Problem in NVT run in Cray machine..

Hi All,

I am simulating polymer bundle by using class II force field on Cray XE6 . There is no any problem with CG minimization. However, when i use NVT program does
not work properly. All i ve got is " NAN" . Even i use low temperature, all i ve got is “NAN” . I used small thermo to see whether there is wild change or not. But from the first step it gave “NAN” .

Dumped positions “NAN”

Thermo output :

Step TotEng Temp Press Pxx Pyy Pzz Volume
3000 -19.99142 0 -0.1371864 -0.15156241 -0.012069375 -0.24792742 15625
3001 nan nan nan nan nan nan 15625
3002 nan nan nan nan nan nan 15625
3003 nan nan nan nan nan nan 15625
3004 nan nan nan nan nan nan 15625
3005 nan nan nan nan nan nan 15625

Now i use newer version of LAMMPS. But then i decided to use old version “lammps-16Jan09” . This time lammps is working without any problem. I ve got good and reliable results. It seems there is no any problem in my input file. So what i believe that there might be some bug or something else in new version.
Any idea?

All the best
Sefa

Hi All,

I am simulating polymer bundle by using class II force field on Cray XE6 .
There is no any problem with CG minimization. However, when i use NVT
program does
not work properly. All i ve got is " NAN" . Even i use low temperature,
all i ve got is "NAN" . I used small thermo to see whether there is wild
change or not. But from the first step it gave "NAN" .

Dumped positions "NAN"

Thermo output :

Step TotEng Temp Press Pxx Pyy Pzz Volume
3000 -19.99142 0 -0.1371864 -0.15156241 -0.012069375
-0.24792742 15625
3001 nan nan nan nan
nan nan 15625
3002 nan nan nan nan
nan nan 15625
3003 nan nan nan nan
nan nan 15625
3004 nan nan nan nan
nan nan 15625
3005 nan nan nan nan
nan nan 15625

Now i use newer version of LAMMPS. But then i decided to use old version
"lammps-16Jan09" . This time lammps is working without any problem. I ve
got good and reliable results. It seems there is no any problem in my input
file. So what i believe that there might be some bug or something else in
new version.
Any idea?

if you believe it is a bug in the newer version, then you have to post
an input deck that reproduces it , so that somebody can debug it
and fix the problem. one cannot debug beliefs.

axel.