Integral Accounting Enterprise Help System

<body leftmargin="0" topmargin="0" rightmargin="0" bottommargin="0" marginwidth="0" marginheight="0" class="body_workspace"> <html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:st1="urn:schemas-microsoft-com:office:smarttags" xmlns="http://www.w3.org/TR/REC-html40"> <head> <font color="#3B64A2" face="Tahoma" size="1"> <meta http-equiv=Content-Type content="text/html; charset=windows-1252"> <meta name=ProgId content=Word.Document> <meta name=Generator content="Microsoft Word 11"> <meta name=Originator content="Microsoft Word 11"> <link rel=File-List href="homepage_files/filelist.xml"> <link rel=Edit-Time-Data href="overview_files/editdata.mso">  <title>Overview</title> <o:SmartTagType namespaceuri="urn:schemas-microsoft-com:office:smarttags" name="PlaceType"/> <o:SmartTagType namespaceuri="urn:schemas-microsoft-com:office:smarttags" name="PlaceName"/> <o:SmartTagType namespaceuri="urn:schemas-microsoft-com:office:smarttags" name="place" downloadurl="http://www.5iantlavalamp.com/"/> </font> <font color="#3B64A2" face="Tahoma" size="1"> <style>  </style>  </head> </font> <body lang=EN-US link=navy vlink=purple> <div class=Section1> <p class=MsoNormal align=center style='text-align:center'><b> <span style='font-family:Tahoma'><font color="#3B64A2" size="4">Overview of Supercomputer <span class=SpellE>Architectures</span>, Hardware and Programming Methods</font><o:p></o:p></span></b></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b> <span style='font-family:Tahoma'><br> </span></b></font> <font color="#3B64A2" face="Tahoma" size="1">The basic concept behind HPC (High Performance Computing) is parallelism. HPC hardware (and software) falls into 3 categories -</font><b><o:p></o:p></b></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <ol style='margin-top:0in' start=1 type=1> <li class=MsoNormal style='mso-list:l0 level1 lfo1;tab-stops:.5in'><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">SMP (Symmetric Multi-Processors)</font><o:p></o:p></span></b></li> <li class=MsoNormal style='mso-list:l0 level1 lfo1;tab-stops:.5in'><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Vector processors </font> <o:p></o:p></span></b></li> <li class=MsoNormal style='mso-list:l0 level1 lfo1;tab-stops:.5in'><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Clustering</font><o:p></o:p></span></b></li> </ol> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'>SMP</span></b></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> is a multiprocessor computer architecture where to or more identical processors are connected with a single uniformly shared memory. Most common commercial systems use SMP for parallel processing.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Almost, all the modern operating systems are SMP capable <span class=SpellE>i.e</span> they can simultaneously execute code on multiple processors to achieve parallelism. Recent <span class=GramE>O(</span>1)<span style='mso-spacerun:yes'> </span>( Read Order 1 or constant order) scheduling capability of Linux<span style='mso-spacerun:yes'> </span>kernel has put it in lime-light.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The scalability matrix of various systems is -</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <ol style='margin-top:0in' start=1 type=1> <li class=MsoNormal style='mso-list:l7 level1 lfo8;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Windows 2000/NT – 2 CPUs (Although they claim 4 CPUs)</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l7 level1 lfo8;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Windows 2003 R2 server – 4/8 CPUs <span class=GramE>( performance</span> doesn't increase after 4 CPUs).</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l7 level1 lfo8;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Linux – 32 CPU <span class=GramE>( commercial</span> servers support 16 CPU servers only).</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l7 level1 lfo8;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">AIX – 64 CPUs</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l7 level1 lfo8;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Solaris – 128 CPUs</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l7 level1 lfo8;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">HP-UX – 256 CPUs <span class=GramE>( Superdome</span> Servers which even have 2TB = 2000Gb Ram !!!)</font><o:p></o:p></span></li> </ol> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Currently IBM <span class=SpellE>ans</span> Sun <span class=GramE>( Sun</span> Fire X series) offer <span class=SpellE>linux</span> SMP systems with 16 64bit AMD <span class=SpellE>Opteron</span> or Intel Zeon/Itanium CPUs. But there are currently some issues with <span class=SpellE>linux</span> scalability on SMP systems mainly attributed to its kernel locking mechanism.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">HP Boasts of its offerings in HP Superdome Servers which has <span class=SpellE>upto</span> 256 64Bit Intel Itanium or HP PA-RISC processors and 2TB or RAM. They only run HP-UX which is again licensed on per CPU bases.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The only discouraging factor against SMP is the high cost of owner ship and maintainability. <span class=GramE>Because even if one CPU or even its cache goes out of order entire system has to be replaced.</span></font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">On the contrary the hot-swappable nodes in a <span class=SpellE>linux</span> compute cluster makes it more popular for HPC market. Also, the very thin Linux layer (if any) running these cluster nodes make all the CPU power available for the application, as compared to a SMP system where OS itself eats up significant resources. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Not <span class=GramE>very<span style='mso-spacerun:yes'> </span>surprisingly</span> a compressed HP UX kernel image is > 180MB and sometimes even 300MB which hogs Huge memory space when running in uncompressed form. A Linux kernel image usually is 1.5 MB and can be made as small as 200 KB.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The product information about SMP systems can be achieved from -</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <ol style='margin-top:0in' start=1 type=1> <li class=MsoNormal style='mso-list:l8 level1 lfo9;tab-stops:.5in'> <span style='font-family:Tahoma'><a href="http://www.sun.com/software/linux/"><font color="#3B64A2" size="1">http://www.sun.com/software/linux/</font></a><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l8 level1 lfo9;tab-stops:.5in'> <span style='font-family:Tahoma'><a href="http://www.ibm.com/servers/eserver/linux/home.html?c=serversintro&n=Linux2001&t=ad"> <font color="#3B64A2" size="1">http://www.ibm.com/servers/eserver/linux/home.html?c=serversintro&n=Linux2001&t=ad</font></a><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l8 level1 lfo9;tab-stops:.5in'> <span style='font-family:Tahoma'><a href="http://www.alinuxservers.com/"><font color="#3B64A2" size="1">http://www.alinuxservers.com/</font></a><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l8 level1 lfo9;tab-stops:.5in'> <span style='font-family:Tahoma'><a href="http://www.ibm.com/linux/"><font color="#3B64A2" size="1">http://www.ibm.com/linux/</font></a><o:p></o:p></span></li> </ol> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span></font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'>Vector processors</span></b></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> which were invented in early 1980s are still a subject of research and still far-away from meeting any enterprise requirements.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'><o:p> </o:p></span></b></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'>Clusters</span></b></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> are groups of loosely bound independent servers (called nodes) that work together closely so that in many aspects they can be viewed as though they are a single large server. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Before cluster-based computing, the typical supercomputer was a vector processor that could typically cost over a million dollars due to the specialized hardware and software. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Fundamentally, both vector and scalar processors execute instructions based on a clock; what sets them apart is the ability of vector processors to parallelize computations involving vectors (such as matrix multiplication), which are commonly used in High Performance Computing. To illustrate this, suppose you have two double arrays <i style='mso-bidi-font-style:normal'>a</i> and <i style='mso-bidi-font-style: normal'>b</i> and want to create a third array <i style='mso-bidi-font-style: normal'>x</i> such that <i style='mso-bidi-font-style:normal'>x[<span class=SpellE>i</span><span class=GramE>]=</span>a[<span class=SpellE>i</span>]+b[<span class=SpellE>i</span>]</i>. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Any floating-point operation like addition or multiplication is achieved in a few discrete steps: </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:35.35pt;text-indent:-14.15pt;mso-list: l2 level1 lfo3;tab-stops:35.35pt'><font color="#3B64A2"><![if !supportLists]> <font face="Tahoma" size="1"> <span style='mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'>·</span></font><span style='font-style:normal; font-variant:normal; font-weight:normal; font-family:Tahoma; mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'><font size="1">       </font> </span><![endif]></font><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Exponents are adjusted </font> <o:p></o:p></span></p> <p class=MsoNormal style='margin-left:35.35pt;text-indent:-14.15pt;mso-list: l2 level1 lfo3;tab-stops:35.35pt'><font color="#3B64A2"><![if !supportLists]> <font face="Tahoma" size="1"> <span style='mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'>·</span></font><span style='font-style:normal; font-variant:normal; font-weight:normal; font-family:Tahoma; mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'><font size="1">       </font> </span><![endif]><span class=SpellE> <span style='font-family:Tahoma'><font size="1">Significants</font></span></span></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> are added </font> <o:p></o:p></span></p> <p class=MsoNormal style='margin-left:35.35pt;text-indent:-14.15pt;mso-list: l2 level1 lfo3;tab-stops:35.35pt'><font color="#3B64A2"><![if !supportLists]> <font face="Tahoma" size="1"> <span style='mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'>·</span></font><span style='font-style:normal; font-variant:normal; font-weight:normal; font-family:Tahoma; mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'><font size="1">       </font> </span><![endif]></font><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The result is checked for rounding errors and so on </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">A vector processor parallelizes these internal steps by using a <i style='mso-bidi-font-style: normal'>pipelining</i> technique. Suppose there are six steps (as in IEEE arithmetic hardware) in a floating-point addition as shown in Figure 1. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" face="Tahoma" size="1"><![if !vml]><span style='mso-ignore:vglayout;position: absolute;z-index:2;left:19px;top:801px;width:254px;height:275px'><img width=254 height=275 src="homepage_files/image002.gif" align=left v:shapes="_x0000_s1029"></span><![endif]><br style='mso-ignore:vglayout' clear=ALL> </font> <span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal> </p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">A vector processor does these six steps in parallel. As you can see, for a six-step floating-point addition, the speedup factor will be very close to six times. A big benefit is that parallelization is happening behind the scenes, and you need not explicitly code this in your programs. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Compared to vector processing, cluster-based computing takes a fundamentally different approach. Instead of using specially optimized vector hardware, it uses standard scalar processors, but uses the <i>software layer</i> to break the given task in parallel independent sub-tasks and compute then on different nodes.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Some features of clusters are as follows:</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:35.35pt;text-indent:-14.15pt;mso-list: l3 level1 lfo4;tab-stops:35.35pt'><font color="#3B64A2"><![if !supportLists]> <font face="Tahoma" size="1"> <span style='mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'>·</span></font><span style='font-style:normal; font-variant:normal; font-weight:normal; font-family:Tahoma; mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'><font size="1">       </font> </span><![endif]></font><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Clusters are built using commodity hardware and cost a fraction of the vector processors. In many cases, the price is lower by more than an order of magnitude. </font> <o:p></o:p></span></p> <p class=MsoNormal style='margin-left:35.35pt;text-indent:-14.15pt;mso-list: l3 level1 lfo4;tab-stops:35.35pt'><font color="#3B64A2"><![if !supportLists]> <font face="Tahoma" size="1"> <span style='mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'>·</span></font><span style='font-style:normal; font-variant:normal; font-weight:normal; font-family:Tahoma; mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'><font size="1">       </font> </span><![endif]></font><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Clusters use a message-passing paradigm for communication, and programs have to be explicitly coded to make use of distributed hardware. </font> <o:p></o:p></span></p> <p class=MsoNormal style='margin-left:35.35pt;text-indent:-14.15pt;mso-list: l3 level1 lfo4;tab-stops:35.35pt'><font color="#3B64A2"><![if !supportLists]> <font face="Tahoma" size="1"> <span style='mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'>·</span></font><span style='font-style:normal; font-variant:normal; font-weight:normal; font-family:Tahoma; mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'><font size="1">       </font> </span><![endif]></font><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">With clusters, you can add more nodes to the cluster based on need. </font> <o:p></o:p></span></p> <p class=MsoNormal style='margin-left:35.35pt;text-indent:-14.15pt;mso-list: l3 level1 lfo4;tab-stops:35.35pt'><font color="#3B64A2"><![if !supportLists]> <font face="Tahoma" size="1"> <span style='mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'>·</span></font><span style='font-style:normal; font-variant:normal; font-weight:normal; font-family:Tahoma; mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'><font size="1">       </font> </span><![endif]></font><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Open source software components and Linux lead to lower software costs. </font> <o:p></o:p></span></p> <p class=MsoNormal style='margin-left:35.35pt;text-indent:-14.15pt;mso-list: l3 level1 lfo4;tab-stops:35.35pt'><font color="#3B64A2"><![if !supportLists]> <font face="Tahoma" size="1"> <span style='mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'>·</span></font><span style='font-style:normal; font-variant:normal; font-weight:normal; font-family:Tahoma; mso-list:Ignore; mso-fareast-font-family:Symbol; mso-bidi-font-family:Symbol'><font size="1">       </font> </span><![endif]></font><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Clusters have a much lower maintenance cost (they take up less space, take less power, and need less cooling). </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Clusters are usually deployed to improve speed and/or reliability over that provided by a single computer. All the nodes inside a cluster are almost always connected by a fast low-latency network, which restricts the cluster nodes to same LAN (local area network).</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">With the introduction of technologies like <span class=SpellE>Infiniband</span>, PVFS <span class=GramE>( parallel</span> virtual file-system), gigabit Ethernet etc. clustering has emerged as the most promising form of supercomputing. You can build powerful clusters with a very small budget and keep adding extra nodes based on need. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Clusters reveal exceptional potential and are recognized as the possible future leader in the arena of high-performance computing</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The first cluster based supercomputer made an entry to top500 supercomputer list in 1997 and figure 1 shows the current share of supercomputing cluster in the top500 supercomputers list [1].</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal align=center style='text-align:center'> <span style='font-family:Tahoma;color:#3B64A2'><font size="1"><![if !vml]><img border=0 width=439 height=319 src="homepage_files/image004.jpg" v:shapes="_x0000_s1031"><![endif]></font></span><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">These clusters have become very popular over the last few years because of the dramatic performance improvements in PC processors, to the point where they are now comparable to high-end UNIX workstations for scientific computations. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Clusters can further be categorized into -</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <ol style='margin-top:0in' start=1 type=1> <li class=MsoNormal style='mso-list:l1 level1 lfo2;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">High Availability (HA) cluster</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l1 level1 lfo2;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Load Balancing cluster</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l1 level1 lfo2;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">High Performance Computing (HPC) cluster</font><o:p></o:p></span></li> </ol> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">This site will primarily be devoted to providing information on the HPC (High Performance Computing) variety of cluster supercomputing. From now on, we will restrict this report to HPC clusters and the term “cl<i>uster</i>” would refer specifically to the HPC cluster.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b> <span style='font-family:Tahoma'><o:p> </o:p></span></b></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">As shown in the diagram below, typically a cluster comprises of – </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <ol style='margin-top:0in' start=1 type=1> <li class=MsoNormal style='mso-list:l4 level1 lfo5;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">head node - which controls the clustering and computing operations</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l4 level1 lfo5;tab-stops:.5in'> <font color="#3B64A2" size="1"><span class=GramE><span style='font-family:Tahoma'>compute</span></span></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> node – which takes in independent compute tasks from the head node and deliver them back.</font><o:p></o:p></span></li> </ol> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" face="Tahoma" size="1"><![if !vml]><span style='mso-ignore:vglayout;position: absolute;z-index:1;left:0px;top:13px;width:208px;height:204px'><img width=208 height=204 src="homepage_files/image006.gif" align=left v:shapes="_x0000_s1026"></span><![endif]></font><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span></font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='margin-left:220.35pt'><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Since the tasks handled by compute nodes are independent these nodes can be added or removed from a cluster even while it’s up and working. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Based on the way these clusters break down a massive tasks, they can be divided into two broad categories - </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <ol style='margin-top:0in' start=1 type=1> <li class=MsoNormal style='mso-list:l5 level1 lfo6;tab-stops:.5in'> <font color="#3B64A2" size="1"><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'>Master-Slave clusters</span></b></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1">.</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l5 level1 lfo6;tab-stops:.5in'> <font color="#3B64A2" size="1"><b> <span style='font-family:Tahoma'>SSI</span></b></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> – Single System Image clusters.</font><o:p></o:p></span></li> </ol> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'>Master-Slave</span></b></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> clusters run specially written applications which by design take advantage of the underlying parallel architecture. Applications run on the Master which distributes parallel/independent the work to the slaves. Such applications make use of parallel compilers which breaks the application code into independent parallel instructions capable of running on a MPI (Message parsing interface) layer. Programs have to be specifically written to take full advantage of this particular architecture.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'>SSI</span></b></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> clusters create a <b style='mso-bidi-font-weight:normal'><u>s</u></b>ingle (operating) <b style='mso-bidi-font-weight:normal'><u>s</u></b>ystem <b style='mso-bidi-font-weight: normal'><i style='mso-bidi-font-style:normal'>i</i></b>mage from the collection of independent nodes. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Here the SSI architecture and modified kernel take care of the implementation of parallelism, rather than the application. All the applications such as database and web-server can run un-modified, but can take full advantage of the cluster only if they are multi-threaded, so that they can be scheduled on different CPU simultaneously. This system has roots from NUMA based MMP or SMP systems.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma;color:#3B64A2'> <font size="1"><![if !vml]><img border=0 width=588 height=240 src="homepage_files/image008.gif" v:shapes="_x0000_s1030"><![endif]></font></span><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Both these approaches have advantages and disadvantages and used by entirely different audience. We shall see this in details now.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">We have specific sections of this site covering the major distributions of Master Slave Clusters and SSI Clusters including a review of the distributions and information on their installations.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><b><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Clustering Hardware Requirements</font><o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">By definition, a cluster must consist of at least two nodes, a master and a slave. It is not necessary that these machines have the same levels of performance. The only requirement is that they both share the same architecture. For instance, the cluster should only consist of all Intel machines or all Apple machines but not a mixture of the two. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Strictly speaking, the only hardware requirement when building a cluster is two computers and some type of networking hardware to connect them with. To maximize the benefits of a cluster, the right hardware must be used. <span class=docemphasis>If at all possible, use identical systems for your nodes</span>. Life will be much simpler. You'll need to develop and test only one configuration. This is due to the fact that one node which takes longer to do its work can slow the entire cluster down as the rest of the nodes must stop what they are doing and wait for the slow node<span style='mso-spacerun:yes'> </span>to catch up. This is not always the case, but it is a consideration that must be made. Having identical hardware specs also simplifies the setup process a great deal as it will allow each hard drive to be imaged from a master instead of configuring each node individually.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">In constructing a cluster, you can scrounge for existing computers, buy assembled computers, or buy the parts and assemble your own. Scrounging is the cheapest way to go, but this approach is often the most time consuming. Usually, using scrounged systems means you'll end up with a wide variety of hardware, which creates both hardware and software problems. With older scrounged systems, you are also more likely to have even more hardware problems. If this is your only option, try to standardize hardware as much as possible.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Buying new, preassembled computers may be the simplest approach if money isn't the primary concern. This is often the best approach for mission-critical applications or when time is a critical factor. Buying new is also the safest way to go if you are uncomfortable assembling computers. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Building your own system is cheaper, provides higher performance and reliability, and allows for customization. Assembling your own computers may seem daunting, but it isn't that difficult. You'll need time, personnel, space, and a few tools. It's a good idea to build a single system and test it for hardware and software compatibility before you commit to a large bulk order. Even if you do buy preassembled computers, you will still need to do some testing and maintenance.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">If you are constructing a <a name=highperlinuxc-CHP-3-ITERM-1373></a><a name=highperlinuxc-CHP-3-ITERM-1372></a>dedicated cluster, you will not need full systems. The more you can leave out of each computer, the more computers you will be able to afford, and the less you will need to maintain on individual computers. For example, with dedicated clusters you can do without monitors, keyboards, and mice for each individual compute node. With a minimal configuration, wiring is usually significantly easier, particularly if you use rack-mounted equipment. (However, heat dissipation can be a serious problem with rack-mounted systems.)</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='mso-outline-level:3'><font color="#3B64A2"> <font face="Tahoma" size="1"><a name="_Toc141024067"></a></font><a name=Step1><span style='mso-bookmark:_Toc141024067'><b style='mso-bidi-font-weight: normal'><span style='font-family:Tahoma'><font size="1">The First Step: Hardware Considerations</font></span></b></span></a></font><b style='mso-bidi-font-weight: normal'><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">To build a cluster, one must have access to computers on which to install the software. Therefore, it makes sense to cover this early in the process. For a dedicated cluster, you determine your needs and there may be a lot you won't need—audio cards and speakers, video capture cards, etc. Beyond these obvious expendables, there are other additional parts you might want to consider omitting such as disk drives, keyboards, mice, and displays.</font><o:p></o:p></span></p> <p class=MsoNormal style='mso-outline-level:3'><font color="#3B64A2" size="1"><a name="_Toc141024068"><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'><o:p> </o:p></span></b></a></font></p> <p class=MsoNormal style='mso-outline-level:3'><font color="#3B64A2" size="1"><span style='mso-bookmark:_Toc141024068'><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'>CPU & Motherboards</span></b></span></font><b style='mso-bidi-font-weight: normal'><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">A choice of CPU should be made from two families: Intel x86 [IA32] compatible (such as the Pentium4) or Compaq Alpha systems [Formerly DEC]. Alternate vendor CPUs are supported by Linux and cluster-able operating systems (such as the IBM Power PC chip and the SUN SPARC), however there is only limited support and accompanying software distributions available for these architectures. In general, clusters are mainly built with Intel or Alpha architectures. Intel based systems are considered as commodity systems because there are multiple sources (Intel, AMD, Cyrix) and obviously ubiquitous. Compaq Alpha on the other hand, is a clear performance winner, but does represent a single source part, and therefore can be hard to source at a good price.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma;color:#3B64A2'> <font size="1">Within the Intel based systems, the Pentium 3 and 4 [23] have the best floating point performance and they support SMP motherboards.</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma;color:#3B64A2'><o:p> <font size="1"> </font></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Selecting a processor is a balancing act. Your choice will be constrained by cost, performance, and compatibility. Remember, the rationale behind a commodity off-the-shelf (COTS) cluster is buying machines that have the most favorable price to performance ratio, not pricey individual machines. Typically you'll get the best ratio by purchasing a CPU that is a generation behind the current cutting edge. This means comparing the numbers. When comparing CPUs, you should look at the increase in performance versus the increase in the <span class=docemphasis>total cost</span> of a node.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Since Linux works with most major chip families, stay mainstream and you shouldn't have any software compatibility problems. Nonetheless, it is a good idea to test a system before committing to a bulk purchase. Since a primary rationale for building your own cluster is the economic advantage, you'll probably want to stay away from the less common chips. While clusters built with <span class=SpellE>UltraSPARC</span> systems may be wonderful performers, few people would describe these as commodity systems. So unless you just happen to have a number of these systems that you aren't otherwise using, you'll probably want to avoid them.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">When comparing motherboards, look to see what is integrated into the board. There are some significant differences. Serial, parallel, and USB ports along with EIDE disk adapters are fairly standard. You may also find motherboards with integrated FireWire ports, a network interface, or even a video interface. You may be able to save money with built-in network or display interfaces (provided they actually meet your needs), If you are really certain that some fully integrated motherboard meets your needs, eliminating the need for daughter cards may allow you to go with a small case.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">So a motherboard with built-in components i.e. AGP, LAN card (preferably with both 100 Mbps and 1 <span class=SpellE>Gbps</span> <span class=SpellE>NICs</span>) etc. is a feasible option because of the following reasons,</font><o:p></o:p></span></p> <ul style='margin-top:0in' type=disc> <li class=MsoNormal style='mso-pagination:widow-orphan;mso-list:l9 level1 lfo10; mso-hyphenate:auto;tab-stops:list .5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">It costs fairly low</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-pagination:widow-orphan;mso-list:l9 level1 lfo10; mso-hyphenate:auto;tab-stops:list .5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">It increases the density of nodes per rack</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-pagination:widow-orphan;mso-list:l9 level1 lfo10; mso-hyphenate:auto;tab-stops:list .5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Nodes can easily be added and removed from the rack.</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-pagination:widow-orphan;mso-list:l9 level1 lfo10; mso-hyphenate:auto;tab-stops:list .5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">No high graphics are required on client nodes therefore a normal onboard VGA would work.</font><o:p></o:p></span></li> </ul> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Example: Intel Entry level Server Board S875WP1-E</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The standard Pentium 4 <span class=GramE>processor(</span>without extended addressing) can only address 4 GB of virtual memory. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Hence Intel Pentium4 with EMT64 or AMD Athlon64 would be better if choosing a low price <span class=GramE>CPU,</span> they can address up-to 2TB of virtual memory.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">If choosing server class CPU Intel Xeon or AMD <span class=SpellE>Optron</span> would be a better choice.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='mso-outline-level:3'><font color="#3B64A2"> <font face="Tahoma" size="1"><a name=Memoryanddisks></a></font><a name="_Toc141024069"><span style='mso-bookmark:Memoryanddisks'><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'><font size="1">Memory and disks</font></span></b></span></a></font><span style='mso-bookmark:Memoryanddisks'><b style='mso-bidi-font-weight:normal'><span style='font-size:10.0pt;font-family: Arial'><o:p></o:p></span></b></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Memory per node is another deciding factor that can affect the through put of the cluster. Amongst the hardware resources on a cluster node, memory is the one that is comparatively cheaper and can be easily upgraded. Increasing any hardware resource on the cluster will definitely affect the performance of the cluster positively. Subject to your budget, the more cache and RAM in your system, the better. Typically, the faster the processor, the more RAM you will need. A very crude rule of thumb is one byte of RAM for every floating-point operation per second. So a processor capable of 100 <span class=SpellE>MFLOPs</span> would need around 100 MB of RAM.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Each node should at least have 2GB of memory to facilitate the process check-pointing techniques various clusters solutions use.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Almost all the clustering solutions allow each node to have either decentralized private storage or a common shared storage.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Decentralized storage – Each node can have any SATA or SCSI disk with speed > 5400. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">PVFS – Parallel virtual file-system would be a good choice, it creates a single file-system across different private disks.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Shared Storage – <span class=GramE>A</span> <span class=SpellE>iSCSI</span> setup would be ideal for shared storage. In this case the (centralized) shared storage would be connected using network to all the nodes. Each node would have a preferably gigabyte Ethernet NIC for communication with <span class=SpellE>iSCSI</span> drives. If using <span class=SpellE>iSCSI</span>, GFS (Global file system) from <span class=SpellE>Redhat</span> can be installed on stared storage, so that all the nodes can access the disks concurrently. GFS comes by default with RHEL installation.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">In both PVFS and GFS all the nodes would same namespace of the file-system. GFS is known to be more secure, but <span class=SpellE>iSCSI</span> setup costs much more.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">If you are buying hard disks, there are three issues: interface type (EIDE vs. SCSI), disk latency (a function of rotational speed), and disk capacity. From a price-performance perspective, EIDE is probably a better choice than SCSI since virtually all motherboards include a built-in EIDE interface. And unless you are willing to pay a premium, you won't have much choice with respect to disk latency. Almost all current drives rotate at 7,200 RPM. While a few 10,000 RPM drives are available, their performance, unlike their price, is typically not all that much higher. With respect to disk capacity, you'll need enough space for the operating system, local paging, and the data sets you will be manipulating. Unless you have extremely large data sets, when recycling older computers a 10 GB disk should be adequate for most uses, however 40 GB IDE disk is recommended mostly because of other non-cluster requirements.<br style='mso-special-character:line-break'> <![if !supportLineBreakNewLine]><br style='mso-special-character:line-break'> <![endif]></font><o:p></o:p></span></p> <p class=MsoNormal style='mso-outline-level:3'><font color="#3B64A2" size="1"><a name="_Toc141024070"><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'>Monitors, keyboards, and mice</span></b></a></font><b style='mso-bidi-font-weight: normal'><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">In many clusters, on the compute nodes, usually no keyboard, mice and keyboard are required. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Its advantage includes low cost, less equipment to maintain, and a smaller equipment footprint. However monitor, keyboard and mouse is required at head node. Also it is recommended that an extra set of these peripherals should be kept for troubleshooting any node.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='mso-outline-level:3'><font color="#3B64A2" size="1"><a name="_Toc141024071"><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'>Adapters, power supplies, and cases</span></b></a></font><b style='mso-bidi-font-weight: normal'><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Each node should include a video adapter. The network adapter is a key component. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Gigabyte Ethernet or <span class=SpellE>Infiniband</span> network would be preferred and each compute node should have at lease two <span class=SpellE><span class=GramE>NICs</span></span><span class=GramE><span style='mso-spacerun:yes'> </span>(</span>network interface cards) both configured for private networks. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The master node should have 3 <span class=SpellE>NICs</span>, two for private network and one for public connection which would facilitate remote administration.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The Linux kernel > 2.6.16 has support for both <span class=SpellE>Infiniband</span> and Gigabyte Ethernet <span class=SpellE>NICs</span>. RHEL 4 Update2 also gave out support of these.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">You must buy an adapter that is compatible with the cluster network. If you are planning to boot a diskless system over the network, you'll need an adapter that supports it. This translates into an adapter with an appropriate network BOOT ROM, i.e., one with <span class=docemphasis>pre-execution environment (PXE)</span> support. Many adapters come with a built-in (but empty) BOOT ROM socket so that the ROM can be added. You can purchase BOOT ROMs for these cards or burn your own. However, it may be cheaper to buy a new card with an installed BOOT ROM than to add the BOOT ROMs.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"><b> <span style='font-family:Tahoma'><o:p> </o:p></span></b></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">A cluster can be easily constructed from cheap commodity computers. The faster they <span class=GramE>are,</span> the better. The two most important requirements of a cluster are -</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <ol style='margin-top:0in' start=1 type=1> <li class=MsoNormal style='mso-list:l6 level1 lfo7;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">High speed and low latency network.</font><o:p></o:p></span></li> <li class=MsoNormal style='mso-list:l6 level1 lfo7;tab-stops:.5in'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">High amount of RAM > 2GB.</font><o:p></o:p></span></li> </ol> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Supercomputing Programming Overview</font><o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">If you will be writing your own applications for your cluster, this information should get you started. After the operating system and other basic system software and selected and installed, you need to select and install core software development tools including libraries that support parallel processing. If you are using the OSCAR or Rocks clustering toolkits, then these libraries and tools are already in your cluster.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='mso-outline-level:2'><font color="#3B64A2" size="1"><a name="_Toc141024100"><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'>Programming Languages</span></b></a></font><b style='mso-bidi-font-weight:normal'><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">While there are hundreds of programming languages available FORTRAN or C/C++ are still reasonable choices for writing code for high performance clusters.<span style='mso-spacerun:yes'> </span>FORTRAN has changed considerably over the years, so the term can mean different things to different people. While there are more recent versions of FORTRAN, your choice will likely be between FORTRAN 77 and FORTRAN 90. For a variety of reasons, FORTRAN 77 is likely to get the nod over FORTRAN 90 despite the greater functionality of FORTRAN 90. First, the GNU implementation of FORTRAN 77 is likely to already be on your machine. If it isn't, it is freely available and easily obtainable. If you really want FORTRAN 90, don't forget to budget for it. But you should also realize that you may face compatibility issues. <span class=GramE>When selecting parallel programming libraries to use with your compiler, your choices will be more limited with FORTRAN 90.</span></font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" face="Tahoma" size="1"><a name=highperlinuxc-CHP-9-ITERM-1810></a><a name=highperlinuxc-CHP-9-ITERM-1809></a></font> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">C and C++ are the obvious alternatives to FORTRAN. For new applications that don't depend on compatibility with legacy FORTRAN applications, C is probably the best choice. In general, you have greater compatibility with libraries. And at this point in time, you are likely to find more programmers trained in C than FORTRAN. So when you need help, you are more likely to find a helpful C than FORTRAN programmer.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal style='mso-outline-level:2'><font color="#3B64A2" size="1"><a name="_Toc141024101"><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'>Selecting a Library</span></b></a></font><b style='mso-bidi-font-weight:normal'><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Most common and widely used parallel programming libraries are –</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Parallel Virtual Machine (PVM)</font><o:p></o:p></span></b></p> <p class=MsoNormal><b style='mso-bidi-font-weight:normal'> <span style='font-family:Tahoma'><font color="#3B64A2" size="1">Message Passing Interface (MPI) library</font><o:p></o:p></span></b></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Development of PVM started in 1989 and continued into 90’s as a <span class=SpellE>joing</span> effort among Oak Ridge National Laboratory, the <st1:PlaceType w:st="on">University</st1:PlaceType> of <st1:PlaceName w:st="on">Tennessee</st1:PlaceName>, <st1:PlaceName w:st="on">Emory</st1:PlaceName> <st1:PlaceType w:st="on">University</st1:PlaceType>, and <st1:place w:st="on"><st1:PlaceName w:st="on">Carnegie-Mellon</st1:PlaceName> <st1:PlaceType w:st="on">University</st1:PlaceType></st1:place>. An implementation of PVM is available from </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'><a href="http://www.netlib.org/pvm3/" target="_blank"> <font color="#3B64A2" size="1">http://www.netlib.org/pvm3/</font></a><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">This PVM implementation provides both libraries and tools based on a message-passing model.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">MPI is a newer standard that seems to be generally preferred over PVM by many users.</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">MPI is an API for parallel programming based on a message-passing model for parallel computing. MPI processes execute in parallel. Each process has a separate address space. Sending processes specify data to be sent and a destination process. The receiving process specifies an area in memory for the message, the identity of the source, etc.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Primarily, MPI can be thought of as a standard that specifies a library. Users can write code in C, C++, or FORTRAN using a standard compiler and then link to the MPI library. The library implements a predefined set of function calls to send and receive messages among collaborating processes on the different machines in the cluster. You write your code using these functions and link the completed code to the library.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The MPI specification was developed by the MPI Forum, a collaborative effort with support from both academia and industry. It is suitable for both small clusters and "big-iron" implementations. It was designed with functionality, portability, and efficiency in mind. By providing a well-designed set of function calls, the library provides a wide range of functionality that can be implemented in an efficient manner. As a clearly defined standard, the library can be implemented on a variety of architectures, allowing code to move easily among machines.</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">MPI has gone through a couple of revisions since it was introduced in the early '90s. MPI-2 is the latest version available. While there are several different implementations of MPI, there are two that are widely used—LAM/MPI and MPICH. Both LAM/MPI and MPICH go beyond simply providing a library. Both include programming and runtime environments providing mechanisms to run programs across the cluster. Both are widely used, robust, well supported, and freely available. Excellent documentation is provided with both</font><o:p></o:p></span></p> <p class=MsoNormal style='mso-outline-level:2'><font color="#3B64A2" size="1"><a name="_Toc141024102"><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'><o:p> </o:p></span></b></a></font></p> <p class=MsoNormal style='mso-outline-level:2'><font color="#3B64A2" size="1"><span style='mso-bookmark:_Toc141024102'><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'>LAM/MPI</span></b></span></font><b style='mso-bidi-font-weight:normal'><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> </font> <o:p></o:p></span></b></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The <i>Local Area <span class=SpellE>Multicomputer</span>/Message Passing Interface (LAM/MPI)</i><a name=highperlinuxc-CHP-9-ITERM-1820></a><a name=highperlinuxc-CHP-9-ITERM-1819></a><a name=highperlinuxc-CHP-9-ITERM-1818></a> was originally developed by the <st1:place w:st="on"><st1:PlaceName w:st="on">Ohio</st1:PlaceName> <st1:PlaceName w:st="on">Supercomputing</st1:PlaceName> <st1:PlaceType w:st="on">Center</st1:PlaceType></st1:place>. It is now maintained by the Open Systems Laboratory at <st1:place w:st="on"><st1:PlaceName w:st="on">Indiana</st1:PlaceName> <st1:PlaceType w:st="on">University</st1:PlaceType></st1:place>. As previously noted, LAM/MPI (or LAM for short) is both an MPI library and an execution environment.<span style='mso-spacerun:yes'> </span>LAM was designed to include an extensible component framework known as <i>System Service Interface (SSI)</i><a name=highperlinuxc-CHP-9-ITERM-1822></a><a name=highperlinuxc-CHP-9-ITERM-1821></a>, one of its major strengths. It works well in a wide variety of environments and supports several methods of inter-process communications using TCP/IP. LAM will run on most UNIX machines (but not Windows). New releases are tested with both Red Hat and Mandrake Linux.</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Documentation can be downloaded from the LAM site, </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'><a href="http://www.lam-mpi.org/" target="_blank"><font size="1" color="#3B64A2">http://www.lam-mpi.org/</font></a><font size="1" color="#3B64A2">. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">There are also tutorials, a FAQ, and archived mailing lists.</font><o:p></o:p></span></p> <p class=MsoNormal style='mso-outline-level:2'><font color="#3B64A2" size="1"><a name="_Toc141024103"><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'><o:p> </o:p></span></b></a></font></p> <p class=MsoNormal style='mso-outline-level:2'><font color="#3B64A2" size="1"><span style='mso-bookmark:_Toc141024103'><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'>MPI</span></b></span></font><b style='mso-bidi-font-weight:normal'><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></b></p> <p class=MsoNormal><font color="#3B64A2"><i> <span style='font-family:Tahoma'><font size="1">Message Passing Interface Chameleon (MPICH)</font></span></i><font face="Tahoma" size="1"><a name=highperlinuxc-CHP-9-ITERM-1862></a><a name=highperlinuxc-CHP-9-ITERM-1861></a><a name=highperlinuxc-CHP-9-ITERM-1860></a></font></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> was developed by<a name=highperlinuxc-CHP-9-ITERM-1863></a> William <span class=SpellE>Gropp</span> and <a name=highperlinuxc-CHP-9-ITERM-1864></a>Ewing Lusk and is freely available from Argonne National Laboratory </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'><a href="http://www-unix.mcs.anl.gov/mpi/mpich/" target="_blank"> <font color="#3B64A2" size="1">http://www-unix.mcs.anl.gov/mpi/mpich/</font></a><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Like LAM, it is both a library and an execution environment. It runs on a wide variety of UNIX platforms and is even available for Windows NT.</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Documentation can be downloaded from the web site. There are separate manuals for each of the communication models.</font><o:p></o:p></span></p> <p class=MsoNormal style='mso-outline-level:2'><font color="#3B64A2" size="1"><a name="_Toc141024104"><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'><o:p> </o:p></span></b></a></font></p> <p class=MsoNormal style='mso-outline-level:2'><font color="#3B64A2" size="1"><span style='mso-bookmark:_Toc141024104'><b style='mso-bidi-font-weight:normal'> <span style='font-family: Tahoma'>Examples Using MPI</span></b></span></font><b style='mso-bidi-font-weight:normal'><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></b></p> <p class=MsoNormal><font color="#3B64A2"> <span style='font-family:Tahoma'><font size="1">While MPI does not offer some of the specialized features available in PVM, it is based on agreed-upon standards, is increasingly used for code development, and has adequate features for most parallel applications. Hence, the coding examples presented here are written in C using MPI. The codes have been tested on a Beowulf cluster using the Gnu C compiler (</font></span><span class=SpellE><code><span style='font-family:Tahoma'><font size="1">gcc</font></span></code></span></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1">) with the MPICH implementation of MP</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Program 1 is a "Hello World!" program that illustrates the basic MPI calls necessary to startup and end an MPI program.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <div align=center> <table class=MsoNormalTable border=1 cellpadding=0 width="90%" style='width:90.0%;mso-cellspacing:1.5pt'> <tr style='mso-yfti-irow:0;mso-yfti-firstrow:yes'> <td style='border:none;background:#BFBFBF;padding:.75pt .75pt .75pt .75pt'> <p class=MsoNormal><font color="#3B64A2"> <font face="Tahoma" size="1"><a name=Program1></a></font><b> <span style='font-family:Tahoma'><font size="1">Program 1: </font> </span></b></font><span class=SpellE><code><b><u> <span style='font-family:Tahoma;color:#3B64A2'><font size="1">hello.c</font></span></u></b></code></span><span style='font-size:10.0pt;font-family:Arial'><o:p></o:p></span></p> </td> </tr> <tr style='mso-yfti-irow:1;mso-yfti-lastrow:yes'> <td style='background:#BFBFBF;padding:.75pt .75pt .75pt .75pt'> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">#include <<span class=SpellE>stdio.h</span>></font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">#include "<span class=SpellE>mpi.h</span>"</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">void main(<span class=SpellE>int</span> <span class=SpellE>argc</span>, char **<span class=SpellE>argv</span>)</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">{</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span><span class=SpellE>int</span> me, <span class=SpellE>nprocs</span>, <span class=SpellE>namelen</span>;</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span>char <span class=SpellE>processor_name</span>[MPI_MAX_PROCESSOR_NAME];</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span><span class=SpellE>MPI_Init</span>(&<span class=SpellE>argc</span>, &<span class=SpellE>argv</span>);</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span><span class=SpellE>MPI_Comm_size</span>(MPI_COMM_WORLD, &<span class=SpellE>nprocs</span>);</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span><span class=SpellE>MPI_Comm_rank</span>(MPI_COMM_WORLD, &me);</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span><span class=SpellE>MPI_Get_processor_name</span>(<span class=SpellE>processor_name</span>, &<span class=SpellE>namelen</span>);</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-tab-count:1'> </span><span class=SpellE><span class=GramE>printf</span></span><span class=GramE>(</span>"Hello World!<span style='mso-spacerun:yes'> </span>I'm process %d of %d on %s\n", me, <span class=SpellE>nprocs</span>,</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-tab-count:2'> </span><span class=SpellE>processor_name</span>);</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1"><span style='mso-spacerun:yes'> </span><span class=SpellE>MPI_Finalize</span>();</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">}</font><o:p></o:p></span></p> </td> </tr> </table> </div> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2"> <span style='font-family:Tahoma'><font size="1">In order to successfully compile the code, the MPI header file (</font></span><span class=SpellE><code><span style='font-family:Tahoma'><font size="1">mpi.h</font></span></code></span><span style='font-family:Tahoma'><font size="1">) must be included at the top of the code. Just inside </font> </span><span class=GramE><code> <span style='font-family:Tahoma'><font size="1">main(</font></span></code></span><code><span style='font-family:Tahoma'><font size="1">)</font></span></code><span style='font-family: Tahoma'><font size="1">, </font> </span><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">MPI_Init</font></span></code></span><code><span style='font-family:Tahoma'><font size="1">()</font></span></code></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> must be called and handed the command line arguments so that the environment is setup correctly for the program to run in parallel.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">The next three MPI routines in Program 1 return information about the parallel environment for use later in the code. In this example, we merely print out the information, but in most parallel codes this information is used to do automatic problem decomposition and to setup communications between processes. </font> <o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2"><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">MPI_Comm_</font><span class=GramE><font size="1">size</font></span></span></code></span><span class=GramE><code><span style='font-family:Tahoma'><font size="1">(</font></span></code></span><code><span style='font-family:Tahoma'><font size="1">)</font></span></code><span style='font-family:Tahoma'><font size="1"> provides the number of processes, which is subsequently stored in </font> </span><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">nprocs</font></span></code></span><span style='font-family:Tahoma'><font size="1">, in the communicator group </font> </span><code> <span style='font-family:Tahoma'><font size="1">MPI_COMM_WORLD</font></span></code><span style='font-family:Tahoma'><font size="1">. </font> </span><code> <span style='font-family:Tahoma'><font size="1">MPI_COMM_WORLD</font></span></code><span style='font-family:Tahoma'><font size="1"> is a special communicator which denotes all of the processes available at initialization. </font> </span><span class=SpellE><code><span style='font-family:Tahoma'><font size="1">MPI_Comm_</font><span class=GramE><font size="1">rank</font></span></span></code></span><span class=GramE><code><span style='font-family:Tahoma'><font size="1">(</font></span></code></span><code><span style='font-family:Tahoma'><font size="1">)</font></span></code><span style='font-family:Tahoma'><font size="1"> provides the rank or process number (ranging from 0 to </font> </span><code><span style='font-family:Tahoma'> <font size="1">nprocs</font></span></code><span style='font-family:Tahoma'><font size="1">-1) of the calling process. The rank is subsequently stored in </font> </span><code> <span style='font-family:Tahoma'><font size="1">me</font></span></code><span style='font-family: Tahoma'><font size="1">. </font> </span><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">MPI_Get_processor_</font><span class=GramE><font size="1">name</font></span></span></code></span><span class=GramE><code><span style='font-family:Tahoma'><font size="1">(</font></span></code></span><code><span style='font-family:Tahoma'><font size="1">)</font></span></code><span style='font-family:Tahoma'><font size="1"> provides the hostname of the node (not the individual processor) being used, stored in </font> </span><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">processor_name</font></span></code></span><span style='font-family:Tahoma'><font size="1">, as well as the length of this hostname, stored in </font> </span><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">namelen</font></span></code></span></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1">.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2"> <span style='font-family:Tahoma'><font size="1">Next, the code prints "Hello World!" and the values of the variables obtained in the three previous MPI calls. Finally, </font> </span><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">MPI_</font><span class=GramE><font size="1">Finalize</font></span></span></code></span><span class=GramE><code><span style='font-family:Tahoma'><font size="1">(</font></span></code></span><code><span style='font-family:Tahoma'><font size="1">)</font></span></code></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1"> is called to terminate the parallel environment.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2"> <span style='font-family:Tahoma'><font size="1">MPI programs can be compiled in many ways, but most MPI implementations provide an easy-to-use script which will set desired compiler flags, point the compiler at the right directory for MPI header files, and include the necessary libraries for the linker. The MPICH implementation provides a script called </font> </span><span class=SpellE><code><span style='font-family:Tahoma'><font size="1">mpicc</font></span></code></span><span style='font-family:Tahoma'><font size="1"> which will use the desired compiler, in this case </font> </span><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">gcc</font></span></code></span><span style='font-family:Tahoma'><font size="1">, and will pass the other command line arguments to it. Output 1 shows how to compile Program 1 with </font> </span><span class=SpellE><code> <span style='font-family:Tahoma'><font size="1">mpicc</font></span></code></span></font><span style='font-family:Tahoma'><font color="#3B64A2" size="1">.</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" face="Tahoma" size="1"><a name=Output1></a> </font> <span lang=IT style='font-family:Tahoma;mso-ansi-language:IT'> <font color="#3B64A2" size="1"># mpicc -O -o hello hello.c</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span lang=IT style='font-family:Tahoma; mso-ansi-language:IT'><o:p> </o:p></span></font></p> <p class=MsoNormal> <span lang=IT style='font-family:Tahoma; mso-ansi-language:IT'><font color="#3B64A2" size="1"># mpirun -np 6 hello</font><o:p></o:p></span></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Hello World!<span style='mso-spacerun:yes'> </span>I'm process 4 of 6 on beowulf005</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Hello World!<span style='mso-spacerun:yes'> </span>I'm process 1 of 6 on beowulf002</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Hello World!<span style='mso-spacerun:yes'> </span>I'm process 5 of 6 on beowulf006</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Hello World!<span style='mso-spacerun:yes'> </span>I'm process 2 of 6 on beowulf003</font><o:p></o:p></span></p> <p class=MsoNormal><span style='font-family:Tahoma'> <font color="#3B64A2" size="1">Hello World!<span style='mso-spacerun:yes'> </span>I'm process 3 of 6 on beowulf004</font><o:p></o:p></span></p> <p class=MsoNormal> <font color="#3B64A2" face="Tahoma" size="1">Hello World!<span style='mso-spacerun:yes'> </span>I'm process 0 of 6 on beowulf001</font><b style='mso-bidi-font-weight:normal'><o:p></o:p></b></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2" size="1"> <span style='font-family:Tahoma'><o:p> </o:p></span></font></p> <p class=MsoNormal><font color="#3B64A2"> <span style='font-family:Tahoma'><o:p><font size="1"> </font></o:p></span></p> </div> </body>