[gmx-users] Platform-specific differences in grompp's .tpr output

[Danny Ngo]

Hi,

This concerns the water simulation from the tutorial and GROMACS 3.2 with 
FFTW 2.1.5.

I've been running the simulation on three different platforms:
Red Hat Linux 7.1 on i686 (i686-pc-linux-gnu)
IRIX64 (mips-sgi-irix6.5)
OSF1 (alphaev68-dec-osf5.1)

The issue is that grompp produces a slightly different .tpr file on each 
of these platforms.  Consequently, the mdrun results (ie. water.trr, 
water_out.gro) from each .tpr file are also slightly different.  Is this 
to be expected from varying platforms?

I ran the same simulation using GROMACS 3.1.4 before, and the odd thing is 
that grompp produced the same .tpr file on IRIX64 and OSF1 (I compared 
their md5 sums).  The one produced on Linux was different, however.  Was 
this just coincidence??

Any comments/suggestions are appreciated.  Thanks.

Please let me know if you want to see the .tpr files and results.

Danny