[gmx-users] 3dc + GPU

[Ondrej Kroutil]

Hi gromacs users and developers,
I have troubles with simulations of the system with surface (slab geometry)
using GPU and Verlet cut-off scheme.
When I run this simulation on CPU everything goes right and ions dissolved
in the water above the surface seems to behave normally. But when I switch
to GPU positively charged ions move in z-direction to the bottom of the
simulation box and negatively charged ions go to opposite side of the box.
I have had this problem with 4.6 version and now even with version 5.0.
Do you have any suggestion what can be wrong? Bad input parameters? Etc.
You can download tar file with input file, .tpr file (if you want to run it
on your system, output log and pictures of initial and final state)

https://drive.google.com/file/d/0Byl8CD_rdzQOZVB2UGVHQzllMms/edit?usp=sharing

Thank you very much and keep developing Gromacs, the best simulation
package ever!

Ondřej

Btw.: How can I set-up in gromacs-5.0 to run simulations only on CPU?

-- 
Ondřej Kroutil
        ,,              Faculty of Health and Social Studies
   ----"))'             University of South Bohemia
     OOO           Jirovcova 24, Ceske Budejovice
      OOO          The Czech Republic
        | OO         E-mail:  okroutil at gmail.com
>------    O         Mobile:  +420 736 537 190