Sorry, I just realized that I missed the MD part. Another not so easy way is to use C/fortran to read/convert the binary MD file, finding the coordinate and the velocity of the - say, 500th step, then to use all that info to create a new XV file. I don't think that you can just directly use the MD file.
Unless you are really concerned of the accuracy and rigor, I would suggest to use the method I suggested previously, which could be done in about 1 minute. Chih-Chien Wang Research Assistance Institute of Atomic and Molecular Science, Academia Sinica, Taipei, Taiwan [email protected] ----- Original Message ----- From: Chih-Chien Wang To: [email protected] Sent: Tuesday, February 03, 2015 3:02 PM Subject: Re: [SIESTA-L] restarting a job In that case, the easiest way might be find the coordinate of the - say 500th step, and just copy that into your new .fdf file. Unfortunately, .XV file, according to the webpage, would be renewed each step. So you probably wouldn't be able to use it unless it is the last step. http://departments.icmab.es/leem/siesta/Documentation/Manuals/manual-2.0/node23.html I am not exactly certain about the DM file, but my understanding is that .DM file is refreshed each step, too. Chih-Chien Wang Research Assistance Institute of Atomic and Molecular Science, Academia Sinica, Taipei, Taiwan [email protected] ----- Original Message ----- From: Ting Zheng To: [email protected] Sent: Tuesday, February 03, 2015 4:46 AM Subject: Re: [SIESTA-L] restarting a job Thanks for your reply. In my way, I just want to restart the job at special point (not the end of the MD run). So I need the .DM, .XV (easy to get it) and .MD file to restart it. So how do I get the .DM and .MD file at special time point? Thanks Ting 2015-02-02 2:52 GMT-06:00 Chih-Chien Wang <[email protected]>: According to the webpage, it seems to me that MD file would only have the coordinate and velocity of atoms. And since .MD is unformatted, one might consider to use fortran/C to convert the MD file. Is velocity very important in this case? If not, there are much easier ways to get around with only atomic coordinates. Chih-Chien Wang Research Assistance Institute of Atomic and Molecular Science, Academia Sinica, Taipei, Taiwan [email protected] ----- Original Message ----- From: Ting Zheng To: [email protected] Sent: Sunday, February 01, 2015 4:29 AM Subject: [SIESTA-L] restarting a job Dear siesta users, I have a problem with restarting a job. As known to all, we need .XV, .DM, .MD files to restart a siesta MD job. It seems that the .MD file seems to have all the messages during the whole Molecular dynamics procedure since the size of that file just increasing all the time. But know I have a special problem that I have to restart the job from a point which is not the final one. Let me free my question as an example. If I have conducted a molecular dynamics job that lasted 1000 steps. After analyzing the data, I want to rerun the job from s special point, say step=500. I can make the special XV file from the out file, but how should I make the special MD file which start at step=500? Thanks Ting
