Good ... only the EVB/PIMD fails. For now, let's not worry about
EVB/PIMD since we are are planning a new scheme to couple to two after
Wei finalizes the atom-first PIMD.
The differences in the output files for evb_dg_cart and evb_dg_cart_irc
could be due to numerical reasons resulting from the single-val decomp.
... but I need to check with some experts first and also to do more
testing. I am aware of the divergence between the output and *.save files.
Thanks, Ross.
-Kim
> evb/evb_nrg_map_pimd_ld - fails with:
>
> [cli_0]: aborting job:
> Fatal error in MPI_Send: Other MPI error, error stack:
> MPI_Send(172): MPI_Send(buf=0x1d79d20, count=36, MPI_DOUBLE_PRECISION,
> dest=0, tag=1, MPI_COMM_WORLD) failed
> MPID_Send(51): DEADLOCK: attempting to send a message to the local process
> without a prior matching receive
> rank 0 in job 391 caffeine.sdsc.edu_39729 caused collective abort of all
> ranks
> exit status of rank 0: killed by signal 9
> ./Run.evb: Program error
> make: *** [test.sander.EVB] Error 1
>
> evb/evb_nrg_umb_pimd_ld - fails with same problem.
> evb_bond_umb_pimd_ld - fails with same problem.
> evb_dbonds_umb_pimd_ld - fails with same problem.
> evb_nrg_map_pimd_nhc - fails with same problem.
> evb_nrg_umb_pimd_nhc - fails with same problem.
> evb_bond_umb_pimd_nhc - fails with same problem.
> evb_dbonds_umb_pimd_nhc - fails with same problem.
> evb_nrg_map_nmpimd_nhc - fails with same problem.
> evb_nrg_umb_nmpimd_nhc - fails with same problem.
> evb_bond_umb_nmpimd_nhc - fails with same problem.
> evb_dbonds_umb_nmpimd_nhc - fails with same problem.
>
> evb_dg_cart - output missing from saved file:
> |ERROR: EVB energy - ab initio energy > toler
> EVB =
> -92.24604260
> ab initio =
> -92.24604260
> |Deviation = 0.14210855E-13
>
> evb_dg_cart_irc - Major differences throughout output file
>
>
Received on Sun Oct 15 2006 - 06:07:05 PDT
