Hello. When I use PETSc, when I use PETSc, the program can run smoothly when calculating the coarse grid(getARGV(“-global”, 128)). When the grid is refined(getARGV(“-global”, 2048)), PETSc calculates very slowly and reports an error. why?
space.edp (1.7 KB)
Use more MPI processes, or switch to a preconditioned iterative method instead of an exact LU solver.
Thank you for your answer. I tried the iterative method, which can be calculated, but when the grid is refined to h = 1 / 256, an error will still be reported; Can you provide the example of adding MPI process under PETSc mentioned above? thank you
ff-mpirun -n X script.edp , with an increasing
Hello, the number of processes has been increased, but the above problems still occur.
That means the problem is too big for your machine, you’ll need to switch to a cluster.