| 查看: 288 | 回复: 2 | |||
| 当前主题已经存档。 | |||
maad木虫 (小有名气)
|
[交流]
【求助】VASP并行出问题
|
||
|
编译用的是如下环境（Centos 5.2） 1、MKL 2、ifort 3、MPICH2.1.0.8 参照的该贴：http://muchong.com/bbs/viewthread.php?tid=1539206 VASP编译成功以后，运行时除了问题, 速度特别慢，查看了一下LOG文件，发现每个进程似乎都在做相同的运算，并不是并行运算。 而且，我们之前编译的可以正常进行并行运算的VASP中的log文件里是 running on 4nodes，这里出现了4个1nodes。 以下为文件： -catch_rsh /opt/gridengine/default/spool/compute-4-1/active_jobs/1009.1/pe_hostfile compute-4-1 compute-4-1 compute-4-1 compute-4-1 Warning: no access to tty (Bad file descriptor). Thus no job control in this shell. Tue Oct 13 09:30:13 CST 2009 LEXCH = CA LEXCH = CA TITEL = US C TITEL = US H running on 1 nodes distr: one band on 1 nodes, 1 groups vasp.4.6.21 23Feb03 complex running on 1 nodes distr: one band on 1 nodes, 1 groups vasp.4.6.21 23Feb03 complex running on 1 nodes distr: one band on 1 nodes, 1 groups running on 1 nodes distr: one band on 1 nodes, 1 groups vasp.4.6.21 23Feb03 complex vasp.4.6.21 23Feb03 complex POSCAR found : 2 types and 60 ions POSCAR found : 2 types and 60 ions POSCAR found : 2 types and 60 ions POSCAR found : 2 types and 60 ions WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 LDA part: xc-table for Ceperly-Alder, standard interpolation LDA part: xc-table for Ceperly-Alder, standard interpolation LDA part: xc-table for Ceperly-Alder, standard interpolation LDA part: xc-table for Ceperly-Alder, standard interpolation POSCAR, INCAR and KPOINTS ok, starting setup POSCAR, INCAR and KPOINTS ok, starting setup POSCAR, INCAR and KPOINTS ok, starting setup POSCAR, INCAR and KPOINTS ok, starting setup FFT: planning ... 9 reading WAVECAR FFT: planning ... 9 reading WAVECAR FFT: planning ... 9 reading WAVECAR FFT: planning ... 9 reading WAVECAR entering main loop N E dE d eps ncg rms rms(c) entering main loop N E dE d eps ncg rms rms(c) entering main loop N E dE d eps ncg rms rms(c) entering main loop N E dE d eps ncg rms rms(c) |
回复此楼» 猜你喜欢
地球科学部D01口青年基金，最低几A几B几C才能有几率中呀。
已经有5人回复
-
航天502所 高瑛珂博士 婚内征婚 欺骗女性开房
已经有27人回复
-
26/27申博
已经有4人回复
-
博士申请
已经有5人回复
anin
木虫 (著名写手)
如来的神掌
- 应助: 4 (幼儿园)
- 贵宾: 0.082
- 金币: 2950.2
- 散金: 20
- 红花: 6
- 帖子: 1833
- 在线: 184.6小时
- 虫号: 69399
- 注册: 2005-05-21
- 专业: 金属材料的微观结构
2楼2009-10-13 17:54:21
maad
木虫 (小有名气)
- 应助: 6 (幼儿园)
- 金币: 1649.2
- 散金: 769
- 红花: 9
- 帖子: 245
- 在线: 470小时
- 虫号: 565636
- 注册: 2008-05-29
- 专业: 膜生物化学与膜生物物理学
|
应该是的。换了个MPICH后，这个问题就消失了，但计算速度依旧没有提高。而且算一段时间后，CPU的负载高的吓人，节点也处于假死的状态，很诡异。再往下就不怎么算得动了 LOG文件如下： compute-4-0 compute-4-0 compute-4-0 compute-4-0 Warning: no access to tty (Bad file descriptor). Thus no job control in this shell. Tue Oct 13 22:35:29 CST 2009 LEXCH = CA LEXCH = CA LEXCH = CA LEXCH = CA LEXCH = CA TITEL = US B TITEL = US C TITEL = US C TITEL = US H TITEL = US N running on 4 nodes distr: one band on 4 nodes, 1 groups vasp.4.6.21 23Feb03 complex POSCAR found : 5 types and 50 ions WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 WARNING: for PREC=h ENMAX is automatically increase by 25 % this was not the case for versions prior to vasp.4.4 LDA part: xc-table for Ceperly-Alder, standard interpolation POSCAR, INCAR and KPOINTS ok, starting setup FFT: planning ... 8 reading WAVECAR entering main loop N E dE d eps ncg rms rms(c) |
3楼2009-10-13 22:39:03