AMG troubles - help?
 

I started a parallel calculation with AMG. After 60 iterations the job is killed. Error message is following:

Process 1 amg_error: amg_trunc : error (1) occurred ierr_amg= 2981 Please contact CD-adapco support. STAR MPI_ABORT executed from processor 1

For an another mesh (same geometry ans settings) the job run until convergence.

What is the problem?

Thanks for your help.

Re: AMG troubles - help?
 

try CG algorithm. If it works, it means the AMG solver diverges.