应《网络安全法》要求，自2017年10月1日起，未进行实名认证将不得使用互联网跟帖服务。为保障您的帐号能够正常使用，请尽快对帐号进行手机号验证，感谢您的理解与支持！

24小时热门版块排行榜    

小木虫论坛-学术科研互动平台 » 计算模拟区 » 第一性原理 » Abinit » 有關聲子eigen value的計算
121/2返回列表12下一页
查看: 1683  |  回复: 11
只看楼主 @他人 存档 新回复提醒 (忽略) 收藏    在APP中查看

mouwat

铁虫 (初入文坛)

[求助] 有關聲子eigen value的計算

我用tnlo_2計算一層平面結構時是OK的  但在改變原子位置使其有點皺摺時就無法計算了  err檔中?K沒有對錯誤的敘述  請問該怎麼辦呢
回复此楼

» 猜你喜欢

» 本主题相关价值贴推荐，对您同样有帮助:
1楼 2012-09-22 23:37:47
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖
回帖支持 ( 显示支持度最高的前 50 名 )

souledge

专家顾问 (著名写手)
★ ★
franch: 金币+2, 专家考核, 谢谢回帖交流 2012-09-24 12:57:28
看楼主的语言习惯是港台的，我还是用英语回复好了。
The error info can be found at the end of xxx.log file if you ran Abinit with command: abinit < xxx.files >& xxx.log, otherwise, it displays in your terminal.
I am not sure which kind of "error" you met, or what means "cannot calculate". By guessing, displacements of ion(s) broke the symmetry, thus the convergence of response function may become more difficult.
赞一下(1人) 回复此楼
思想重于技巧，内涵重于表象
2楼2012-09-23 00:16:06
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖

souledge

专家顾问 (著名写手)
???????:
3?: Originally posted by mouwat at 2012-09-23 00:42:40
???w??????  ?x?x????N??

sorry?????o?????Y??

err files:
mpiexec: Warning: tasks 2,4-15 exited with status 1.

tnlo_2????????岽??r???????????

input files??????
xred         ...

Actually, the error info you provided was not really the proper one, since it was just a stop signal from MPI code, instead of that from Abinit.
Why don't you attach your full log or output file here? You can edit your first post and upload the file as accessory.
Further, you may download my tutorial about Abinit from the ??????? broad. That will give you some idea for finding error info from Abinit.
?????????δ???????????????????????MPI??????????????????Abinit???
??????log???????????????????????Σ?????????????????????????
????????????????????????????Abinit??????????????β???Abinit??????????
赞一下(1人) 回复此楼
思想重于技巧，内涵重于表象
4楼2012-09-23 10:11:45
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖

souledge

专家顾问 (著名写手)
mouwat: 回帖置顶 2012-09-25 06:59:03
引用回帖:
5楼: Originally posted by mouwat at 2012-09-23 11:04:38
附上log檔
http://www.putlocker.com/file/507AF85E4CF0552D

感謝...

I have checked it, the error info was as:
引用回帖:
p15_14241:  p4_error: interrupt SIGSEGV: 11
rm_l_15_14245: (13.175781) net_send: could not write to fd=5, errno = 32
p1_10669:  p4_error: net_recv read:  probable EOF on socket: 1
p2_10670:  p4_error: net_recv read:  probable EOF on socket: 1
p4_25239:  p4_error: net_recv read:  probable EOF on socket: 1
p6_25241:  p4_error: net_recv read:  probable EOF on socket: 1
p5_25240:  p4_error: net_recv read:  probable EOF on socket: 1
p7_25242:  p4_error: net_recv read:  probable EOF on socket: 1

Therefore, I wanna ask a question, which version of MPI you are using now, MPICH1? Since I can just find p4_error from MPICH1 or P4 but not from modern MPICH2 and OpenMPI. Would you like to take a look at your mpi version by : mpiexec -info ?
Of course, some person reported that p4_error may be caused by network problem among nodes, I am not sure whether you are using cluster or workstation (16 cores Xeon platform could be multi-CPU workstation). But if you were using the later one, it does not make sense to discuss about network problem.
赞一下(1人) 回复此楼
思想重于技巧，内涵重于表象
6楼2012-09-23 13:19:33
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖

souledge

专家顾问 (著名写手)
★ ★
franch: 金币+2, 谢谢回帖交流 2012-09-25 10:02:27
franch: 专家考核 2012-09-25 10:02:35
???????:
7?: Originally posted by mouwat at 2012-09-23 14:22:34
???????殅?I??????^?????а???r????????\??  ?????????????????
???????f?W·???????f?^abinit???@???????????????}
?????o???B?j????  ?????@?Y?????????P????o???鵽???P
????????????? ...

It could be an Abinit problem, even though I still think it is higher possible raised from the old version of MPI.
If you wanna calculate the 3rd-order deviation of energy ( anharmonic phonon), why not giving a try to Quantum-ESPRESSO? It is normally stable for response function issues.
赞一下(1人) 回复此楼
思想重于技巧，内涵重于表象
8楼2012-09-23 22:25:19
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖

oxox6085

专家顾问 (正式写手)

【答案】应助回帖

★ ★
感谢参与，应助指数 +1
franch: 金币+2, 谢谢回帖交流 2012-09-24 12:59:43
我有一个想法：
本身能量的三阶导即要求之前结构优化要严格的收敛，随意改动原子的位置后，原来收敛的体系现在必然就不收敛了。建议楼主尝试重新优化具有褶皱的体系，可以利用一些优化技巧，如：只充分优化晶胞不改变原子相对位置从而达到收敛，或者锁定原子z方向的坐标，等等。
赞一下(1人) 回复此楼
有些事情可能只有计算机模拟能告诉我们答案
9楼2012-09-24 00:03:38
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖
普通回帖

mouwat

铁虫 (初入文坛)
???????:
2?: Originally posted by souledge at 2012-09-23 00:16:06
??????????????????????????????????????
The error info can be found at the end of xxx.log file if you ran Abinit with command: abinit < xxx.files >& xxx.log, otherwise, it display ...

???w??????  ?x?x????N??

sorry?????o?????Y??

err files:
mpiexec: Warning: tasks 2,4-15 exited with status 1.

tnlo_2????????岽??r???????????

input files??????
xred        0.33333 0.66667  0.5
               0.66667 0.33333  0.5
to
xred        0.33333 0.66667  0.5026208779
               0.66667 0.33333  0.4973791221
赞一下 回复此楼
3楼2012-09-23 00:42:40
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖

mouwat

铁虫 (初入文坛)
引用回帖:
4楼: Originally posted by souledge at 2012-09-23 10:11:45
Actually, the error info you provided was not really the proper one, since it was just a stop signal from MPI code, instead of that from Abinit.
Why don't you attach your full log or output file he ...

附上log檔
http://www.putlocker.com/file/507AF85E4CF0552D

感謝
赞一下 回复此楼
5楼2012-09-23 11:04:38
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖

mouwat

铁虫 (初入文坛)
???????:
6?: Originally posted by souledge at 2012-09-23 13:19:33
I have checked it, the error info was as:

Therefore, I wanna ask a question, which version of MPI you are using now, MPICH1? Since I can just find p4_error from MPICH1 or P4 but not from modern M ...

???????殅?I??????^?????а???r????????\??  ?????????????????
???????f?W·???????f?^abinit???@???????????????}
?????o???B?j????  ?????@?Y?????????P????o???鵽???P
?????????????  ???????????????????岽????????????????岽??????  ????s???????岽??r??o?????....
赞一下 回复此楼
7楼2012-09-23 14:22:34
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖

mouwat

铁虫 (初入文坛)
???????:
6?: Originally posted by souledge at 2012-09-23 13:19:33
I have checked it, the error info was as:

Therefore, I wanna ask a question, which version of MPI you are using now, MPICH1? Since I can just find p4_error from MPICH1 or P4 but not from modern M ...

?x?x????V??some person reported that p4_error may be caused by network problem among nodes?@??,???????????????CPU????]????????}??
赞一下 回复此楼
10楼2012-09-25 06:54:55
已阅   回复此楼   关注TA 给TA发消息 送TA红花 TA的回帖
相关版块跳转 我要订阅楼主 mouwat 的主题更新
121/2返回列表12下一页
普通表情 龙 兔 虎 猫
信息提示
关闭
请填处理意见
关闭