Linux Distributed Memory Parallel (DMP) error forrtl: severe (168)

- July 5, 2021 at 6:46 pm
  
  vcompan
  Subscriber
  
  Hello All,
  I have just setup a HPC (2 machines) with CentOS 7 (Distributed Memory Parallel) and I find an error when I execute the mpitest program.
  [ ]$ /usr/local/ansys_inc/v195/ansys/bin/mpitest -machines ladon:2:nodo01:2
  forrtl: severe (168): Program Exception - illegal instruction
  Image             PC               Routine           Line       Source
  mpitestintelmpi.e 000000000040D2EA Unknown              Unknown Unknown
  libpthread-2.17.s 00007F2EF5B2C630 Unknown              Unknown Unknown
  libibverbs.so.1.5 00007F2EF38C480E Unknown              Unknown Unknown
  libibverbs.so.1.5 00007F2EF38C56A1 Unknown              Unknown Unknown
  libibverbs.so.1.5 00007F2EF38C3238 ibv_get_device_li    Unknown Unknown
  libdaploscm.so.2. 00007F2EF3AE5FF9 Unknown              Unknown Unknown
  libdaploscm.so.2. 00007F2EF3AE7200 Unknown              Unknown Unknown
  libdat2.so.2.0.0  00007F2EF3CF57BF Unknown              Unknown Unknown
  libdat2.so.2.0.0  00007F2EF3CF7EBF dat_extension_op     Unknown Unknown
  libmpi.so.12.0    00007F2EF695B1BB Unknown              Unknown Unknown
  libmpi.so.12.0    00007F2EF69B7212 Unknown              Unknown Unknown
  libmpi.so.12.0    00007F2EF6AD6C60 Unknown              Unknown Unknown
  libmpi.so.12.0    00007F2EF6AD5003 Unknown              Unknown Unknown
  libmpi.so.12.0    00007F2EF68B27F1 Unknown              Unknown Unknown
  libmpi.so.12.0    00007F2EF6AC08BE Unknown              Unknown Unknown
  libmpi.so.12.0    00007F2EF6A6414A Unknown              Unknown Unknown
  libmpi.so.12.0    00007F2EF6A51903 MPI_Init             Unknown Unknown
  libmpifort.so.12. 00007F2EF64B8240 MPI_INIT             Unknown Unknown
  mpitestintelmpi.e 0000000000404031 Unknown              Unknown Unknown
  mpitestintelmpi.e 00000000004031CE Unknown              Unknown Unknown
  libc-2.17.so      00007F2EF5771555 __libc_start_main    Unknown Unknown
  mpitestintelmpi.e 00000000004030A9 Unknown              Unknown Unknown
  
  I have followed the "Linux_Installation_Guide for Ansys195". The files .hosts, machines and hosts195.ans are OK!
  Both machines have Selinux and Firewalld desactivated.
  I have also setup ssh without password.
  The installation folder are shared with NFS and the network is Infiniband Mellanox (100 Gb/s)
  Could you please help us?. Thanks.!!..
- July 9, 2021 at 12:50 pm
  
  George Karnos
  Ansys Employee
  
  Can you verify that all of the prerequisites are installed?
  Chapter 2 of the Linux Installation Guide