mirror of
https://github.com/ml-explore/mlx.git
synced 2025-09-20 12:08:14 +08:00
rebase
This commit is contained in:
147
docs/build/html/loader__general_8h_source.html
vendored
147
docs/build/html/loader__general_8h_source.html
vendored
@@ -3,7 +3,7 @@
|
||||
<head>
|
||||
<meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/>
|
||||
<meta http-equiv="X-UA-Compatible" content="IE=11"/>
|
||||
<meta name="generator" content="Doxygen 1.12.0"/>
|
||||
<meta name="generator" content="Doxygen 1.13.1"/>
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1"/>
|
||||
<title>MLX: mlx/backend/metal/kernels/steel/conv/loaders/loader_general.h Source File</title>
|
||||
<link href="tabs.css" rel="stylesheet" type="text/css"/>
|
||||
@@ -11,11 +11,18 @@
|
||||
<script type="text/javascript" src="dynsections.js"></script>
|
||||
<script type="text/javascript" src="clipboard.js"></script>
|
||||
<link href="navtree.css" rel="stylesheet" type="text/css"/>
|
||||
<script type="text/javascript" src="navtreedata.js"></script>
|
||||
<script type="text/javascript" src="navtree.js"></script>
|
||||
<script type="text/javascript" src="resize.js"></script>
|
||||
<script type="text/javascript" src="cookie.js"></script>
|
||||
<link href="search/search.css" rel="stylesheet" type="text/css"/>
|
||||
<script type="text/javascript" src="search/searchdata.js"></script>
|
||||
<script type="text/javascript" src="search/search.js"></script>
|
||||
<script type="text/javascript">
|
||||
/* @license magnet:?xt=urn:btih:d3d9a9a6595521f9666a5e94cc830dab83b65699&dn=expat.txt MIT */
|
||||
$(function() { init_search(); });
|
||||
/* @license-end */
|
||||
</script>
|
||||
<link href="doxygen.css" rel="stylesheet" type="text/css" />
|
||||
</head>
|
||||
<body>
|
||||
@@ -28,12 +35,24 @@
|
||||
<div id="projectname">MLX
|
||||
</div>
|
||||
</td>
|
||||
<td> <div id="MSearchBox" class="MSearchBoxInactive">
|
||||
<span class="left">
|
||||
<span id="MSearchSelect" onmouseover="return searchBox.OnSearchSelectShow()" onmouseout="return searchBox.OnSearchSelectHide()"> </span>
|
||||
<input type="text" id="MSearchField" value="" placeholder="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.svg" alt=""/></a>
|
||||
</span>
|
||||
</div>
|
||||
</td>
|
||||
</tr>
|
||||
</tbody>
|
||||
</table>
|
||||
</div>
|
||||
<!-- end header part -->
|
||||
<!-- Generated by Doxygen 1.12.0 -->
|
||||
<!-- Generated by Doxygen 1.13.1 -->
|
||||
<script type="text/javascript">
|
||||
/* @license magnet:?xt=urn:btih:d3d9a9a6595521f9666a5e94cc830dab83b65699&dn=expat.txt MIT */
|
||||
var searchBox = new SearchBox("searchBox", "search/",'.html');
|
||||
@@ -44,22 +63,23 @@ var searchBox = new SearchBox("searchBox", "search/",'.html');
|
||||
$(function() { codefold.init(0); });
|
||||
/* @license-end */
|
||||
</script>
|
||||
<script type="text/javascript" src="menudata.js"></script>
|
||||
<script type="text/javascript" src="menu.js"></script>
|
||||
</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">
|
||||
/* @license magnet:?xt=urn:btih:d3d9a9a6595521f9666a5e94cc830dab83b65699&dn=expat.txt MIT */
|
||||
$(function() {
|
||||
initMenu('',true,false,'search.php','Search',false);
|
||||
$(function() { init_search(); });
|
||||
});
|
||||
/* @license-end */
|
||||
</script>
|
||||
<div id="main-nav"></div>
|
||||
<script type="text/javascript">
|
||||
/* @license magnet:?xt=urn:btih:d3d9a9a6595521f9666a5e94cc830dab83b65699&dn=expat.txt MIT */
|
||||
$(function(){ initResizable(false); });
|
||||
$(function(){initNavTree('loader__general_8h_source.html',''); initResizable(true); });
|
||||
/* @license-end */
|
||||
</script>
|
||||
<div id="doc-content">
|
||||
<!-- window showing the filter options -->
|
||||
<div id="MSearchSelectWindow"
|
||||
onmouseover="return searchBox.OnSearchSelectShow()"
|
||||
@@ -81,12 +101,6 @@ $(function(){ initResizable(false); });
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<div id="nav-path" class="navpath">
|
||||
<ul>
|
||||
<li class="navelem"><a class="el" href="dir_938ab0ecf10b8b860ff766c820f665fd.html">mlx</a></li><li class="navelem"><a class="el" href="dir_1d446c9bd3c99228254c9484e0bc5c06.html">backend</a></li><li class="navelem"><a class="el" href="dir_d0c977ea65824390717cdb7efc36c157.html">metal</a></li><li class="navelem"><a class="el" href="dir_70a37effa88bcbd6b791977fa1e64356.html">kernels</a></li><li class="navelem"><a class="el" href="dir_76215a6c54e2b67053e723fc2395583c.html">steel</a></li><li class="navelem"><a class="el" href="dir_df9494e83ef22ae6150a0e080d9709ed.html">conv</a></li><li class="navelem"><a class="el" href="dir_ba4426224ef60f409462a2a12fa18f06.html">loaders</a></li> </ul>
|
||||
</div>
|
||||
</div><!-- top -->
|
||||
<div id="doc-content">
|
||||
<div class="header">
|
||||
<div class="headertitle"><div class="title">loader_general.h</div></div>
|
||||
</div><!--header-->
|
||||
@@ -96,11 +110,11 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00003" name="l00003"></a><span class="lineno"> 3</span><span class="preprocessor">#pragma once</span></div>
|
||||
<div class="line"><a id="l00004" name="l00004"></a><span class="lineno"> 4</span> </div>
|
||||
<div class="line"><a id="l00005" name="l00005"></a><span class="lineno"> 5</span><span class="preprocessor">#include "<a class="code" href="steel_2defines_8h.html">mlx/backend/metal/kernels/steel/defines.h</a>"</span></div>
|
||||
<div class="line"><a id="l00006" name="l00006"></a><span class="lineno"> 6</span> </div>
|
||||
<div class="line"><a id="l00006" name="l00006"></a><span class="lineno"> 6</span></div>
|
||||
<div class="line"><a id="l00008" name="l00008"></a><span class="lineno"> 8</span><span class="comment">// Loading helper</span></div>
|
||||
<div class="line"><a id="l00010" name="l00010"></a><span class="lineno"> 10</span> </div>
|
||||
<div class="line"><a id="l00011" name="l00011"></a><span class="lineno"> 11</span><span class="keyword">namespace </span><a class="code hl_namespace" href="namespacemlx.html">mlx</a> {</div>
|
||||
<div class="line"><a id="l00012" name="l00012"></a><span class="lineno"> 12</span><span class="keyword">namespace </span>steel {</div>
|
||||
<div class="line"><a id="l00012" name="l00012"></a><span class="lineno"> 12</span><span class="keyword">namespace </span><a class="code hl_namespace" href="namespacemlx_1_1steel.html">steel</a> {</div>
|
||||
<div class="line"><a id="l00013" name="l00013"></a><span class="lineno"> 13</span> </div>
|
||||
<div class="line"><a id="l00014" name="l00014"></a><span class="lineno"> 14</span><span class="keyword">template</span> <</div>
|
||||
<div class="line"><a id="l00015" name="l00015"></a><span class="lineno"> 15</span> <span class="keyword">typename</span> T,</div>
|
||||
@@ -110,7 +124,7 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00019" name="l00019"></a><span class="lineno"> 19</span> <span class="keywordtype">short</span> tgp_size,</div>
|
||||
<div class="line"><a id="l00020" name="l00020"></a><span class="lineno"> 20</span> <span class="keywordtype">short</span> tgp_padding = 0></div>
|
||||
<div class="foldopen" id="foldopen00021" data-start="{" data-end="};">
|
||||
<div class="line"><a id="l00021" name="l00021"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html"> 21</a></span><span class="keyword">struct </span><a class="code hl_struct" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html">Conv2DInputBlockLoaderGeneral</a> {</div>
|
||||
<div class="line"><a id="l00021" name="l00021"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html"> 21</a></span><span class="keyword">struct </span><a class="code hl_function" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a1d83af561a483432bf8dcb42e734b23b">Conv2DInputBlockLoaderGeneral</a> {</div>
|
||||
<div class="line"><a id="l00022" name="l00022"></a><span class="lineno"> 22</span> <span class="comment">// Destination dimensions</span></div>
|
||||
<div class="line"><a id="l00023" name="l00023"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#aba1e1c8012e4e50f0e9bcfb9486c1781"> 23</a></span> <a class="code hl_define" href="steel_2defines_8h.html#a90b91c866313ffa46eff6d9cc944ad2b">STEEL_CONST</a> <span class="keywordtype">short</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#aba1e1c8012e4e50f0e9bcfb9486c1781">BROWS</a> = BM;</div>
|
||||
<div class="line"><a id="l00024" name="l00024"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a3d6272d000f8ea79d9b3b5228bdca20f"> 24</a></span> <a class="code hl_define" href="steel_2defines_8h.html#a90b91c866313ffa46eff6d9cc944ad2b">STEEL_CONST</a> <span class="keywordtype">short</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a3d6272d000f8ea79d9b3b5228bdca20f">BCOLS</a> = BK;</div>
|
||||
@@ -164,7 +178,7 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00071" name="l00071"></a><span class="lineno"> 71</span> : <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a401f0c7cf1588552556603c7ffba2316">thread_idx</a>(simd_group_id * 32 + simd_lane_id),</div>
|
||||
<div class="line"><a id="l00072" name="l00072"></a><span class="lineno"> 72</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a32a3a91fa715b82f36e05ceb10933d09">bi</a>(<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a401f0c7cf1588552556603c7ffba2316">thread_idx</a> / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ae25c676b7318d78462ee89bcd80dc805">TCOLS</a>),</div>
|
||||
<div class="line"><a id="l00073" name="l00073"></a><span class="lineno"> 73</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ace16704025bc6e6204c306a357f3a8b8">bj</a>(<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a1587047caa339cf5b2c06adc4b332ab8">vec_size</a> * (<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a401f0c7cf1588552556603c7ffba2316">thread_idx</a> % <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ae25c676b7318d78462ee89bcd80dc805">TCOLS</a>)),</div>
|
||||
<div class="line"><a id="l00074" name="l00074"></a><span class="lineno"> 74</span> dst(dst_ + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a32a3a91fa715b82f36e05ceb10933d09">bi</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a07c85eab8cbf7b02c60df29cf32031ef">dst_ld</a> + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ace16704025bc6e6204c306a357f3a8b8">bj</a>),</div>
|
||||
<div class="line"><a id="l00074" name="l00074"></a><span class="lineno"> 74</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#aa84c4ad43a5defb83ba1a5f49a7adb2a">dst</a>(dst_ + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a32a3a91fa715b82f36e05ceb10933d09">bi</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a07c85eab8cbf7b02c60df29cf32031ef">dst_ld</a> + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ace16704025bc6e6204c306a357f3a8b8">bj</a>),</div>
|
||||
<div class="line"><a id="l00075" name="l00075"></a><span class="lineno"> 75</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>(params_),</div>
|
||||
<div class="line"><a id="l00076" name="l00076"></a><span class="lineno"> 76</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>(jump_params_),</div>
|
||||
<div class="line"><a id="l00077" name="l00077"></a><span class="lineno"> 77</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#aca37adba6f148579eb1cd0a7800a5cfe">base_wh</a>(base_wh_),</div>
|
||||
@@ -174,22 +188,22 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00081" name="l00081"></a><span class="lineno"> 81</span> <a class="code hl_define" href="steel_2defines_8h.html#a5a5c3095b132a7589bc19cd5cb80e2c6">STEEL_PRAGMA_UNROLL</a></div>
|
||||
<div class="line"><a id="l00082" name="l00082"></a><span class="lineno"> 82</span> <span class="keywordflow">for</span> (<span class="keywordtype">short</span> i = 0; i < <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#abff29c5d96645d9113314c9a997dd7a8">n_rows</a>; ++i) {</div>
|
||||
<div class="line"><a id="l00083" name="l00083"></a><span class="lineno"> 83</span> <span class="keywordtype">int</span> offset_nhw = offsets.y + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a32a3a91fa715b82f36e05ceb10933d09">bi</a> + i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a2aff22af70f685f858adea73f5575cf7">TROWS</a>;</div>
|
||||
<div class="line"><a id="l00084" name="l00084"></a><span class="lineno"> 84</span> <span class="keywordtype">int</span> n = offset_nhw / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#aed0ffd63fbc85fd5d5c4cc7b43f68363">adj_out_hw</a>;</div>
|
||||
<div class="line"><a id="l00085" name="l00085"></a><span class="lineno"> 85</span> <span class="keywordtype">int</span> hw = offset_nhw % <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#aed0ffd63fbc85fd5d5c4cc7b43f68363">adj_out_hw</a>;</div>
|
||||
<div class="line"><a id="l00084" name="l00084"></a><span class="lineno"> 84</span> <span class="keywordtype">int</span> n = offset_nhw / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>->adj_out_hw;</div>
|
||||
<div class="line"><a id="l00085" name="l00085"></a><span class="lineno"> 85</span> <span class="keywordtype">int</span> hw = offset_nhw % <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>->adj_out_hw;</div>
|
||||
<div class="line"><a id="l00086" name="l00086"></a><span class="lineno"> 86</span> <span class="keywordtype">int</span> oh =</div>
|
||||
<div class="line"><a id="l00087" name="l00087"></a><span class="lineno"> 87</span> (hw / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#ab971bf879079895189331fbeaf33c211">adj_out_w</a>) * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a78d48b55cf182f000abece0e5e7fadcb">f_out_jump_h</a> + offsets.z;</div>
|
||||
<div class="line"><a id="l00087" name="l00087"></a><span class="lineno"> 87</span> (hw / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>->adj_out_w) * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>->f_out_jump_h + offsets.z;</div>
|
||||
<div class="line"><a id="l00088" name="l00088"></a><span class="lineno"> 88</span> <span class="keywordtype">int</span> ow =</div>
|
||||
<div class="line"><a id="l00089" name="l00089"></a><span class="lineno"> 89</span> (hw % <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#ab971bf879079895189331fbeaf33c211">adj_out_w</a>) * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a0fd755691482cb03ea4534b4a556c197">f_out_jump_w</a> + offsets.w;</div>
|
||||
<div class="line"><a id="l00089" name="l00089"></a><span class="lineno"> 89</span> (hw % <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>->adj_out_w) * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>->f_out_jump_w + offsets.w;</div>
|
||||
<div class="line"><a id="l00090" name="l00090"></a><span class="lineno"> 90</span> </div>
|
||||
<div class="line"><a id="l00091" name="l00091"></a><span class="lineno"> 91</span> <span class="keywordtype">int</span> ih = oh * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#a862191e8ab1bc8a47aa1396b36d46058">str</a>[0] - <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#ae84a9afb3a95b57e0b763bb0ebda0753">pad</a>[0];</div>
|
||||
<div class="line"><a id="l00092" name="l00092"></a><span class="lineno"> 92</span> <span class="keywordtype">int</span> iw = ow * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#a862191e8ab1bc8a47aa1396b36d46058">str</a>[1] - <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#ae84a9afb3a95b57e0b763bb0ebda0753">pad</a>[1];</div>
|
||||
<div class="line"><a id="l00091" name="l00091"></a><span class="lineno"> 91</span> <span class="keywordtype">int</span> ih = oh * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->str[0] - <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->pad[0];</div>
|
||||
<div class="line"><a id="l00092" name="l00092"></a><span class="lineno"> 92</span> <span class="keywordtype">int</span> iw = ow * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->str[1] - <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->pad[1];</div>
|
||||
<div class="line"><a id="l00093" name="l00093"></a><span class="lineno"> 93</span> </div>
|
||||
<div class="line"><a id="l00094" name="l00094"></a><span class="lineno"> 94</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#afe5caaf38b574d3380533856c493dd92">read_n</a>[i] = n;</div>
|
||||
<div class="line"><a id="l00095" name="l00095"></a><span class="lineno"> 95</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a35a010c3819df6667339d37a5e8f5b43">read_ih</a>[i] = ih;</div>
|
||||
<div class="line"><a id="l00096" name="l00096"></a><span class="lineno"> 96</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a53a683adf280e4806363020754525261">read_iw</a>[i] = iw;</div>
|
||||
<div class="line"><a id="l00097" name="l00097"></a><span class="lineno"> 97</span> </div>
|
||||
<div class="line"><a id="l00098" name="l00098"></a><span class="lineno"> 98</span> <span class="comment">// Read from input if in bounds</span></div>
|
||||
<div class="line"><a id="l00099" name="l00099"></a><span class="lineno"> 99</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a1ee2922961b5fcb1db577928c4d9d731">src</a>[i] = src_ + n * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#ab25eade6573784985dbea1216f9068cf">in_strides</a>[0] + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ace16704025bc6e6204c306a357f3a8b8">bj</a>;</div>
|
||||
<div class="line"><a id="l00099" name="l00099"></a><span class="lineno"> 99</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a1ee2922961b5fcb1db577928c4d9d731">src</a>[i] = src_ + n * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->in_strides[0] + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ace16704025bc6e6204c306a357f3a8b8">bj</a>;</div>
|
||||
<div class="line"><a id="l00100" name="l00100"></a><span class="lineno"> 100</span> }</div>
|
||||
<div class="line"><a id="l00101" name="l00101"></a><span class="lineno"> 101</span> }</div>
|
||||
</div>
|
||||
@@ -202,23 +216,23 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00107" name="l00107"></a><span class="lineno"> 107</span> <span class="comment">// Find bounds</span></div>
|
||||
<div class="line"><a id="l00108" name="l00108"></a><span class="lineno"> 108</span> <span class="keywordtype">int</span> n = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#afe5caaf38b574d3380533856c493dd92">read_n</a>[i];</div>
|
||||
<div class="line"><a id="l00109" name="l00109"></a><span class="lineno"> 109</span> </div>
|
||||
<div class="line"><a id="l00110" name="l00110"></a><span class="lineno"> 110</span> <span class="keywordtype">int</span> h_flip = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#a8b30cda15eda20f84f12db868f21d0ef">flip</a> ? <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#aba2074189644b1b59567d018409277a9">wS</a>[0] - <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a397412909eb955babc935a35d97c3fd4">weight_h</a> - 1 : <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a397412909eb955babc935a35d97c3fd4">weight_h</a>;</div>
|
||||
<div class="line"><a id="l00111" name="l00111"></a><span class="lineno"> 111</span> <span class="keywordtype">int</span> w_flip = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#a8b30cda15eda20f84f12db868f21d0ef">flip</a> ? <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#aba2074189644b1b59567d018409277a9">wS</a>[1] - <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">weight_w</a> - 1 : <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">weight_w</a>;</div>
|
||||
<div class="line"><a id="l00110" name="l00110"></a><span class="lineno"> 110</span> <span class="keywordtype">int</span> h_flip = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->flip ? <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->wS[0] - <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a397412909eb955babc935a35d97c3fd4">weight_h</a> - 1 : <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a397412909eb955babc935a35d97c3fd4">weight_h</a>;</div>
|
||||
<div class="line"><a id="l00111" name="l00111"></a><span class="lineno"> 111</span> <span class="keywordtype">int</span> w_flip = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->flip ? <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->wS[1] - <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">weight_w</a> - 1 : <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">weight_w</a>;</div>
|
||||
<div class="line"><a id="l00112" name="l00112"></a><span class="lineno"> 112</span> </div>
|
||||
<div class="line"><a id="l00113" name="l00113"></a><span class="lineno"> 113</span> <span class="keywordtype">int</span> ih_dil = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a35a010c3819df6667339d37a5e8f5b43">read_ih</a>[i] + h_flip * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#a7611db8f1621c7e09fc685ed44073b14">kdil</a>[0];</div>
|
||||
<div class="line"><a id="l00114" name="l00114"></a><span class="lineno"> 114</span> <span class="keywordtype">int</span> iw_dil = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a53a683adf280e4806363020754525261">read_iw</a>[i] + w_flip * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#a7611db8f1621c7e09fc685ed44073b14">kdil</a>[1];</div>
|
||||
<div class="line"><a id="l00113" name="l00113"></a><span class="lineno"> 113</span> <span class="keywordtype">int</span> ih_dil = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a35a010c3819df6667339d37a5e8f5b43">read_ih</a>[i] + h_flip * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->kdil[0];</div>
|
||||
<div class="line"><a id="l00114" name="l00114"></a><span class="lineno"> 114</span> <span class="keywordtype">int</span> iw_dil = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a53a683adf280e4806363020754525261">read_iw</a>[i] + w_flip * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->kdil[1];</div>
|
||||
<div class="line"><a id="l00115" name="l00115"></a><span class="lineno"> 115</span> </div>
|
||||
<div class="line"><a id="l00116" name="l00116"></a><span class="lineno"> 116</span> <span class="keywordtype">int</span> ih = ih_dil / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#af900fdb4c4d4ea35eed02940dee8d4d1">idil</a>[0];</div>
|
||||
<div class="line"><a id="l00117" name="l00117"></a><span class="lineno"> 117</span> <span class="keywordtype">int</span> iw = iw_dil / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#af900fdb4c4d4ea35eed02940dee8d4d1">idil</a>[1];</div>
|
||||
<div class="line"><a id="l00116" name="l00116"></a><span class="lineno"> 116</span> <span class="keywordtype">int</span> ih = ih_dil / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->idil[0];</div>
|
||||
<div class="line"><a id="l00117" name="l00117"></a><span class="lineno"> 117</span> <span class="keywordtype">int</span> iw = iw_dil / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->idil[1];</div>
|
||||
<div class="line"><a id="l00118" name="l00118"></a><span class="lineno"> 118</span> </div>
|
||||
<div class="line"><a id="l00119" name="l00119"></a><span class="lineno"> 119</span> <span class="keywordtype">size_t</span> offset = ih * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#ab25eade6573784985dbea1216f9068cf">in_strides</a>[1] + iw * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#ab25eade6573784985dbea1216f9068cf">in_strides</a>[2];</div>
|
||||
<div class="line"><a id="l00119" name="l00119"></a><span class="lineno"> 119</span> <span class="keywordtype">size_t</span> offset = ih * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->in_strides[1] + iw * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">params</a>->in_strides[2];</div>
|
||||
<div class="line"><a id="l00120" name="l00120"></a><span class="lineno"> 120</span> </div>
|
||||
<div class="line"><a id="l00121" name="l00121"></a><span class="lineno"> 121</span> <span class="comment">// Read from input if in bounds</span></div>
|
||||
<div class="line"><a id="l00122" name="l00122"></a><span class="lineno"> 122</span> <span class="keywordflow">if</span> ((n < params->N) && (ih_dil >= 0 && ih < params->iS[0]) &&</div>
|
||||
<div class="line"><a id="l00123" name="l00123"></a><span class="lineno"> 123</span> (iw_dil >= 0 && iw < params->iS[1])) {</div>
|
||||
<div class="line"><a id="l00124" name="l00124"></a><span class="lineno"> 124</span> <a class="code hl_define" href="steel_2defines_8h.html#a5a5c3095b132a7589bc19cd5cb80e2c6">STEEL_PRAGMA_UNROLL</a></div>
|
||||
<div class="line"><a id="l00125" name="l00125"></a><span class="lineno"> 125</span> <span class="keywordflow">for</span> (<span class="keywordtype">short</span> j = 0; j < <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a1587047caa339cf5b2c06adc4b332ab8">vec_size</a>; ++j) {</div>
|
||||
<div class="line"><a id="l00126" name="l00126"></a><span class="lineno"> 126</span> dst[is * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a07c85eab8cbf7b02c60df29cf32031ef">dst_ld</a> + j] = (<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a1ee2922961b5fcb1db577928c4d9d731">src</a>[i])[offset + j];</div>
|
||||
<div class="line"><a id="l00126" name="l00126"></a><span class="lineno"> 126</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#aa84c4ad43a5defb83ba1a5f49a7adb2a">dst</a>[is * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a07c85eab8cbf7b02c60df29cf32031ef">dst_ld</a> + j] = (<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a1ee2922961b5fcb1db577928c4d9d731">src</a>[i])[offset + j];</div>
|
||||
<div class="line"><a id="l00127" name="l00127"></a><span class="lineno"> 127</span> }</div>
|
||||
<div class="line"><a id="l00128" name="l00128"></a><span class="lineno"> 128</span> }</div>
|
||||
<div class="line"><a id="l00129" name="l00129"></a><span class="lineno"> 129</span> </div>
|
||||
@@ -226,7 +240,7 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00131" name="l00131"></a><span class="lineno"> 131</span> <span class="keywordflow">else</span> {</div>
|
||||
<div class="line"><a id="l00132" name="l00132"></a><span class="lineno"> 132</span> <a class="code hl_define" href="steel_2defines_8h.html#a5a5c3095b132a7589bc19cd5cb80e2c6">STEEL_PRAGMA_UNROLL</a></div>
|
||||
<div class="line"><a id="l00133" name="l00133"></a><span class="lineno"> 133</span> <span class="keywordflow">for</span> (<span class="keywordtype">short</span> j = 0; j < <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a1587047caa339cf5b2c06adc4b332ab8">vec_size</a>; ++j) {</div>
|
||||
<div class="line"><a id="l00134" name="l00134"></a><span class="lineno"> 134</span> dst[is * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a07c85eab8cbf7b02c60df29cf32031ef">dst_ld</a> + j] = T(0);</div>
|
||||
<div class="line"><a id="l00134" name="l00134"></a><span class="lineno"> 134</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#aa84c4ad43a5defb83ba1a5f49a7adb2a">dst</a>[is * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a07c85eab8cbf7b02c60df29cf32031ef">dst_ld</a> + j] = T(0);</div>
|
||||
<div class="line"><a id="l00135" name="l00135"></a><span class="lineno"> 135</span> }</div>
|
||||
<div class="line"><a id="l00136" name="l00136"></a><span class="lineno"> 136</span> }</div>
|
||||
<div class="line"><a id="l00137" name="l00137"></a><span class="lineno"> 137</span> }</div>
|
||||
@@ -236,14 +250,14 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00140" name="l00140"></a><span class="lineno"> 140</span> <span class="comment">/* Iteration helper */</span></div>
|
||||
<div class="foldopen" id="foldopen00141" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00141" name="l00141"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a3e5ee68ed0ee43f7e979dd4222f76a8c"> 141</a></span> METAL_FUNC <span class="keywordtype">void</span> <a class="code hl_function" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a3e5ee68ed0ee43f7e979dd4222f76a8c">next</a>() {</div>
|
||||
<div class="line"><a id="l00142" name="l00142"></a><span class="lineno"> 142</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">weight_w</a> += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a568435a612574ab19a051a48055d4cfc">f_wgt_jump_w</a>;</div>
|
||||
<div class="line"><a id="l00142" name="l00142"></a><span class="lineno"> 142</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">weight_w</a> += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>->f_wgt_jump_w;</div>
|
||||
<div class="line"><a id="l00143" name="l00143"></a><span class="lineno"> 143</span> <span class="keywordflow">if</span> (<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">weight_w < params-></a>wS[1]) {</div>
|
||||
<div class="line"><a id="l00144" name="l00144"></a><span class="lineno"> 144</span> <span class="keywordflow">return</span>;</div>
|
||||
<div class="line"><a id="l00145" name="l00145"></a><span class="lineno"> 145</span> }</div>
|
||||
<div class="line"><a id="l00146" name="l00146"></a><span class="lineno"> 146</span> </div>
|
||||
<div class="line"><a id="l00147" name="l00147"></a><span class="lineno"> 147</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">weight_w</a> = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a32d020c6715d06f7de360877fcb7b6e4">base_ww</a>;</div>
|
||||
<div class="line"><a id="l00148" name="l00148"></a><span class="lineno"> 148</span> </div>
|
||||
<div class="line"><a id="l00149" name="l00149"></a><span class="lineno"> 149</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a397412909eb955babc935a35d97c3fd4">weight_h</a> += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a198ba0c2740ab4ded99345edf58917a7">f_wgt_jump_h</a>;</div>
|
||||
<div class="line"><a id="l00149" name="l00149"></a><span class="lineno"> 149</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a397412909eb955babc935a35d97c3fd4">weight_h</a> += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a21b9ee9168dad4af84a611f861519e77">jump_params</a>->f_wgt_jump_h;</div>
|
||||
<div class="line"><a id="l00150" name="l00150"></a><span class="lineno"> 150</span> <span class="keywordflow">if</span> (<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a397412909eb955babc935a35d97c3fd4">weight_h < params-></a>wS[0]) {</div>
|
||||
<div class="line"><a id="l00151" name="l00151"></a><span class="lineno"> 151</span> <span class="keywordflow">return</span>;</div>
|
||||
<div class="line"><a id="l00152" name="l00152"></a><span class="lineno"> 152</span> }</div>
|
||||
@@ -267,7 +281,7 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00168" name="l00168"></a><span class="lineno"> 168</span> <span class="keywordtype">short</span> tgp_size,</div>
|
||||
<div class="line"><a id="l00169" name="l00169"></a><span class="lineno"> 169</span> <span class="keywordtype">short</span> tgp_padding = 0></div>
|
||||
<div class="foldopen" id="foldopen00170" data-start="{" data-end="};">
|
||||
<div class="line"><a id="l00170" name="l00170"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html"> 170</a></span><span class="keyword">struct </span><a class="code hl_struct" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html">Conv2DWeightBlockLoaderGeneral</a> {</div>
|
||||
<div class="line"><a id="l00170" name="l00170"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html"> 170</a></span><span class="keyword">struct </span><a class="code hl_function" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ad0550fabbdc9297559381a5b488e9af1">Conv2DWeightBlockLoaderGeneral</a> {</div>
|
||||
<div class="line"><a id="l00171" name="l00171"></a><span class="lineno"> 171</span> <span class="comment">// Destination dimensions</span></div>
|
||||
<div class="line"><a id="l00172" name="l00172"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a015a0c56de74a0c4d51953a7e94fbba8"> 172</a></span> <a class="code hl_define" href="steel_2defines_8h.html#a90b91c866313ffa46eff6d9cc944ad2b">STEEL_CONST</a> <span class="keywordtype">short</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a015a0c56de74a0c4d51953a7e94fbba8">BROWS</a> = BN;</div>
|
||||
<div class="line"><a id="l00173" name="l00173"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a1843921cd67926002bb0dcccf3048eb6"> 173</a></span> <a class="code hl_define" href="steel_2defines_8h.html#a90b91c866313ffa46eff6d9cc944ad2b">STEEL_CONST</a> <span class="keywordtype">short</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a1843921cd67926002bb0dcccf3048eb6">BCOLS</a> = BK;</div>
|
||||
@@ -323,7 +337,7 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00222" name="l00222"></a><span class="lineno"> 222</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08a517bc50caf41155b98be0690bfe44">thread_idx</a>(simd_group_id * 32 + simd_lane_id),</div>
|
||||
<div class="line"><a id="l00223" name="l00223"></a><span class="lineno"> 223</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a4c91f848856ab0872bdfd37c62d4b0ba">bi</a>(<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08a517bc50caf41155b98be0690bfe44">thread_idx</a> / <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aff021a6fae860b4ac01fb593b2720457">TCOLS</a>),</div>
|
||||
<div class="line"><a id="l00224" name="l00224"></a><span class="lineno"> 224</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#acec010e10d5733654963407af38d4f67">bj</a>(<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08dba753ec7c8ea2892775746933b3e7">vec_size</a> * (<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08a517bc50caf41155b98be0690bfe44">thread_idx</a> % <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aff021a6fae860b4ac01fb593b2720457">TCOLS</a>)),</div>
|
||||
<div class="line"><a id="l00225" name="l00225"></a><span class="lineno"> 225</span> dst(dst_ + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a4c91f848856ab0872bdfd37c62d4b0ba">bi</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aae121ca6016fc6c7255027b3641f3a09">dst_ld</a> + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#acec010e10d5733654963407af38d4f67">bj</a>),</div>
|
||||
<div class="line"><a id="l00225" name="l00225"></a><span class="lineno"> 225</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a8474daf268013e138a84fc1c4bff7352">dst</a>(dst_ + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a4c91f848856ab0872bdfd37c62d4b0ba">bi</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aae121ca6016fc6c7255027b3641f3a09">dst_ld</a> + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#acec010e10d5733654963407af38d4f67">bj</a>),</div>
|
||||
<div class="line"><a id="l00226" name="l00226"></a><span class="lineno"> 226</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a50f458dbb74d61be2ed24727d8d43614">src</a>(src_ + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a4c91f848856ab0872bdfd37c62d4b0ba">bi</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6918c1df7712c4e408e2871467ea7987">src_ld</a> + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#acec010e10d5733654963407af38d4f67">bj</a>),</div>
|
||||
<div class="line"><a id="l00227" name="l00227"></a><span class="lineno"> 227</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ae4759d18c0e5cc3530b3da8493008419">params</a>(params_),</div>
|
||||
<div class="line"><a id="l00228" name="l00228"></a><span class="lineno"> 228</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aa5611e9a84bebaee966d2b339c214ff5">jump_params</a>(jump_params_),</div>
|
||||
@@ -337,28 +351,28 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00235" name="l00235"></a><span class="lineno"> 235</span> <span class="comment">/* Load from device memory into threadgroup memory - without bound checking */</span></div>
|
||||
<div class="foldopen" id="foldopen00236" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00236" name="l00236"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a8f078982186421f5b484c0b53af9c655"> 236</a></span> METAL_FUNC <span class="keywordtype">void</span> <a class="code hl_function" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a8f078982186421f5b484c0b53af9c655">load_unsafe</a>()<span class="keyword"> const </span>{</div>
|
||||
<div class="line"><a id="l00237" name="l00237"></a><span class="lineno"> 237</span> <span class="keyword">const</span> device T* curr_src = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a50f458dbb74d61be2ed24727d8d43614">src</a> + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a5997fd8ef249e4cd3df7dad7b251d8d5">weight_h</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ae4759d18c0e5cc3530b3da8493008419">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#a887fee0da1494d038526fb0f59faff45">wt_strides</a>[1] +</div>
|
||||
<div class="line"><a id="l00238" name="l00238"></a><span class="lineno"> 238</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6efa6268a37f18f4d225674bf1780cf6">weight_w</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ae4759d18c0e5cc3530b3da8493008419">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#a887fee0da1494d038526fb0f59faff45">wt_strides</a>[2];</div>
|
||||
<div class="line"><a id="l00237" name="l00237"></a><span class="lineno"> 237</span> <span class="keyword">const</span> device T* curr_src = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a50f458dbb74d61be2ed24727d8d43614">src</a> + <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a5997fd8ef249e4cd3df7dad7b251d8d5">weight_h</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ae4759d18c0e5cc3530b3da8493008419">params</a>->wt_strides[1] +</div>
|
||||
<div class="line"><a id="l00238" name="l00238"></a><span class="lineno"> 238</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6efa6268a37f18f4d225674bf1780cf6">weight_w</a> * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ae4759d18c0e5cc3530b3da8493008419">params</a>->wt_strides[2];</div>
|
||||
<div class="line"><a id="l00239" name="l00239"></a><span class="lineno"> 239</span> </div>
|
||||
<div class="line"><a id="l00240" name="l00240"></a><span class="lineno"> 240</span> <span class="keywordflow">if</span> ((<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a220e033b689c8d6a6f319dae02b38334">start_row</a> + BN <= params->O)) {</div>
|
||||
<div class="line"><a id="l00241" name="l00241"></a><span class="lineno"> 241</span> <a class="code hl_define" href="steel_2defines_8h.html#a5a5c3095b132a7589bc19cd5cb80e2c6">STEEL_PRAGMA_UNROLL</a></div>
|
||||
<div class="line"><a id="l00242" name="l00242"></a><span class="lineno"> 242</span> <span class="keywordflow">for</span> (<span class="keywordtype">short</span> i = 0; i < BN; i += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#acbc28f364381166faaeec2783dc88e10">TROWS</a>) {</div>
|
||||
<div class="line"><a id="l00243" name="l00243"></a><span class="lineno"> 243</span> <a class="code hl_define" href="steel_2defines_8h.html#a5a5c3095b132a7589bc19cd5cb80e2c6">STEEL_PRAGMA_UNROLL</a></div>
|
||||
<div class="line"><a id="l00244" name="l00244"></a><span class="lineno"> 244</span> <span class="keywordflow">for</span> (<span class="keywordtype">short</span> j = 0; j < <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08dba753ec7c8ea2892775746933b3e7">vec_size</a>; j++) {</div>
|
||||
<div class="line"><a id="l00245" name="l00245"></a><span class="lineno"> 245</span> dst[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aae121ca6016fc6c7255027b3641f3a09">dst_ld</a> + j] = curr_src[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6918c1df7712c4e408e2871467ea7987">src_ld</a> + j];</div>
|
||||
<div class="line"><a id="l00245" name="l00245"></a><span class="lineno"> 245</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a8474daf268013e138a84fc1c4bff7352">dst</a>[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aae121ca6016fc6c7255027b3641f3a09">dst_ld</a> + j] = curr_src[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6918c1df7712c4e408e2871467ea7987">src_ld</a> + j];</div>
|
||||
<div class="line"><a id="l00246" name="l00246"></a><span class="lineno"> 246</span> }</div>
|
||||
<div class="line"><a id="l00247" name="l00247"></a><span class="lineno"> 247</span> }</div>
|
||||
<div class="line"><a id="l00248" name="l00248"></a><span class="lineno"> 248</span> } <span class="keywordflow">else</span> {</div>
|
||||
<div class="line"><a id="l00249" name="l00249"></a><span class="lineno"> 249</span> <span class="keywordflow">for</span> (<span class="keywordtype">short</span> i = 0; i < BN; i += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#acbc28f364381166faaeec2783dc88e10">TROWS</a>) {</div>
|
||||
<div class="line"><a id="l00250" name="l00250"></a><span class="lineno"> 250</span> <span class="keywordflow">if</span> ((<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a220e033b689c8d6a6f319dae02b38334">start_row</a> + i) < <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ae4759d18c0e5cc3530b3da8493008419">params</a>-><a class="code hl_variable" href="struct_m_l_x_conv_params.html#ad55ff586d30072d8154865f9dfe92d97">O</a>) {</div>
|
||||
<div class="line"><a id="l00250" name="l00250"></a><span class="lineno"> 250</span> <span class="keywordflow">if</span> ((<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a220e033b689c8d6a6f319dae02b38334">start_row</a> + i) < <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ae4759d18c0e5cc3530b3da8493008419">params</a>->O) {</div>
|
||||
<div class="line"><a id="l00251" name="l00251"></a><span class="lineno"> 251</span> <a class="code hl_define" href="steel_2defines_8h.html#a5a5c3095b132a7589bc19cd5cb80e2c6">STEEL_PRAGMA_UNROLL</a></div>
|
||||
<div class="line"><a id="l00252" name="l00252"></a><span class="lineno"> 252</span> <span class="keywordflow">for</span> (<span class="keywordtype">short</span> j = 0; j < <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08dba753ec7c8ea2892775746933b3e7">vec_size</a>; j++) {</div>
|
||||
<div class="line"><a id="l00253" name="l00253"></a><span class="lineno"> 253</span> dst[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aae121ca6016fc6c7255027b3641f3a09">dst_ld</a> + j] = curr_src[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6918c1df7712c4e408e2871467ea7987">src_ld</a> + j];</div>
|
||||
<div class="line"><a id="l00253" name="l00253"></a><span class="lineno"> 253</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a8474daf268013e138a84fc1c4bff7352">dst</a>[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aae121ca6016fc6c7255027b3641f3a09">dst_ld</a> + j] = curr_src[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6918c1df7712c4e408e2871467ea7987">src_ld</a> + j];</div>
|
||||
<div class="line"><a id="l00254" name="l00254"></a><span class="lineno"> 254</span> }</div>
|
||||
<div class="line"><a id="l00255" name="l00255"></a><span class="lineno"> 255</span> } <span class="keywordflow">else</span> {</div>
|
||||
<div class="line"><a id="l00256" name="l00256"></a><span class="lineno"> 256</span> <a class="code hl_define" href="steel_2defines_8h.html#a5a5c3095b132a7589bc19cd5cb80e2c6">STEEL_PRAGMA_UNROLL</a></div>
|
||||
<div class="line"><a id="l00257" name="l00257"></a><span class="lineno"> 257</span> <span class="keywordflow">for</span> (<span class="keywordtype">short</span> j = 0; j < <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08dba753ec7c8ea2892775746933b3e7">vec_size</a>; j++) {</div>
|
||||
<div class="line"><a id="l00258" name="l00258"></a><span class="lineno"> 258</span> dst[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aae121ca6016fc6c7255027b3641f3a09">dst_ld</a> + j] = T(0);</div>
|
||||
<div class="line"><a id="l00258" name="l00258"></a><span class="lineno"> 258</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a8474daf268013e138a84fc1c4bff7352">dst</a>[i * <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aae121ca6016fc6c7255027b3641f3a09">dst_ld</a> + j] = T(0);</div>
|
||||
<div class="line"><a id="l00259" name="l00259"></a><span class="lineno"> 259</span> }</div>
|
||||
<div class="line"><a id="l00260" name="l00260"></a><span class="lineno"> 260</span> }</div>
|
||||
<div class="line"><a id="l00261" name="l00261"></a><span class="lineno"> 261</span> }</div>
|
||||
@@ -369,14 +383,14 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00265" name="l00265"></a><span class="lineno"> 265</span> <span class="comment">/* Iteration helper */</span></div>
|
||||
<div class="foldopen" id="foldopen00266" data-start="{" data-end="}">
|
||||
<div class="line"><a id="l00266" name="l00266"></a><span class="lineno"><a class="line" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a11743cb1c108f42ccdc6e59204a5b3e8"> 266</a></span> METAL_FUNC <span class="keywordtype">void</span> <a class="code hl_function" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a11743cb1c108f42ccdc6e59204a5b3e8">next</a>() {</div>
|
||||
<div class="line"><a id="l00267" name="l00267"></a><span class="lineno"> 267</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6efa6268a37f18f4d225674bf1780cf6">weight_w</a> += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aa5611e9a84bebaee966d2b339c214ff5">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a568435a612574ab19a051a48055d4cfc">f_wgt_jump_w</a>;</div>
|
||||
<div class="line"><a id="l00267" name="l00267"></a><span class="lineno"> 267</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6efa6268a37f18f4d225674bf1780cf6">weight_w</a> += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aa5611e9a84bebaee966d2b339c214ff5">jump_params</a>->f_wgt_jump_w;</div>
|
||||
<div class="line"><a id="l00268" name="l00268"></a><span class="lineno"> 268</span> <span class="keywordflow">if</span> (<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6efa6268a37f18f4d225674bf1780cf6">weight_w < params-></a>wS[1]) {</div>
|
||||
<div class="line"><a id="l00269" name="l00269"></a><span class="lineno"> 269</span> <span class="keywordflow">return</span>;</div>
|
||||
<div class="line"><a id="l00270" name="l00270"></a><span class="lineno"> 270</span> }</div>
|
||||
<div class="line"><a id="l00271" name="l00271"></a><span class="lineno"> 271</span> </div>
|
||||
<div class="line"><a id="l00272" name="l00272"></a><span class="lineno"> 272</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a6efa6268a37f18f4d225674bf1780cf6">weight_w</a> = <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a230f0e581f9b8227b9ee68760b3b1503">base_ww</a>;</div>
|
||||
<div class="line"><a id="l00273" name="l00273"></a><span class="lineno"> 273</span> </div>
|
||||
<div class="line"><a id="l00274" name="l00274"></a><span class="lineno"> 274</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a5997fd8ef249e4cd3df7dad7b251d8d5">weight_h</a> += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aa5611e9a84bebaee966d2b339c214ff5">jump_params</a>-><a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a198ba0c2740ab4ded99345edf58917a7">f_wgt_jump_h</a>;</div>
|
||||
<div class="line"><a id="l00274" name="l00274"></a><span class="lineno"> 274</span> <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a5997fd8ef249e4cd3df7dad7b251d8d5">weight_h</a> += <a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aa5611e9a84bebaee966d2b339c214ff5">jump_params</a>->f_wgt_jump_h;</div>
|
||||
<div class="line"><a id="l00275" name="l00275"></a><span class="lineno"> 275</span> <span class="keywordflow">if</span> (<a class="code hl_variable" href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a5997fd8ef249e4cd3df7dad7b251d8d5">weight_h < params-></a>wS[0]) {</div>
|
||||
<div class="line"><a id="l00276" name="l00276"></a><span class="lineno"> 276</span> <span class="keywordflow">return</span>;</div>
|
||||
<div class="line"><a id="l00277" name="l00277"></a><span class="lineno"> 277</span> }</div>
|
||||
@@ -391,28 +405,13 @@ $(function(){ initResizable(false); });
|
||||
<div class="line"><a id="l00284" name="l00284"></a><span class="lineno"> 284</span> </div>
|
||||
<div class="line"><a id="l00285" name="l00285"></a><span class="lineno"> 285</span>} <span class="comment">// namespace steel</span></div>
|
||||
<div class="line"><a id="l00286" name="l00286"></a><span class="lineno"> 286</span>} <span class="comment">// namespace mlx</span></div>
|
||||
<div class="ttc" id="anamespacemlx_1_1steel_html"><div class="ttname"><a href="namespacemlx_1_1steel.html">mlx::steel</a></div><div class="ttdef"><b>Definition</b> attn.h:19</div></div>
|
||||
<div class="ttc" id="anamespacemlx_html"><div class="ttname"><a href="namespacemlx.html">mlx</a></div><div class="ttdef"><b>Definition</b> allocator.h:7</div></div>
|
||||
<div class="ttc" id="asteel_2defines_8h_html"><div class="ttname"><a href="steel_2defines_8h.html">defines.h</a></div></div>
|
||||
<div class="ttc" id="asteel_2defines_8h_html_a5a5c3095b132a7589bc19cd5cb80e2c6"><div class="ttname"><a href="steel_2defines_8h.html#a5a5c3095b132a7589bc19cd5cb80e2c6">STEEL_PRAGMA_UNROLL</a></div><div class="ttdeci">#define STEEL_PRAGMA_UNROLL</div><div class="ttdef"><b>Definition</b> defines.h:4</div></div>
|
||||
<div class="ttc" id="asteel_2defines_8h_html_a90b91c866313ffa46eff6d9cc944ad2b"><div class="ttname"><a href="steel_2defines_8h.html#a90b91c866313ffa46eff6d9cc944ad2b">STEEL_CONST</a></div><div class="ttdeci">#define STEEL_CONST</div><div class="ttdef"><b>Definition</b> defines.h:3</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html"><div class="ttname"><a href="struct_m_l_x_conv_params.html">MLXConvParams</a></div><div class="ttdef"><b>Definition</b> params.h:6</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_a7611db8f1621c7e09fc685ed44073b14"><div class="ttname"><a href="struct_m_l_x_conv_params.html#a7611db8f1621c7e09fc685ed44073b14">MLXConvParams::kdil</a></div><div class="ttdeci">const int kdil[NDIM]</div><div class="ttdef"><b>Definition</b> params.h:15</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_a862191e8ab1bc8a47aa1396b36d46058"><div class="ttname"><a href="struct_m_l_x_conv_params.html#a862191e8ab1bc8a47aa1396b36d46058">MLXConvParams::str</a></div><div class="ttdeci">const int str[NDIM]</div><div class="ttdef"><b>Definition</b> params.h:13</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_a887fee0da1494d038526fb0f59faff45"><div class="ttname"><a href="struct_m_l_x_conv_params.html#a887fee0da1494d038526fb0f59faff45">MLXConvParams::wt_strides</a></div><div class="ttdeci">const size_t wt_strides[NDIM+2]</div><div class="ttdef"><b>Definition</b> params.h:18</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_a8b30cda15eda20f84f12db868f21d0ef"><div class="ttname"><a href="struct_m_l_x_conv_params.html#a8b30cda15eda20f84f12db868f21d0ef">MLXConvParams::flip</a></div><div class="ttdeci">const bool flip</div><div class="ttdef"><b>Definition</b> params.h:21</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_ab25eade6573784985dbea1216f9068cf"><div class="ttname"><a href="struct_m_l_x_conv_params.html#ab25eade6573784985dbea1216f9068cf">MLXConvParams::in_strides</a></div><div class="ttdeci">const size_t in_strides[NDIM+2]</div><div class="ttdef"><b>Definition</b> params.h:17</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_aba2074189644b1b59567d018409277a9"><div class="ttname"><a href="struct_m_l_x_conv_params.html#aba2074189644b1b59567d018409277a9">MLXConvParams::wS</a></div><div class="ttdeci">const int wS[NDIM]</div><div class="ttdef"><b>Definition</b> params.h:11</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_ad55ff586d30072d8154865f9dfe92d97"><div class="ttname"><a href="struct_m_l_x_conv_params.html#ad55ff586d30072d8154865f9dfe92d97">MLXConvParams::O</a></div><div class="ttdeci">const int O</div><div class="ttdef"><b>Definition</b> params.h:9</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_ae84a9afb3a95b57e0b763bb0ebda0753"><div class="ttname"><a href="struct_m_l_x_conv_params.html#ae84a9afb3a95b57e0b763bb0ebda0753">MLXConvParams::pad</a></div><div class="ttdeci">const int pad[NDIM]</div><div class="ttdef"><b>Definition</b> params.h:14</div></div>
|
||||
<div class="ttc" id="astruct_m_l_x_conv_params_html_af900fdb4c4d4ea35eed02940dee8d4d1"><div class="ttname"><a href="struct_m_l_x_conv_params.html#af900fdb4c4d4ea35eed02940dee8d4d1">MLXConvParams::idil</a></div><div class="ttdeci">const int idil[NDIM]</div><div class="ttdef"><b>Definition</b> params.h:16</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_general_jump_params_html"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html">mlx::steel::Conv2DGeneralJumpParams</a></div><div class="ttdef"><b>Definition</b> params.h:43</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_general_jump_params_html_a0fd755691482cb03ea4534b4a556c197"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a0fd755691482cb03ea4534b4a556c197">mlx::steel::Conv2DGeneralJumpParams::f_out_jump_w</a></div><div class="ttdeci">const int f_out_jump_w</div><div class="ttdef"><b>Definition</b> params.h:48</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_general_jump_params_html_a198ba0c2740ab4ded99345edf58917a7"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a198ba0c2740ab4ded99345edf58917a7">mlx::steel::Conv2DGeneralJumpParams::f_wgt_jump_h</a></div><div class="ttdeci">const int f_wgt_jump_h</div><div class="ttdef"><b>Definition</b> params.h:44</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_general_jump_params_html_a568435a612574ab19a051a48055d4cfc"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a568435a612574ab19a051a48055d4cfc">mlx::steel::Conv2DGeneralJumpParams::f_wgt_jump_w</a></div><div class="ttdeci">const int f_wgt_jump_w</div><div class="ttdef"><b>Definition</b> params.h:45</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_general_jump_params_html_a78d48b55cf182f000abece0e5e7fadcb"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#a78d48b55cf182f000abece0e5e7fadcb">mlx::steel::Conv2DGeneralJumpParams::f_out_jump_h</a></div><div class="ttdeci">const int f_out_jump_h</div><div class="ttdef"><b>Definition</b> params.h:47</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_general_jump_params_html_ab971bf879079895189331fbeaf33c211"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#ab971bf879079895189331fbeaf33c211">mlx::steel::Conv2DGeneralJumpParams::adj_out_w</a></div><div class="ttdeci">const int adj_out_w</div><div class="ttdef"><b>Definition</b> params.h:51</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_general_jump_params_html_aed0ffd63fbc85fd5d5c4cc7b43f68363"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_general_jump_params.html#aed0ffd63fbc85fd5d5c4cc7b43f68363">mlx::steel::Conv2DGeneralJumpParams::adj_out_hw</a></div><div class="ttdeci">const int adj_out_hw</div><div class="ttdef"><b>Definition</b> params.h:52</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_input_block_loader_general_html"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html">mlx::steel::Conv2DInputBlockLoaderGeneral</a></div><div class="ttdef"><b>Definition</b> loader_general.h:21</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_input_block_loader_general_html_a0261d0349a0a95ca1a02a959b73e9352"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a0261d0349a0a95ca1a02a959b73e9352">mlx::steel::Conv2DInputBlockLoaderGeneral::weight_w</a></div><div class="ttdeci">short weight_w</div><div class="ttdef"><b>Definition</b> loader_general.h:52</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_input_block_loader_general_html_a07c85eab8cbf7b02c60df29cf32031ef"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a07c85eab8cbf7b02c60df29cf32031ef">mlx::steel::Conv2DInputBlockLoaderGeneral::dst_ld</a></div><div class="ttdeci">STEEL_CONST short dst_ld</div><div class="ttdef"><b>Definition</b> loader_general.h:27</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_input_block_loader_general_html_a09fd92c74ef57c20b48bc780153365ba"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#a09fd92c74ef57c20b48bc780153365ba">mlx::steel::Conv2DInputBlockLoaderGeneral::params</a></div><div class="ttdeci">const constant MLXConvParams< 2 > * params</div><div class="ttdef"><b>Definition</b> loader_general.h:45</div></div>
|
||||
@@ -437,7 +436,6 @@ $(function(){ initResizable(false); });
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_input_block_loader_general_html_ace16704025bc6e6204c306a357f3a8b8"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ace16704025bc6e6204c306a357f3a8b8">mlx::steel::Conv2DInputBlockLoaderGeneral::bj</a></div><div class="ttdeci">const short bj</div><div class="ttdef"><b>Definition</b> loader_general.h:40</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_input_block_loader_general_html_ae25c676b7318d78462ee89bcd80dc805"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#ae25c676b7318d78462ee89bcd80dc805">mlx::steel::Conv2DInputBlockLoaderGeneral::TCOLS</a></div><div class="ttdeci">STEEL_CONST short TCOLS</div><div class="ttdef"><b>Definition</b> loader_general.h:31</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_input_block_loader_general_html_afe5caaf38b574d3380533856c493dd92"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_input_block_loader_general.html#afe5caaf38b574d3380533856c493dd92">mlx::steel::Conv2DInputBlockLoaderGeneral::read_n</a></div><div class="ttdeci">int read_n[n_rows]</div><div class="ttdef"><b>Definition</b> loader_general.h:56</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_weight_block_loader_general_html"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html">mlx::steel::Conv2DWeightBlockLoaderGeneral</a></div><div class="ttdef"><b>Definition</b> loader_general.h:170</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_weight_block_loader_general_html_a015a0c56de74a0c4d51953a7e94fbba8"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a015a0c56de74a0c4d51953a7e94fbba8">mlx::steel::Conv2DWeightBlockLoaderGeneral::BROWS</a></div><div class="ttdeci">STEEL_CONST short BROWS</div><div class="ttdef"><b>Definition</b> loader_general.h:172</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_weight_block_loader_general_html_a08a517bc50caf41155b98be0690bfe44"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08a517bc50caf41155b98be0690bfe44">mlx::steel::Conv2DWeightBlockLoaderGeneral::thread_idx</a></div><div class="ttdeci">const short thread_idx</div><div class="ttdef"><b>Definition</b> loader_general.h:191</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_weight_block_loader_general_html_a08dba753ec7c8ea2892775746933b3e7"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#a08dba753ec7c8ea2892775746933b3e7">mlx::steel::Conv2DWeightBlockLoaderGeneral::vec_size</a></div><div class="ttdeci">STEEL_CONST short vec_size</div><div class="ttdef"><b>Definition</b> loader_general.h:177</div></div>
|
||||
@@ -462,10 +460,13 @@ $(function(){ initResizable(false); });
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_weight_block_loader_general_html_ae4759d18c0e5cc3530b3da8493008419"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#ae4759d18c0e5cc3530b3da8493008419">mlx::steel::Conv2DWeightBlockLoaderGeneral::params</a></div><div class="ttdeci">const constant MLXConvParams< 2 > * params</div><div class="ttdef"><b>Definition</b> loader_general.h:199</div></div>
|
||||
<div class="ttc" id="astructmlx_1_1steel_1_1_conv2_d_weight_block_loader_general_html_aff021a6fae860b4ac01fb593b2720457"><div class="ttname"><a href="structmlx_1_1steel_1_1_conv2_d_weight_block_loader_general.html#aff021a6fae860b4ac01fb593b2720457">mlx::steel::Conv2DWeightBlockLoaderGeneral::TCOLS</a></div><div class="ttdeci">STEEL_CONST short TCOLS</div><div class="ttdef"><b>Definition</b> loader_general.h:181</div></div>
|
||||
</div><!-- fragment --></div><!-- contents -->
|
||||
<!-- start footer part -->
|
||||
<hr class="footer"/><address class="footer"><small>
|
||||
Generated by <a href="https://www.doxygen.org/index.html"><img class="footer" src="doxygen.svg" width="104" height="31" alt="doxygen"/></a> 1.12.0
|
||||
</small></address>
|
||||
</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_938ab0ecf10b8b860ff766c820f665fd.html">mlx</a></li><li class="navelem"><a class="el" href="dir_1d446c9bd3c99228254c9484e0bc5c06.html">backend</a></li><li class="navelem"><a class="el" href="dir_d0c977ea65824390717cdb7efc36c157.html">metal</a></li><li class="navelem"><a class="el" href="dir_70a37effa88bcbd6b791977fa1e64356.html">kernels</a></li><li class="navelem"><a class="el" href="dir_76215a6c54e2b67053e723fc2395583c.html">steel</a></li><li class="navelem"><a class="el" href="dir_df9494e83ef22ae6150a0e080d9709ed.html">conv</a></li><li class="navelem"><a class="el" href="dir_ba4426224ef60f409462a2a12fa18f06.html">loaders</a></li><li class="navelem"><a class="el" href="loader__general_8h.html">loader_general.h</a></li>
|
||||
<li class="footer">Generated by <a href="https://www.doxygen.org/index.html"><img class="footer" src="doxygen.svg" width="104" height="31" alt="doxygen"/></a> 1.13.1 </li>
|
||||
</ul>
|
||||
</div>
|
||||
</body>
|
||||
</html>
|
||||
|
Reference in New Issue
Block a user