running fix atom/swap in parallel

LAMMPS LAMMPS Mailing List Mirror
1

Dear all,

I’m testing new the feature “fix atom/swap”. It seems to work very well on a single core. When I run it on 4 cores it returned error:

ERROR: Delta_mu not allowed when not using semi-grand in fix atom/swap command (…/fix_atom_swap.cpp:216)

Then I initialized the variable:
ndeltamutypes=0;

in the function void FixAtomSwap::options(int narg, char **arg)

and now had another error:

Setting up run …
Memory usage per processor = 4.14107 Mbytes
Step Temp PotEng Lx Ly Lz Volume
10000 302.79298 -5.2033227 45.222532 45.143987 37.613579 76789.077
MPI Application rank 0 killed before MPI_Finalize() with signal 11

Now I have no idea how to fix this. I’m using the latest git version.

Regards,
Konstantin

2

Dear all,

I'm testing new the feature "fix atom/swap". It seems to work very well on a
single core. When I run it on 4 cores it returned error:

ERROR: Delta_mu not allowed when not using semi-grand in fix atom/swap
command (../fix_atom_swap.cpp:216)

Then I initialized the variable:
  ndeltamutypes=0;

in the function void FixAtomSwap::options(int narg, char **arg)

and now had another error:

Setting up run ...
Memory usage per processor = 4.14107 Mbytes
Step Temp PotEng Lx Ly Lz Volume
   10000 302.79298 -5.2033227 45.222532 45.143987 37.613579
76789.077
MPI Application rank 0 killed before MPI_Finalize() with signal 11

Now I have no idea how to fix this. I'm using the latest git version.

...which has been updated since you posted your e-mail. please update
and try again.

axel.

3

Hi Axel,

Yes, I’ve seen and updated it today. The bug is still there.

ERROR: Delta_mu not allowed when not using semi-grand in fix atom/swap command (…/fix_atom_swap.cpp:215)

Also, option ke=yes does’t conserve momentum. A structure starts to move with a constant speed…
I understood that fix has been added just two weeks ago, so many updates will appear.

Cheers,
Konstantin

4

Hi Axel,

Yes, I've seen and updated it today. The bug is still there.

ERROR: Delta_mu not allowed when not using semi-grand in fix atom/swap
command (../fix_atom_swap.cpp:215)

Also, option ke=yes does't conserve momentum. A structure starts to move
with a constant speed...
I understood that fix has been added just two weeks ago, so many updates
will appear.

and they are likely to appear faster, if you'd provide paul with a
suitable test case. :wink:

axel.

5

Here is my input. Also this bug appears on a single core if you comment command thermo_style in my script.

Co3Pt_swap.zip (346 KB)

6

Thanks for the test case. I’ll take a look and let you know what I find.

Paul

7

I just checked in a fix for this issue.

Paul

8

will be in next patch

Steve

9

On the non-conservation of momentum, note that you can use fix atom/swap’s “ke=no” option to conserve momentum (but not kinetic energy). See: http://lammps.sandia.gov/doc/fix_atom_swap.html . Also note that LAMMPS has a fix for re-zeroing momentum: http://lammps.sandia.gov/doc/fix_momentum.html . And LAMMPS also has various thermostat fixes for maintaining T and hence kinetic energy.

The other issues you previously mentioned have been addressed and the patch is now available.

Paul