[CPMD-list] Trying to get MPI to work

mkosmows mkosmows at mailbox.syr.edu
Wed Aug 6 02:39:20 CEST 2003


Dear CPMD community:
 
I have gotten the MPI version of CPMD 3.7.2 (PGI f77, GCC cc) to run.  
However, I get an error, described below.  I am running MPICH 1.2.5 and 
CPMD 3.7.2.  I have not tried other versions of CPMD.  I am using a "cluster" 
of two Athlon 1.3GHz workstations with 1Gb RAM each.  Both computers are 
running Mandrake 9.1 linux.
 
In the terminal where the program was running:
 
[mark at linux SP]$ mpirun -np 2 ~/work/bin/cpmd.x BH3NH3.in >BH3NH3.out
    p4_error: latest msg from perror: No route to host 
Killed by signal 2. 
/home/mark/work/mpich-1.2.5/bin/mpirun: line 1:  8167 Broken pipe             
/home/mark/work/bin/cpmd.x "BH3NH3.in" -p4pg 
/home/mark/work/cpmd/BH3NH3/lda/25K8.8.8/SP/PI8082 -p4wd 
/home/mark/work/cpmd/BH3NH3/lda/25K8.8.8/SP 
[mark at linux SP]$ mpirun -np 2 ~/work/bin/cpmd.x BH3NH3.in >BH3NH3.out
    p4_error: latest msg from perror: No route to host 
Killed by signal 2. 
/home/mark/work/mpich-1.2.5/bin/mpirun: line 1: 10028 Broken pipe             
/home/mark/work/bin/cpmd.x "BH3NH3.in" -p4pg 
/home/mark/work/cpmd/BH3NH3/lda/25K8.8.8/SP/PI9943 -p4wd 
/home/mark/work/cpmd/BH3NH3/lda/25K8.8.8/SP 
[mark at linux SP]$
 
And in the output file:

 NFI      GEMAX       CNORM           ETOT        DETOT      TCPU
   1  2.112E-01   1.667E-02     -29.677204    0.000E+00     88.83
   2  1.074E-01   6.307E-03     -31.056725   -1.380E+00     90.80
   3  4.252E-02   2.703E-03     -31.253852   -1.971E-01     91.64
   4  3.083E-02   1.413E-03     -31.281939   -2.809E-02     93.56 
p0_10028: (5378.632967) net_send: could not write to fd=5, errno = 113 
p0_10028:  p4_error: net_send write: -1
 
The first time this happened, only two lines after the NFI line were printed.  
Also, 
the second workstation (not the one that mpirun ... cpmd.x was invoked on) 
stops giving a video signal and is unresponsive to ssh or webmin from the 
first 
workstation.  Is this something that mpi is doing, or should I be looking for 
a 
hardware problem?
 
Thank you,
 
Mark Kosmowski

Chemistry Department
Syracuse University
mkosmows at syr.edu




More information about the CPMD-list mailing list