[mpich-discuss] MPICH Trouble:failed to connect. Seek help

songjing songjing at mail.ustc.edu.cn
Sun Dec 19 23:00:34 CST 2010


Dear Sir:
First of all, thank you for your excellent MPICH! It helps us a lot.
But I have some trouble puzzles me.
I compiled a parallel code system with MPICH NT 1.2.3 x32 on Windows XP x32.And installed MPICH NT 1.2.3 x 32 on Windows Server 2003,64 bit system ,a computational cluster with 12 nodes.And I ran the code system on the cluster with MPICH NT 1.2.3.Sometimes it's OK,but sometimes it failed and show the information: failed to connect nodeX and there is no process of this code on nodeX.After I restart the server of MPICH it can work well.
And my question is why it failed sometimes and how can I make it stable and don't need to restart the server  of MPICH?
Thank you very much!
                                                                                      SongJing

                                                                                                                    
**********************************************************************************************************
University of Science and Technology of China,Nuclear Science And Technology
E-mail:   songjing at mail.ustc.edu.cn
Website:www.fds.org.cn
     CellPhone : 13645698849 
**********************************************************************************************************
2010-12-20 



songjing 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20101220/57a66b53/attachment-0001.htm>


More information about the mpich-discuss mailing list