| 查看: 789 | 回复: 3 | ||||
| 当前主题已经存档。 | ||||
| 当前只显示满足指定条件的回帖，点击这里查看本话题的所有回帖 | ||||
[交流]
谁有WIEN2K在Linux下的安装过程说明？
|
||||
|
以前听人说有位高人在SUSE上安装成功后，写过一个安装过程说明，不知谁有这个文件，共享一下多谢！ email: cougarlz@lut.cn [ Last edited by wuli8 on 2008-11-25 at 16:02 ] |
回复此楼» 收录本帖的淘帖专辑推荐
 第一性原理相关文档 |
» 猜你喜欢
国自然评审意见
已经有3人回复
-
有多少人是今天查系统知道结果的？
已经有18人回复
-
为什么资助数各大高校都创新高，自己申请怎么就这么难
已经有12人回复
-
有没有仍没收到信息的
已经有3人回复
-
导师吐槽：我怎么摊上了这么个极品研究生！
已经有8人回复
-
基金不中，共勉
已经有11人回复
-
麻烦专家们看看评委们的意见（F口面上）
已经有9人回复
-
2026年叶企孙基金
已经有6人回复
-
国自然面上复盘~欢迎讨论
已经有13人回复
-
基金系统什么内容也没有
已经有9人回复
4楼2007-04-06 06:08:05
bluce
木虫 (正式写手)
- 应助: 0 (幼儿园)
- 金币: 4258.1
- 红花: 1
- 帖子: 440
- 在线: 183.6小时
- 虫号: 293064
- 注册: 2006-11-04
- 性别: GG
- 专业: 凝聚态物性 II ：电子结构
2楼2006-11-24 16:20:11
haiw201
木虫 (正式写手)
- 应助: 0 (幼儿园)
- 贵宾: 0.2
- 金币: 2860.3
- 散金: 1
- 红花: 1
- 帖子: 635
- 在线: 36.6小时
- 虫号: 147102
- 注册: 2005-12-27
- 性别: GG
- 专业: 凝聚态物性 II ：电子结构
★ ★
wuli8(金币+2,VIP+0):3X!(*^__^*) ……
wuli8(金币+2,VIP+0):3X!(*^__^*) ……
|
Howto-WIEN2k v 0.12 May 8 2006 ============================= 0. make sure you got some good compiler and libraries, e.g. ifort9 and mkl8 or later (free for academic use from intel) on your computer If you are the administrator of the computer update '/etc/ld.so.conf' with paths to the libraries you are using: /opt/intel/fc/9.0/lib /opt/intel/mkl/8.0.1/lib/32 (just an example above) and run 'ldconfig' check that the compiler is in your path (if you enter 'ifort' is it found?) 1. copy a recent distribution of WIEN2k, for example 'WIEN2k_06.2_060309.tar' located at '/HOME/ikeno' on "earth" or at '/home/weine' on "earth" For the following steps also refer to the WIEN2k manual ======================================================= also check http://www.wien2k.at/reg_user/index.html for F.A.Q. and searching the WIEN2k mailing list 2. create a directory to install WIEN2k, for example 'WIEN2k/06.2' move your tar-file there, expand files etc: $ tar xvf WIEN2k_06.2_060309.tar $ gzip -d *.gz $ ./expand_lapw clean up a bit by: $ mkdir tars $ mv *.tar tars 3. configure and compile WIEN2k by running $ ./siteconfig you can choose for example system 'J' modify the settings 'O', 'L' and 'R', I use the following compiler flags on a 64-bit Opteron system: O (FOPT) -FR -w -mp1 -prec_div -pad -ip -DINTEL_VML -O3 -xW L (LDFLAGS) -L/opt/intel/fce/9.0/lib -i-static -lguide_stats -lsvml -lpthread R (R_LIBS) -L/opt/intel/mkl/8.0.1/lib/em64t -lmkl_lapack -lmkl_em64t -lguide -lpthread on a 32-bit Pentium 4 system I just modify: -FR -w -mp1 -prec_div -pad -ip -DINTEL_VML -O3 -xW -L/opt/intel/fc/9.0/lib -i-static -lguide_stats -lsvml -lpthread -L/opt/intel/mkl/8.0.1/lib/32 -lmkl_lapack -lmkl_ia32 -lguide -lpthread you can also ask Ikeno-san for good compiler flags on your system you might also try the default flags for system 'J' and check if they work use the appropriate version of ifort, depending if you use 32-bit or 64-bit system for itanium systems the compiler and flags would be a bit different I found the flags I'm using by reading 'Fecher_CompileIntel.pdf' the parameters NMATMAX and NMATMUL should reflect the RAM available on your computer, e.g. on my 8 GB RAM opteron I used (though I'm no expert) NMATMAX = 80000 NMATMUL = 8000 this also depends on 4. configure the user environment $ ./userconfig Questions about running WIEN2k ============================== - first ask people in the group who are using it (Yoshioka-san, Kuwabara-san and me, as far as I know) - refer to the WIEN2k manual, website (the F.A.Q. and mailinglist) Supercell calculations ====================== One easy method for creating supercells: (1) download appropriate experimental .cif file from the open database: http://sdpd.univ-lemans.fr/cod/index.html (2) import the .cif file in the 'MS modeling' program for 'Windows' 'File' ->'Import' a supercell is easily created by using: 'Build' ->'Symmetry' ->'Supercell' to export back to another .cif file, just use 'File' ->'Export' (3) before uploading the .cif file into WIEN2k one need to undosify the file, use $ perl -pi -e 's/\r\n/\n/g' name_of_file (4) copy the .cif file to the workdir and use 'StructGenTM' beware that when you load the .cif file into WIEN2k, the last digits will be set to zeros, e.g. '0.22222000' instead of '0.22222222' this should be fixed - by hand, or much more preferably by a script (I use a script called 'mksr' which search and replace patterns in files) (5) for 'StructGenTM' to understand that you want to make a supercell calc you must break the symmetry somehow (if you're not doing impurity calc) Running jobs directly (without the webinterface) ================================================ you may not always be able to run the webinterface for WIEN2k on the machine you want to run your jobs on. Fortunately it is very easy to use anyway, e.g.: (1) copy the .struct file to your workdir (2) $ init_lapw follow the instructions, it might be useful to have 'export EDITOR="vi"' in .bashrc (3) $ run_lapw -ec 0.0001 -i 30 start calcs, flags here just as example. A (better?) alternative might be to run the jobs from a script instead, e.g. in the file 'run': ------------------------------------- # my WIEN2k script-file, 8 May 2006 cd /home/path_to_my_workdir run_lapw -ec 0.0001 -i 30 cd /home/path_to_my_workdir2 run_lapw -ec 0.0001 -i 30 exit ------------------------------------- $ chmod 744 run $ nohup ./run & by using 'nohup' the script will continue to run even after you logout from the computer In the program 'MS modeling' from Accelrys available on Windows systems one can create supercells Using the open database it is possible to download .cif files with experimental information on structures One can use MSm to import .cif files and with them as basis create supercells, later one can export the supercell as a .cif file, which can be read by WIEN2k OBS: when using .cif as a template for creating a WIEN2k .struct, the last three decimals in .struct will be zeros, you have to put in the correct decimals by hand (or with some script) The freely available program xcrysden can read WIEN2k .struct-files and show the structure. |
3楼2007-04-05 18:51:00
投票:
