[gmx-users] The trajectory and energy in by different nodes

[Tanping Li]

Dear all,

I am using the gromacs to calculate protein solution.
I use 8 nodes in mdrun in Itanium cluster. I am
totally stuck know. My problems is:

1.) I find that my trajectory and energy file from 8
nodes is different from the files from one node, even
if they have the totally same initial condition. But
the one node result is the same with the local machine
I am not sure where the difference comes from, the
compiler or the gromacs itself? Following are the
energy files from one node and 8 nodes trajectories.

The energy file from one node trajecory.
@ s0 legend "LJ (SR)"
@ s1 legend "Coulomb (SR)"
@ s2 legend "Coulomb (LR)"
@ s3 legend "Potential"
@ s4 legend "Kinetic En."

    0.000000  43551.857565  -225851.063464
-67543.206515  -225815.900842  38278.991420  
    0.200000  43233.031293  -225401.171795 
-67514.909219  -225581.998109  38045.903177  
    0.400000  43042.908425  -224988.939329 
-67403.010519  -225375.287501  37838.405258  


The energy file from 8 nodes trajectory.


    0.000000  43551.857565  -225851.063464 
-67539.349853  -225812.044180  38278.931445  
    0.200000  43229.472406  -225409.095771 
-67507.866436  -225587.560497  38055.247171  
    0.400000  43083.080600  -225160.286370 
-67396.841904  -225499.357111  37966.139550  




2) When I use the mdrun -rerun function, even with the
totally same trr file, the tpr files with different
nodes give me different Coulomb_longrange energy, and
the other energy terms are same. Does anybody have the
same experience? Following is the energy files from
one and 8 nodes rerun function with the same
trajectory.

The energy file from one node trajecory.
@ s0 legend "LJ (SR)"
@ s1 legend "Coulomb (SR)"
@ s2 legend "Coulomb (LR)"
@ s3 legend "Potential"
@ s4 legend "Kinetic En."

    0.000000  43551.857565  -225851.063464 
-67543.206515  -225815.900842  38278.991420  
    0.200000  43229.472406  -225409.095771 
-67517.080772  -225596.774833  38055.182028  
    0.400000  43083.080600  -225160.286370 
-67403.853960  -225506.369167  37966.062540  

The energy file from 8 nodes trajecory.

    0.000000  43551.857565  -225851.063464 
-67539.349853  -225812.044180  38278.931445  
    0.200000  43229.472406  -225409.095771 
-67507.866436  -225587.560497  38055.247171  
    0.400000  43083.080600  -225160.286370 
-67396.841904  -225499.357111  37966.139550  




I have already get 200 trajectories by 8 nodes. I hope
they are OK.

Best
Tanping