Bug Reports for CFDEM®coupling, LIGGGHTS®, and ParScale

(Possible) bugs / suspicious behavior should be reported as a new thread here, not in the user forum. This should give both you and the developers a clear idea about the status of your bug report (submitted/assigned/fixed)

parallel calc crash

Submitted by zamir on

Hello all.

I run the attached input script using the standard liggghts-stable < in.triaxial_remix, and it runs just fine. When I run it with mpi using the command line: mpirun -np 4 liggghts-stable < in.triaxial_remix, it gives me a segmentation fault.

What is extremely puzzling is that I ran it on a 32 bit laptop with mpi just fine, and now on my workstation, it crashes.

Bug Report: Issues with torque calculation due to p_ref

Submitted by ryan.houlihan on

Hello,

Recently I noticed that using fix/mesh/gran/move and fix/gran/stressanalysis the torques were quite off. A stable force will lead to a ever increasing torque which is clearly immpossible. Under further investigation I found that the variable, p_ref, which is the center of the mesh, is not updated throughout the execution cycle.

Torque is computed as follows:

vectorSubtract3D(contactPoint,p_ref,tmp);
vectorCross3D(tmp,frc,tmp2);

wall/gran wiggle z not working?

Submitted by ericparteli on

Dear Sirs,

I can't set z-plane walls to wiggle in LIGGGHTS using fix/wall/gran.
When I use walls from a CAD file (mesh/gran) then wiggle works fine.

Please see the attached in.packing_wiggle file, which is nothing but the in.packing file from the EXAMPLES of LIGGGHTS with some small modifications, i.e. I just added gravity in z direction, changed the position of the z walls and made them wiggle by an amplitude of 1. And although it should cause an oscillation with a huge amplitude - if I did everything correctly - I don't see (?) the expected change due to the wiggle.

gran/hooke/history/stiffness crash

Submitted by joesatch on

Hi,

I am trying this out this pair style:

pair_style gran/hooke/history/stiffness
pair_coeff * *
fix m1 all property/global kn peratomtypepair 1 1.e8
fix m2 all property/global kt peratomtypepair 1 1.e8
fix m3 all property/global gamman peratomtypepair 1 0.1
fix m4 all property/global gammat peratomtypepair 1 0.1
fix m5 all property/global coefficientFriction peratomtypepair 1 0. # DOES NOT WORK!

The last property causes this crash:

lmp_mac_mpi(2028) malloc: *** error for object 0x10066c230: pointer being freed was not allocated

segfault when running MPI simulation

Submitted by msbentley on

Hi all,

I'm having a problem which only occurs when I run a simulation on multiple processors using MPI. My script is a hybrid granular/molecular simulation, which simply reads in a datafile defining ~2k aggregates (each of 32 spheres), and lets these settle under gravity. Running without MPI this works fine. With MPI (no matter how many processors I try to use) it consistently crashes with errors like:

Segmentation fault liggghts 1.5.3

Submitted by SergeiD on

I got segmentation fault with liggghts 1.5.3 while with 1.5.1 works fine. I attached test script.
Setting up run ...
[sega:20198] *** Process received signal ***
[sega:20198] Signal: Segmentation fault (11)
[sega:20198] Signal code: Address not mapped (1)
[sega:20198] Failing at address: 0x17e4a63e0
[sega:20198] [ 0] /lib/x86_64-linux-gnu/libc.so.6(+0x364c0) [0x7f3d5e9084c0]
[sega:20198] [ 1] lmp_openmpi(_ZN9LAMMPS_NS17FixContactHistory12pre_exchangeEv+0x1c8) [0x5804a8]
[sega:20198] [ 2] lmp_openmpi(_ZN9LAMMPS_NS6Modify18setup_pre_exchangeEv+0x43) [0x684903]

liggghts 1.5.2: ERROR: Dump local compute does not calculate local array

Submitted by SergeiD on

Hello.
When I run attached script with liggghts 1.5.2 I got the error:

ERROR: Dump local compute does not calculate local array

while the liggghts 1.5.1 there was no error.

UPD: may be it is related with bug node/758

wall/gran/local saves the out of date forces

Submitted by SergeiD on

I've noticed what sometimes wall/gran/local saves a contact's forces even if the contact is don't exists anymore. I attach screenshot. The red point (the particle) have id 71079. The red bar (the contact) also have id 71079. But the particle and the contact is detached from one another. Also one can see other contacts without particles, marked yellow arrows.

wall/gran/local and removing mesh/gran

Submitted by SergeiD on

If I want remove some mesh/gran from a simulation I use unfix/fix wall/gran procedure. But, as it turned out, it invalidates compute wall/gran/local for all wall/gran, not just removed. So, after doing unfix/fix wall/gran all dump forces form wall/gran/local is invalidated. If it is not a bug, seems, need to add WARNING about it.

Segmentation error

Submitted by SergeiD on

I need to remove the one mesh/gran after finish particles pouring and to continue simulation. Also I need to compute wall/gran/local and do several simulations with different parameters. But I got the segmentation fault when jump to next case if I remove mesh/gran by unfix/fix procedure. I attach example script. Please, tell me how to fix my script?