Reimar/templeos-info
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
152be33191
templeos-info/public/Wb/Demo/MultiCore/MPRadix.HC.HTML

174 lines
13 KiB
Plaintext
Raw Normal View History

Use HTML for Holyc files as well
2024-03-24 21:24:44 +00:00
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="generator" content="TempleOS V5.03">
<meta name="viewport" content="width=device-width">
<link rel="stylesheet" href="/style/templeos.css">
<script src="/script/templeos.js"></script>
<style type="text/css">
.cF0{color:#000000;background-color:#ffffff;}
.cF1{color:#0000aa;background-color:#ffffff;}
.cF2{color:#00aa00;background-color:#ffffff;}
.cF3{color:#00aaaa;background-color:#ffffff;}
.cF4{color:#aa0000;background-color:#ffffff;}
.cF5{color:#aa00aa;background-color:#ffffff;}
.cF6{color:#aa5500;background-color:#ffffff;}
.cF7{color:#aaaaaa;background-color:#ffffff;}
.cF8{color:#555555;background-color:#ffffff;}
.cF9{color:#5555ff;background-color:#ffffff;}
.cFA{color:#55ff55;background-color:#ffffff;}
.cFB{color:#55ffff;background-color:#ffffff;}
.cFC{color:#ff5555;background-color:#ffffff;}
.cFD{color:#ff55ff;background-color:#ffffff;}
.cFE{color:#ffff55;background-color:#ffffff;}
.cFF{color:#ffffff;background-color:#ffffff;}
</style>
</head>
<body>
<pre id="content">
<a name="l1"></a><span class=cF2>/*</span><span class=cF0>
<a name="l2"></a></span><span class=cF2>On an 8-core machine, this takes the top 3-bits</span><span class=cF0>
<a name="l3"></a></span><span class=cF2>of random numbers and distributes them to the 8 cores</span><span class=cF0>
<a name="l4"></a></span><span class=cF2>for sorting. Then, it merge sorts them.</span><span class=cF0>
<a name="l5"></a></span><span class=cF2>*/</span><span class=cF0>
<a name="l6"></a>
<a name="l7"></a>#</span><span class=cF1>define</span><span class=cF0> NUM 1000000
<a name="l8"></a>
<a name="l9"></a></span><span class=cF9>I64</span><span class=cF0> my_mp_cnt=1&lt;&lt;</span><span class=cF5>Bsr</span><span class=cF0>(</span><span class=cFB>mp_cnt</span><span class=cF0>);</span><span class=cF2>//Power of 2</span><span class=cF0>
<a name="l10"></a>
<a name="l11"></a></span><span class=cF9>I32</span><span class=cF0> *arg1,*arg2;
<a name="l12"></a></span><span class=cF9>I32</span><span class=cF0> *b[my_mp_cnt],bn[my_mp_cnt];
<a name="l13"></a></span><span class=cF9>I64</span><span class=cF0> mp_not_done_flags;
<a name="l14"></a>
<a name="l15"></a></span><span class=cF9>I64</span><span class=cF0> Compare(</span><span class=cF9>I32</span><span class=cF0> *e1,</span><span class=cF9>I32</span><span class=cF0> *e2)
<a name="l16"></a>{
<a name="l17"></a> </span><span class=cF1>return</span><span class=cF0> *e1-*e2;
<a name="l18"></a>}
<a name="l19"></a>
<a name="l20"></a></span><span class=cF1>U0</span><span class=cF0> QSortU32(</span><span class=cF9>I32</span><span class=cF0> *base,</span><span class=cF9>I64</span><span class=cF0> num)
<a name="l21"></a>{</span><span class=cF2>//By customizing, we dramatically improve it!</span><span class=cF0>
<a name="l22"></a></span><span class=cF2>//Cut and paste from </span><a href="/Wb/Kernel/QSort.HC.HTML#l1"><span class=cF4>QSortI64</span></a><span class=cF2>().</span><span class=cF0>
<a name="l23"></a> </span><span class=cF9>I64</span><span class=cF0> i;
<a name="l24"></a> </span><span class=cF9>I32</span><span class=cF0> *less,*greater,pivot;
<a name="l25"></a> </span><span class=cF1>if</span><span class=cF0> (num&gt;1) </span><span class=cF7>{</span><span class=cF0>
<a name="l26"></a> </span><span class=cF1>do</span><span class=cF0> {
<a name="l27"></a> less=base;
<a name="l28"></a> greater=base+num;
<a name="l29"></a> pivot=base[num/2];
<a name="l30"></a> </span><span class=cF1>while</span><span class=cF0> (less&lt;greater) </span><span class=cF7>{</span><span class=cF0>
<a name="l31"></a> </span><span class=cF1>if</span><span class=cF0> (*less&lt;=pivot)
<a name="l32"></a> less++;
<a name="l33"></a> </span><span class=cF1>else</span><span class=cF0> {
<a name="l34"></a> greater--;
<a name="l35"></a> </span><span class=cF5>SwapU32</span><span class=cF0>(less,greater);
<a name="l36"></a> }
<a name="l37"></a> </span><span class=cF7>}</span><span class=cF0>
<a name="l38"></a> i=less-base;
<a name="l39"></a> </span><span class=cF1>if</span><span class=cF0> (i==num) </span><span class=cF7>{</span><span class=cF2>//All less or equ to pivot</span><span class=cF0>
<a name="l40"></a>
<a name="l41"></a> </span><span class=cF2>//Point greater to first less</span><span class=cF0>
<a name="l42"></a> </span><span class=cF1>do</span><span class=cF0> greater--;
<a name="l43"></a> </span><span class=cF1>while</span><span class=cF0> (--i &amp;&amp; *greater==pivot);
<a name="l44"></a>
<a name="l45"></a> </span><span class=cF1>if</span><span class=cF0> (i) {
<a name="l46"></a> less=base+num/2; </span><span class=cF2>//Pivot was not moved, point to it</span><span class=cF0>
<a name="l47"></a> </span><span class=cF1>if</span><span class=cF0> (less&lt;greater)
<a name="l48"></a> </span><span class=cF5>SwapU32</span><span class=cF0>(less,greater);
<a name="l49"></a> num=i;
<a name="l50"></a> } </span><span class=cF1>else</span><span class=cF0> </span><span class=cF2>//All equ</span><span class=cF0>
<a name="l51"></a> </span><span class=cF1>break</span><span class=cF0>;
<a name="l52"></a> </span><span class=cF7>}</span><span class=cF0> </span><span class=cF1>else</span><span class=cF0> </span><span class=cF1>if</span><span class=cF0> (i&lt;num/2) </span><span class=cF7>{</span><span class=cF0>
<a name="l53"></a> QSortU32(base,i);
<a name="l54"></a> num-=i;
<a name="l55"></a> base=greater;
<a name="l56"></a> </span><span class=cF7>}</span><span class=cF0> </span><span class=cF1>else</span><span class=cF0> </span><span class=cF7>{</span><span class=cF0>
<a name="l57"></a> QSortU32(greater,num-i);
<a name="l58"></a> num=i;
<a name="l59"></a> </span><span class=cF7>}</span><span class=cF0>
<a name="l60"></a> } </span><span class=cF1>while</span><span class=cF0> (num&gt;1);
<a name="l61"></a> </span><span class=cF7>}</span><span class=cF0>
<a name="l62"></a>}
<a name="l63"></a>
<a name="l64"></a></span><span class=cF1>U0</span><span class=cF0> MPSort(</span><span class=cF9>I64</span><span class=cF0> dummy=0)
<a name="l65"></a>{
<a name="l66"></a> </span><span class=cF1>no_warn</span><span class=cF0> dummy;
<a name="l67"></a> QSortU32(b[</span><span class=cF5>Gs</span><span class=cF0>-&gt;num],bn[</span><span class=cF5>Gs</span><span class=cF0>-&gt;num]);
<a name="l68"></a> </span><span class=cF5>LBtr</span><span class=cF0>(&amp;mp_not_done_flags,</span><span class=cF5>Gs</span><span class=cF0>-&gt;num);
<a name="l69"></a>}
<a name="l70"></a>
<a name="l71"></a></span><span class=cF1>U0</span><span class=cF0> MPRadixSortDemo(</span><span class=cF9>I64</span><span class=cF0> dummy=0)
<a name="l72"></a>{
<a name="l73"></a> </span><span class=cF1>no_warn</span><span class=cF0> dummy;
<a name="l74"></a> </span><span class=cF9>I64</span><span class=cF0> i,j,k1,k2;
<a name="l75"></a> </span><span class=cF1>F64</span><span class=cF0> t0;
<a name="l76"></a> arg1=</span><span class=cF5>MAlloc</span><span class=cF0>(NUM*</span><span class=cF1>sizeof</span><span class=cF7>(</span><span class=cF9>I32</span><span class=cF7>)</span><span class=cF0>);
<a name="l77"></a> </span><span class=cF1>for</span><span class=cF0> (i=0;i&lt;NUM;i++)
<a name="l78"></a> arg1[i]=</span><span class=cF5>RandI32</span><span class=cF0>;
<a name="l79"></a>
<a name="l80"></a> arg2=</span><span class=cF5>MAlloc</span><span class=cF0>(NUM*</span><span class=cF1>sizeof</span><span class=cF7>(</span><span class=cF9>I32</span><span class=cF7>)</span><span class=cF0>);
<a name="l81"></a>
<a name="l82"></a> </span><span class=cF6>&quot;$GREEN$QSort$FG$\n&quot;</span><span class=cF0>;
<a name="l83"></a> t0=</span><span class=cF5>tS</span><span class=cF0>;
<a name="l84"></a> </span><span class=cF5>MemCpy</span><span class=cF0>(arg2,arg1,</span><span class=cF1>sizeof</span><span class=cF7>(</span><span class=cF9>I32</span><span class=cF7>)</span><span class=cF0>*NUM);
<a name="l85"></a> </span><span class=cF5>QSort</span><span class=cF0>(arg2,NUM,</span><span class=cF1>sizeof</span><span class=cF7>(</span><span class=cF9>I32</span><span class=cF7>)</span><span class=cF0>,&amp;Compare);
<a name="l86"></a> </span><span class=cF6>&quot;Time:%9.6f\n&quot;</span><span class=cF0>,</span><span class=cF5>tS</span><span class=cF0>-t0;
<a name="l87"></a> </span><span class=cF5>D</span><span class=cF0>(arg2+NUM/4);
<a name="l88"></a>
<a name="l89"></a> </span><span class=cF6>&quot;$GREEN$QSortU32$FG$\n&quot;</span><span class=cF0>;
<a name="l90"></a> t0=</span><span class=cF5>tS</span><span class=cF0>;
<a name="l91"></a> </span><span class=cF5>MemCpy</span><span class=cF0>(arg2,arg1,</span><span class=cF1>sizeof</span><span class=cF7>(</span><span class=cF9>I32</span><span class=cF7>)</span><span class=cF0>*NUM);
<a name="l92"></a> QSortU32(arg2,NUM);
<a name="l93"></a> </span><span class=cF6>&quot;Time:%9.6f\n&quot;</span><span class=cF0>,</span><span class=cF5>tS</span><span class=cF0>-t0;
<a name="l94"></a> </span><span class=cF5>D</span><span class=cF0>(arg2+NUM/4);
<a name="l95"></a>
<a name="l96"></a> </span><span class=cF1>for</span><span class=cF0> (i=0;i&lt;my_mp_cnt;i++) </span><span class=cF7>{</span><span class=cF0>
<a name="l97"></a></span><span class=cF2>//We must do full size, just in case.</span><span class=cF0>
<a name="l98"></a> </span><span class=cF2>//There will be uneven split between cores</span><span class=cF0>
<a name="l99"></a> </span><span class=cF2>//depending on the distribution of rand numbers.</span><span class=cF0>
<a name="l100"></a> b[i]=</span><span class=cF5>MAlloc</span><span class=cF0>(NUM*</span><span class=cF1>sizeof</span><span class=cF7>(</span><span class=cF9>I32</span><span class=cF7>)</span><span class=cF0>);
<a name="l101"></a> bn[i]=0;
<a name="l102"></a> </span><span class=cF7>}</span><span class=cF0>
<a name="l103"></a>
<a name="l104"></a> </span><span class=cF1>if</span><span class=cF0> (my_mp_cnt&lt;2) </span><span class=cF5>throw</span><span class=cF0>(</span><span class=cF6>'MultCore'</span><span class=cF0>);
<a name="l105"></a>
<a name="l106"></a> </span><span class=cF6>&quot;$GREEN$MP Radix QSortU32$FG$\n&quot;</span><span class=cF0>;
<a name="l107"></a> t0=</span><span class=cF5>tS</span><span class=cF0>;
<a name="l108"></a> k1=32-</span><span class=cF5>Bsr</span><span class=cF0>(my_mp_cnt);
<a name="l109"></a> k2=my_mp_cnt/2;
<a name="l110"></a> </span><span class=cF1>for</span><span class=cF0> (i=0;i&lt;NUM;i++) </span><span class=cF7>{</span><span class=cF0>
<a name="l111"></a> j=arg1[i]&gt;&gt;k1+k2; </span><span class=cF2>//This is a preliminary radix sort.</span><span class=cF0>
<a name="l112"></a> b[j][bn[j]++]=arg1[i];
<a name="l113"></a> </span><span class=cF7>}</span><span class=cF0>
<a name="l114"></a> mp_not_done_flags=1&lt;&lt;my_mp_cnt-1;
<a name="l115"></a> </span><span class=cF1>for</span><span class=cF0> (i=0;i&lt;my_mp_cnt;i++)
<a name="l116"></a> </span><span class=cF5>Spawn</span><span class=cF0>(&amp;MPSort,</span><span class=cF3>NULL</span><span class=cF0>,</span><span class=cF3>NULL</span><span class=cF0>,i);
<a name="l117"></a> </span><span class=cF1>while</span><span class=cF0> (mp_not_done_flags)
<a name="l118"></a> </span><span class=cF5>Yield</span><span class=cF0>;
<a name="l119"></a> j=0;
<a name="l120"></a> </span><span class=cF1>for</span><span class=cF0> (i=0;i&lt;my_mp_cnt;i++) </span><span class=cF7>{</span><span class=cF0>
<a name="l121"></a> </span><span class=cF5>MemCpy</span><span class=cF0>(&amp;arg2[j],b[i],bn[i]*</span><span class=cF1>sizeof</span><span class=cF7>(</span><span class=cF9>I32</span><span class=cF7>)</span><span class=cF0>);
<a name="l122"></a> j+=bn[i];
<a name="l123"></a> </span><span class=cF7>}</span><span class=cF0>
<a name="l124"></a> </span><span class=cF6>&quot;Time:%9.6f\n&quot;</span><span class=cF0>,</span><span class=cF5>tS</span><span class=cF0>-t0;
<a name="l125"></a> </span><span class=cF5>D</span><span class=cF0>(arg2+NUM/4);
<a name="l126"></a>
<a name="l127"></a> </span><span class=cF5>Free</span><span class=cF0>(arg1);
<a name="l128"></a> </span><span class=cF5>Free</span><span class=cF0>(arg2);
<a name="l129"></a> </span><span class=cF1>for</span><span class=cF0> (i=0;i&lt;my_mp_cnt;i++)
<a name="l130"></a> </span><span class=cF5>Free</span><span class=cF0>(b[i]);
<a name="l131"></a>}
<a name="l132"></a>
<a name="l133"></a>MPRadixSortDemo;
<a name="l134"></a>
<a name="l135"></a></span><span class=cF2>/*</span><span class=cF0> Results on 8 Cores 3.397GHz Core i7:
<a name="l136"></a></span><span class=cF2>QSort</span><span class=cF0>
<a name="l137"></a>Time: 0.759998
<a name="l138"></a></span><span class=cF2>QSortU32</span><span class=cF0>
<a name="l139"></a>Time: 0.093684
<a name="l140"></a></span><span class=cF2>MP Radix QSortU32</span><span class=cF0>
<a name="l141"></a>Time: 0.045450
<a name="l142"></a></span><span class=cF2>*/</span><span class=cF0>
</span></pre></body>
</html>
Reference in New Issue Copy Permalink