g98l problems

[Jean-Christophe Ducom]

I think it's a shared memory issue (+Linda library missing). You should 
increase the maximum shared memory segment for instance:
echo 134217728 >/proc/sys/kernel/shmmax (depending on the total physical 
ram you have)
You also can add it in the /etc/sysctl.conf so that it does it at every 
reboot).
Do you use Linda for the "parallel" library? If so you have to put 
NProcL in your .com file...otherwise I don't know how you can get g98l 
work without it.
Hope this helps.

		JC



luis.licon at cimav.edu.mx wrote:
> Hi you all guys, I have a problem while running g98l on a small beowulf (4
> single nodes interconnected by a fast ethernet running on Red Hat Linux
> 7.2)
> 
> The error occurs as follows:
> 
> 
> /////////////////////////O U T P U T////////////////////////////////////
> 
> 
>  ***********************************************
>  Gaussian 98:  x86-Linux-G98RevA.11.1 1-Nov-2001
>                    26-Nov-2002
>  ***********************************************
>  %NProc=4
>  Will use up to    4 processors via shared memory.
>  %mem=8MW
>  ----------------------------
>  #p opt freq=noraman rpm3 zdo
>  ----------------------------
>  1/14=-1,18=50,19=11,26=3,38=1/1,3;
>  2/9=110,17=6,18=5,40=1/2;
>  3/5=2,11=1,12=1,25=1,30=1/1;
>  4/5=3,7=1,11=1,20=7,22=1,24=1,35=1/1,2;
>  6/7=2,8=2,9=2,10=2/1;
>  7//16;
>  1/14=-1,18=50,19=11/3(1);
>  99//99;
>  2/9=110/2;
>  3/5=2,11=1,12=1,25=1,30=1/1;
>  4/5=5,7=1,11=1,16=2,20=7,22=1,24=1,35=1/1,2;
>  7//16;
>  1/14=-1,18=50,19=11/3(-4);
>  2/9=110/2;
>  6/7=2,8=2,9=2,10=2/1;
>  99//99;
>  Leave Link    1 at Tue Nov 26 18:46:49 2002, MaxMem=    8388608 cpu:     
>  0.1
> shmget failed.
> shmget failed.: Invalid argument
> 
> ///////////////////////////////////////////////////////////////////////////
> 
> 
> 
> Is there any idea of what is going on?
> 
> 
> 
> 
> Thanks,
> 
> 
> Lycon
> 
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
>