[lammps-users] PPPM and FFTW?

LAMMPS LAMMPS Mailing List Mirror
1

Hi,

I intall the lammps with fftw-2.1.5. There isnot error as I install it. However, when I do the test with /examples/peptide in the lammps software. The strange errors appear, which are listed as follows:

The output (if any) follows:

LAMMPS (9 Jan 2009)
Scanning data file …
3 = max bonds/atom
6 = max angles/atom
14 = max dihedrals/atom
1 = max impropers/atom
Reading data file …
orthogonal box = (36.8402 41.0137 29.7681) to (64.2116 68.3851 57.1395)
2 by 2 by 4 processor grid
2004 atoms
2004 velocities
1365 bonds
786 angles
207 dihedrals
12 impropers
Finding 1-2 1-3 1-4 neighbors …
4 = max # of 1-2 neighbors
7 = max # of 1-3 neighbors
14 = max # of 1-4 neighbors
18 = max # of special neighbors
Finding SHAKE clusters …
19 = # of size 2 clusters
6 = # of size 3 clusters
3 = # of size 4 clusters
640 = # of frozen angles
84 atoms in group peptide
PPPM initialization …
G vector = 0.268721
grid = 15 15 15
stencil order = 5
RMS precision = 4.86052e-05
WARNING: Reducing PPPM order b/c stencil extends beyond neighbor processor
G vector = 0.266021
grid = 18 18 18
stencil order = 4
RMS precision = 5.61544e-05
brick FFT buffer size/proc = 1960 648 2352
Setting up run …
SHAKE stats (type/ave/delta) on step 0
4 1.111 1.44264e-05
6 0.996998 7.26967e-06
8 1.08 1.32536e-05
10 1.111 1.22749e-05
12 1.08 1.11767e-05
14 0.96 0
18 0.957206 4.37979e-05
31 104.519 0.00396029
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
WARNING: Shake determinant < 0.0
Memory usage per processor = 7.53338 Mbytes
---------------- Step 0 ----- CPU = 0.0000 (sec) ----------------
TotEng = 69910992907538569926244983351885057838284966357914337004224512.0000 KinEng = 1134.9186 Temp = 282.1005
PotEng = 69910992907538569926244983351885057838284966357914337004224512.0000 E_bond = 16.5572 E_angle = 36.3726
E_dihed = 15.5190 E_impro = 1.9426 E_vdwl = 692.8945
E_coul = 69910992907538569926244983351885057838284966357914337004224512.0000 E_long = -33567.2858 Press = nan
Exit code -5 signaled from b326
MPI process terminated unexpectedly
Killing remote processes…ERROR: Out of range atoms - cannot compute PPPM
DONE
Signal 15 received.

In addition, the errors always appear as long as I use the PPPM arithmetic. Why?

2

Hi,

I intall the lammps with fftw-2.1.5. There isnot error as I install it.
However, when I do the test with /examples/peptide in the lammps software.
The strange errors appear, which are listed as follows:

The output (if any) follows:

LAMMPS (9 Jan 2009)

this is an obsolete version of lammps. before reporting
any problems try with the current version first.

Scanning data file ...
3 = max bonds/atom
6 = max angles/atom
14 = max dihedrals/atom
1 = max impropers/atom
Reading data file ...

[...]

---------------- Step 0 ----- CPU = 0.0000 (sec)
----------------
TotEng =
69910992907538569926244983351885057838284966357914337004224512.0000 KinEng
= 1134.9186 Temp = 282.1005
PotEng =

this is a typical indication of an executable with a miscompiled tabulation.
this has been fixed a while ago and is properly documented in the installation
notes for the older version.

69910992907538569926244983351885057838284966357914337004224512.0000 E_bond
= 16.5572 E_angle = 36.3726
E_dihed = 15.5190 E_impro = 1.9426 E_vdwl =
692.8945
E_coul =
69910992907538569926244983351885057838284966357914337004224512.0000 E_long
= -33567.2858 Press = nan
Exit code -5 signaled from b326
MPI process terminated unexpectedly
Killing remote processes...ERROR: Out of range atoms - cannot compute PPPM
DONE
Signal 15 received.

In addition, the errors always appear as long as I use the PPPM arithmetic.
Why?

it is not the PPPM, but the tabulation used to speed up "/long" pair styles.

this very same symptom has been reported repeatedly over the last years.
it can be avoided by using compiler flags that tell gcc 4.x (and other
compilers)
that the older LAMMPS code does not 100% comply to ANSI-C++ standard,
or - better - use the current version of LAMMPS which does comply now.

cheers,
    axel.

3

I would guess you have some bad initial config that is
causing SHAKE to have problems. And that is leading
to bad initial energies. Can you run fine without shake?

Steve

4

69910992907538569926244983351885057838284966357914337004224512.0000 KinEng
= 1134.9186 Temp = 282.1005
PotEng =

this is a typical indication of an executable with a miscompiled tabulation.
this has been fixed a while ago and is properly documented in the installation
notes for the older version.

yes - I'd forgotten about this issue with the older version.
Please use/try the current version.

Steve