问题描述
cbb@WORKSTATION-DEV:~/lammps/lammps-29Sep2021/examples/shear$ mpirun -np 6 lammps/lammps-29Sep2021/src/lmp_mpi < in.shear
[mpiexec@WORKSTATION-DEV] wait_proxies_to_terminate (../../../../../src/pm/i_hydra/mpiexec/intel/i_mpiexec.c:558): downstream from host WORKSTATION-DEV was killed by signal 13 (Broken pipe)
[mpiexec@WORKSTATION-DEV] main (../../../../../src/pm/i_hydra/mpiexec/mpiexec.c:2162): assert (pg->intel.exitcodes != NULL) failed
[mpiexec@WORKSTATION-DEV] HYD_sock_write (../../../../../src/pm/i_hydra/libhydra/sock/hydra_sock_intel.c:362): write error (Bad file descriptor)
原因分析:
错误的调用mpirun
直接写mpirun会调用自己在/opt下面的intel mpi去执行
正常情况下是调用/usr/local/bin/下的
解决方案:
cbb@WORKSTATION-DEV:~/lammps/lammps-29Sep2021/examples/shear$ /usr/local/bin/mpirun -np 2 ../../src/lmp_mpi < in.shear