by running different test examples i noticed that this error always comes after this line
1) this is example of benzene in tests InitMesh: MESH = 80 x 80 x 48 = 307200 InitMesh: Mesh cutoff (required, used) = 120.000 127.546 Ry * Maximum dynamic memory allocated = 6 MB rank 18 in job 3 hcg_59877 caused collective abort of all ranks exit status of rank 18: return code 1 2) example of h2O InitMesh: MESH = 32 x 30 x 24 = 23040 InitMesh: Mesh cutoff (required, used) = 50.000 50.384 Ry * Maximum dynamic memory allocated = 3 MB rank 19 in job 2 hcg_59877 caused collective abort of all ranks exit status of rank 19: return code 1 and similarly in other examples. any help from experts will be greatly appricated. ________________________________ From: karan deep <[email protected]> To: karan deep <[email protected]> Cc: ""[email protected]"" <[email protected]> Sent: Monday, 25 June 2012 1:45 PM Subject: Fw: [SIESTA-L] mpi error while running siesta ----- Forwarded Message ----- From: karan deep <[email protected]> To: Abraham Hmiel <[email protected]> Sent: Sunday, 24 June 2012 1:59 PM Subject: Re: [SIESTA-L] mpi error while running siesta Dear Sir here are answers to ur questions siesta was successfully compiled with no errors. 1) mpirun -np 20 siesta < input.fdf | tee Output.fdf 2 i was running an FC type of run so mpi works fine with CG. Then suddenly i got this error. then i try to run force constants in tests directory with mpirun same error comes. then i try to test batio3 directory same error. siesta runs successfully during serial runs in these tests. 3) i m using ifort 11.1.072 version with mpich2-1.2 on 64-bit machine ubuntu lucid siesta 3.1. my arch.make file is attached. 4) i try to make changes in arch.make file then somewhere in mailing list i found after updates u how reinstall siesta. i never remember last time i updated. but still recompile mpich2 and siesta. i try siesta 2.0.2 still same error Thanks ________________________________ From: Abraham Hmiel <[email protected]> To: karan deep <[email protected]> Cc: "[email protected]" <[email protected]> Sent: Saturday, 23 June 2012 9:55 PM Subject: Re: [SIESTA-L] mpi error while running siesta Dear Karan, We are going to need a lot more information than the error message you provided to solve your problem. What is the command you used to invoke the program? What are the contents of your arch.make file and the input fdf file? What few things have you tried, exactly? Does the invocation of siesta work in serial mode? Has siesta with mpi ever successfully run? Is this a chronic problem or does it happen seemingly at random? The more information you put forward to the listserv, the more able we are to help you. Best, On Sat, Jun 23, 2012 at 10:19 AM, karan deep <[email protected]> wrote: Dear users i m suddenly getting this error while running siesta > > > Fatal error in MPI_Comm_group: Invalid communicator, error stack: >MPI_Comm_group(150): MPI_Comm_group(comm=0x0, group=0x7fff0c620494) failed >MPI_Comm_group(74).: Invalid communicator > > >i try few things but nothing is working -- Abraham Hmiel Katherine Belz Groves Fellow in Nanoscience Xue Group, College of Nanoscale Science and Engineering at SUNY Albany
