aaru-dps/libaaruformat
1
0
Fork 0
mirror of https://github.com/aaru-dps/libaaruformat.git synced 2025-12-16 19:24:40 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
ced56fad6c562b5950668596c284dd0ac82cf1ce
libaaruformat/docs/html/arm__vmull_8c_source.html

275 lines
30 KiB
HTML
Raw Blame History

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "https://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" lang="en-US">
<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 1.14.0"/>
<meta name="viewport" content="width=device-width, initial-scale=1"/>
<title>libaaruformat: src/crc64/arm_vmull.c Source File</title>
<link href="tabs.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript" src="jquery.js"></script>
<script type="text/javascript" src="dynsections.js"></script>
<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="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>
<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 id="projectrow">
<td id="projectalign">
<div id="projectname">libaaruformat<span id="projectnumber">&#160;1.0</span>
</div>
<div id="projectbrief">Aaru Data Preservation Suite - Format Library</div>
</td>
</tr>
</tbody>
</table>
</div>
<!-- end header part -->
<!-- Generated by Doxygen 1.14.0 -->
<script type="text/javascript">
var searchBox = new SearchBox("searchBox", "search/",'.html');
</script>
<script type="text/javascript">
$(function() { codefold.init(); });
</script>
<script type="text/javascript" src="menudata.js"></script>
<script type="text/javascript" src="menu.js"></script>
<script type="text/javascript">
$(function() {
initMenu('',true,false,'search.php','Search',true);
$(function() { init_search(); });
});
</script>
<div id="main-nav"></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">
$(function(){initNavTree('arm__vmull_8c_source.html','',''); });
</script>
<div id="container">
<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">
<div id="MSearchResults">
<div class="SRPage">
<div id="SRIndex">
<div id="SRResults"></div>
<div class="SRStatus" id="Loading">Loading...</div>
<div class="SRStatus" id="Searching">Searching...</div>
<div class="SRStatus" id="NoMatches">No Matches</div>
</div>
</div>
</div>
</div>
<div class="header">
<div class="headertitle"><div class="title">arm_vmull.c</div></div>
</div><!--header-->
<div class="contents">
<a href="arm__vmull_8c.html">Go to the documentation of this file.</a><div class="fragment"><div class="line"><a id="l00001" name="l00001"></a><span class="lineno"> 1</span><span class="comment">/*</span></div>
<div class="line"><a id="l00002" name="l00002"></a><span class="lineno"> 2</span><span class="comment"> * This file is part of the Aaru Data Preservation Suite.</span></div>
<div class="line"><a id="l00003" name="l00003"></a><span class="lineno"> 3</span><span class="comment"> * Copyright (c) 2019-2025 Natalia Portillo.</span></div>
<div class="line"><a id="l00004" name="l00004"></a><span class="lineno"> 4</span><span class="comment"> * Copyright sse2neon.h contributors</span></div>
<div class="line"><a id="l00005" name="l00005"></a><span class="lineno"> 5</span><span class="comment"> *</span></div>
<div class="line"><a id="l00006" name="l00006"></a><span class="lineno"> 6</span><span class="comment"> * sse2neon is freely redistributable under the MIT License.</span></div>
<div class="line"><a id="l00007" name="l00007"></a><span class="lineno"> 7</span><span class="comment"> *</span></div>
<div class="line"><a id="l00008" name="l00008"></a><span class="lineno"> 8</span><span class="comment"> * Permission is hereby granted, free of charge, to any person obtaining a copy</span></div>
<div class="line"><a id="l00009" name="l00009"></a><span class="lineno"> 9</span><span class="comment"> * of this software and associated documentation files (the &quot;Software&quot;), to deal</span></div>
<div class="line"><a id="l00010" name="l00010"></a><span class="lineno"> 10</span><span class="comment"> * in the Software without restriction, including without limitation the rights</span></div>
<div class="line"><a id="l00011" name="l00011"></a><span class="lineno"> 11</span><span class="comment"> * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell</span></div>
<div class="line"><a id="l00012" name="l00012"></a><span class="lineno"> 12</span><span class="comment"> * copies of the Software, and to permit persons to whom the Software is</span></div>
<div class="line"><a id="l00013" name="l00013"></a><span class="lineno"> 13</span><span class="comment"> * furnished to do so, subject to the following conditions:</span></div>
<div class="line"><a id="l00014" name="l00014"></a><span class="lineno"> 14</span><span class="comment"> *</span></div>
<div class="line"><a id="l00015" name="l00015"></a><span class="lineno"> 15</span><span class="comment"> * The above copyright notice and this permission notice shall be included in</span></div>
<div class="line"><a id="l00016" name="l00016"></a><span class="lineno"> 16</span><span class="comment"> * all copies or substantial portions of the Software.</span></div>
<div class="line"><a id="l00017" name="l00017"></a><span class="lineno"> 17</span><span class="comment"> *</span></div>
<div class="line"><a id="l00018" name="l00018"></a><span class="lineno"> 18</span><span class="comment"> * THE SOFTWARE IS PROVIDED &quot;AS IS&quot;, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR</span></div>
<div class="line"><a id="l00019" name="l00019"></a><span class="lineno"> 19</span><span class="comment"> * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,</span></div>
<div class="line"><a id="l00020" name="l00020"></a><span class="lineno"> 20</span><span class="comment"> * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE</span></div>
<div class="line"><a id="l00021" name="l00021"></a><span class="lineno"> 21</span><span class="comment"> * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER</span></div>
<div class="line"><a id="l00022" name="l00022"></a><span class="lineno"> 22</span><span class="comment"> * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,</span></div>
<div class="line"><a id="l00023" name="l00023"></a><span class="lineno"> 23</span><span class="comment"> * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE</span></div>
<div class="line"><a id="l00024" name="l00024"></a><span class="lineno"> 24</span><span class="comment"> * SOFTWARE.</span></div>
<div class="line"><a id="l00025" name="l00025"></a><span class="lineno"> 25</span><span class="comment"> */</span></div>
<div class="line"><a id="l00026" name="l00026"></a><span class="lineno"> 26</span> </div>
<div class="line"><a id="l00027" name="l00027"></a><span class="lineno"> 27</span><span class="preprocessor">#if defined(__aarch64__) || defined(_M_ARM64) || defined(__arm__) || defined(_M_ARM)</span></div>
<div class="line"><a id="l00028" name="l00028"></a><span class="lineno"> 28</span> </div>
<div class="line"><a id="l00029" name="l00029"></a><span class="lineno"> 29</span><span class="preprocessor">#include &lt;arm_neon.h&gt;</span></div>
<div class="line"><a id="l00030" name="l00030"></a><span class="lineno"> 30</span> </div>
<div class="line"><a id="l00031" name="l00031"></a><span class="lineno"> 31</span><span class="preprocessor">#include &lt;<a class="code" href="aaruformat_8h.html">aaruformat.h</a>&gt;</span></div>
<div class="line"><a id="l00032" name="l00032"></a><span class="lineno"> 32</span> </div>
<div class="line"><a id="l00033" name="l00033"></a><span class="lineno"> 33</span><span class="preprocessor">#include &quot;<a class="code" href="arm__vmull_8h.html">arm_vmull.h</a>&quot;</span></div>
<div class="line"><a id="l00034" name="l00034"></a><span class="lineno"> 34</span> </div>
<div class="line"><a id="l00035" name="l00035"></a><span class="lineno"> 35</span><span class="preprocessor">#if !defined(__MINGW32__) &amp;&amp; (!defined(__ANDROID__) || !defined(__arm__))</span></div>
<div class="line"><a id="l00036" name="l00036"></a><span class="lineno"> 36</span>TARGET_WITH_CRYPTO <span class="keyword">static</span> uint64x2_t sse2neon_vmull_p64_crypto(uint64x1_t _a, uint64x1_t _b)</div>
<div class="line"><a id="l00037" name="l00037"></a><span class="lineno"> 37</span>{</div>
<div class="line"><a id="l00038" name="l00038"></a><span class="lineno"> 38</span> poly64_t a = vget_lane_p64(vreinterpret_p64_u64(_a), 0);</div>
<div class="line"><a id="l00039" name="l00039"></a><span class="lineno"> 39</span> poly64_t b = vget_lane_p64(vreinterpret_p64_u64(_b), 0);</div>
<div class="line"><a id="l00040" name="l00040"></a><span class="lineno"> 40</span> <span class="keywordflow">return</span> vreinterpretq_u64_p128(vmull_p64(a, b));</div>
<div class="line"><a id="l00041" name="l00041"></a><span class="lineno"> 41</span>}</div>
<div class="line"><a id="l00042" name="l00042"></a><span class="lineno"> 42</span><span class="preprocessor">#endif</span></div>
<div class="line"><a id="l00043" name="l00043"></a><span class="lineno"> 43</span> </div>
<div class="line"><a id="l00044" name="l00044"></a><span class="lineno"> 44</span>TARGET_WITH_SIMD uint64x2_t sse2neon_vmull_p64(uint64x1_t _a, uint64x1_t _b)</div>
<div class="line"><a id="l00045" name="l00045"></a><span class="lineno"> 45</span>{</div>
<div class="line"><a id="l00046" name="l00046"></a><span class="lineno"> 46</span><span class="preprocessor">#if !defined(__MINGW32__) &amp;&amp; (!defined(__ANDROID__) || !defined(__arm__))</span></div>
<div class="line"><a id="l00047" name="l00047"></a><span class="lineno"> 47</span> <span class="comment">// Wraps vmull_p64</span></div>
<div class="line"><a id="l00048" name="l00048"></a><span class="lineno"> 48</span> <span class="keywordflow">if</span>(have_arm_crypto()) <span class="keywordflow">return</span> sse2neon_vmull_p64_crypto(_a, _b);</div>
<div class="line"><a id="l00049" name="l00049"></a><span class="lineno"> 49</span><span class="preprocessor">#endif</span></div>
<div class="line"><a id="l00050" name="l00050"></a><span class="lineno"> 50</span> </div>
<div class="line"><a id="l00051" name="l00051"></a><span class="lineno"> 51</span> <span class="comment">// ARMv7 polyfill</span></div>
<div class="line"><a id="l00052" name="l00052"></a><span class="lineno"> 52</span> <span class="comment">// ARMv7/some A64 lacks vmull_p64, but it has vmull_p8.</span></div>
<div class="line"><a id="l00053" name="l00053"></a><span class="lineno"> 53</span> <span class="comment">//</span></div>
<div class="line"><a id="l00054" name="l00054"></a><span class="lineno"> 54</span> <span class="comment">// vmull_p8 calculates 8 8-bit-&gt;16-bit polynomial multiplies, but we need a</span></div>
<div class="line"><a id="l00055" name="l00055"></a><span class="lineno"> 55</span> <span class="comment">// 64-bit-&gt;128-bit polynomial multiply.</span></div>
<div class="line"><a id="l00056" name="l00056"></a><span class="lineno"> 56</span> <span class="comment">//</span></div>
<div class="line"><a id="l00057" name="l00057"></a><span class="lineno"> 57</span> <span class="comment">// It needs some work and is somewhat slow, but it is still faster than all</span></div>
<div class="line"><a id="l00058" name="l00058"></a><span class="lineno"> 58</span> <span class="comment">// known scalar methods.</span></div>
<div class="line"><a id="l00059" name="l00059"></a><span class="lineno"> 59</span> <span class="comment">//</span></div>
<div class="line"><a id="l00060" name="l00060"></a><span class="lineno"> 60</span> <span class="comment">// Algorithm adapted to C from</span></div>
<div class="line"><a id="l00061" name="l00061"></a><span class="lineno"> 61</span> <span class="comment">// https://www.workofard.com/2017/07/ghash-for-low-end-cores/, which is adapted</span></div>
<div class="line"><a id="l00062" name="l00062"></a><span class="lineno"> 62</span> <span class="comment">// from &quot;Fast Software Polynomial Multiplication on ARM Processors Using the</span></div>
<div class="line"><a id="l00063" name="l00063"></a><span class="lineno"> 63</span> <span class="comment">// NEON Engine&quot; by Danilo Camara, Conrado Gouvea, Julio Lopez and Ricardo Dahab</span></div>
<div class="line"><a id="l00064" name="l00064"></a><span class="lineno"> 64</span> <span class="comment">// (https://hal.inria.fr/hal-01506572)</span></div>
<div class="line"><a id="l00065" name="l00065"></a><span class="lineno"> 65</span> </div>
<div class="line"><a id="l00066" name="l00066"></a><span class="lineno"> 66</span> poly8x8_t a = vreinterpret_p8_u64(_a);</div>
<div class="line"><a id="l00067" name="l00067"></a><span class="lineno"> 67</span> poly8x8_t b = vreinterpret_p8_u64(_b);</div>
<div class="line"><a id="l00068" name="l00068"></a><span class="lineno"> 68</span> </div>
<div class="line"><a id="l00069" name="l00069"></a><span class="lineno"> 69</span> <span class="comment">// Masks</span></div>
<div class="line"><a id="l00070" name="l00070"></a><span class="lineno"> 70</span> uint8x16_t k48_32 = vcombine_u8(vcreate_u8(0x0000ffffffffffff), vcreate_u8(0x00000000ffffffff));</div>
<div class="line"><a id="l00071" name="l00071"></a><span class="lineno"> 71</span> uint8x16_t k16_00 = vcombine_u8(vcreate_u8(0x000000000000ffff), vcreate_u8(0x0000000000000000));</div>
<div class="line"><a id="l00072" name="l00072"></a><span class="lineno"> 72</span> </div>
<div class="line"><a id="l00073" name="l00073"></a><span class="lineno"> 73</span> <span class="comment">// Do the multiplies, rotating with vext to get all combinations</span></div>
<div class="line"><a id="l00074" name="l00074"></a><span class="lineno"> 74</span> uint8x16_t d = vreinterpretq_u8_p16(vmull_p8(a, b)); <span class="comment">// D = A0 * B0</span></div>
<div class="line"><a id="l00075" name="l00075"></a><span class="lineno"> 75</span> uint8x16_t e = vreinterpretq_u8_p16(vmull_p8(a, vext_p8(b, b, 1))); <span class="comment">// E = A0 * B1</span></div>
<div class="line"><a id="l00076" name="l00076"></a><span class="lineno"> 76</span> uint8x16_t f = vreinterpretq_u8_p16(vmull_p8(vext_p8(a, a, 1), b)); <span class="comment">// F = A1 * B0</span></div>
<div class="line"><a id="l00077" name="l00077"></a><span class="lineno"> 77</span> uint8x16_t g = vreinterpretq_u8_p16(vmull_p8(a, vext_p8(b, b, 2))); <span class="comment">// G = A0 * B2</span></div>
<div class="line"><a id="l00078" name="l00078"></a><span class="lineno"> 78</span> uint8x16_t h = vreinterpretq_u8_p16(vmull_p8(vext_p8(a, a, 2), b)); <span class="comment">// H = A2 * B0</span></div>
<div class="line"><a id="l00079" name="l00079"></a><span class="lineno"> 79</span> uint8x16_t i = vreinterpretq_u8_p16(vmull_p8(a, vext_p8(b, b, 3))); <span class="comment">// I = A0 * B3</span></div>
<div class="line"><a id="l00080" name="l00080"></a><span class="lineno"> 80</span> uint8x16_t j = vreinterpretq_u8_p16(vmull_p8(vext_p8(a, a, 3), b)); <span class="comment">// J = A3 * B0</span></div>
<div class="line"><a id="l00081" name="l00081"></a><span class="lineno"> 81</span> uint8x16_t k = vreinterpretq_u8_p16(vmull_p8(a, vext_p8(b, b, 4))); <span class="comment">// L = A0 * B4</span></div>
<div class="line"><a id="l00082" name="l00082"></a><span class="lineno"> 82</span> </div>
<div class="line"><a id="l00083" name="l00083"></a><span class="lineno"> 83</span> <span class="comment">// Add cross products</span></div>
<div class="line"><a id="l00084" name="l00084"></a><span class="lineno"> 84</span> uint8x16_t l = veorq_u8(e, f); <span class="comment">// L = E + F</span></div>
<div class="line"><a id="l00085" name="l00085"></a><span class="lineno"> 85</span> uint8x16_t m = veorq_u8(g, h); <span class="comment">// M = G + H</span></div>
<div class="line"><a id="l00086" name="l00086"></a><span class="lineno"> 86</span> uint8x16_t n = veorq_u8(i, j); <span class="comment">// N = I + J</span></div>
<div class="line"><a id="l00087" name="l00087"></a><span class="lineno"> 87</span> </div>
<div class="line"><a id="l00088" name="l00088"></a><span class="lineno"> 88</span> <span class="comment">// Interleave. Using vzip1 and vzip2 prevents Clang from emitting TBL</span></div>
<div class="line"><a id="l00089" name="l00089"></a><span class="lineno"> 89</span> <span class="comment">// instructions.</span></div>
<div class="line"><a id="l00090" name="l00090"></a><span class="lineno"> 90</span><span class="preprocessor">#if defined(__aarch64__)</span></div>
<div class="line"><a id="l00091" name="l00091"></a><span class="lineno"> 91</span> uint8x16_t lm_p0 = vreinterpretq_u8_u64(vzip1q_u64(vreinterpretq_u64_u8(l), vreinterpretq_u64_u8(m)));</div>
<div class="line"><a id="l00092" name="l00092"></a><span class="lineno"> 92</span> uint8x16_t lm_p1 = vreinterpretq_u8_u64(vzip2q_u64(vreinterpretq_u64_u8(l), vreinterpretq_u64_u8(m)));</div>
<div class="line"><a id="l00093" name="l00093"></a><span class="lineno"> 93</span> uint8x16_t nk_p0 = vreinterpretq_u8_u64(vzip1q_u64(vreinterpretq_u64_u8(n), vreinterpretq_u64_u8(k)));</div>
<div class="line"><a id="l00094" name="l00094"></a><span class="lineno"> 94</span> uint8x16_t nk_p1 = vreinterpretq_u8_u64(vzip2q_u64(vreinterpretq_u64_u8(n), vreinterpretq_u64_u8(k)));</div>
<div class="line"><a id="l00095" name="l00095"></a><span class="lineno"> 95</span><span class="preprocessor">#else</span></div>
<div class="line"><a id="l00096" name="l00096"></a><span class="lineno"> 96</span> uint8x16_t lm_p0 = vcombine_u8(vget_low_u8(l), vget_low_u8(m));</div>
<div class="line"><a id="l00097" name="l00097"></a><span class="lineno"> 97</span> uint8x16_t lm_p1 = vcombine_u8(vget_high_u8(l), vget_high_u8(m));</div>
<div class="line"><a id="l00098" name="l00098"></a><span class="lineno"> 98</span> uint8x16_t nk_p0 = vcombine_u8(vget_low_u8(n), vget_low_u8(k));</div>
<div class="line"><a id="l00099" name="l00099"></a><span class="lineno"> 99</span> uint8x16_t nk_p1 = vcombine_u8(vget_high_u8(n), vget_high_u8(k));</div>
<div class="line"><a id="l00100" name="l00100"></a><span class="lineno"> 100</span><span class="preprocessor">#endif</span></div>
<div class="line"><a id="l00101" name="l00101"></a><span class="lineno"> 101</span> <span class="comment">// t0 = (L) (P0 + P1) &lt;&lt; 8</span></div>
<div class="line"><a id="l00102" name="l00102"></a><span class="lineno"> 102</span> <span class="comment">// t1 = (M) (P2 + P3) &lt;&lt; 16</span></div>
<div class="line"><a id="l00103" name="l00103"></a><span class="lineno"> 103</span> uint8x16_t t0t1_tmp = veorq_u8(lm_p0, lm_p1);</div>
<div class="line"><a id="l00104" name="l00104"></a><span class="lineno"> 104</span> uint8x16_t t0t1_h = vandq_u8(lm_p1, k48_32);</div>
<div class="line"><a id="l00105" name="l00105"></a><span class="lineno"> 105</span> uint8x16_t t0t1_l = veorq_u8(t0t1_tmp, t0t1_h);</div>
<div class="line"><a id="l00106" name="l00106"></a><span class="lineno"> 106</span> </div>
<div class="line"><a id="l00107" name="l00107"></a><span class="lineno"> 107</span> <span class="comment">// t2 = (N) (P4 + P5) &lt;&lt; 24</span></div>
<div class="line"><a id="l00108" name="l00108"></a><span class="lineno"> 108</span> <span class="comment">// t3 = (K) (P6 + P7) &lt;&lt; 32</span></div>
<div class="line"><a id="l00109" name="l00109"></a><span class="lineno"> 109</span> uint8x16_t t2t3_tmp = veorq_u8(nk_p0, nk_p1);</div>
<div class="line"><a id="l00110" name="l00110"></a><span class="lineno"> 110</span> uint8x16_t t2t3_h = vandq_u8(nk_p1, k16_00);</div>
<div class="line"><a id="l00111" name="l00111"></a><span class="lineno"> 111</span> uint8x16_t t2t3_l = veorq_u8(t2t3_tmp, t2t3_h);</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="comment">// De-interleave</span></div>
<div class="line"><a id="l00114" name="l00114"></a><span class="lineno"> 114</span><span class="preprocessor">#if defined(__aarch64__)</span></div>
<div class="line"><a id="l00115" name="l00115"></a><span class="lineno"> 115</span> uint8x16_t t0 = vreinterpretq_u8_u64(vuzp1q_u64(vreinterpretq_u64_u8(t0t1_l), vreinterpretq_u64_u8(t0t1_h)));</div>
<div class="line"><a id="l00116" name="l00116"></a><span class="lineno"> 116</span> uint8x16_t t1 = vreinterpretq_u8_u64(vuzp2q_u64(vreinterpretq_u64_u8(t0t1_l), vreinterpretq_u64_u8(t0t1_h)));</div>
<div class="line"><a id="l00117" name="l00117"></a><span class="lineno"> 117</span> uint8x16_t t2 = vreinterpretq_u8_u64(vuzp1q_u64(vreinterpretq_u64_u8(t2t3_l), vreinterpretq_u64_u8(t2t3_h)));</div>
<div class="line"><a id="l00118" name="l00118"></a><span class="lineno"> 118</span> uint8x16_t t3 = vreinterpretq_u8_u64(vuzp2q_u64(vreinterpretq_u64_u8(t2t3_l), vreinterpretq_u64_u8(t2t3_h)));</div>
<div class="line"><a id="l00119" name="l00119"></a><span class="lineno"> 119</span><span class="preprocessor">#else</span></div>
<div class="line"><a id="l00120" name="l00120"></a><span class="lineno"> 120</span> uint8x16_t t1 = vcombine_u8(vget_high_u8(t0t1_l), vget_high_u8(t0t1_h));</div>
<div class="line"><a id="l00121" name="l00121"></a><span class="lineno"> 121</span> uint8x16_t t0 = vcombine_u8(vget_low_u8(t0t1_l), vget_low_u8(t0t1_h));</div>
<div class="line"><a id="l00122" name="l00122"></a><span class="lineno"> 122</span> uint8x16_t t3 = vcombine_u8(vget_high_u8(t2t3_l), vget_high_u8(t2t3_h));</div>
<div class="line"><a id="l00123" name="l00123"></a><span class="lineno"> 123</span> uint8x16_t t2 = vcombine_u8(vget_low_u8(t2t3_l), vget_low_u8(t2t3_h));</div>
<div class="line"><a id="l00124" name="l00124"></a><span class="lineno"> 124</span><span class="preprocessor">#endif</span></div>
<div class="line"><a id="l00125" name="l00125"></a><span class="lineno"> 125</span> <span class="comment">// Shift the cross products</span></div>
<div class="line"><a id="l00126" name="l00126"></a><span class="lineno"> 126</span> uint8x16_t t0_shift = vextq_u8(t0, t0, 15); <span class="comment">// t0 &lt;&lt; 8</span></div>
<div class="line"><a id="l00127" name="l00127"></a><span class="lineno"> 127</span> uint8x16_t t1_shift = vextq_u8(t1, t1, 14); <span class="comment">// t1 &lt;&lt; 16</span></div>
<div class="line"><a id="l00128" name="l00128"></a><span class="lineno"> 128</span> uint8x16_t t2_shift = vextq_u8(t2, t2, 13); <span class="comment">// t2 &lt;&lt; 24</span></div>
<div class="line"><a id="l00129" name="l00129"></a><span class="lineno"> 129</span> uint8x16_t t3_shift = vextq_u8(t3, t3, 12); <span class="comment">// t3 &lt;&lt; 32</span></div>
<div class="line"><a id="l00130" name="l00130"></a><span class="lineno"> 130</span> </div>
<div class="line"><a id="l00131" name="l00131"></a><span class="lineno"> 131</span> <span class="comment">// Accumulate the products</span></div>
<div class="line"><a id="l00132" name="l00132"></a><span class="lineno"> 132</span> uint8x16_t cross1 = veorq_u8(t0_shift, t1_shift);</div>
<div class="line"><a id="l00133" name="l00133"></a><span class="lineno"> 133</span> uint8x16_t cross2 = veorq_u8(t2_shift, t3_shift);</div>
<div class="line"><a id="l00134" name="l00134"></a><span class="lineno"> 134</span> uint8x16_t mix = veorq_u8(d, cross1);</div>
<div class="line"><a id="l00135" name="l00135"></a><span class="lineno"> 135</span> uint8x16_t r = veorq_u8(mix, cross2);</div>
<div class="line"><a id="l00136" name="l00136"></a><span class="lineno"> 136</span> <span class="keywordflow">return</span> vreinterpretq_u64_u8(r);</div>
<div class="line"><a id="l00137" name="l00137"></a><span class="lineno"> 137</span>}</div>
<div class="line"><a id="l00138" name="l00138"></a><span class="lineno"> 138</span> </div>
<div class="line"><a id="l00139" name="l00139"></a><span class="lineno"> 139</span>TARGET_WITH_SIMD uint64x2_t mm_shuffle_epi8(uint64x2_t a, uint64x2_t b)</div>
<div class="line"><a id="l00140" name="l00140"></a><span class="lineno"> 140</span>{</div>
<div class="line"><a id="l00141" name="l00141"></a><span class="lineno"> 141</span> uint8x16_t tbl = vreinterpretq_u8_u64(a); <span class="comment">// input a</span></div>
<div class="line"><a id="l00142" name="l00142"></a><span class="lineno"> 142</span> uint8x16_t idx = vreinterpretq_u8_u64(b); <span class="comment">// input b</span></div>
<div class="line"><a id="l00143" name="l00143"></a><span class="lineno"> 143</span> uint8x16_t idx_masked = vandq_u8(idx, vdupq_n_u8(0x8F)); <span class="comment">// avoid using meaningless bits</span></div>
<div class="line"><a id="l00144" name="l00144"></a><span class="lineno"> 144</span><span class="preprocessor">#if defined(__aarch64__)</span></div>
<div class="line"><a id="l00145" name="l00145"></a><span class="lineno"> 145</span> <span class="keywordflow">return</span> vreinterpretq_u64_u8(vqtbl1q_u8(tbl, idx_masked));</div>
<div class="line"><a id="l00146" name="l00146"></a><span class="lineno"> 146</span><span class="preprocessor">#else</span></div>
<div class="line"><a id="l00147" name="l00147"></a><span class="lineno"> 147</span> <span class="comment">// use this line if testing on aarch64</span></div>
<div class="line"><a id="l00148" name="l00148"></a><span class="lineno"> 148</span> uint8x8x2_t a_split = {vget_low_u8(tbl), vget_high_u8(tbl)};</div>
<div class="line"><a id="l00149" name="l00149"></a><span class="lineno"> 149</span> <span class="keywordflow">return</span> vreinterpretq_u64_u8(</div>
<div class="line"><a id="l00150" name="l00150"></a><span class="lineno"> 150</span> vcombine_u8(vtbl2_u8(a_split, vget_low_u8(idx_masked)), vtbl2_u8(a_split, vget_high_u8(idx_masked))));</div>
<div class="line"><a id="l00151" name="l00151"></a><span class="lineno"> 151</span><span class="preprocessor">#endif</span></div>
<div class="line"><a id="l00152" name="l00152"></a><span class="lineno"> 152</span>}</div>
<div class="line"><a id="l00153" name="l00153"></a><span class="lineno"> 153</span> </div>
<div class="line"><a id="l00154" name="l00154"></a><span class="lineno"> 154</span>TARGET_WITH_SIMD uint64x2_t mm_srli_si128(uint64x2_t a, <span class="keywordtype">int</span> imm)</div>
<div class="line"><a id="l00155" name="l00155"></a><span class="lineno"> 155</span>{</div>
<div class="line"><a id="l00156" name="l00156"></a><span class="lineno"> 156</span> uint8x16_t tmp[2] = {vreinterpretq_u8_u64(a), vdupq_n_u8(0)};</div>
<div class="line"><a id="l00157" name="l00157"></a><span class="lineno"> 157</span> <span class="keywordflow">return</span> vreinterpretq_u64_u8(vld1q_u8(((uint8_t <span class="keyword">const</span> *)tmp) + imm));</div>
<div class="line"><a id="l00158" name="l00158"></a><span class="lineno"> 158</span>}</div>
<div class="line"><a id="l00159" name="l00159"></a><span class="lineno"> 159</span> </div>
<div class="line"><a id="l00160" name="l00160"></a><span class="lineno"> 160</span>TARGET_WITH_SIMD uint64x2_t mm_slli_si128(uint64x2_t a, <span class="keywordtype">int</span> imm)</div>
<div class="line"><a id="l00161" name="l00161"></a><span class="lineno"> 161</span>{</div>
<div class="line"><a id="l00162" name="l00162"></a><span class="lineno"> 162</span> uint8x16_t tmp[2] = {vdupq_n_u8(0), vreinterpretq_u8_u64(a)};</div>
<div class="line"><a id="l00163" name="l00163"></a><span class="lineno"> 163</span> <span class="keywordflow">return</span> vreinterpretq_u64_u8(vld1q_u8(((uint8_t <span class="keyword">const</span> *)tmp) + (16 - imm)));</div>
<div class="line"><a id="l00164" name="l00164"></a><span class="lineno"> 164</span>}</div>
<div class="line"><a id="l00165" name="l00165"></a><span class="lineno"> 165</span> </div>
<div class="line"><a id="l00166" name="l00166"></a><span class="lineno"> 166</span><span class="preprocessor">#endif</span></div>
<div class="ttc" id="aaaruformat_8h_html"><div class="ttname"><a href="aaruformat_8h.html">aaruformat.h</a></div></div>
<div class="ttc" id="aarm__vmull_8h_html"><div class="ttname"><a href="arm__vmull_8h.html">arm_vmull.h</a></div></div>
</div><!-- fragment --></div><!-- contents -->
</div><!-- doc-content -->
</div><!-- container -->
<!-- start footer part -->
<div id="nav-path" class="navpath"><!-- id is needed for treeview function! -->
<ul>
<li class="navelem"><a href="dir_68267d1309a1af8e8297ef4c3efbcdba.html">src</a></li><li class="navelem"><a href="dir_b86b57f9d61b104fa8483aa475f863ae.html">crc64</a></li><li class="navelem"><a href="arm__vmull_8c.html">arm_vmull.c</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> 1.14.0 </li>
</ul>
</div>
</body>
</html>
Reference in New Issue View Git Blame Copy Permalink