[Wien] question about paralled computation on distributed file system

[白宇]

Dear users: 
when I am doing paralled computation on distributed file system, the lapw1 always crashed. 

I read the case.output1up_1(2,3,...), I found that some of them have the eigenvalues, while other doesn\'t. I list two differents output1 files on this mail(just the end of them). 

I run SCF several times and found that every time there are some cpu can not run lapw1, however, every time the failed cpus are different. 

Can you give me some suggestions?
-------------------------------------------------------
**************************************************** 
K=    0    1    1  IND= 2
           1. WAVE=    1    0   -1    TAUP=   1.00000   0.00000
                                    WARPING=   0.01535  -0.01640
           2. WAVE=    0    1    1    TAUP=   1.00000   0.00000
                                    WARPING=   0.01535  -0.01640
 K=   -1   -1    0  IND= 1


----------------------------------------------------------
*********    *********    eigenvalue    *********    *********
1.4686104    1.4744913    1.4764108    1.4883047    1.4998610
       0 EIGENVALUES BELOW THE ENERGY   -7.00000
********************************************************
NUMBER OF K-POINTS:           1
 ===> TOTAL CPU       TIME:    359.0 (INIT =      4.0 + K-POINTS =    355.0)
        > SUM OF WALL CLOCK TIMES:    360.2 (INIT =      4.1 + K-POINTS =    356.0)
           Maximum WALL clock time:    360.730095863342
           Maximum CPU time:           359.180000000000