problem with dump command

Hi,
I’m having a problem with the ave/atom command combined with the dump command. In the latter, I’m using the * character so that the timestep appears in the file name, but it seems to me that it doesn’t keep track of the timestep value correctly. Here is the part of my script that causes me problem, it comes right after initializing the system and reading a configuration file :

fix 1 all npt temp 300 300 0.1 iso 0.0 0.0 1.0 drag 2

compute POSITION all property/atom xu yu zu
fix P1 all ave/atom 50 100 25000 c_POSITION[1] c_POSITION[2] c_POSITION[3]
dump avP1 all custom 25000 out_test.* id f_P1[1] f_P1[2] f_P1[3]

run 100000

unfix 1
unfix P1
undump avP1

fix 2 all npt temp 300 1000 0.1 iso 0.0 0.0 1.0 drag 2
run 5000
unfix 2

fix P2 all ave/atom 50 100 25000 c_POSITION[1] c_POSITION[2] c_POSITION[3]
dump avP2 all custom 25000 out_test.* id f_P2[1] f_P2[2] f_P2[3]

fix 3 all npt temp 1000 1000 0.1 iso 0.0 0.0 1.0 drag 2
run 100000

Normally, I should get files with names out_test.0, out_test.25000, … , out_test.100000, out_test.100005, out_test.130000, …, out_test.205000. Instead, I get out_test.0, … out_test.100000, out_test.125000, …, out_test.200000. It’s as if the 5000 timesteps performed under fix 2 have not been added to the time steps value in the file names.

Is there a problem with how I’m using the command ? Thank you for your help.

Paule

The doc page for the dump command explains that it outputs
on multiples of the Nevery factor, which is 25000. It doesn’t
matter what timestep it is when you start, it will
always be multiples of 25K.

Fix ave/atom will also not produce output the first time
if you start it on a timestep (like 105000) that is not
a multiple of its output, which is 25K in your script.

Steve