Hi,
The bug that caused "exploding" particles when they interact with mesh corners seems to be still there in version 2.1. I attach a case to illustrate the same. As you run the input file "Bug.in" and monitor the kinetic energy you will notice that it jumps between steps 432000 to 433000. The particle that explodes is particle id 58. The bug is processor dependent, similar to what Eric had reported earlier. If you use one, two or three processors everything is fine. The bug only occurs when I specify 4 processors.
Best,
Tapan
### Christoph.....I am having problems uploading the files on here, can I email them to you instead?? ############
cstoltz | Mon, 09/03/2012 - 04:07
Having similar problems
I'm still having a similar problem with one of the examples I have been running. I am observing a particle that falls onto a pyramid-type structure and suddenly accelerates. I only observe this, however, when I use 24 cores; the acceleration doesn't occur for this particular simulation when I use 12 or 36 cores (our machine is comprised of 12-core nodes), though it does occur for other particles at different times. I'm finding in this particular simulation that the mixer ends up 'leaking' a few particles more often than not due to this issue.
Regards,
Chris
ckloss | Mon, 09/03/2012 - 08:29
Hi Tapan and Chris, could you
Hi Tapan and Chris,
could you send these cases to Philippe Seil - you find his email on the project members page. Hopefully we'll get rid of that soon...
Thanks!
Christoph
tapsab | Wed, 09/05/2012 - 18:29
Any updates on the problem??
Any updates on the problem??
ckloss | Wed, 09/05/2012 - 23:52
We are working hard to
We are working hard (also in our spare time) to resolve this and other issues brought up by the community. Please be patient and rest assured that we'll let you know in due time when there is something to report
Christoph