[fluentfreak]

I am experiencing strange things using my upgraded 64-bit Fluent version. I had it installed on a linux machine; but moved it to a win64 machine to make sure it's not the OS fault. But the problem is still there.......now the problem:

3D Turbulent external flow (modeled as if it was in wind tunnel), 2.5 million grid cells, turb model:any, ran on parallel mode to use all 4-cores of the AMD QUAD CORE Processor

BUG: I get errors from AMG solver; the same thing used to run perfect on 32bit machine (parallel Quad Core). It also works well on my 64bit machine when I run as Single processor.

HAS ANYONE EXPERIENCED THE SAME THING?
ANY COMMENTS?

Thanks!!

[Shamoon Jamshed]

AMG solver is not the error of an operating system, you may have diveregence some wheare, try to see if your Boundary constiitons are correct.

[fluentfreak]

Originally Posted by Shamoon Jamshed

AMG solver is not the error of an operating system, you may have diveregence some wheare, try to see if your Boundary constiitons are correct.

Thanks for your reply.

I have the same model on a windows machine....everything is the same...all the settings.....the model...grid....

but the one on linux gives me this error message after about 1000timesteps.

I think that's the installation.....but how can I fix it?

[Shamoon Jamshed]

Try to reduce under relaxation factor
If it doesnt work then try to run first order with only flow equations solved, then also solve the energy equation and after some stable residuals start second order.
IT should nt be a problem in density based solver. If yes start with a lower courant number.