);</div><div class="line"><a name="l00560"></a><span class="lineno"> 560</span> } <span class="keywordflow">while</span> (assumed != old);</div><div class="line"><a name="l00561"></a><span class="lineno"> 561</span> }</div><div class="line"><a name="l00562"></a><span class="lineno"> 562</span> </div><div class="line"><a name="l00563"></a><span class="lineno"> 563</span> <span class="keyword">static</span> <span class="keyword">inline</span> __device__ <span class="keywordtype">void</span> atomicAdd(int8_t *address, int8_t val) {</div><div class="line"><a name="l00564"></a><span class="lineno"> 564</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> * address_as_ui = (<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> *) (address - ((<span class="keywordtype">size_t</span>)address & 0x3));</div><div class="line"><a name="l00565"></a><span class="lineno"> 565</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> old = *address_as_ui;</div><div class="line"><a name="l00566"></a><span class="lineno"> 566</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> shift = (((size_t)address & 0x3) << 3);</div><div class="line"><a name="l00567"></a><span class="lineno"> 567</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="namespacemxnet_1_1cpp.html#acb4fccaa546283e233c15b46f6465443">sum</a>;</div><div class="line"><a name="l00568"></a><span class="lineno"> 568</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> assumed;</div><div class="line"><a name="l00569"></a><span class="lineno"> 569</span> </div><div class="line"><a name="l00570"></a><span class="lineno"> 570</span> <span class="keywordflow">do</span> {</div><div class="line"><a name="l00571"></a><span class="lineno"> 571</span> assumed = old;</div><div class="line"><a name="l00572"></a><span class="lineno"> 572</span> sum = val + <span class="keyword">static_cast<</span>int8_t<span class="keyword">></span>((old >> shift) & 0xff);</div><div class="line"><a name="l00573"></a><span class="lineno"> 573</span> old = (old & ~(0x000000ff << shift)) | (sum << shift);</div><div class="line"><a name="l00574"></a><span class="lineno"> 574</span> old = atomicCAS(address_as_ui, assumed, old);</div><div class="line"><a name="l00575"></a><span class="lineno"> 575</span> } <span class="keywordflow">while</span> (assumed != old);</div><div class="line"><a name="l00576"></a><span class="lineno"> 576</span> }</div><div class="line"><a name="l00577"></a><span class="lineno"> 577</span> </div><div class="line"><a name="l00578"></a><span class="lineno"> 578</span> <span class="comment">// Overload atomicAdd to work for signed int64 on all architectures</span></div><div class="line"><a name="l00579"></a><span class="lineno"> 579</span> <span class="keyword">static</span> <span class="keyword">inline</span> __device__ <span class="keywordtype">void</span> atomicAdd(int64_t *address, int64_t val) {</div><div class="line"><a name="l00580"></a><span class="lineno"> 580</span> atomicAdd(reinterpret_cast<unsigned long long*>(address), static_cast<unsigned long long>(val)); <span class="comment">// NOLINT</span></div><div class="line"><a name="l00581"></a><span class="lineno"> 581</span> }</div><div class="line"><a name="l00582"></a><span class="lineno"> 582</span> </div><div class="line"><a name="l00583"></a><span class="lineno"> 583</span> <span class="keyword">template</span> <<span class="keyword">typename</span> DType></div><div class="line"><a name="l00584"></a><span class="lineno"> 584</span> __device__ <span class="keyword">inline</span> DType ldg(<span class="keyword">const</span> DType* address) {</div><div class="line"><a name="l00585"></a><span class="lineno"> 585</span> <span class="preprocessor">#if __CUDA_ARCH__ >= 350</span></div><div class="line"><a name="l00586"></a><span class="lineno"> 586</span> <span class="keywordflow">return</span> __ldg(address);</div><div class="line"><a name="l00587"></a><span class="lineno"> 587</span> <span class="preprocessor">#else</span></div><div class="line"><a name="l00588"></a><span class="lineno"> 588</span> <span class="keywordflow">return</span> *address;</div><div class="line"><a name="l00589"></a><span class="lineno"> 589</span> <span class="preprocessor">#endif</span></div><div class="line"><a name="l00590"></a><span class="lineno"> 590</span> }</div><div class="line"><a name="l00591"></a><span class="lineno"> 591</span> <span class="preprocessor">#endif</span></div><div class="line"><a name="l00592"></a><span class="lineno"> 592</span> </div><div class="line"><a name="l00593"></a><span class="lineno"> 593</span> <span class="preprocessor">#endif // MXNET_COMMON_CUDA_UTILS_H_</span></div><div class="ttc" id="cuda__utils_8h_html_a685d7ca3c9370ff471665abcacdeb381"><div class="ttname"><a href="cuda__utils_8h.html#a685d7ca3c9370ff471665abcacdeb381">CUBLAS_CALL</a></div><div class="ttdeci">#define CUBLAS_CALL(func)</div><div class="ttdoc">Protected cuBLAS call. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:100</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_aa79f548df23452162de37663f171e99d"><div class="ttname"><a href="cuda__utils_8h.html#aa79f548df23452162de37663f171e99d">ComputeCapabilityMajor</a></div><div class="ttdeci">int ComputeCapabilityMajor(int device_id)</div><div class="ttdoc">Determine major version number of the gpu&#39;s cuda compute architecture. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:330</div></div> |
| <div class="ttc" id="classmxnet_1_1common_1_1cuda_1_1DeviceStore_html"><div class="ttname"><a href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html">mxnet::common::cuda::DeviceStore</a></div><div class="ttdef"><b>Definition:</b> cuda_utils.h:286</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_ad77e70546b7f35ecba0098caa2d07523"><div class="ttname"><a href="cuda__utils_8h.html#ad77e70546b7f35ecba0098caa2d07523">GetEnvAllowTensorCoreConversion</a></div><div class="ttdeci">bool GetEnvAllowTensorCoreConversion()</div><div class="ttdoc">Returns global policy for TensorCore implicit type casting. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:419</div></div> |
| <div class="ttc" id="namespacemxnet_html"><div class="ttname"><a href="namespacemxnet.html">mxnet</a></div><div class="ttdoc">namespace of mxnet </div><div class="ttdef"><b>Definition:</b> base.h:118</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_a464dee13053e3b0b1006c6307069196c"><div class="ttname"><a href="cuda__utils_8h.html#a464dee13053e3b0b1006c6307069196c">GetEnvAllowTensorCore</a></div><div class="ttdeci">bool GetEnvAllowTensorCore()</div><div class="ttdoc">Returns global policy for TensorCore algo use. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:396</div></div> |
| <div class="ttc" id="classmxnet_1_1common_1_1cuda_1_1DeviceStore_html_ad9878a09a93d4fcaf9d0639b3613d9f7"><div class="ttname"><a href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#ad9878a09a93d4fcaf9d0639b3613d9f7">mxnet::common::cuda::DeviceStore::DeviceStore</a></div><div class="ttdeci">DeviceStore(int requested_device=-1, bool restore=true)</div><div class="ttdoc">default constructor- only optionally restores previous device </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:289</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_a9779e3ad0efd0faec7fbe431c0db896d"><div class="ttname"><a href="cuda__utils_8h.html#a9779e3ad0efd0faec7fbe431c0db896d">SMArch</a></div><div class="ttdeci">int SMArch(int device_id)</div><div class="ttdoc">Return the integer SM architecture (e.g. Volta = 70). </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:354</div></div> |
| <div class="ttc" id="namespacemxnet_1_1common_1_1cuda_html_a03888f252f813f6d052ae84bf8801498"><div class="ttname"><a href="namespacemxnet_1_1common_1_1cuda.html#a03888f252f813f6d052ae84bf8801498">mxnet::common::cuda::CudaMin</a></div><div class="ttdeci">DType __device__ CudaMin(DType a, DType b)</div><div class="ttdef"><b>Definition:</b> cuda_utils.h:282</div></div> |
| <div class="ttc" id="classmxnet_1_1common_1_1cuda_1_1DeviceStore_html_a01163fd4915e74bdd81dd7305917f0e4"><div class="ttname"><a href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#a01163fd4915e74bdd81dd7305917f0e4">mxnet::common::cuda::DeviceStore::SetDevice</a></div><div class="ttdeci">void SetDevice(int device)</div><div class="ttdef"><b>Definition:</b> cuda_utils.h:308</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_afb4268417c1d8886a39142c85c8f188f"><div class="ttname"><a href="cuda__utils_8h.html#afb4268417c1d8886a39142c85c8f188f">SupportsFloat16Compute</a></div><div class="ttdeci">bool SupportsFloat16Compute(int device_id)</div><div class="ttdoc">Determine whether a cuda-capable gpu&#39;s architecture supports float16 math. Assume not if device_id is...</div><div class="ttdef"><b>Definition:</b> cuda_utils.h:366</div></div> |
| <div class="ttc" id="namespacemxnet_1_1common_1_1cuda_html_a6f3ee04eb382c57e10916108db3efd80"><div class="ttname"><a href="namespacemxnet_1_1common_1_1cuda.html#a6f3ee04eb382c57e10916108db3efd80">mxnet::common::cuda::CudaMax</a></div><div class="ttdeci">DType __device__ CudaMax(DType a, DType b)</div><div class="ttdef"><b>Definition:</b> cuda_utils.h:277</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_aa16d34c218441b0d4074baa8c66a5521"><div class="ttname"><a href="cuda__utils_8h.html#aa16d34c218441b0d4074baa8c66a5521">MXNET_CUDA_TENSOR_OP_MATH_ALLOW_CONVERSION_DEFAULT</a></div><div class="ttdeci">#define MXNET_CUDA_TENSOR_OP_MATH_ALLOW_CONVERSION_DEFAULT</div><div class="ttdef"><b>Definition:</b> cuda_utils.h:414</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_af7e22ce6d80d61e8ca37df23880ff1a9"><div class="ttname"><a href="cuda__utils_8h.html#af7e22ce6d80d61e8ca37df23880ff1a9">SupportsTensorCore</a></div><div class="ttdeci">bool SupportsTensorCore(int device_id)</div><div class="ttdoc">Determine whether a cuda-capable gpu&#39;s architecture supports Tensor Core math. Assume not if device_i...</div><div class="ttdef"><b>Definition:</b> cuda_utils.h:383</div></div> |
| <div class="ttc" id="namespacemxnet_1_1common_1_1cuda_html_abf9bcb4cb696e9ae61b818510dac39c8"><div class="ttname"><a href="namespacemxnet_1_1common_1_1cuda.html#abf9bcb4cb696e9ae61b818510dac39c8">mxnet::common::cuda::CusolverGetErrorString</a></div><div class="ttdeci">const char * CusolverGetErrorString(cusolverStatus_t error)</div><div class="ttdoc">Get string representation of cuSOLVER errors. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:215</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_aa7ba00b841d6b7ba443b0e58dac9ab88"><div class="ttname"><a href="cuda__utils_8h.html#aa7ba00b841d6b7ba443b0e58dac9ab88">MXNET_CUDA_ALLOW_TENSOR_CORE_DEFAULT</a></div><div class="ttdeci">#define MXNET_CUDA_ALLOW_TENSOR_CORE_DEFAULT</div><div class="ttdef"><b>Definition:</b> cuda_utils.h:390</div></div> |
| <div class="ttc" id="namespacemxnet_1_1common_1_1cuda_html_a97c06b2f4d26445a7386b0f54fae1feb"><div class="ttname"><a href="namespacemxnet_1_1common_1_1cuda.html#a97c06b2f4d26445a7386b0f54fae1feb">mxnet::common::cuda::CurandGetErrorString</a></div><div class="ttdeci">const char * CurandGetErrorString(curandStatus_t status)</div><div class="ttdoc">Get string representation of cuRAND errors. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:244</div></div> |
| <div class="ttc" id="classmxnet_1_1common_1_1cuda_1_1DeviceStore_html_a701d38ae493688ee2136995fe8611aa0"><div class="ttname"><a href="classmxnet_1_1common_1_1cuda_1_1DeviceStore.html#a701d38ae493688ee2136995fe8611aa0">mxnet::common::cuda::DeviceStore::~DeviceStore</a></div><div class="ttdeci">~DeviceStore()</div><div class="ttdef"><b>Definition:</b> cuda_utils.h:300</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_a7c16e8770e4f399cabed1fc231ffd9b6"><div class="ttname"><a href="cuda__utils_8h.html#a7c16e8770e4f399cabed1fc231ffd9b6">ComputeCapabilityMinor</a></div><div class="ttdeci">int ComputeCapabilityMinor(int device_id)</div><div class="ttdoc">Determine minor version number of the gpu&#39;s cuda compute architecture. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:342</div></div> |
| <div class="ttc" id="cuda__utils_8h_html_a06cc7d24ca66505e69f5ad40009f5e8d"><div class="ttname"><a href="cuda__utils_8h.html#a06cc7d24ca66505e69f5ad40009f5e8d">CUDA_CALL</a></div><div class="ttdeci">#define CUDA_CALL(func)</div><div class="ttdoc">Protected CUDA call. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:87</div></div> |
| <div class="ttc" id="namespacemxnet_1_1common_1_1cuda_html_a9feee613a4f16a954dd68e55345a72ac"><div class="ttname"><a href="namespacemxnet_1_1common_1_1cuda.html#a9feee613a4f16a954dd68e55345a72ac">mxnet::common::cuda::CublasGetErrorString</a></div><div class="ttdeci">const char * CublasGetErrorString(cublasStatus_t error)</div><div class="ttdoc">Get string representation of cuBLAS errors. </div><div class="ttdef"><b>Definition:</b> cuda_utils.h:184</div></div> |
| <div class="ttc" id="namespacemxnet_1_1cpp_html_acb4fccaa546283e233c15b46f6465443"><div class="ttname"><a href="namespacemxnet_1_1cpp.html#acb4fccaa546283e233c15b46f6465443">mxnet::cpp::sum</a></div><div class="ttdeci">Symbol sum(const std::string &symbol_name, Symbol data, dmlc::optional< Shape > axis=dmlc::optional< Shape >(), bool keepdims=false, bool exclude=false)</div><div class="ttdef"><b>Definition:</b> op.h:2567</div></div> |
| </div><!-- fragment --></div><!-- contents --> |
| <!-- start footer part --> |
| <hr class="footer"/><address class="footer"><small> |
| Generated on Thu Sep 19 2019 12:37:33 for mxnet by <a href="http://www.doxygen.org/index.html"> |
| <img alt="doxygen" class="footer" src="doxygen.png"/> |
| </a> 1.8.11 |
| </small></address> |
| </body> |
| </html> |
