Dear users,
We are currently struggling to compile Amber v18 on our cluster with
openmpi an cuda, our specifications are the following:
- CentOS 6.5
- double socket Intel Xeon(R) CPU E5-2660
- gcc/4.8.2
- openmpi/1.8.1 and also tried openmpi/2.1.1
- cuda-8.0.61
./configure -mpi -noX11 -openmp -cuda gnu
...
mpif90 -DMPI -DBINTRAJ -DEMIL -DPUBFFT -DGNU_HACKS -O3 -mtune=native
-fopenmp -D_OPENMP_ -DCUDA -DGTI -DMPI -DMPICH_IGNORE_CXX_SEEK
-I/opt/amber-18/amber18/include -c gb_force.F90
gb_force.F90:288.59:
call gbsa_ene(crd, gbsafrc, pot_ene%surf ,atm_cnt, jj, r2x,
belly_atm_cnt
1
Error: Name 'jj' at (1) is an ambiguous reference to 'jj' from module
'gb_ene_hybrid_mod'
gb_force.F90:335.59:
call gbsa_ene(crd, gbsafrc, pot_ene%surf, atm_cnt, jj, r2x,
belly_atm_cnt
1
Error: Name 'jj' at (1) is an ambiguous reference to 'jj' from module
'gb_ene_hybrid_mod'
I have found a similar unsolved issue here:
http://dev-archive.ambermd.org/201603/0044.html
We have successfully installed amber on another cluster without gpus but we have not been able to sort this problem.
We were wondering if you could tell us what we are doing wrong.
Thank you very much in advance.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Sep 28 2018 - 01:30:02 PDT
