[Smeagol-discuss] MPI error

[Xi Zhu]

have you tried MPICH version 1.x.x ?




2014-07-10 14:57 GMT+08:00 joyce79928cc . <joyce79928 at gmail.com>:

> Dear:
>
> Is there somebody can help with my problem?
> It really makes a lot of trouble for me...
> or did I explain not good or...?
> thank you!
>
> --
> Best Regards
>
> 邱芳瑜  Chiu Fang Yu
> 國立成功大學 材料科學與工程學系碩一
> MOBILE：0930287221(中華)
> GMAIL：joyce79928 at gmail.com <wisdom4640 at gmail.com>
> HOTMAIL：joyce79928 at hotmail.com <wisdom4640 at hotmail.com>
>
>
>
> 2014-07-05 23:12 GMT+08:00 joyce79928cc . <joyce79928 at gmail.com>:
>
> Dear:
>>
>> Every time I want to use MPI to calculate using SMEAGOL,
>> I need to select a "proper" number of core for the model to do a complete
>> calculation.
>> sometimes I need to try for several times to get the proper number of
>> core,
>> otherwise it will have some error below:
>> (1)Terminating due to failed diagonalisation
>> (2)forrtl: severe (174): SIGSEGV, segmentation fault occurred (when the
>> core is too many)
>> (3)the file attached (output1.out)
>> for error (1) I have search on the Internet and set Diag.DivideAndConquer
>> .false., but the error still occur.
>>
>> I think maybe the errors above caused by "non-proper" number of core for
>> MPI run?
>> and is it normal for SMEAGOL user to choose the "proper" number of core
>> every time?
>>
>> thank you!
>>
>>  --
>> Best Regards
>>
>> 邱芳瑜  Chiu Fang Yu
>> 國立成功大學 材料科學與工程學系碩一
>> MOBILE：0930287221(中華)
>> GMAIL：joyce79928 at gmail.com <wisdom4640 at gmail.com>
>> HOTMAIL：joyce79928 at hotmail.com <wisdom4640 at hotmail.com>
>>
>>
>
> _______________________________________________
> Smeagol-discuss mailing list
> Smeagol-discuss at lists.tchpc.tcd.ie
> http://lists.tchpc.tcd.ie/listinfo/smeagol-discuss
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.tchpc.tcd.ie/pipermail/smeagol-discuss/attachments/20140710/c06f1d8e/attachment-0001.html>