**ERROR**: Particle with unique tag 190470 has NaN for its position.
373 views
Skip to first unread message
JM Kim
Mar 21, 2021, 9:56:39 AM3/21/21
to hoomd-users
Hello, I am having trouble with running HOOMD codes on GPU.
My system in interest is A31B7 block copolymer micelle in solvent particles and I slightly modified the DPD example provided in tutorial. (Tutorial-MD-DPD polymers.ipython)
It works wonderfully in CPU mode, but when I run it on Google Colab GPU mode, it shows
**ERROR**: Particle with unique tag 190470 has NaN for its position.
I searched for similiar cases.
However changing the particle position slightly and removing spherical spaces did not work on me. The number 190470 changes in different run.
I also tried to activate GPU on cmake but it did not work as well.
$ cmake /content -DENABLE_GPU=ON -Wno-dev
Best Regards,
Jaemin Kim
P.S. A31B7 code contains installing miniconda, hoomd, gsd, cmake etc.. cause i am working on Google Colab. Please ignore them.
Joshua Anderson
Mar 24, 2021, 1:57:49 PM3/24/21
to hoomd...@googlegroups.com
Jaemin,
This is informally known as the system "blowing up", when it becomes unstable and particles travel several simulation boxes in a single time step. It usually occurs when the step size dt is too large or particles are overlapping leading to high forces. However, you have a DPD simulation which has bounded forces even for overlapping particles and the viscous damping term tends to keep even unstable systems with very high dt from blowing up.
I don't have time to try running your script, maybe someone else on the mailing list can try it and see if they reproduce your behavior. I would suggest trying smaller values of dt, writing a trajectory out every frame and trying to identify which particles are gaining a lot of momentum and why.
------
Joshua A. Anderson, Ph.D.
Research Area Specialist, Chemical Engineering, University of Michigan
> --
> You received this message because you are subscribed to the Google Groups "hoomd-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to hoomd-users...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/hoomd-users/10ef06c6-1ef8-4dee-a27f-cc1a656a1360n%40googlegroups.com.
> <a31b7__50_25_50_15e4.py><unique tag has NaN for its position.PNG>
This is informally known as the system "blowing up", when it becomes unstable and particles travel several simulation boxes in a single time step. It usually occurs when the step size dt is too large or particles are overlapping leading to high forces. However, you have a DPD simulation which has bounded forces even for overlapping particles and the viscous damping term tends to keep even unstable systems with very high dt from blowing up.
I don't have time to try running your script, maybe someone else on the mailing list can try it and see if they reproduce your behavior. I would suggest trying smaller values of dt, writing a trajectory out every frame and trying to identify which particles are gaining a lot of momentum and why.
------
Joshua A. Anderson, Ph.D.
Research Area Specialist, Chemical Engineering, University of Michigan
> You received this message because you are subscribed to the Google Groups "hoomd-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to hoomd-users...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/hoomd-users/10ef06c6-1ef8-4dee-a27f-cc1a656a1360n%40googlegroups.com.
> <a31b7__50_25_50_15e4.py><unique tag has NaN for its position.PNG>
Reply all
Reply to author
Forward
0 new messages