probs running mpqc

zolia zolia at lydys.sc-uni.ktu.lt
Mon Aug 14 00:30:13 EDT 2000


hello, 

in this case i'm stuck not with calculations, but with establishing
connection with other then localhost. As i mentioned in the earlier
posting , sys call trace shows that 753 port is tried to be opened, but
thera are nothing on that end. Maybe there is something very bad with my
env configurationt and starting method ? My env for running mpqc on two
nodes is :


----------- script to run-----------

#!/bin/sh

export Nproc=2, export MessageGrp=MPIMessageGrp
export NPROC=2;
export MESSAGEGRP="<MPIMessageGrp>:(
	debug = 1
        argv = [  
                -execer_id DQS 
                -master_host lydys
                -my_hostname lydys
                -my_nodenum 0 
                -my_numprocs 2 
                -total_numnodes 2 # ????
 	        -p4pg .p4pg  # ????
                -job_id 100 ] 
        )";
export GIVENMESSAGEGRP="mpich";
export SCLIBDIR=/usr/src/mpqc/mpqc-1.2.3/lib

/usr/src/mpqc/mpqc-1.2.3/src/bin/mpqc/mpqc $*

------------- end of script ------------

and run it to test smth like :

lydys:/usr/src/mpqc/mpqc-1.2.3$ mpirun -np 2 ./runmpqc-mpipro -d -f
src/bin/mpqc/sample/scfopt.in

but this doesn't start any proces on other machine. When i create .p4pg
file:
lydys 0
perlas 1

[occures that problem with 753 port:]

mpqc: unrecognized option -p4pg
MPIMessageGrp::init: entered
Calling MPI_Init with -execer_id DQS -master_host lydys -my_hostname lydys
-my_nodenum 0 -my_numprocs 2 -total_numnodes 2 -p4pg .p4pg -job_id 100
rcmd: perlas.sc-uni.ktu.lt: Success
p0_12456:  p4_error: Timeout in making connection to remote process on
perlas.sc-uni.ktu.lt: 0
bm_list_12457:  p4_error: interrupt SIGINT: 2

[ and sys trace again: ]

connect(8, {sin_family=AF_INET, sin_port=htons(753),
sin_addr=inet_addr("193.219.61.66")}}, 16) = -1 ECONNREFUSED (Connection
refused)
fork()                                  = 12654
rt_sigaction(SIGALRM, NULL, {SIG_DFL}, 8) = 0
rt_sigaction(SIGALRM, {0x40023f04, [ALRM], 0x4000000}, NULL, 8) = 0
setitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={300, 0}}, NULL) = 0
accept(6, rcmd: perlas.sc-uni.ktu.lt: Success
0xbfffdb6c, [16])             = ? ERESTARTSYS (To be restarted)
--- SIGCHLD (Child exited) ---
accept(6, 

and hangs here, some SIG problems ???

any ideas?


====================================================================
Antanas Masevicius             Kaunas University of Technology
Studentu 48a-101               Computer Center
LT-3028 Kaunas                 LITNET NOC UNIX Systems Administrator
Lithuania                      E-mail: zolia at sc.ktu.lt

On 14 Aug 2000, Curtis L. Janssen wrote:

> I mainly use MPI/Pro on Linux machines since it is the only MPI
> that will support the MP2 gradient algorithm in MPQC.  However,
> for Hartree-Fock, MPICH should work.  It would help to know what
> kind of calculation you are doing and how far it got.
> 
> Curt
> -----------------------------------------------------------------------
> Curtis Janssen
> cljanss at ca.sandia.gov                http://aros.ca.sandia.gov/~cljanss
> Tel: +1 925-294-1509                 Fax: +1 925-294-2234
> Sandia National Labs, MS 9217, PO Box 969, Livermore, CA 94551, USA
> 





_______________________________________________
Beowulf mailing list
Beowulf at beowulf.org
http://www.beowulf.org/mailman/listinfo/beowulf



More information about the Beowulf mailing list