blob: e083207aec8917e9bea2ec121d49496452e1ab53 [file] [log] [blame]
<!-- HTML header for doxygen 1.8.9.1-->
<!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 1.8.11"/>
<meta name="robots" content="NOINDEX, NOFOLLOW" /> <!-- Prevent indexing by search engines -->
<title>Compute Library: src/core/CL/cl_kernels/convolution_layer.cl File Reference</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="navtree.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="resize.js"></script>
<script type="text/javascript" src="navtreedata.js"></script>
<script type="text/javascript" src="navtree.js"></script>
<script type="text/javascript">
$(document).ready(initResizable);
$(window).load(resizeHeight);
</script>
<link href="search/search.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="search/searchdata.js"></script>
<script type="text/javascript" src="search/search.js"></script>
<script type="text/javascript">
$(document).ready(function() { init_search(); });
</script>
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
extensions: ["tex2jax.js"],
jax: ["input/TeX","output/HTML-CSS"],
});
</script><script type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js"></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">Compute Library
&#160;<span id="projectnumber">v17.06</span>
</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 id="navrow1" class="tabs">
<ul class="tablist">
<li><a href="index.xhtml"><span>Main&#160;Page</span></a></li>
<li><a href="pages.xhtml"><span>Related&#160;Pages</span></a></li>
<li><a href="namespaces.xhtml"><span>Namespaces</span></a></li>
<li><a href="annotated.xhtml"><span>Data&#160;Structures</span></a></li>
<li class="current"><a href="files.xhtml"><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.xhtml"><span>File&#160;List</span></a></li>
<li><a href="globals.xhtml"><span>Globals</span></a></li>
</ul>
</div>
</div><!-- top -->
<div id="side-nav" class="ui-resizable side-nav-resizable">
<div id="nav-tree">
<div id="nav-tree-contents">
<div id="nav-sync" class="sync"></div>
</div>
</div>
<div id="splitbar" style="-moz-user-select:none;"
class="ui-resizable-handle">
</div>
</div>
<script type="text/javascript">
$(document).ready(function(){initNavTree('convolution__layer_8cl.xhtml','');});
</script>
<div id="doc-content">
<!-- window showing the filter options -->
<div id="MSearchSelectWindow"
onmouseover="return searchBox.OnSearchSelectShow()"
onmouseout="return searchBox.OnSearchSelectHide()"
onkeydown="return searchBox.OnSearchSelectKey(event)">
</div>
<!-- iframe showing the search results (closed by default) -->
<div id="MSearchResultsWindow">
<iframe src="javascript:void(0)" frameborder="0"
name="MSearchResults" id="MSearchResults">
</iframe>
</div>
<div class="header">
<div class="summary">
<a href="#func-members">Functions</a> </div>
<div class="headertitle">
<div class="title">convolution_layer.cl File Reference</div> </div>
</div><!--header-->
<div class="contents">
<div class="textblock"><code>#include &quot;<a class="el" href="helpers_8h_source.xhtml">helpers.h</a>&quot;</code><br />
</div>
<p><a href="convolution__layer_8cl_source.xhtml">Go to the source code of this file.</a></p>
<table class="memberdecls">
<tr class="heading"><td colspan="2"><h2 class="groupheader"><a name="func-members"></a>
Functions</h2></td></tr>
<tr class="memitem:ab6358b28d1e7bc946950075d94412013"><td class="memItemLeft" align="right" valign="top">__kernel void&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="convolution__layer_8cl.xhtml#ab6358b28d1e7bc946950075d94412013">reshape_to_columns</a> (__global uchar *src_ptr, uint src_stride_x, uint src_step_x, uint src_stride_y, uint src_step_y, uint src_stride_z, uint src_step_z, uint src_offset_first_element_in_bytes, __global uchar *dst_ptr, uint dst_stride_x, uint dst_step_x, uint dst_stride_y, uint dst_step_y, uint dst_offset_first_element_in_bytes, __global uchar *bias_ptr, uint bias_stride_x, uint bias_step_x, uint bias_offset_first_element_in_bytes, uint width, uint height, uint depth, uint total_filters)</td></tr>
<tr class="memdesc:ab6358b28d1e7bc946950075d94412013"><td class="mdescLeft">&#160;</td><td class="mdescRight">This kernel reshapes the tensor's low three dimensions to single column. <a href="#ab6358b28d1e7bc946950075d94412013">More...</a><br /></td></tr>
<tr class="separator:ab6358b28d1e7bc946950075d94412013"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:afaddfccc926506181141b2dc86cb73f1"><td class="memItemLeft" align="right" valign="top">__kernel void&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="convolution__layer_8cl.xhtml#afaddfccc926506181141b2dc86cb73f1">im2col_generic</a> (__global uchar *src_ptr, uint src_stride_x, uint src_step_x, uint src_stride_y, uint src_step_y, uint src_stride_z, uint src_step_z, uint src_offset_first_element_in_bytes, __global uchar *dst_ptr, uint dst_stride_x, uint dst_step_x, uint dst_stride_y, uint dst_step_y, uint dst_offset_first_element_in_bytes, int kernel_size, int kernel_depth, int width, int2 input_dims, int2 strides, int2 paddings)</td></tr>
<tr class="memdesc:afaddfccc926506181141b2dc86cb73f1"><td class="mdescLeft">&#160;</td><td class="mdescRight">This kernel performs a reshaping of the input tensor to a tensor used to perform convolution using GEMM. <a href="#afaddfccc926506181141b2dc86cb73f1">More...</a><br /></td></tr>
<tr class="separator:afaddfccc926506181141b2dc86cb73f1"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:af11bad06b2f7c26cc6ee2f93add5dcbf"><td class="memItemLeft" align="right" valign="top">__kernel void&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="convolution__layer_8cl.xhtml#af11bad06b2f7c26cc6ee2f93add5dcbf">col2im</a> (__global uchar *src_ptr, uint src_stride_x, uint src_step_x, uint src_stride_y, uint src_step_y, uint src_offset_first_element_in_bytes, __global uchar *dst_ptr, uint dst_stride_x, uint dst_step_x, uint dst_stride_y, uint dst_step_y, uint dst_stride_z, uint dst_step_z, uint dst_offset_first_element_in_bytes, uint width)</td></tr>
<tr class="memdesc:af11bad06b2f7c26cc6ee2f93add5dcbf"><td class="mdescLeft">&#160;</td><td class="mdescRight">This kernel performs a reshaping of the output of the convolution layer. <a href="#af11bad06b2f7c26cc6ee2f93add5dcbf">More...</a><br /></td></tr>
<tr class="separator:af11bad06b2f7c26cc6ee2f93add5dcbf"><td class="memSeparator" colspan="2">&#160;</td></tr>
<tr class="memitem:a6f31f2c8d606e1b57d9bdfacd416024c"><td class="memItemLeft" align="right" valign="top">__kernel void&#160;</td><td class="memItemRight" valign="bottom"><a class="el" href="convolution__layer_8cl.xhtml#a6f31f2c8d606e1b57d9bdfacd416024c">im2col_reduced</a> (__global uchar *src_ptr, uint src_stride_x, uint src_step_x, uint src_stride_y, uint src_step_y, uint src_stride_z, uint src_step_z, uint src_offset_first_element_in_bytes, __global uchar *dst_ptr, uint dst_stride_x, uint dst_step_x, uint dst_offset_first_element_in_bytes, uint width, uint height)</td></tr>
<tr class="memdesc:a6f31f2c8d606e1b57d9bdfacd416024c"><td class="mdescLeft">&#160;</td><td class="mdescRight">This kernel reshapes the tensor's low three dimensions to single row for GEMM operation. <a href="#a6f31f2c8d606e1b57d9bdfacd416024c">More...</a><br /></td></tr>
<tr class="separator:a6f31f2c8d606e1b57d9bdfacd416024c"><td class="memSeparator" colspan="2">&#160;</td></tr>
</table>
<h2 class="groupheader">Function Documentation</h2>
<a class="anchor" id="af11bad06b2f7c26cc6ee2f93add5dcbf"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">__kernel void col2im </td>
<td>(</td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>src_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>dst_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_stride_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_step_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_stride_z</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_step_z</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>width</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>This kernel performs a reshaping of the output of the convolution layer. </p>
<dl class="section note"><dt>Note</dt><dd>The data type must be passed at compile time using -DDATA_TYPE: e.g. -DDATA_TYPE=float</dd></dl>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramdir">[in]</td><td class="paramname">src_ptr</td><td>Pointer to the source tensor. Supported data types: F16, F32 </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_x</td><td>Stride of the source tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_x</td><td>src_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_y</td><td>Stride of the source tensor in Y dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_y</td><td>src_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_offset_first_element_in_bytes</td><td>The offset of the first element in the source tensor </td></tr>
<tr><td class="paramdir">[out]</td><td class="paramname">dst_ptr</td><td>Pointer to the destination tensor. Supported data types: F16, F32 </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_stride_x</td><td>Stride of the destination tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_step_x</td><td>dst_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_stride_y</td><td>Stride of the destination tensor in Y dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_step_y</td><td>dst_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_stride_z</td><td>Stride of the destination tensor in Z dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_step_z</td><td>dst_stride_z * number of elements along Z processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_offset_first_element_in_bytes</td><td>The offset of the first element in the destination tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">width</td><td>The output tensor width </td></tr>
</table>
</dd>
</dl>
<p>Definition at line <a class="el" href="convolution__layer_8cl_source.xhtml#l00187">187</a> of file <a class="el" href="convolution__layer_8cl_source.xhtml">convolution_layer.cl</a>.</p>
<p>References <a class="el" href="helpers_8h_source.xhtml#l00073">CONVERT_TO_IMAGE_STRUCT</a>, <a class="el" href="helpers_8h_source.xhtml#l00083">CONVERT_TO_TENSOR3D_STRUCT_NO_STEP</a>, <a class="el" href="convolution3x3_8cl_source.xhtml#l00027">DATA_TYPE</a>, <a class="el" href="helpers_8h_source.xhtml#l00097">Image::ptr</a>, <a class="el" href="helpers_8h_source.xhtml#l00106">Tensor3D::ptr</a>, <a class="el" href="helpers_8h_source.xhtml#l00108">Tensor3D::stride_x</a>, <a class="el" href="helpers_8h_source.xhtml#l00109">Tensor3D::stride_y</a>, and <a class="el" href="helpers_8h_source.xhtml#l00110">Tensor3D::stride_z</a>.</p>
<div class="fragment"><div class="line"><a name="l00191"></a><span class="lineno"> 191</span>&#160;{</div><div class="line"><a name="l00192"></a><span class="lineno"> 192</span>&#160; <a class="code" href="struct_image.xhtml">Image</a> src = <a class="code" href="helpers_8h.xhtml#aebe814363556c244be043b13e7969197">CONVERT_TO_IMAGE_STRUCT</a>(src);</div><div class="line"><a name="l00193"></a><span class="lineno"> 193</span>&#160; <a class="code" href="struct_tensor3_d.xhtml">Tensor3D</a> dst = <a class="code" href="helpers_8h.xhtml#a79e8e562daa6599317d2d1cd86ef1bf2">CONVERT_TO_TENSOR3D_STRUCT_NO_STEP</a>(dst);</div><div class="line"><a name="l00194"></a><span class="lineno"> 194</span>&#160;</div><div class="line"><a name="l00195"></a><span class="lineno"> 195</span>&#160; <span class="keywordtype">int</span> idx = get_global_id(0) * dst.<a class="code" href="struct_tensor3_d.xhtml#ad5ff7a2b2bd0eec50fe09c254b127d1c">stride_z</a> + (get_global_id(1) / width) * dst.<a class="code" href="struct_tensor3_d.xhtml#a4f0b90c9ecd6e57ceb3f37332fefe8f1">stride_y</a> + (get_global_id(1) % width) * dst.<a class="code" href="struct_tensor3_d.xhtml#ae01febbfd0689ef709f3ff6fdd2abc7e">stride_x</a>;</div><div class="line"><a name="l00196"></a><span class="lineno"> 196</span>&#160; __global uchar *tmp_out_ptr = dst.<a class="code" href="struct_tensor3_d.xhtml#acf52c23cbd7424606c10a606524e3e32">ptr</a> + idx;</div><div class="line"><a name="l00197"></a><span class="lineno"> 197</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)tmp_out_ptr) = *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)(src.<a class="code" href="struct_image.xhtml#acf52c23cbd7424606c10a606524e3e32">ptr</a>));</div><div class="line"><a name="l00198"></a><span class="lineno"> 198</span>&#160;}</div><div class="ttc" id="convolution3x3_8cl_xhtml_afb8c72ce35c4a1f4a2588d6573e54aa1"><div class="ttname"><a href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a></div><div class="ttdeci">#define DATA_TYPE</div><div class="ttdef"><b>Definition:</b> <a href="convolution3x3_8cl_source.xhtml#l00027">convolution3x3.cl:27</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml_ad5ff7a2b2bd0eec50fe09c254b127d1c"><div class="ttname"><a href="struct_tensor3_d.xhtml#ad5ff7a2b2bd0eec50fe09c254b127d1c">Tensor3D::stride_z</a></div><div class="ttdeci">int stride_z</div><div class="ttdoc">Stride of the image in Z dimension (in bytes) </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00110">helpers.h:110</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml_ae01febbfd0689ef709f3ff6fdd2abc7e"><div class="ttname"><a href="struct_tensor3_d.xhtml#ae01febbfd0689ef709f3ff6fdd2abc7e">Tensor3D::stride_x</a></div><div class="ttdeci">int stride_x</div><div class="ttdoc">Stride of the image in X dimension (in bytes) </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00108">helpers.h:108</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml"><div class="ttname"><a href="struct_tensor3_d.xhtml">Tensor3D</a></div><div class="ttdoc">Structure to hold 3D tensor information. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00104">helpers.h:104</a></div></div>
<div class="ttc" id="helpers_8h_xhtml_aebe814363556c244be043b13e7969197"><div class="ttname"><a href="helpers_8h.xhtml#aebe814363556c244be043b13e7969197">CONVERT_TO_IMAGE_STRUCT</a></div><div class="ttdeci">#define CONVERT_TO_IMAGE_STRUCT(name)</div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00073">helpers.h:73</a></div></div>
<div class="ttc" id="struct_image_xhtml"><div class="ttname"><a href="struct_image.xhtml">Image</a></div><div class="ttdoc">Structure to hold Image information. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00095">helpers.h:95</a></div></div>
<div class="ttc" id="struct_image_xhtml_acf52c23cbd7424606c10a606524e3e32"><div class="ttname"><a href="struct_image.xhtml#acf52c23cbd7424606c10a606524e3e32">Image::ptr</a></div><div class="ttdeci">__global uchar * ptr</div><div class="ttdoc">Pointer to the starting postion of the buffer. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00097">helpers.h:97</a></div></div>
<div class="ttc" id="helpers_8h_xhtml_a79e8e562daa6599317d2d1cd86ef1bf2"><div class="ttname"><a href="helpers_8h.xhtml#a79e8e562daa6599317d2d1cd86ef1bf2">CONVERT_TO_TENSOR3D_STRUCT_NO_STEP</a></div><div class="ttdeci">#define CONVERT_TO_TENSOR3D_STRUCT_NO_STEP(name)</div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00083">helpers.h:83</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml_acf52c23cbd7424606c10a606524e3e32"><div class="ttname"><a href="struct_tensor3_d.xhtml#acf52c23cbd7424606c10a606524e3e32">Tensor3D::ptr</a></div><div class="ttdeci">__global uchar * ptr</div><div class="ttdoc">Pointer to the starting postion of the buffer. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00106">helpers.h:106</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml_a4f0b90c9ecd6e57ceb3f37332fefe8f1"><div class="ttname"><a href="struct_tensor3_d.xhtml#a4f0b90c9ecd6e57ceb3f37332fefe8f1">Tensor3D::stride_y</a></div><div class="ttdeci">int stride_y</div><div class="ttdoc">Stride of the image in Y dimension (in bytes) </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00109">helpers.h:109</a></div></div>
</div><!-- fragment -->
</div>
</div>
<a class="anchor" id="afaddfccc926506181141b2dc86cb73f1"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">__kernel void im2col_generic </td>
<td>(</td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>src_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_z</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_z</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>dst_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_stride_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_step_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>kernel_size</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>kernel_depth</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int&#160;</td>
<td class="paramname"><em>width</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int2&#160;</td>
<td class="paramname"><em>input_dims</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int2&#160;</td>
<td class="paramname"><em>strides</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">int2&#160;</td>
<td class="paramname"><em>paddings</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>This kernel performs a reshaping of the input tensor to a tensor used to perform convolution using GEMM. </p>
<dl class="section note"><dt>Note</dt><dd>The data type must be passed at compile time using -DDATA_TYPE: e.g. -DDATA_TYPE=float </dd>
<dd>
In case biases will be added to the convolution -DHAS_BIAS has to be passed to append the final matrix with 1 in each row.</dd></dl>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramdir">[in]</td><td class="paramname">src_ptr</td><td>Pointer to the source tensor. Supported data types: F16, F32 </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_x</td><td>Stride of the source tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_x</td><td>src_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_y</td><td>Stride of the source tensor in Y dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_y</td><td>src_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_z</td><td>Stride of the source tensor in Z dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_z</td><td>src_stride_z * number of elements along Z processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_offset_first_element_in_bytes</td><td>The offset of the first element in the source tensor </td></tr>
<tr><td class="paramdir">[out]</td><td class="paramname">dst_ptr</td><td>Pointer to the destination tensor. Supported data types: F16, F32 </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_stride_x</td><td>Stride of the destination tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_step_x</td><td>dst_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_stride_y</td><td>Stride of the destination tensor in Y dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_step_y</td><td>dst_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_offset_first_element_in_bytes</td><td>The offset of the first element in the destination tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">kernel_size</td><td>The convolution kernel size </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">kernel_depth</td><td>The kernel depth </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">width</td><td>The output tensor width </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">input_dims</td><td>The input tensor dimensions </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">strides</td><td>The strides of the im2col operation </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">paddings</td><td>The input tensor paddings </td></tr>
</table>
</dd>
</dl>
<p>Definition at line <a class="el" href="convolution__layer_8cl_source.xhtml#l00122">122</a> of file <a class="el" href="convolution__layer_8cl_source.xhtml">convolution_layer.cl</a>.</p>
<p>References <a class="el" href="helpers_8h_source.xhtml#l00076">CONVERT_TO_IMAGE_STRUCT_NO_STEP</a>, <a class="el" href="helpers_8h_source.xhtml#l00079">CONVERT_TO_TENSOR3D_STRUCT</a>, <a class="el" href="convolution3x3_8cl_source.xhtml#l00027">DATA_TYPE</a>, <a class="el" href="helpers_8h_source.xhtml#l00097">Image::ptr</a>, <a class="el" href="helpers_8h_source.xhtml#l00099">Image::stride_x</a>, <a class="el" href="helpers_8h_source.xhtml#l00100">Image::stride_y</a>, and <a class="el" href="helpers_8h_source.xhtml#l00213">tensor3D_offset()</a>.</p>
<div class="fragment"><div class="line"><a name="l00131"></a><span class="lineno"> 131</span>&#160;{</div><div class="line"><a name="l00132"></a><span class="lineno"> 132</span>&#160; <a class="code" href="struct_tensor3_d.xhtml">Tensor3D</a> src = <a class="code" href="helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a>(src);</div><div class="line"><a name="l00133"></a><span class="lineno"> 133</span>&#160; <a class="code" href="struct_image.xhtml">Image</a> dst = <a class="code" href="helpers_8h.xhtml#a4334a4a76f8e9628c0fb9e1acf616e2a">CONVERT_TO_IMAGE_STRUCT_NO_STEP</a>(dst);</div><div class="line"><a name="l00134"></a><span class="lineno"> 134</span>&#160;</div><div class="line"><a name="l00135"></a><span class="lineno"> 135</span>&#160; <span class="comment">// Determine output index</span></div><div class="line"><a name="l00136"></a><span class="lineno"> 136</span>&#160; uint idx = (get_global_id(1) * width + get_global_id(0)) * dst.<a class="code" href="struct_image.xhtml#a4f0b90c9ecd6e57ceb3f37332fefe8f1">stride_y</a>;</div><div class="line"><a name="l00137"></a><span class="lineno"> 137</span>&#160; __global uchar *output_ptr = dst.<a class="code" href="struct_image.xhtml#acf52c23cbd7424606c10a606524e3e32">ptr</a> + idx;</div><div class="line"><a name="l00138"></a><span class="lineno"> 138</span>&#160;</div><div class="line"><a name="l00139"></a><span class="lineno"> 139</span>&#160; <span class="comment">// Determine current input index</span></div><div class="line"><a name="l00140"></a><span class="lineno"> 140</span>&#160; <span class="keyword">const</span> <span class="keywordtype">int</span> top_left_x = get_global_id(0) * strides.x - paddings.x;</div><div class="line"><a name="l00141"></a><span class="lineno"> 141</span>&#160; <span class="keyword">const</span> <span class="keywordtype">int</span> top_left_y = get_global_id(1) * strides.y - paddings.y;</div><div class="line"><a name="l00142"></a><span class="lineno"> 142</span>&#160;</div><div class="line"><a name="l00143"></a><span class="lineno"> 143</span>&#160; <span class="comment">// Linearize convolution elements</span></div><div class="line"><a name="l00144"></a><span class="lineno"> 144</span>&#160; <span class="keywordflow">for</span>(<span class="keywordtype">int</span> d = 0; d &lt; kernel_depth; ++d)</div><div class="line"><a name="l00145"></a><span class="lineno"> 145</span>&#160; {</div><div class="line"><a name="l00146"></a><span class="lineno"> 146</span>&#160; <span class="keywordflow">for</span>(<span class="keywordtype">int</span> y = top_left_y, y_e = top_left_y + kernel_size; y &lt; y_e; ++y)</div><div class="line"><a name="l00147"></a><span class="lineno"> 147</span>&#160; {</div><div class="line"><a name="l00148"></a><span class="lineno"> 148</span>&#160; <span class="keywordflow">for</span>(<span class="keywordtype">int</span> x = top_left_x, x_e = top_left_x + kernel_size; x &lt; x_e; ++x, output_ptr += dst.<a class="code" href="struct_image.xhtml#ae01febbfd0689ef709f3ff6fdd2abc7e">stride_x</a>)</div><div class="line"><a name="l00149"></a><span class="lineno"> 149</span>&#160; {</div><div class="line"><a name="l00150"></a><span class="lineno"> 150</span>&#160; <span class="keywordflow">if</span>(x &lt; 0 || x &gt;= input_dims.x || y &lt; 0 || y &gt;= input_dims.y)</div><div class="line"><a name="l00151"></a><span class="lineno"> 151</span>&#160; {</div><div class="line"><a name="l00152"></a><span class="lineno"> 152</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)output_ptr) = 0;</div><div class="line"><a name="l00153"></a><span class="lineno"> 153</span>&#160; }</div><div class="line"><a name="l00154"></a><span class="lineno"> 154</span>&#160; <span class="keywordflow">else</span></div><div class="line"><a name="l00155"></a><span class="lineno"> 155</span>&#160; {</div><div class="line"><a name="l00156"></a><span class="lineno"> 156</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)output_ptr) = *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)(<a class="code" href="helpers_8h.xhtml#a2101b2fe0193ce227ae4e0945e321d85">tensor3D_offset</a>(&amp;src, x, y, d)));</div><div class="line"><a name="l00157"></a><span class="lineno"> 157</span>&#160; }</div><div class="line"><a name="l00158"></a><span class="lineno"> 158</span>&#160; }</div><div class="line"><a name="l00159"></a><span class="lineno"> 159</span>&#160; }</div><div class="line"><a name="l00160"></a><span class="lineno"> 160</span>&#160; }</div><div class="line"><a name="l00161"></a><span class="lineno"> 161</span>&#160;</div><div class="line"><a name="l00162"></a><span class="lineno"> 162</span>&#160;<span class="preprocessor">#if defined HAS_BIAS</span></div><div class="line"><a name="l00163"></a><span class="lineno"> 163</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)output_ptr) = 1;</div><div class="line"><a name="l00164"></a><span class="lineno"> 164</span>&#160;<span class="preprocessor">#endif</span></div><div class="line"><a name="l00165"></a><span class="lineno"> 165</span>&#160;}</div><div class="ttc" id="helpers_8h_xhtml_a31c8c760f08fb1a331b16b7c204321dc"><div class="ttname"><a href="helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a></div><div class="ttdeci">#define CONVERT_TO_TENSOR3D_STRUCT(name)</div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00079">helpers.h:79</a></div></div>
<div class="ttc" id="convolution3x3_8cl_xhtml_afb8c72ce35c4a1f4a2588d6573e54aa1"><div class="ttname"><a href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a></div><div class="ttdeci">#define DATA_TYPE</div><div class="ttdef"><b>Definition:</b> <a href="convolution3x3_8cl_source.xhtml#l00027">convolution3x3.cl:27</a></div></div>
<div class="ttc" id="helpers_8h_xhtml_a4334a4a76f8e9628c0fb9e1acf616e2a"><div class="ttname"><a href="helpers_8h.xhtml#a4334a4a76f8e9628c0fb9e1acf616e2a">CONVERT_TO_IMAGE_STRUCT_NO_STEP</a></div><div class="ttdeci">#define CONVERT_TO_IMAGE_STRUCT_NO_STEP(name)</div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00076">helpers.h:76</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml"><div class="ttname"><a href="struct_tensor3_d.xhtml">Tensor3D</a></div><div class="ttdoc">Structure to hold 3D tensor information. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00104">helpers.h:104</a></div></div>
<div class="ttc" id="helpers_8h_xhtml_a2101b2fe0193ce227ae4e0945e321d85"><div class="ttname"><a href="helpers_8h.xhtml#a2101b2fe0193ce227ae4e0945e321d85">tensor3D_offset</a></div><div class="ttdeci">__global const uchar * tensor3D_offset(const Tensor3D *tensor, int x, int y, int z)</div><div class="ttdoc">Get the pointer position of a Tensor3D. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00213">helpers.h:213</a></div></div>
<div class="ttc" id="struct_image_xhtml"><div class="ttname"><a href="struct_image.xhtml">Image</a></div><div class="ttdoc">Structure to hold Image information. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00095">helpers.h:95</a></div></div>
<div class="ttc" id="struct_image_xhtml_acf52c23cbd7424606c10a606524e3e32"><div class="ttname"><a href="struct_image.xhtml#acf52c23cbd7424606c10a606524e3e32">Image::ptr</a></div><div class="ttdeci">__global uchar * ptr</div><div class="ttdoc">Pointer to the starting postion of the buffer. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00097">helpers.h:97</a></div></div>
<div class="ttc" id="struct_image_xhtml_a4f0b90c9ecd6e57ceb3f37332fefe8f1"><div class="ttname"><a href="struct_image.xhtml#a4f0b90c9ecd6e57ceb3f37332fefe8f1">Image::stride_y</a></div><div class="ttdeci">int stride_y</div><div class="ttdoc">Stride of the image in Y dimension (in bytes) </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00100">helpers.h:100</a></div></div>
<div class="ttc" id="struct_image_xhtml_ae01febbfd0689ef709f3ff6fdd2abc7e"><div class="ttname"><a href="struct_image.xhtml#ae01febbfd0689ef709f3ff6fdd2abc7e">Image::stride_x</a></div><div class="ttdeci">int stride_x</div><div class="ttdoc">Stride of the image in X dimension (in bytes) </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00099">helpers.h:99</a></div></div>
</div><!-- fragment -->
</div>
</div>
<a class="anchor" id="a6f31f2c8d606e1b57d9bdfacd416024c"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">__kernel void im2col_reduced </td>
<td>(</td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>src_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_z</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_z</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>dst_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>width</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>height</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>This kernel reshapes the tensor's low three dimensions to single row for GEMM operation. </p>
<dl class="section note"><dt>Note</dt><dd>Datatype should be given as a preprocessor argument using -DDATA_TYPE=type. e.g. -DDATA_TYPE=float </dd>
<dd>
In case biases will be added in late stage, -DHAS_BIAS has to be passed to append the final matrix with 1 in each row.</dd></dl>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramdir">[in]</td><td class="paramname">src_ptr</td><td>Pointer to the source tensor. Supported data types: F16, F32 </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_x</td><td>Stride of the source tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_x</td><td>src_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_y</td><td>Stride of the source tensor in Y dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_y</td><td>src_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_z</td><td>Stride of the source tensor in Z dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_z</td><td>src_stride_z * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_offset_first_element_in_bytes</td><td>The offset of the first element in the source tensor </td></tr>
<tr><td class="paramdir">[out]</td><td class="paramname">dst_ptr</td><td>Pointer to the destination tensor. Same as input. </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_stride_x</td><td>Stride of the destination tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_step_x</td><td>dst_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_offset_first_element_in_bytes</td><td>The offset of the first element in the destination tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">width</td><td>The width of the input tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">height</td><td>The height of the input tensor </td></tr>
</table>
</dd>
</dl>
<p>Definition at line <a class="el" href="convolution__layer_8cl_source.xhtml#l00220">220</a> of file <a class="el" href="convolution__layer_8cl_source.xhtml">convolution_layer.cl</a>.</p>
<p>References <a class="el" href="helpers_8h_source.xhtml#l00079">CONVERT_TO_TENSOR3D_STRUCT</a>, <a class="el" href="convolution3x3_8cl_source.xhtml#l00027">DATA_TYPE</a>, and <a class="el" href="helpers_8h_source.xhtml#l00106">Tensor3D::ptr</a>.</p>
<div class="fragment"><div class="line"><a name="l00224"></a><span class="lineno"> 224</span>&#160;{</div><div class="line"><a name="l00225"></a><span class="lineno"> 225</span>&#160; <a class="code" href="struct_tensor3_d.xhtml">Tensor3D</a> src = <a class="code" href="helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a>(src);</div><div class="line"><a name="l00226"></a><span class="lineno"> 226</span>&#160;</div><div class="line"><a name="l00227"></a><span class="lineno"> 227</span>&#160; <span class="keyword">const</span> uint image_size = width * height;</div><div class="line"><a name="l00228"></a><span class="lineno"> 228</span>&#160;</div><div class="line"><a name="l00229"></a><span class="lineno"> 229</span>&#160; __global uchar *tmp_out_ptr = dst_ptr + dst_offset_first_element_in_bytes + (get_global_id(0) + get_global_id(1) * width + get_global_id(2) * image_size) * dst_stride_x;</div><div class="line"><a name="l00230"></a><span class="lineno"> 230</span>&#160;</div><div class="line"><a name="l00231"></a><span class="lineno"> 231</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)tmp_out_ptr) = *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)src.ptr);</div><div class="line"><a name="l00232"></a><span class="lineno"> 232</span>&#160;</div><div class="line"><a name="l00233"></a><span class="lineno"> 233</span>&#160;<span class="preprocessor">#if defined HAS_BIAS</span></div><div class="line"><a name="l00234"></a><span class="lineno"> 234</span>&#160; <span class="comment">// If it is the last thread in the 3 dimensional workgroup</span></div><div class="line"><a name="l00235"></a><span class="lineno"> 235</span>&#160; <span class="keywordflow">if</span>(get_global_id(0) == (get_global_size(0) - 1) &amp;&amp; get_global_id(1) == (get_global_size(1) - 1) &amp;&amp; get_global_id(2) == (get_global_size(2) - 1))</div><div class="line"><a name="l00236"></a><span class="lineno"> 236</span>&#160; {</div><div class="line"><a name="l00237"></a><span class="lineno"> 237</span>&#160; tmp_out_ptr += dst_stride_x;</div><div class="line"><a name="l00238"></a><span class="lineno"> 238</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)tmp_out_ptr) = (<a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a>)1;</div><div class="line"><a name="l00239"></a><span class="lineno"> 239</span>&#160; }</div><div class="line"><a name="l00240"></a><span class="lineno"> 240</span>&#160;<span class="preprocessor">#endif</span></div><div class="line"><a name="l00241"></a><span class="lineno"> 241</span>&#160;}</div><div class="ttc" id="helpers_8h_xhtml_a31c8c760f08fb1a331b16b7c204321dc"><div class="ttname"><a href="helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a></div><div class="ttdeci">#define CONVERT_TO_TENSOR3D_STRUCT(name)</div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00079">helpers.h:79</a></div></div>
<div class="ttc" id="convolution3x3_8cl_xhtml_afb8c72ce35c4a1f4a2588d6573e54aa1"><div class="ttname"><a href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a></div><div class="ttdeci">#define DATA_TYPE</div><div class="ttdef"><b>Definition:</b> <a href="convolution3x3_8cl_source.xhtml#l00027">convolution3x3.cl:27</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml"><div class="ttname"><a href="struct_tensor3_d.xhtml">Tensor3D</a></div><div class="ttdoc">Structure to hold 3D tensor information. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00104">helpers.h:104</a></div></div>
</div><!-- fragment -->
</div>
</div>
<a class="anchor" id="ab6358b28d1e7bc946950075d94412013"></a>
<div class="memitem">
<div class="memproto">
<table class="memname">
<tr>
<td class="memname">__kernel void reshape_to_columns </td>
<td>(</td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>src_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_stride_z</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_step_z</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>src_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>dst_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_stride_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_step_y</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>dst_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">__global uchar *&#160;</td>
<td class="paramname"><em>bias_ptr</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>bias_stride_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>bias_step_x</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>bias_offset_first_element_in_bytes</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>width</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>height</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>depth</em>, </td>
</tr>
<tr>
<td class="paramkey"></td>
<td></td>
<td class="paramtype">uint&#160;</td>
<td class="paramname"><em>total_filters</em>&#160;</td>
</tr>
<tr>
<td></td>
<td>)</td>
<td></td><td></td>
</tr>
</table>
</div><div class="memdoc">
<p>This kernel reshapes the tensor's low three dimensions to single column. </p>
<dl class="section note"><dt>Note</dt><dd>Datatype should be given as a preprocessor argument using -DDATA_TYPE=type. e.g. -DDATA_TYPE=short</dd></dl>
<dl class="params"><dt>Parameters</dt><dd>
<table class="params">
<tr><td class="paramdir">[in]</td><td class="paramname">src_ptr</td><td>Pointer to the source tensor. Supported data types: F16, F32 </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_x</td><td>Stride of the source tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_x</td><td>src_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_y</td><td>Stride of the source tensor in Y dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_y</td><td>src_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_stride_z</td><td>Stride of the source tensor in Z dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_step_z</td><td>src_stride_z * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">src_offset_first_element_in_bytes</td><td>The offset of the first element in the source tensor </td></tr>
<tr><td class="paramdir">[out]</td><td class="paramname">dst_ptr</td><td>Pointer to the destination tensor. Same as input </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_stride_x</td><td>Stride of the destination tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_step_x</td><td>dst_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_stride_y</td><td>Stride of the destination tensor in Y dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_step_y</td><td>dst_stride_y * number of elements along Y processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">dst_offset_first_element_in_bytes</td><td>The offset of the first element in the destination tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">bias_ptr</td><td>Pointer to the bias tensor. Same as input </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">bias_stride_x</td><td>Stride of the bias tensor in X dimension (in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">bias_step_x</td><td>bias_stride_x * number of elements along X processed per workitem(in bytes) </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">bias_offset_first_element_in_bytes</td><td>The offset of the first element in the source tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">width</td><td>The width of the input tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">height</td><td>The height of the input tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">depth</td><td>The depth of the input tensor </td></tr>
<tr><td class="paramdir">[in]</td><td class="paramname">total_filters</td><td>Total number of filters. 4th dimension of the weights matrix </td></tr>
</table>
</dd>
</dl>
<p>Definition at line <a class="el" href="convolution__layer_8cl_source.xhtml#l00053">53</a> of file <a class="el" href="convolution__layer_8cl_source.xhtml">convolution_layer.cl</a>.</p>
<p>References <a class="el" href="helpers_8h_source.xhtml#l00079">CONVERT_TO_TENSOR3D_STRUCT</a>, <a class="el" href="convolution3x3_8cl_source.xhtml#l00027">DATA_TYPE</a>, and <a class="el" href="helpers_8h_source.xhtml#l00106">Tensor3D::ptr</a>.</p>
<div class="fragment"><div class="line"><a name="l00060"></a><span class="lineno"> 60</span>&#160;{</div><div class="line"><a name="l00061"></a><span class="lineno"> 61</span>&#160; <a class="code" href="struct_tensor3_d.xhtml">Tensor3D</a> src = <a class="code" href="helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a>(src);</div><div class="line"><a name="l00062"></a><span class="lineno"> 62</span>&#160; <span class="keywordtype">bool</span> is_last_thread = (get_global_id(0) == (get_global_size(0) - 1) &amp;&amp; get_global_id(1) == (get_global_size(1) - 1) &amp;&amp; get_global_id(2) == (get_global_size(2) - 1));</div><div class="line"><a name="l00063"></a><span class="lineno"> 63</span>&#160;</div><div class="line"><a name="l00064"></a><span class="lineno"> 64</span>&#160; __global uchar *tmp_src_ptr = src.<a class="code" href="struct_tensor3_d.xhtml#acf52c23cbd7424606c10a606524e3e32">ptr</a>;</div><div class="line"><a name="l00065"></a><span class="lineno"> 65</span>&#160; __global uchar *tmp_dst_ptr = dst_ptr + dst_offset_first_element_in_bytes + get_global_id(0) * dst_stride_y + get_global_id(1) * width * dst_stride_y + get_global_id(</div><div class="line"><a name="l00066"></a><span class="lineno"> 66</span>&#160; 2) * width * height * dst_stride_y;</div><div class="line"><a name="l00067"></a><span class="lineno"> 67</span>&#160;<span class="preprocessor">#if defined HAS_BIAS</span></div><div class="line"><a name="l00068"></a><span class="lineno"> 68</span>&#160; __global uchar *tmp_bias_ptr = bias_ptr + bias_offset_first_element_in_bytes;</div><div class="line"><a name="l00069"></a><span class="lineno"> 69</span>&#160;<span class="preprocessor">#endif</span></div><div class="line"><a name="l00070"></a><span class="lineno"> 70</span>&#160;</div><div class="line"><a name="l00071"></a><span class="lineno"> 71</span>&#160; <span class="keywordflow">if</span>(is_last_thread)</div><div class="line"><a name="l00072"></a><span class="lineno"> 72</span>&#160; {</div><div class="line"><a name="l00073"></a><span class="lineno"> 73</span>&#160; <span class="keywordflow">for</span>(uint i = 0; i &lt; total_filters; ++i)</div><div class="line"><a name="l00074"></a><span class="lineno"> 74</span>&#160; {</div><div class="line"><a name="l00075"></a><span class="lineno"> 75</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)tmp_dst_ptr) = *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)tmp_src_ptr);</div><div class="line"><a name="l00076"></a><span class="lineno"> 76</span>&#160;</div><div class="line"><a name="l00077"></a><span class="lineno"> 77</span>&#160;<span class="preprocessor">#if defined HAS_BIAS</span></div><div class="line"><a name="l00078"></a><span class="lineno"> 78</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)(tmp_dst_ptr + dst_stride_y)) = *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)(tmp_bias_ptr));</div><div class="line"><a name="l00079"></a><span class="lineno"> 79</span>&#160; tmp_bias_ptr += bias_stride_x;</div><div class="line"><a name="l00080"></a><span class="lineno"> 80</span>&#160;<span class="preprocessor">#endif</span></div><div class="line"><a name="l00081"></a><span class="lineno"> 81</span>&#160; tmp_src_ptr += depth * src_stride_z;</div><div class="line"><a name="l00082"></a><span class="lineno"> 82</span>&#160; tmp_dst_ptr += dst_stride_x;</div><div class="line"><a name="l00083"></a><span class="lineno"> 83</span>&#160; }</div><div class="line"><a name="l00084"></a><span class="lineno"> 84</span>&#160; }</div><div class="line"><a name="l00085"></a><span class="lineno"> 85</span>&#160; <span class="keywordflow">else</span></div><div class="line"><a name="l00086"></a><span class="lineno"> 86</span>&#160; {</div><div class="line"><a name="l00087"></a><span class="lineno"> 87</span>&#160; <span class="keywordflow">for</span>(uint i = 0; i &lt; total_filters; ++i)</div><div class="line"><a name="l00088"></a><span class="lineno"> 88</span>&#160; {</div><div class="line"><a name="l00089"></a><span class="lineno"> 89</span>&#160; *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)tmp_dst_ptr) = *((__global <a class="code" href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a> *)tmp_src_ptr);</div><div class="line"><a name="l00090"></a><span class="lineno"> 90</span>&#160; tmp_src_ptr += depth * src_stride_z;</div><div class="line"><a name="l00091"></a><span class="lineno"> 91</span>&#160; tmp_dst_ptr += dst_stride_x;</div><div class="line"><a name="l00092"></a><span class="lineno"> 92</span>&#160; }</div><div class="line"><a name="l00093"></a><span class="lineno"> 93</span>&#160; }</div><div class="line"><a name="l00094"></a><span class="lineno"> 94</span>&#160;}</div><div class="ttc" id="helpers_8h_xhtml_a31c8c760f08fb1a331b16b7c204321dc"><div class="ttname"><a href="helpers_8h.xhtml#a31c8c760f08fb1a331b16b7c204321dc">CONVERT_TO_TENSOR3D_STRUCT</a></div><div class="ttdeci">#define CONVERT_TO_TENSOR3D_STRUCT(name)</div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00079">helpers.h:79</a></div></div>
<div class="ttc" id="convolution3x3_8cl_xhtml_afb8c72ce35c4a1f4a2588d6573e54aa1"><div class="ttname"><a href="convolution3x3_8cl.xhtml#afb8c72ce35c4a1f4a2588d6573e54aa1">DATA_TYPE</a></div><div class="ttdeci">#define DATA_TYPE</div><div class="ttdef"><b>Definition:</b> <a href="convolution3x3_8cl_source.xhtml#l00027">convolution3x3.cl:27</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml"><div class="ttname"><a href="struct_tensor3_d.xhtml">Tensor3D</a></div><div class="ttdoc">Structure to hold 3D tensor information. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00104">helpers.h:104</a></div></div>
<div class="ttc" id="struct_tensor3_d_xhtml_acf52c23cbd7424606c10a606524e3e32"><div class="ttname"><a href="struct_tensor3_d.xhtml#acf52c23cbd7424606c10a606524e3e32">Tensor3D::ptr</a></div><div class="ttdeci">__global uchar * ptr</div><div class="ttdoc">Pointer to the starting postion of the buffer. </div><div class="ttdef"><b>Definition:</b> <a href="helpers_8h_source.xhtml#l00106">helpers.h:106</a></div></div>
</div><!-- fragment -->
</div>
</div>
</div><!-- contents -->
</div><!-- doc-content -->
<!-- start footer part -->
<div id="nav-path" class="navpath"><!-- id is needed for treeview function! -->
<ul>
<li class="navelem"><a class="el" href="dir_68267d1309a1af8e8297ef4c3efbcdba.xhtml">src</a></li><li class="navelem"><a class="el" href="dir_aebb8dcc11953d78e620bbef0b9e2183.xhtml">core</a></li><li class="navelem"><a class="el" href="dir_8c278f79c760e5c5fbd911f9870614c1.xhtml">CL</a></li><li class="navelem"><a class="el" href="dir_25885286e9dad4fa105b7b25a8031bbf.xhtml">cl_kernels</a></li><li class="navelem"><a class="el" href="convolution__layer_8cl.xhtml">convolution_layer.cl</a></li>
<li class="footer">Generated on Fri Jun 23 2017 15:44:34 for Compute Library by
<a href="http://www.doxygen.org/index.html">
<img class="footer" src="doxygen.png" alt="doxygen"/></a> 1.8.11 </li>
</ul>
</div>
</body>
</html>