CCM+ limitations
Hi
It would be interesting to hear what your experience is with CCM+ regarding maximum number of mesh elements and solver capability. So I would like your all to add a comment where you first of all state the config of your machine(s) (OS, Memory, single/parallel) Next to that I would like you to state the maximum number of cells that you have reached without program crashes. (How is volume meshing working for you, polyhedral, trimmer) Finally I would like to hear about any solver experience (segreagated, coupled, when does it crash if it crashes at all) The reason for these questions is that currently I'm having problems with the volume mesher to go beyond approx 1.000.000 polyhedra cells. My config is windows, 2Gb, 2.2GHz dual. Crashes due to lack of virtual memory even though I have increased the limit of virtual memory to max. BR Kasper Skriver |
Re: CCM+ limitations
get rid of windows and move to suse linux
|
Re: CCM+ limitations
Serveral million cells without any problem on a similar machine with 4GB of memory and SUSE 10.2
|
Re: CCM+ limitations
Yes, but it is not possible due to company support limitations.
Still I think it is interesting for all of us to share experience on these topics. |
Re: CCM+ limitations
Do you have 32bit or 64bit windows? If its only 32bit then you are really killing yourself. You can't talk about big models with a 32bit OS.
|
Re: CCM+ limitations
Currently it is 32bits.
With in 3-4 months we are changing to 64 bits and win XP. The company where I'm working has decided to wait this long, and it is really killing me... |
Re: CCM+ limitations
I'm pretty much interested in using CCM+ in a Linux environment. Do you run OpenSuse or Suse Enterprise Linux?
|
Re: CCM+ limitations
OpenSuse 10.2
NVIDIA Quadro FX 550 , Core 2 Duo 2.4 GHz It's a very nice desktop machine. Also the Z-Buffer problems with ccm+ disappeared. |
Re: CCM+ limitations
I've found that with a 32 bit machine and Win XP the max amount of cells for a CCM+ simulation is somewhere around 1.2 million cells. I've since moved to a 64 bit machine with 10 Gb of RAM and Linux Suse Enterprise 10. Since I've done this my models are limited by CPU time rather than mesh size. I've generated meshes up to about 4 million cells since then.
I strongly suggest moving away from Windows. If CFD work is a primary component of your job / research, I suggest twisting your IT guy's arm, Linux is a much better environment. |
Re: CCM+ limitations
Thank you for your answer.
I see you get almost the same amount of polyhedral cells as I do. There are many issues preventing us from going to Linux, so it will be 64 bit win xp in the end. |
All times are GMT -4. The time now is 05:29. |