<?xml version='1.0' encoding='UTF-8'?><?xml-stylesheet href="http://www.blogger.com/styles/atom.css" type="text/css"?><feed xmlns='http://www.w3.org/2005/Atom' xmlns:openSearch='http://a9.com/-/spec/opensearchrss/1.0/' xmlns:georss='http://www.georss.org/georss' xmlns:gd='http://schemas.google.com/g/2005' xmlns:thr='http://purl.org/syndication/thread/1.0'><id>tag:blogger.com,1999:blog-29480451</id><updated>2011-04-21T18:16:32.938-07:00</updated><title type='text'>Lochmeters</title><subtitle type='html'>All about Q</subtitle><link rel='http://schemas.google.com/g/2005#feed' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/posts/default'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default?max-results=100'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/'/><link rel='hub' href='http://pubsubhubbub.appspot.com/'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><generator version='7.00' uri='http://www.blogger.com'>Blogger</generator><openSearch:totalResults>33</openSearch:totalResults><openSearch:startIndex>1</openSearch:startIndex><openSearch:itemsPerPage>100</openSearch:itemsPerPage><entry><id>tag:blogger.com,1999:blog-29480451.post-116085120627865670</id><published>2006-10-14T11:21:00.000-07:00</published><updated>2006-10-14T11:40:06.496-07:00</updated><title type='text'>Pandora的混音</title><content type='html'>&lt;a href="http://www.pandora.com/"&gt;Pandora&lt;/a&gt;确实是一个伟大的音乐盒子. 它可以通过用户提供的搜索, 给出风格类似的音乐或者歌手, 也即其网页里宣称的"音乐基因". 同时它还提供"喜欢","讨厌"来选择收听的歌曲是否重复播放. &lt;br /&gt;&lt;br /&gt;   但是这样的选择会导致某一风格歌曲始终只有那么几首, 即使如何欣赏的歌曲也会在一遍又一遍的播放中产生腻烦. Pandora先前只能通过选择来设定一种风格歌曲的播放,也就所谓的"station", 所以久而久之,我们需要手动地选择或者重新搜索新歌来调换口味.  现在Pandora新提供了QuickMix的功能, 也就是可以在几个stations之间选择播放歌曲, 一定程度上扩展了大家的选择..&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-116085120627865670?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/116085120627865670/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=116085120627865670' title='3 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/116085120627865670'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/116085120627865670'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/10/pandora.html' title='Pandora的混音'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>3</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-116033785736417128</id><published>2006-10-08T13:04:00.000-07:00</published><updated>2006-10-08T13:04:17.463-07:00</updated><title type='text'>NEGF Vs Master Equation</title><content type='html'>最近在读主方程master equation 相关的文献. 以前不太了解处理开放体系的数学工具还包括&lt;br /&gt;Master Equation, 花了比较大的精力读得都是NEGF相关的文献. 仅仅是因为想当然地认为Weak coupling的器件在实际应用中没有什么用处, 这个想法其实挺愚蠢的, 而实际上来说主方程方法是目前处理开放体系最主要的工具, 特别是在量子光学, 量子测量方面显得特别重要. &lt;br /&gt;   在凝聚态体系中, 当然非平衡Green函数NEGF由于在数学上的简洁得到广泛的应用, 很多量子化学计算的也借着所谓的分子电子学之名 推波助澜. 不过, 在R.Car最近一篇文章来看, 处理非平衡态的开放体系,利用基于DFT的NEGF方法在理论上是没有验证的, 他们认为基于TDDFT的主方程方法至少在物理基础上来看是更有前途的.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-116033785736417128?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/116033785736417128/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=116033785736417128' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/116033785736417128'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/116033785736417128'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/10/negf-vs-master-equation.html' title='NEGF Vs Master Equation'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115895725361903633</id><published>2006-09-22T13:34:00.000-07:00</published><updated>2006-09-22T13:39:28.706-07:00</updated><title type='text'>不为别的, 就为这些话!</title><content type='html'>这期的Science 围绕丘田事件评论中国兼职教授掠取资源的问题.&lt;br /&gt;   &lt;br /&gt;   有些话, 不管怎么说, 是很有道理的. 就像当初chern说的一样, 只有在本土培养人, 才有希望.&lt;br /&gt; &lt;br /&gt;   但是, 在国内, 人力资源是廉价的, 如何可能重视人力资源呢 ?  &lt;br /&gt;&lt;br /&gt;  &lt;br /&gt; &lt;br /&gt;   &lt;blockquote&gt; But Yau and other critics insist that the popularity of these programs does not justify the expense. Rather than lavish money on part-time academics, they argue, Chinese institutions should raise stipends of&lt;br /&gt;students and young researchers from their present paltry levels of $30 to $160 a&lt;br /&gt;month. “The Chinese government does not pay enough attention to young people,” Yau&lt;br /&gt;says. As long as the brightest young minds seek greener pastures outside China, the&lt;br /&gt;brain drain—and the hunger for overseas talents—will continue.&lt;/blockquote&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115895725361903633?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115895725361903633/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115895725361903633' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115895725361903633'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115895725361903633'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/blog-post_115895725361903633.html' title='不为别的, 就为这些话!'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115894188151059957</id><published>2006-09-22T09:05:00.000-07:00</published><updated>2006-09-22T09:18:23.876-07:00</updated><title type='text'>量子点, 真是太困难了</title><content type='html'>没办法, 现在是逼上梁山了. &lt;br /&gt;   老板真是无聊的要紧, 除了关心啥最小, 啥尺度之外基本上没有啥物理了.&lt;br /&gt;   只好想着法子敷衍他了, 所以,现在觉得跟王胜合作做量子点或许有些意思. &lt;br /&gt;   这几天看了量子点的综述和几个人的博士论文, 感觉这个领域真是博大精深啊.&lt;br /&gt;   &lt;br /&gt;   还好, 现在基于carbon nanotube的 量子点 刚做不久, 也有很多比较大的进展, 不过都是荷兰Delft和哈佛macrus组在做, 利用nanotube来define 量子点, 研究所谓的coupling effects Or electron-correlation对体系输运性质的影响.&lt;br /&gt;&lt;br /&gt;   今天读的综述, 是汉堡一个学者写的, 奇怪的是在isi上搜索他的文章,不到10篇. 不知道是杂志的问题呢, 还是人家就只在德国灌水?  其中, 单纯的Coulomb Blockade效应已经理解的很好, 现在一些open的问题是多量子点间的相互作用, 电-电相互作用对量子点态的作用.   他特别强调所谓的intermediate tunnel coupling的量子点体系的问题很open,而我在查他的文章也发现他最近的工作也是集中这一领域.&lt;br /&gt;   但是困难的问题在于他介绍的工具我不太熟悉, 什么Fermi's Golden rule, 看似很简单,但是实际计算是会要人的命诶. 还利用了Keldysh Contour的技术,虽然对这一方法有所了解, 同样是在计算上, 我还没有多少经验. &lt;br /&gt;&lt;br /&gt;   困难啊...&lt;br /&gt;&lt;br /&gt;   让人激奋的是,这个领域有很多open的问题, 当然需要勤奋和运气去做些事情了.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115894188151059957?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115894188151059957/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115894188151059957' title='8 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115894188151059957'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115894188151059957'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/blog-post_22.html' title='量子点, 真是太困难了'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>8</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115851066119835913</id><published>2006-09-17T09:23:00.000-07:00</published><updated>2006-09-17T09:31:33.446-07:00</updated><title type='text'>texlive 2005的中文配置</title><content type='html'>按照&lt;a href="http://www.newsmth.net/pc/pcshowcom.php?cid=152822"&gt;FerretL&lt;/a&gt; 配置说明, 我尝试在RedHat 9.0下配置texlive2005的中文环境. &lt;br /&gt;  很顺利地就搞定了,诸如路径等的设置. 但是不能像FerretL声称地实现texexec --pdf *.tex 直接得到中文输出. 正确的中文输出只可以在texexec --output=dvipdf *.tex 实现. 真是诡异,不知道哪里出了问题.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115851066119835913?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115851066119835913/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115851066119835913' title='2 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115851066119835913'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115851066119835913'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/texlive-2005.html' title='texlive 2005的中文配置'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>2</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115847473111491675</id><published>2006-09-16T23:26:00.002-07:00</published><updated>2006-09-16T23:35:21.213-07:00</updated><title type='text'>Basic Notions of Condensed Matter Physics的简单书评</title><content type='html'>&lt;p&gt;在Yamada K 所著Electron Correlation in metals一书中,对P.W. Anderson的名著 Basic notions of condensed matter physics做了几个简短的评论, 很有Anderson的味道.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;他提到Anderson在凝聚态物理中的深远影响可以分为两个原理: 一是,对称破缺, 这是相变理论的核心原理, 体系在低温情况出现铁磁态和超导态,即可由该原理解释. 另外一基本原理则是绝热连续, 即我们可以参照简单系统来讨论复杂系统的性质: when we study a generally complicated physical system we can refer to a simple system that contains the esential nature of the real system and understand the complicated systemn on the basis of knowledge of the simple system.&lt;/p&gt;&lt;br /&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;  &lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;  &lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a href="http://technorati.com/tag/BookReview" class="ztag" rel="tag"&gt;BookReview&lt;/a&gt;, &lt;a href="http://technorati.com/tag/condensed%20matter%20physics" class="ztag" rel="tag"&gt;condensed matter physics&lt;/a&gt;&lt;/span&gt; &lt;br /&gt;&lt;/p&gt;&lt;p class="poweredbyzoundry"&gt;Powered by &lt;a href="http://www.zoundry.com" class="poweredbyzoundry_link" rel="nofollow"&gt;Zoundry&lt;/a&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115847473111491675?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115847473111491675/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115847473111491675' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115847473111491675'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115847473111491675'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/basic-notions-of-condensed-matter.html' title='Basic Notions of Condensed Matter Physics的简单书评'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115778606796691554</id><published>2006-09-09T00:08:00.000-07:00</published><updated>2006-09-09T04:01:52.846-07:00</updated><title type='text'>美国物理学会宣布开放获取物理评论系列杂志</title><content type='html'>&lt;a href="http://photos1.blogger.com/blogger/2519/3142/1600/FreeToRead.png"&gt;&lt;img style="FLOAT: left; MARGIN: 0px 10px 10px 0px; CURSOR: hand" alt="" src="http://photos1.blogger.com/blogger/2519/3142/320/FreeToRead.png" border="0" /&gt;&lt;/a&gt;&lt;br /&gt;&lt;p&gt;近日美国物理学会(APS)宣布物理评论系列杂志在今年9月份采取&lt;a href="http://publish.aps.org/OpenAccessAnnounce.html"&gt;开放获取&lt;/a&gt;的政策,具体细节如下:&lt;/p&gt;&lt;br /&gt;&lt;p&gt;The American Physical Society (APS) is pleased to announce that it will soon expand its Open Access (OA) offerings to articles published in &lt;em&gt;Physical Review A-E&lt;/em&gt;, &lt;em&gt;Physical Review Letters&lt;/em&gt;, and &lt;em&gt;Reviews of Modern Physics&lt;/em&gt;. This OA initiative is called &lt;span class="style3"&gt;&lt;strong&gt;&lt;span style="color:#006666;"&gt;FREE TO READ&lt;/span&gt;&lt;/strong&gt;&lt;/span&gt; and, when released in early September 2006, can be applied to any article or group of articles published in the Journals of the American Physical Society back to 1893. Anyone (authors, readers, institutions, funding agencies, etc.) may, by paying a &lt;a href="http://forms.aps.org/author/Free2Read_2.pdf"&gt;one-time fee&lt;/a&gt;, make articles published in our journals available on our sites to all readers at no cost and without a subscription. Readers will have access to PDF and postscript versions of the FREE TO READ articles through the APS online journals.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;For years APS has been a leader in OA with its early and continued support of arXiv.org and with its exemplary copyright agreement form. The agreement allows authors to make available their APS publications on their own or their institution's website. APS introduced its first OA journal, &lt;em&gt;Physical Review Special Topics&lt;/em&gt; - &lt;em&gt;Accelerators and Beams&lt;/em&gt;, in 1998. Based on a sponsorship model, this journal has steadily grown over the past 8 years and is now supported by an international group of accelerator laboratories. APS introduced a second OA journal in 2005 called &lt;em&gt;Physical Review Special Topics&lt;/em&gt; - &lt;em&gt;Physics Education Research&lt;/em&gt;. This freely available journal is financed by publication charges to the authors or the authors' institutions. The introduction of FREE TO READ extends OA to the articles for all of APS's journals.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;The FREE TO READ fees will initially be $975 for articles in &lt;em&gt;Physical Review A-E&lt;/em&gt; and $1300 for Letters in PRL. Articles in RMP, due to their large size and the limited number published annually, will be considered on a case-by-case basis. The higher price associated with PRL is due to its higher cost per published Letter (because of its stringent acceptance rate).&lt;/p&gt;&lt;br /&gt;&lt;p&gt;The fees will augment revenues for the APS, since they will not be replacing subscriptions, but have been set well below the current amount per article needed to recover costs in the absence of subscriptions. The fees will therefore be adjusted as necessary to maintain APS's ability to sustain this initiative. Additional revenues from FREE TO READ will primarily be used to lower the current subscription rates of the smallest (lowest tier) institutions.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;The FREE TO READ initiative represents a path by which APS could gradually transition to full Open Access. If the community (especially institutions and funding agencies) shows continued support for this initiative, a sustainable level may be reached in which the APS can recover its costs, offset its risks, and eliminate subscriptions for some or all of its journals.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;The APS is determined to extend every effort to make this model successful. Martin Blume, the Editor-in-Chief, states that "APS is a financially stable organization willing to take risks to support the community," and it is with the community in mind that APS is offering FREE TO READ.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;For additional information, please go to the FREE TO READ FAQ at &lt;a href="http://publish.aps.org/FREETOREAD_FAQ.html"&gt;http://publish.aps.org/FREETOREAD_FAQ.html&lt;/a&gt;.&lt;/p&gt;&lt;br /&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;&lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;&lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a class="ztag" href="http://technorati.com/tag/American%20Physical%20Society" rel="tag"&gt;American Physical Society&lt;/a&gt;, &lt;a class="ztag" href="http://technorati.com/tag/Open%20access" rel="tag"&gt;Open access&lt;/a&gt;, &lt;a class="ztag" href="http://technorati.com/tag/Physical%20Review" rel="tag"&gt;Physical Review&lt;/a&gt;&lt;/span&gt;&lt;br /&gt;&lt;/p&gt;&lt;p class="poweredbyzoundry"&gt;Powered by &lt;a class="poweredbyzoundry_link" href="http://www.zoundry.com" rel="nofollow"&gt;Zoundry&lt;/a&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115778606796691554?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115778606796691554/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115778606796691554' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115778606796691554'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115778606796691554'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/blog-post_09.html' title='美国物理学会宣布开放获取物理评论系列杂志'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115772717780379741</id><published>2006-09-08T07:47:00.000-07:00</published><updated>2006-09-08T07:57:27.963-07:00</updated><title type='text'>ZGEMM的手册页</title><content type='html'>&lt;p&gt;数值计算中, 常用到lapack库zgemm函数,它主要是进行矩阵变换与矩阵乘积的计算. 它的手册页如下 :&lt;/p&gt;&lt;br /&gt;&lt;p&gt;ZGEMM performs one of the matrix-matrix operations&lt;br /&gt;*&lt;br /&gt;* C := alpha*op( A )*op( B ) + beta*C,&lt;br /&gt;*&lt;br /&gt;* where op( X ) is one of&lt;br /&gt;*&lt;br /&gt;* op( X ) = X or op( X ) = X' or op( X ) = conjg( X' ),&lt;br /&gt;*&lt;br /&gt;* alpha and beta are scalars, and A, B and C are matrices, with op( A )&lt;br /&gt;* an m by k matrix, op( B ) a k by n matrix and C an m by n matrix.&lt;br /&gt;*&lt;br /&gt;* Parameters&lt;br /&gt;* ==========&lt;br /&gt;*&lt;br /&gt;* TRANSA - CHARACTER*1.&lt;br /&gt;* On entry, TRANSA specifies the form of op( A ) to be used in&lt;br /&gt;* the matrix multiplication as follows:&lt;br /&gt;*&lt;br /&gt;* TRANSA = 'N' or 'n', op( A ) = A.&lt;br /&gt;*&lt;br /&gt;* TRANSA = 'T' or 't', op( A ) = A'.&lt;br /&gt;*&lt;br /&gt;* TRANSA = 'C' or 'c', op( A ) = conjg( A' ).&lt;br /&gt;*&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* TRANSB - CHARACTER*1.&lt;br /&gt;* On entry, TRANSB specifies the form of op( B ) to be used in&lt;br /&gt;* the matrix multiplication as follows:&lt;br /&gt;*&lt;br /&gt;* TRANSB = 'N' or 'n', op( B ) = B.&lt;br /&gt;*&lt;br /&gt;* TRANSB = 'T' or 't', op( B ) = B'.&lt;br /&gt;*&lt;br /&gt;* TRANSB = 'C' or 'c', op( B ) = conjg( B' ).&lt;br /&gt;*&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* M - INTEGER.&lt;br /&gt;* On entry, M specifies the number of rows of the matrix&lt;br /&gt;* op( A ) and of the matrix C. M must be at least zero.&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* N - INTEGER.&lt;br /&gt;* On entry, N specifies the number of columns of the matrix&lt;br /&gt;* op( B ) and the number of columns of the matrix C. N must be&lt;br /&gt;* at least zero.&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* K - INTEGER.&lt;br /&gt;* On entry, K specifies the number of columns of the matrix&lt;br /&gt;* op( A ) and the number of rows of the matrix op( B ). K must&lt;br /&gt;* be at least zero.&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* ALPHA - COMPLEX*16 .&lt;br /&gt;* On entry, ALPHA specifies the scalar alpha.&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* A - COMPLEX*16 array of DIMENSION ( LDA, ka ), where ka is&lt;br /&gt;* k when TRANSA = 'N' or 'n', and is m otherwise.&lt;br /&gt;* Before entry with TRANSA = 'N' or 'n', the leading m by k&lt;br /&gt;* part of the array A must contain the matrix A, otherwise&lt;br /&gt;* the leading k by m part of the array A must contain the&lt;br /&gt;* matrix A.&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* LDA - INTEGER.&lt;br /&gt;* On entry, LDA specifies the first dimension of A as declared&lt;br /&gt;* in the calling (sub) program. When TRANSA = 'N' or 'n' then&lt;br /&gt;* LDA must be at least max( 1, m ), otherwise LDA must be at&lt;br /&gt;* least max( 1, k ).&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* B - COMPLEX*16 array of DIMENSION ( LDB, kb ), where kb is&lt;br /&gt;* n when TRANSB = 'N' or 'n', and is k otherwise.&lt;br /&gt;* Before entry with TRANSB = 'N' or 'n', the leading k by n&lt;br /&gt;* part of the array B must contain the matrix B, otherwise&lt;br /&gt;* the leading n by k part of the array B must contain the&lt;br /&gt;* matrix B.&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* LDB - INTEGER.&lt;br /&gt;* On entry, LDB specifies the first dimension of B as declared&lt;br /&gt;* in the calling (sub) program. When TRANSB = 'N' or 'n' then&lt;br /&gt;* LDB must be at least max( 1, k ), otherwise LDB must be at&lt;br /&gt;* least max( 1, n ).&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* BETA - COMPLEX*16 .&lt;br /&gt;* On entry, BETA specifies the scalar beta. When BETA is&lt;br /&gt;* supplied as zero then C need not be set on input.&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;*&lt;br /&gt;* C - COMPLEX*16 array of DIMENSION ( LDC, n ).&lt;br /&gt;* Before entry, the leading m by n part of the array C must&lt;br /&gt;* contain the matrix C, except when beta is zero, in which&lt;br /&gt;* case C need not be set on entry.&lt;br /&gt;* On exit, the array C is overwritten by the m by n matrix&lt;br /&gt;* ( alpha*op( A )*op( B ) + beta*C ).&lt;br /&gt;*&lt;br /&gt;* LDC - INTEGER.&lt;br /&gt;* On entry, LDC specifies the first dimension of C as declared&lt;br /&gt;* in the calling (sub) program. LDC must be at least&lt;br /&gt;* max( 1, m ).&lt;br /&gt;* Unchanged on exit.&lt;br /&gt;&lt;/p&gt;&lt;br /&gt;其中开始的两个参数TRANSA/TRANSB用于指定矩阵的操作: 转置,复共厄.&lt;br /&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;&lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;&lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a class="ztag" href="http://technorati.com/tag/lapack" rel="tag"&gt;lapack&lt;/a&gt;, &lt;a class="ztag" href="http://technorati.com/tag/zgemm" rel="tag"&gt;zgemm&lt;/a&gt;&lt;/span&gt;&lt;br /&gt;&lt;/p&gt;&lt;p class="poweredbyzoundry"&gt;Powered by &lt;a class="poweredbyzoundry_link" href="http://www.zoundry.com" rel="nofollow"&gt;Zoundry&lt;/a&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115772717780379741?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115772717780379741/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115772717780379741' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115772717780379741'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115772717780379741'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/zgemm.html' title='ZGEMM的手册页'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115771863424780668</id><published>2006-09-08T05:24:00.000-07:00</published><updated>2006-09-08T05:30:34.303-07:00</updated><title type='text'>一个函数调用的手记</title><content type='html'>&lt;br /&gt;&lt;p&gt;ZGESV是lapack最简单的一个函数, 即求解线性方程组A*X=B , 子程序调用的格式是zgesv(n, nrh,A, LDA,IPIV,B,LDB,info).&lt;br/&gt;各个参数的意义如下 :&lt;br/&gt;n , 线性方程组的方程数, 也就是矩阵A的阶数;&lt;br/&gt;nrh, 方程组右边的数目, 也即矩阵B的列数;&lt;br/&gt;A,其维度定义为A(LDA,N), 输入时,为方程组的系数矩阵;输出则是LU分解的分式;在计算中基本上不关心A的结果.&lt;br/&gt;LDA,是所谓数组A的起始维数(这里应当认为是矩阵A的第一列的维数),也即矩阵A的行数.&lt;br/&gt;IPIV, 输出参量, 用来定义排列矩阵P的指标.&lt;br/&gt;B, 输入是方程组右边的矩阵B, 维度为B(LDB,NRH);输出是矩阵X(N,NRH)的解, 如果info=0.&lt;br/&gt;LDB, 同LDA, 矩阵B的行数.&lt;br/&gt;info, info=0,表示调用成功返回;非零,则是调用失败.&lt;br/&gt;以下是 zgesv的手册页:&lt;/p&gt;&lt;br /&gt;&lt;p&gt;NAME&lt;br/&gt;ZGESV - compute the solution to a complex system of linear equations A * X = B,&lt;/p&gt;&lt;br /&gt;&lt;p&gt;SYNOPSIS&lt;br/&gt;SUBROUTINE ZGESV( N, NRHS, A, LDA, IPIV, B, LDB, INFO )&lt;br/&gt;INTEGER INFO, LDA, LDB, N, NRHS&lt;br/&gt;INTEGER IPIV( * )&lt;br/&gt;COMPLEX*16 A( LDA, * ), B( LDB, * )&lt;br/&gt;PURPOSE&lt;br/&gt;ZGESV computes the solution to a complex system of linear equations A * X = B, where A is an N-by-N matrix and X and B are N-by-NRHS matrices.&lt;br/&gt;The LU decomposition with partial pivoting and row interchanges is used to factor A as&lt;br/&gt;A=P * L * U, where P is a permutation matrix, L is unit lower triangular, and U is upper triangular. The factored form of A is then used to solve the system of equations A * X = B.&lt;br/&gt;&lt;/p&gt;&lt;br /&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;  &lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;  &lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a href="http://technorati.com/tag/lapack" class="ztag" rel="tag"&gt;lapack&lt;/a&gt;, &lt;a href="http://technorati.com/tag/subroutines" class="ztag" rel="tag"&gt;subroutines&lt;/a&gt;&lt;/span&gt; &lt;br /&gt;&lt;/p&gt;&lt;p class="poweredbyzoundry"&gt;Powered by &lt;a href="http://www.zoundry.com" class="poweredbyzoundry_link" rel="nofollow"&gt;Zoundry&lt;/a&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115771863424780668?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115771863424780668/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115771863424780668' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115771863424780668'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115771863424780668'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/blog-post_08.html' title='一个函数调用的手记'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115771476127884888</id><published>2006-09-08T04:20:00.000-07:00</published><updated>2006-09-08T04:26:01.330-07:00</updated><title type='text'>Lapack的命名规则</title><content type='html'>&lt;br /&gt;&lt;p&gt;最近读了读siesta/WanT的源代码, 需要用到Lapack数学库函数库, 随手记下几个常用关键词的含义(也就是所谓的命名规则).&lt;/p&gt;&lt;br /&gt;&lt;p&gt;Lapack中有单精度与双精度两种函数选择, 单精度下用s来标志实数,而c标志复数; 双精度下,用d来标志实数, z来标志复数.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;其中最主要的关键词是涉及矩阵类型 :&lt;/p&gt;&lt;br /&gt;&lt;p&gt;GE , general, 广义矩阵 , 也就是最普遍的矩阵类型;&lt;br/&gt;DI, diagonal, 对角矩阵;&lt;br/&gt;GB, general band,广义带矩阵 ; GT,general triangular,广义三角阵;&lt;br/&gt;HB, Hermitian Band,厄密带矩阵; HE, Hermitian, 厄密阵;&lt;br/&gt;OR, Orthogonal, 正交矩阵;SY,Symmetric, 对称矩阵;&lt;br/&gt;SB, Symmetric Band,对称带矩阵;&lt;br/&gt;ST, Symmetric Triangle,对称三角矩阵;&lt;br/&gt;UN, Unitary,么正矩阵 ....&lt;br/&gt;还有其他很多类型, 基本上这几类的组合,比如general, triangular, symmetric, hermitian, band等...&lt;br/&gt;关键是需要了解这些不同类型矩阵函数之间有什么特点(应该是在计算上可以达到速度等方面的提高).&lt;/p&gt;&lt;br /&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;  &lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;  &lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a href="http://technorati.com/tag/keywords" class="ztag" rel="tag"&gt;keywords&lt;/a&gt;, &lt;a href="http://technorati.com/tag/lapack" class="ztag" rel="tag"&gt;lapack&lt;/a&gt;&lt;/span&gt; &lt;br /&gt;&lt;/p&gt;&lt;p class="poweredbyzoundry"&gt;Powered by &lt;a href="http://www.zoundry.com" class="poweredbyzoundry_link" rel="nofollow"&gt;Zoundry&lt;/a&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115771476127884888?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115771476127884888/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115771476127884888' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115771476127884888'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115771476127884888'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/lapack.html' title='Lapack的命名规则'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115770520691195856</id><published>2006-09-08T01:40:00.000-07:00</published><updated>2006-09-08T01:46:46.950-07:00</updated><title type='text'>稀疏矩阵的压缩存储</title><content type='html'>&lt;br /&gt;&lt;p&gt;Siesta 对hamiltonian 和overlap hamiltonian 的存储采用稀疏矩阵的压缩格式. 这几天翻了翻稀疏矩阵存储的方案, 了解到最主要的格式是 链表方式, 这样处理对增减非零矩阵元的操作相对容易. 也有采用直接的mapping方案, 但是缺点是增减矩阵元的操作相对复杂.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;稀疏矩阵(n X n)的链表存储具体来说就是将矩阵的非零矩阵元(m)按照链表格式存储, 用一个数组(n)按列(或行)的顺序连续存储第一个非零矩阵元的行地址(或列地址), 而非零矩阵元的值/列号(行号)/以及下个非零矩阵元的行(列)地址 则存储在一个3m的数组中. 也就是说一个nXn的矩阵可以map到3m+n的数组上, 从而达到压缩存储目的.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;现在不太了解的是, 存储的行/列地址是采用物理地址呢, 还是自己定义一个序列来描述? 印象中在fortran里用指针来处理问题还是不太稳定的.&lt;/p&gt;&lt;br /&gt;&lt;br/&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;  &lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;  &lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a href="http://technorati.com/tag/sparse%20matrix" class="ztag" rel="tag"&gt;sparse matrix&lt;/a&gt;, &lt;a href="http://technorati.com/tag/store%20scheme" class="ztag" rel="tag"&gt;store scheme&lt;/a&gt;&lt;/span&gt; &lt;br /&gt;&lt;/p&gt;&lt;p class="poweredbyzoundry"&gt;Powered by &lt;a href="http://www.zoundry.com" class="poweredbyzoundry_link" rel="nofollow"&gt;Zoundry&lt;/a&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115770520691195856?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115770520691195856/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115770520691195856' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115770520691195856'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115770520691195856'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/09/blog-post.html' title='稀疏矩阵的压缩存储'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115564247641190933</id><published>2006-08-15T04:42:00.000-07:00</published><updated>2006-08-15T04:47:56.416-07:00</updated><title type='text'>放弃Live Writer</title><content type='html'>&lt;br /&gt;&lt;p&gt;刚听说有live writer, 挺兴奋的.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;刚才尝试安装, 但是失败了. 因为需要.NET支持!&lt;/p&gt;&lt;br /&gt;&lt;p&gt;其实对于用户来说, 安装是一件轻松快活的事情, live writer却是让用户自找麻烦. 并不是每一个用户都有.NET framework, 或者了解.NET.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;这点 Zoundry就挺不错的, 简简单单地安装, 而且支持的blog系统恐怕不会比live writer少吧,虽然发布不了 live space(还不能确定是我的问题,还是Zoundry的问题).&lt;/p&gt;&lt;br /&gt;&lt;p&gt;所以呢, 还是把live writer 放弃了, 继续用Zoundry吧..&lt;/p&gt;&lt;br /&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;  &lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;  &lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a href="http://technorati.com/tag/live" class="ztag" rel="tag"&gt;live&lt;/a&gt;, &lt;a href="http://technorati.com/tag/wirter" class="ztag" rel="tag"&gt;wirter&lt;/a&gt;&lt;/span&gt; &lt;br /&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115564247641190933?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115564247641190933/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115564247641190933' title='4 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115564247641190933'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115564247641190933'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/08/live-writer_15.html' title='放弃Live Writer'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>4</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115564170523927708</id><published>2006-08-15T04:29:00.000-07:00</published><updated>2006-08-15T04:35:05.373-07:00</updated><title type='text'>Live Writer</title><content type='html'>&lt;br /&gt;&lt;p&gt;昨天在 keso那里看到live也发布了一个Blog离线编辑与发布工具, 虽然keso指出live最近的作为没有看出它有所专注,不过我觉得微软live能够推出这样的发布工具,且不说live space上已经有1亿多的用户, 单从用户选择来说, 这使得广大bloggers又多了一个备份与编辑的工具, 而且live-writer也提供多种Blog系统的API接口, 不仅仅针对live space. 从这点来看, 微软这次做的确实挺好的, 一点都没有体现出微软的霸气.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;刚好前一天我装好了Zoundry, 虽然Zoundry挺简洁易用的, 也号称支持live space,但是我尝试了半天没有发现我可以成功发布live space,所以现在我正在由于我要不要安装live writer...毕竟微软的东西, 好像都挺大的...&lt;/p&gt;&lt;br /&gt;&lt;br/&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;  &lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;  &lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a href="http://technorati.com/tag/Blog" class="ztag" rel="tag"&gt;Blog&lt;/a&gt;, &lt;a href="http://technorati.com/tag/Zoundry" class="ztag" rel="tag"&gt;Zoundry&lt;/a&gt;, &lt;a href="http://technorati.com/tag/live" class="ztag" rel="tag"&gt;live&lt;/a&gt;, &lt;a href="http://technorati.com/tag/writer" class="ztag" rel="tag"&gt;writer&lt;/a&gt;&lt;/span&gt; &lt;br /&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115564170523927708?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115564170523927708/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115564170523927708' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115564170523927708'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115564170523927708'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/08/live-writer.html' title='Live Writer'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115539975346310233</id><published>2006-08-12T09:17:00.000-07:00</published><updated>2006-08-12T09:22:33.480-07:00</updated><title type='text'>Zoundry试用</title><content type='html'>&lt;br /&gt;&lt;p&gt;下载了传说中的ZoundryBlog发布工具, 非常不错.&lt;/p&gt;&lt;br /&gt;&lt;p&gt;界面简洁清新,而且可以提供的多个blog 系统发布. 这么好的本地发布系统, 确实值得赞~~~!&lt;/p&gt;&lt;br /&gt;&lt;p class="zoundry_bw_tags"&gt;&lt;br /&gt;  &lt;!-- Tag links generated by Zoundry Blog Writer. Do not manually edit. http://www.zoundry.com --&gt;&lt;br /&gt;  &lt;span class="ztags"&gt;&lt;span class="ztagspace"&gt;Technorati&lt;/span&gt; : &lt;a href="http://technorati.com/tag/Blog" class="ztag" rel="tag"&gt;Blog&lt;/a&gt;, &lt;a href="http://technorati.com/tag/%E5%8F%91%E5%B8%83%E5%B7%A5%E5%85%B7" class="ztag" rel="tag"&gt;发布工具&lt;/a&gt;&lt;/span&gt; &lt;br /&gt;&lt;/p&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115539975346310233?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115539975346310233/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115539975346310233' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115539975346310233'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115539975346310233'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/08/zoundry.html' title='Zoundry试用'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115211119899850051</id><published>2006-07-05T07:53:00.000-07:00</published><updated>2006-07-05T07:53:19.076-07:00</updated><title type='text'></title><content type='html'>&lt;P&gt;&lt;STRONG&gt;金属体系计算中k取点与smearing方法&lt;/STRONG&gt;&lt;/P&gt; &lt;P&gt;&lt;STRONG&gt;&lt;/STRONG&gt; &lt;/P&gt; &lt;P&gt;&lt;STRONG&gt; Phys.Rev.Lett 82 3296&lt;/STRONG&gt;&lt;/P&gt; &lt;P&gt;         &lt;EM&gt;&lt;STRONG&gt;Thermal Contraction and Disordering of the Al(110) Surface&lt;/STRONG&gt;&lt;/EM&gt;&lt;/P&gt; &lt;P&gt;    利用Ensemble-DFT, 以及cold smearing方法&lt;STRONG&gt;(a technical improvement for the Brillouin zone integrations,具体方法描述见下篇论文)&lt;/STRONG&gt; . Applying this scheme to the case of Al(110), we provide the first theoretical confirmation of a negative thermal expansion for this surface, and excellent agreement with the experiments for the temperature-dependent multilayer relaxations.  &lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt;   In first-principles calculations for metals it is customary to introduce a fictitious electronic temperature sigma to broaden the density of states and to smooth the discontinuities at the Fermi energy mu, greatly improving  the sampling accuracy of a given set of k points. &lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt;&lt;STRONG&gt;Phys.Rev.Lett 79 1337&lt;/STRONG&gt;&lt;/P&gt; &lt;P&gt;&lt;EM&gt;&lt;STRONG&gt;    Ensemble Density-Functional Theory for Ab Initio Molecular Dynamics of Metals and Finite-Temperature Insulators&lt;/STRONG&gt;&lt;/EM&gt;&lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt;   在处理金属体系会遇到不同于半导体和绝缘体体系的问题. The application of these methods and techniques to metallic systems has nonetheless encountered several difficulties that have made progress slower than for the case of semiconductors and insulators. The discontinuous variation of the orbital occupancies across the Brillouin zone (BZ) makes the occupation numbers rather illconditioned variables, and the self-consistent solution of the screening problem can suffer from several instabilities. 轨道在过BZ区域的占据不连续性使得占据数是一个病态的变量, 而且自洽求解屏蔽问题会出现难以避免的问题.  The absence of a gap in the energy spectrum and the requirement of an exact diagonalization for the Hamiltonian matrix everywhere in the BZ (in order to assign the occupation numbers) introduce “slow frequencies” in the evolution of the orbitals towards the ground state and preclude the straightforward extension to metals of algorithms which performed well for insulators. 即不能直接推广绝缘体体系的算法至金属体系.  对Fermi Surface 采用有限温度 smearing可以提高BZ的取点, 但是只能部分解决金属体系的问题.  作者在本文给出了一个通用,有效的理论框架得到有限温度下体系的基态Hamiltonian矩阵.&lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt;&lt;STRONG&gt;Phys.Rev.B  51 6773&lt;/STRONG&gt;&lt;/P&gt; &lt;P&gt;&lt;STRONG&gt;  Lattice Dynamics of Metals From Density Functional Perturbation Theory&lt;/STRONG&gt;&lt;/P&gt; &lt;P&gt;&lt;STRONG&gt;&lt;/STRONG&gt; &lt;/P&gt; &lt;P&gt; The smearing technique to metallic systems is  very general and simple. In smearing approach, the results are independent of the assumed crystal geometry, as long as equivalent sampling points are used . This makes the smearing technique the method of choice  in the first-principle molecular dynamical simulation  of metal systems. 即在金属的第一性动力学计算中应采用smearing方法.  主要的smearing functions 有Fermi-Dirac Broadening function , Lorentzian , Gaussian or Gaussian combined with polynomials.   smearing approach 即 局域态密度与一个smearing function 做卷积. 选取不同的smearing function, 其smearing width 对结果有不同的影响.  本文Fig 1给出了 总能与 smearing width的关系, 结果表明一阶Hermit-Gaussian 函数, 其smearing width对总能影响不大,而相应的Simple-Gaussian函数 其smearing width对 总能有~0.03Ryd的影响. 并且注意到smearing width与 electronic temperature直接相连. 也即smearing width 越大, 相应的electronic temperature 越高.&lt;/P&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115211119899850051?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115211119899850051/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115211119899850051' title='4 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115211119899850051'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115211119899850051'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/07/ksmearing-phys.html' title=''/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>4</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115203611362061851</id><published>2006-07-04T11:01:00.000-07:00</published><updated>2006-07-05T04:53:01.400-07:00</updated><title type='text'></title><content type='html'>&lt;P&gt;  &lt;/P&gt; &lt;P&gt;  怎么利用Writely呢 ?&lt;/P&gt; &lt;P&gt; 据说这个工具可以替代 &lt;FONT color=#66ffff&gt;office&lt;/FONT&gt;, 对于像我这样对office比较怵的人, 实在是一件好事!&lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt;  &lt;/P&gt; &lt;P&gt;在&lt;FONT color=#ffcc66&gt;PWSCF&lt;/FONT&gt;计算中, 用relax需要考虑到BFGS优化算法, 但是在AMD-Operon机器上这个算法出现计算错误, &lt;/P&gt; &lt;P&gt;不知道这是espresso的问题, 还是系统本身的问题? 两次测试来看, 这个错误的出现不仅仅是一个进程出现的错误, 因为给出的CRASH信息, 有两种不同的错误. 目前对3.1和3.0的计算都会出现同样的错误,  不知道如何解决这个问题, 因为我现在需要的计算就是做relax,出现这样的错误, 让人好生麻烦.  进行Pd step (66原子)体系的弛豫计算, 如果 ecut设置为 30 Ryd 需要的内存达到 8G,这是一个非常恐怖的数值. &lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt; &lt;/P&gt; &lt;P&gt;this document written by Writely&lt;/P&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115203611362061851?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115203611362061851/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115203611362061851' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115203611362061851'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115203611362061851'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/07/writely-office-office.html' title=''/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115185460684698714</id><published>2006-07-02T07:43:00.000-07:00</published><updated>2006-07-02T08:36:46.856-07:00</updated><title type='text'>Siesta 求解的默认值</title><content type='html'>今天对一个体系进行计算的时候才发现 siesta默认求解方法会根据体系的大小来设置. 当体系的原子数小于100个的时候, SolutionMethod会设置成diagon,而大于100个以上则 SolutionMethod默认设置成OrderN.&lt;br /&gt;&lt;br /&gt;    最保险的方法是, 在计算输入文件中把最主要的参数值写出来, 即使只需要给出默认值的, 一方面容易知道体系的计算内容, 另外在整理数据的时候也容易知道计算所设置的数值.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115185460684698714?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115185460684698714/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115185460684698714' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115185460684698714'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115185460684698714'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/07/siesta.html' title='Siesta 求解的默认值'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115169126150487253</id><published>2006-06-30T10:41:00.000-07:00</published><updated>2006-06-30T11:14:21.526-07:00</updated><title type='text'>如何来调整siesta的收敛速度</title><content type='html'>对于金属的计算, 需要调整各种参数来达到或者加速收敛的速度.&lt;br /&gt;在siesta中, 其中控制计算收敛的参数有OccupationFunction, ElectronicTemperature, DM.MixingWeights, DM.NumberPulay, DM.NumberKick, DM.KickMixingWeight 等.&lt;br /&gt;&lt;br /&gt;其中OccupationFunction有两种占据函数可选择, FD/MP 即Fermi-Dirac 分布和 Methfessel-Paxton分布. 对于金属体系来说,MP占据函数可以更精确地对Fermi面取点, 可以达到加速收敛的目的. 对于MP占据函数, 还有关键词OccupationMPOrder来控制MP占据函数的阶数.在手册上提到, 高阶的展开可以得到好的基态收敛结果.  因此在金属体系相关的计算中, 一般来说选择MP占据函数来达到加速收敛的目的.&lt;br /&gt;&lt;br /&gt;同时升高ElectronicTemperature也可以有助于加快体系计算的收敛速度, 但是需要了解一下, 升高温度对体系计算结果的影响.&lt;br /&gt;&lt;br /&gt;而剩下的几个参数, DM.MixingWeights 和DM.NumberPulay 同样是可以达到加速收敛的目的, 如果DM.NumberPulay 设置为0或1 则在计算中是采取线性混合, 为了加快收敛, 可以将DM.NumberPulay设置为4-10的样子(这个数目应当是根据中间结果分析来设置, 如果收敛很慢可以增加.) . 有的时候试图增加或者降低DM.MixingWeights, 其默认是0.25,  在收敛慢的情形下可以尝试将其设置在0.50的样子.&lt;br /&gt;&lt;br /&gt;这些都是今天初步分析的结果, 仍然需要进一步分析如何加快体系计算的收敛速度各种参数的设置. 一般来说如果SCF迭代次数超过30样子就应当试图改变DM.MixingWeights, DM.NumberPulay的默认值.  比较合理的速度是SCF迭代10次左右.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115169126150487253?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115169126150487253/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115169126150487253' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115169126150487253'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115169126150487253'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/siesta_30.html' title='如何来调整siesta的收敛速度'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115151751048118170</id><published>2006-06-28T10:38:00.000-07:00</published><updated>2006-06-28T11:00:01.036-07:00</updated><title type='text'>Physical Review Online Archive 新增外刊CrossRef</title><content type='html'>&lt;a href="http://www.aps.org"&gt;美国物理学会&lt;/a&gt;主办的物理评论在线档案数据库(&lt;a href="http://prola.aps.org/"&gt;PROA&lt;/a&gt;) 是物理学等领域内最主要的系列学术刊物数据库. 长期以来PROA的论文只提供其库内论文间的CrossRef. 最近PROA改版, 我发现它在论文引用方面已经有了很大的改进, 不仅扩展了引用的库, 而且还给出了年度引用数. 这两者对于评判论文影响都有很大的作用. 这是PROA的Announcement:&lt;br /&gt;&lt;br /&gt;CrossRef Forward Linking Now in PROLA.&lt;br /&gt;PROLA now displays citing articles from publishers that participate in the &lt;a href="http://www.crossref.org/"&gt;CrossRef&lt;/a&gt; forward linking program. The interface for citing articles has also been completely redone to allow convenient browsing by journal and year.&lt;br /&gt;&lt;br /&gt;就Phys.Rev.B相关文章我注意到新增的期刊有 &lt;a href="http://prola.aps.org/forward/PRB/v40/i6/p3979_1?journal=Journal+of+Physics+Condensed+Matter"&gt;Journal of Physics Condensed Matter&lt;/a&gt; ,&lt;a href="http://prola.aps.org/forward/PRB/v40/i6/p3979_1?journal=physica+status+solidi+(b)"&gt;physica status solidi (b)&lt;/a&gt;, &lt;a href="http://prola.aps.org/forward/PRB/v40/i6/p3979_1?journal=International+Journal+of+Quantum+Chemistry"&gt;International Journal of Quantum Chemistry&lt;/a&gt; ,&lt;a href="http://prola.aps.org/forward/PRB/v40/i6/p3979_1?journal=Nanotechnology"&gt;Nanotechnology&lt;/a&gt;, &lt;a href="http://prola.aps.org/forward/PRB/v40/i6/p3979_1?journal=Reports+on+Progress+in+Physics"&gt;Reports on Progress in Physics&lt;/a&gt;等. 也就是说以后所有参加CrossRef的期刊的引用都可以在PROA杂志上看到, 而不用老Refer to ISI. 或许这是打破sci崇拜的重要一步.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115151751048118170?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115151751048118170/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115151751048118170' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115151751048118170'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115151751048118170'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/physical-review-online-archive.html' title='Physical Review Online Archive 新增外刊CrossRef'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115125462118697746</id><published>2006-06-25T09:44:00.000-07:00</published><updated>2006-06-25T09:57:01.196-07:00</updated><title type='text'>什么叫semicore in valence</title><content type='html'>Siesta方法里有所谓的semicore in valence . 今天总算明白了一些,  所谓的semicore 就是在赝势生成计算中考虑核区外壳层电子, 其相应的基组采用最小基即可. &lt;br /&gt;&lt;br /&gt;     通常在TM计算中需要考虑到semicore的作用, 相应的赝势和基组要相对应.  这就意味着需要细致考虑TM的电子组态, 以及基组每个壳层的半径截断. 这是一个细致活, 需要调研和对比. 现在的方法是读一读文献和siesta主页给出的赝势和基组文件.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115125462118697746?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115125462118697746/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115125462118697746' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115125462118697746'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115125462118697746'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/semicore-in-valence.html' title='什么叫semicore in valence'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115125373957633614</id><published>2006-06-25T05:12:00.000-07:00</published><updated>2006-06-25T09:42:19.626-07:00</updated><title type='text'>Siesta 基组的构建和评估(1)</title><content type='html'>由于计算上出现lapack routine错误, 因此有必要自己构建优化好的基组.   刚刚用别人的基组做了一下, 应该是基组的问题, 而不是Lapack routine的问题.  所以只能把basis的设置好好学习设置, 并且做相关的调研.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115125373957633614?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115125373957633614/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115125373957633614' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115125373957633614'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115125373957633614'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/siesta-1.html' title='Siesta 基组的构建和评估(1)'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115123173528030780</id><published>2006-06-25T01:52:00.000-07:00</published><updated>2006-06-25T03:35:35.290-07:00</updated><title type='text'>libgoto V.S. ACML</title><content type='html'>用siesta 测试了一下 在AMD-Operon 248 /RHEL-AS 4.0 平台上.&lt;br /&gt;其中BLAS用libgoto-operon64, lapack系统自带.   对Pt bulk体系进行Relax, 串行需要时间250s . 而直接链acml库, 需要的时间是150s的样子.&lt;br /&gt;&lt;br /&gt;    值得注意的是, libgoto需要的内存非常大, 达到500M的样子. 而且内存占用非常不稳定, 一直振荡.  相应acml库下, 内存的占用不过20M.    这是很有意思的比较,  通常都说libgoto库是优化和速度最好的BLAS库 ,但是对于siesta并不如此.  或许还可以考虑lapack库用siesta自带.有空再做一番比较.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115123173528030780?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115123173528030780/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115123173528030780' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115123173528030780'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115123173528030780'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/libgoto-vs-acml.html' title='libgoto V.S. ACML'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115122506400549201</id><published>2006-06-25T01:29:00.000-07:00</published><updated>2006-06-25T01:44:24.013-07:00</updated><title type='text'>Siesta计算 :对角化错误</title><content type='html'>今天检验Pt赝势,测试DZP基组出现同于Pd DZP基组计算的错误. 这个错误在&lt;br /&gt;siesta-2.0 与siesta-1.3 同样出现.  Google 搜索得到的问题是在于 lapack数学库函数LAPACK -ZHEGV的出错. &lt;br /&gt;&lt;blockquote&gt;&lt;/blockquote&gt;&lt;strong&gt;  ZHEGV&lt;/strong&gt; computes all the eigenvalues, and optionally, the eigenvectors of a complex generalized Hermitian-definite eigenproblem, of the form  A*x=(lambda)*B*x,  A*Bx=(lambda)*x,  or B*A*x=(lambda)*x.  Here A and B are assumed to be Hermitian and B is also  positive definite.&lt;br /&gt;&lt;br /&gt;也就是说广义复Hermitian 矩阵的本征值和本征矢的计算出错. 看来得把ACML数学库更新一下, 再做一番测试. 如果不行的话, 或许再把siesta的基组调试一下进行测试或许有可能成功.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115122506400549201?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115122506400549201/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115122506400549201' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115122506400549201'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115122506400549201'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/siesta_25.html' title='Siesta计算 :对角化错误'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115110637584353076</id><published>2006-06-23T16:32:00.000-07:00</published><updated>2006-06-25T01:27:05.593-07:00</updated><title type='text'>检验赝势的初步计算</title><content type='html'>赝势的好坏是后验的. 虽然可以通过ATOM提供的一些程序来做检验, 但是其优劣仍然需要通过对一些已知体系进行计算来检验.&lt;br /&gt;&lt;br /&gt;今天对Pd金属的赝势进行了初步的检验计算, 基本的收获是cutoff的选择, k-points的选择, 占据函数的选择等的熟悉. 在我一系列的计算中选取的cutoff能量是200 Ryd, 后来查阅一些文献发现150 Ryd 的截断就足够的, 而300 Ryd的能量是算非常精细的计算了. k点的选取也非常重要的, 今天最开始疏忽没有设置k格点,导致一系列计算都仅包含gamma点, 因此可以从计算的进程看到收敛非常振荡,总能出现正值.&lt;br /&gt;&lt;br /&gt;对于金属, 占据函数的选取也有考虑的,由于金属的Fermi面比较复杂, 因此对Fermi的精确取点需要比较好的占据函数. 在siesta中有fermi-dirac和MP两种占据函数, 高阶情形的MP占据函数对金属情形有好的处理. 实际上在Fermi 面如何取点有多种处理, 特别是对金属情形. 需要读一些相关文献进行,补充这方面的知识.&lt;br /&gt;&lt;br /&gt;计算得到的晶格长度跟实验比较,差0.2A的量级, 大概是6%的误差(LDA),这个是在误差允许范围内么? (未必, 另外的计算显示晶格变长, 不知道是什么问题)&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115110637584353076?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115110637584353076/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115110637584353076' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115110637584353076'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115110637584353076'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/blog-post.html' title='检验赝势的初步计算'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115110523523314586</id><published>2006-06-23T16:24:00.000-07:00</published><updated>2006-06-23T16:29:30.543-07:00</updated><title type='text'>Siesta计算的一个典型错误</title><content type='html'>今天做Pd赝势的检测计算, 想用DZP基组来算晶格的优化计算, siesta计算中出现这样的诡异错误:&lt;br /&gt; The factorization of B could not be completed and no eigenvalues or  eigenvectors were computed.&lt;br /&gt;Terminating due to failed diagonalisation.&lt;br /&gt;&lt;br /&gt;把基组设置为DZ则不会出现这个错误, 计算可以顺利地完成.  不知道这个问题如何解决, 已经向邮件组询问了这个问题,晚上看看有没有人回答?&lt;br /&gt; 另外准备在intel的机器上进行检测.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115110523523314586?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115110523523314586/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115110523523314586' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115110523523314586'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115110523523314586'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/siesta.html' title='Siesta计算的一个典型错误'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115102587850255971</id><published>2006-06-22T17:32:00.000-07:00</published><updated>2006-06-23T09:55:41.320-07:00</updated><title type='text'>Gnuplot的学习笔记(2)</title><content type='html'>Gnuplot可以通过scripts来实现个性化的设置. 通过命令set可以实现对图形各个参数进行有效的设置, 如title, xlabel, ylabel等. 以下就是一个典型的例子:&lt;br /&gt;&lt;br /&gt;set grid nopolar&lt;br /&gt;set grid xtics nomxtics ytics nomytics noztics nomztics nox2tics nomx2tics noy2tics nomy2tics nocbtics nomcbtics&lt;br /&gt;set grid layerdefault&lt;br /&gt;set samples 21, 21&lt;br /&gt;set isosamples 11, 11&lt;br /&gt;set title "3D gnuplot demo" 0.000000,0.000000 font ""&lt;br /&gt;set xlabel "X axis" -3.000000,-2.000000 font ""&lt;br /&gt;set xrange [ -10.0000 : 10.0000 ] noreverse nowriteback&lt;br /&gt;set ylabel "Y axis" 3.000000,-2.000000 font ""&lt;br /&gt;set yrange [ -10.0000 : 10.0000 ] noreverse nowriteback &lt;a href="http://lochmeters.bloggerspaces.com/uploaded_images/surface1.5-780537.png"&gt;&lt;/a&gt;&lt;br /&gt;set zlabel "Z axis" -5.000000,0.000000 font ""&lt;br /&gt;splot x**2+y**2, x**2-y**2, x*y &lt;a href="http://lochmeters.bloggerspaces.com/uploaded_images/surface1.3-757317.png"&gt;&lt;img style="FLOAT: right; MARGIN: 0px 0px 10px 10px; WIDTH: 421px; CURSOR: hand; HEIGHT: 296px" height="262" alt="" src="http://lochmeters.bloggerspaces.com/uploaded_images/surface1.3-752501.png" width="312" border="0" /&gt;&lt;/a&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115102587850255971?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115102587850255971/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115102587850255971' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115102587850255971'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115102587850255971'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/gnuplot2.html' title='Gnuplot的学习笔记(2)'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115102113213596076</id><published>2006-06-22T16:19:00.000-07:00</published><updated>2006-06-22T17:27:08.853-07:00</updated><title type='text'>Gnuplot的学习笔记(1)</title><content type='html'>Gnuplot是一个很有效,简单的画图和数据处理工具. 基本上在各种不同的发行版本的Linux上都带有该工具. 长期以来只知道用gnuplot来简单地处理数据, 如简单地可视化x-y数组. 其实gnuplot功能是比较强大的, 因此, 在这里主要是记录一些有趣和有效的功能.&lt;br /&gt;&lt;br /&gt;Gnuplot提供一系列数学函数, 这些函数基本上可以涵盖Fortran/C语言所提供的数学函数功能, 而在gnuplot中是按照C语言的规范来给出数学函数的, 如 sin(x)/cos(x),log(x)/log10(x),tanh(x)/tan(x)等. 在Linux X11环境下, 直接命令行输入gnuplot 即可进入gnuplot-x11环境, 命令plot sin(x)/cos(x)即可实现函数的可视化.&lt;br /&gt;&lt;br /&gt;gnuplot 中plot 和splot两个函数分别处理二维函数/数据和三维表面图形/数据.&lt;br /&gt;plot的语法 &lt;strong&gt;plot {[ranges]} {[function][data]} {axes } { title-spec} {,[function] ...} .&lt;/strong&gt; 如前所述, 简单地plot sin(x)即可实现 sin(x)函数的图形, 而&lt;br /&gt;plot sin(x), tan(x) 则可实现两个函数图形可视化, 也可以通过title来标记图形, 如plot [t=1:10] sin(x) title 'sin' 即实现sin(x)在[1,10]的范围的图形, 并用sin标记. 可以用help plot来查询plot的功能.  同样可以处理数据文件, 假设数据为3列数据存在file.dat中, 我们则可以用 plot "file.dat"  来绘画该数据对应的图形. gnuplot可以实现多种图形格式, 如lines, points, linespoints, impulses, steps等, 如 plot 'file.dat'  with lines 即实现线图形.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115102113213596076?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115102113213596076/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115102113213596076' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115102113213596076'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115102113213596076'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/gnuplot1.html' title='Gnuplot的学习笔记(1)'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115101203679704708</id><published>2006-06-22T14:21:00.000-07:00</published><updated>2006-06-22T15:20:32.826-07:00</updated><title type='text'>利用ATOM生成SIESTA赝势(3)</title><content type='html'>&lt;div align="left"&gt;有了赝势的生成和评测标准, 我们接下来就是要利用atom程序来生成我们所需要的赝势. Atom程序本身提供了赝势的检测功能, 因此输入文件有ae, pg, pe, pt四个关键词来标明计算的类型. ae是全电子计算, 对应的输入文件如下 &lt;/div&gt;&lt;div align="left"&gt;&lt;strong&gt;ae  Ti&lt;br /&gt;Ti car&lt;br /&gt;0.0&lt;br /&gt;3  4&lt;br /&gt;3  0 2.0 0.0&lt;br /&gt;3  1 6.0 0.0&lt;br /&gt;3  2 2.0 0.0&lt;/strong&gt;&lt;br /&gt;&lt;/div&gt;&lt;div align="left"&gt;第一行 ae是指明计算任务, Ti则是这个任务的标记. 第二行中Ti是元素符号, 而car则是计算中的泛函, r是考虑了元素的相对论效应. 而第三行中的0.0是程序默认,在大多数计算中设定为0.0即可. 第四行则是标明原子的组态, 3 是核区态结构, 其意义是1S2/2S2/2P6 三个壳层电子, 而4 则是计算中考虑的价态壳层, 在这里是 3S2/3P6/3d2. 显然第五行开始, 第一列是主量子数n, 第二列是角动量量子数, 第三列和第四列则是指定该壳层中的电子数,第四列的作用是考虑自旋极化下的配对.&lt;br /&gt;&lt;strong&gt;&lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;&lt;/strong&gt;用 pg关键字生成赝势, pg计算中, 需要在ae计算输入文件的基础上加入核区截断半径, 以及考虑激化情况加入的电子壳层. 典型的pg输入文件为&lt;br /&gt;&lt;strong&gt;pg Ti&lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;tm2 2.00&lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;Ti car&lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;0.0 0.0 0.0&lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;3 4 &lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;3 0 2.0 0.0&lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;3 1 6.0 0.0&lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;3 2 2.0 0.0&lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;4 3 0.0 0.0 &lt;/strong&gt;&lt;br /&gt;&lt;strong&gt;1.3 1.3 1.3 2.00 0.0 0.0&lt;/strong&gt;&lt;br /&gt;可见pg计算的输入文件与ae计算的输入文件大体上是接近的, 只是增加了赝势生成方案和核区截断半径的数值. 在ATOM中提供hsc/ker/tm2三种方案生成赝势, 通常选择tm2方案. 截断半径的设置顺序是按照s/p/d/f的顺序以原子单位(bohr)给出. 对于pe计算, 则是对应核区修正, 其与pg计算的差别除了计算关键词不同之外, 还在于最后一行截断半径的最后两列数值的指定上. 我们可以将其标记为rcore_flag/rcore . 如果rcore_flag输入为负值, 则在核区计算中考虑整个核区电荷. rcore用来指定电子密度二阶微分连续处的半径(shy, 这点我还有点糊涂,现在基本的认知是相当于指定一个核区的半径,一般数值越大核区修正越大).&lt;br /&gt;&lt;br /&gt;显然最后一个pt是用来指定赝势的检测计算的. atom程序认为具有合适转移特性的赝势, 可以将全电子的能级和波函数在任意环境下重复出来. 这就可以在不同的原子组态情形下考虑其本征值和激发能量来评测赝势的转移特性. 其输入文件格式是指定不同原子组态情形的ae计算与之对应的pt计算. pt计算输入格式同于ae计算, 除了关键词pt不同之外.  通过检验ae计算和pt计算的能级差值来判断赝势的优劣, 一般来说能量差值小于1mRyd即说明该赝势有好的转移特性.&lt;br /&gt;&lt;strong&gt;&lt;/strong&gt;&lt;br /&gt;&lt;/div&gt;&lt;strong&gt;&lt;/strong&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115101203679704708?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115101203679704708/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115101203679704708' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115101203679704708'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115101203679704708'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/atomsiesta3.html' title='利用ATOM生成SIESTA赝势(3)'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115100905595827632</id><published>2006-06-22T13:39:00.000-07:00</published><updated>2006-06-22T14:17:55.046-07:00</updated><title type='text'>利用ATOM生成SIESTA赝势(2)</title><content type='html'>&lt;strong&gt;对于怎么样来生成赝势,&lt;a href="http://www.nest.sns.it/~giannozz"&gt; Paolo Giannozzi&lt;/a&gt; 给出如下几个步骤:&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;&lt;/strong&gt;&lt;strong&gt;&lt;/strong&gt;&lt;strong&gt;&lt;/strong&gt;&lt;strong&gt;&lt;blockquote&gt;&lt;/blockquote&gt;1. choose a suitable density functional&lt;br /&gt;2. choose valence and core states&lt;br /&gt;3. choose a reference configuration&lt;br /&gt;4. generate all-electron wavefunctions&lt;br /&gt;5. choose the matching radii&lt;br /&gt;6. choose if you want the core corrections&lt;br /&gt;7. generate the pseudopotential&lt;br /&gt;8. check its transferability&lt;br /&gt;9. check the required cutoff&lt;br /&gt;10. check its separable form&lt;br /&gt;&lt;br /&gt;&lt;blockquote&gt;&lt;/blockquote&gt;&lt;strong&gt;显然, 不同的泛函会有不同的赝势, ATOM中有CA,WI,HL,GL,BH&lt;/strong&gt; 四种可选的LDA泛函, 但是手册里明确指出LDA最好的选择是CA,此外也有PB GGA泛函. 而自旋极化和相对论效应则是通过s/r来指定, 其中r包含了自旋极化效应. 即CAR 对应的是考虑相对论效应的CA泛函.&lt;br /&gt;&lt;br /&gt;赝势的目的就是得到一个势来最好地描述其对价电子的行为, 因此如何选择价态与核区态对赝势行为有本质的影响. 基本上的做法是根据原子物理的原子组态来选择, 也就是所谓的壳层结构. 在确定了价态与核区态之后, 我们需要选择核区半径来截断核区电子的波函数,也就是所谓match radii, 这个半径选取直接影响生成赝势的行为,也就是所谓转移特性(transferability)和效率. 这是后验的, 因此需要在后续的检测里不断调整来得到transferability 和effectivity的平衡.&lt;br /&gt;&lt;br /&gt;此外有些元素还需要考虑核区的修正,一般情况下在截断半径外'赝'-电子密度应与电子密度相等, 并有一个平滑形式(三角函数形式). 考虑核区修正时,则拟合一个 e指数函数, 并要求二次导数连续.&lt;br /&gt;&lt;br /&gt;做上述考虑之后, 我们即可以利用atom程序来生成赝势, 继而做检测其转移特性和效率. 赝势的优劣是通过对已知体系的计算来评测的.&lt;br /&gt;&lt;br /&gt;&lt;br /&gt;&lt;/strong&gt;&lt;/strong&gt;&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115100905595827632?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115100905595827632/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115100905595827632' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115100905595827632'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115100905595827632'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/atomsiesta2.html' title='利用ATOM生成SIESTA赝势(2)'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115098731246899988</id><published>2006-06-22T07:38:00.000-07:00</published><updated>2006-06-22T09:08:49.950-07:00</updated><title type='text'>利用ATOM生成SIESTA赝势(1)</title><content type='html'>&lt;strong&gt;Siesta&lt;/strong&gt;在Pseudo目录下提供了赝势生成程序. Siesta采用的是模守恒赝势(Norm-conserving Pseudopotentials).&lt;br /&gt;在这里主要是讨论根据其提供的赝势生成程序进行赝势的生成和评测检验.&lt;br /&gt;Pseudo/atom/Tutorial目录下,已经有ae.sh, pg.sh, pt.sh等赝势生成和检测的scripts, 我们需要根据其所在的目录来设置一些环境变量,如程序atm的绝对路径, 以及gnuplot的一些scripts即*.gps/*.gplot所在的路径. 根据ae.sh/pg.sh/pt.sh所涉及的环境变量, 我们也可以在$HOME/.bashrc文件下设置ATOM_PROGRAM/ATOM_UTILS_DIR. 在这里需要注意的是程序本身提供的scripts采用sh环境, 在我的系统中sh环境下识别不出通配符*,因此运行scripts会出现不能拷贝*.gp*文件的错误. 为了解决这个问题, 我把shell环境改为bash,即可以避免这个问题. 设置好环境变量之后,我们就可以利用程序所提供的pg.sh/ae.sh/pt.sh来生成赝势和评测赝势了.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115098731246899988?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115098731246899988/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115098731246899988' title='1 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115098731246899988'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115098731246899988'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/atomsiesta1.html' title='利用ATOM生成SIESTA赝势(1)'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>1</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-115040889042747443</id><published>2006-06-15T14:55:00.000-07:00</published><updated>2006-06-15T15:16:31.963-07:00</updated><title type='text'>Done, 简陋的Blog</title><content type='html'>折腾一晚上, 而且还看了html代码, 才把blogger的发布模板改成现在的样子.&lt;br /&gt;其实以前也学过一些html语言,当时也是图新鲜,学了些皮毛, 也不怎么用这种scripts来做事情, 到现在自然就淡忘了. 话说过来,很多新鲜的事情, 开始都会让你激动, 而实际中的操作却是繁复与沉闷的. 这些年虽然没怎么做过事情, 可是这种教训却是应当记住的,因为很多时候这是所谓个人执行力的体现, 年龄大了, 很多时候应当知道自己的如何去面对新事物, 如何去学习新知识.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-115040889042747443?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/115040889042747443/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=115040889042747443' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115040889042747443'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/115040889042747443'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/done-blog.html' title='Done, 简陋的Blog'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-114986985143977804</id><published>2006-06-09T09:16:00.000-07:00</published><updated>2006-06-09T09:32:52.543-07:00</updated><title type='text'>有趣的blogger</title><content type='html'>这样挺好的, 只要自己有了空间,可以随时把blog发布到自己的空间.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-114986985143977804?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/114986985143977804/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=114986985143977804' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/114986985143977804'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/114986985143977804'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/blogger.html' title='有趣的blogger'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry><entry><id>tag:blogger.com,1999:blog-29480451.post-114986953921775727</id><published>2006-06-09T09:11:00.000-07:00</published><updated>2006-06-15T14:18:21.560-07:00</updated><title type='text'>My logo</title><content type='html'>&lt;a href="http://lochmeters.512j.com/uploaded_images/0-709441.jpg"&gt;&lt;img style="FLOAT: right; MARGIN: 0px 0px 10px 10px; CURSOR: hand" alt="" src="http://lochmeters.512j.com/uploaded_images/0-705779.jpg" border="0" /&gt;&lt;/a&gt;&lt;br /&gt;这个logo是利用MSN cartoon工具得到的.&lt;br /&gt;挺有趣的, 当然比本人是帅多了.&lt;div class="blogger-post-footer"&gt;&lt;img width='1' height='1' src='https://blogger.googleusercontent.com/tracker/29480451-114986953921775727?l=lochmeter.blogspot.com' alt='' /&gt;&lt;/div&gt;</content><link rel='replies' type='application/atom+xml' href='http://lochmeter.blogspot.com/feeds/114986953921775727/comments/default' title='Post Comments'/><link rel='replies' type='text/html' href='http://www.blogger.com/comment.g?blogID=29480451&amp;postID=114986953921775727' title='0 Comments'/><link rel='edit' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/114986953921775727'/><link rel='self' type='application/atom+xml' href='http://www.blogger.com/feeds/29480451/posts/default/114986953921775727'/><link rel='alternate' type='text/html' href='http://lochmeter.blogspot.com/2006/06/my-logo.html' title='My logo'/><author><name>lochmeters</name><uri>http://www.blogger.com/profile/12356534625704941585</uri><email>noreply@blogger.com</email><gd:image rel='http://schemas.google.com/g/2005#thumbnail' width='16' height='16' src='http://img2.blogblog.com/img/b16-rounded.gif'/></author><thr:total>0</thr:total></entry></feed>
