Page 1 of 1

Maybe something wrong with the mpi implement in GW

Posted: Sun Jun 09, 2013 5:11 am
by Zhang Xiaole
Dear users,
When i perform the calculation of t42 in libxc, which is a test to test the implement of HSE06 functional, everything goes fine with the sequential run. However, when i run the job with mpirun -n nprocs, although the job can run without errors, the results goes very strange. I am wondering there maybe something wrong with the mpi implement in hybrid fuctionals calculations.



Sheleon