Von: lc
Gesendet: Freitag, 18. Januar, 15:18
Betreff: Re: [DuMuX] set_singular_limit and parallel execution
An: dumux@listserv.uni-stuttgart.de


Hello,
which Dune version are you using? DuMux 2.12 is only compatible with the 2.6 
release not master.
I'm using DuMux 2.12 and using compatible branch. I use Dumux 3.0.0 I do not 
encouter such problem, but I would like to fix it on 2.12 also.


The fix for 2.12 is to use Dune 2.6. Dune 2.7 is not supported. You are on your 
own if you want to use Dumux 2.12 with Dune 2.7.


I assume you follow the instructions in the DuMux handbook (3.0) and the 
parallel section. Do you have more than 2 cores? If yes can you post some time 
measurements over the number of cores that we can look at? The kuvette test has 
only 120 cells, that's too small of a grid to run it efficiently on many cores. 
Did you try the same with more cells? Does anything change?
I follow instructions of DuMux handbook 2.12. Here some number:
I ran in /porousmediumflow/2p2c/sequential
and in test_adaptive2p2c2d.input I only changed the grid: Cells = 100 100
mpirun -np 1 ./test_adaptive2p2c2d --> Wall time 3856 sec.
mpirun -np 4 ./test_adaptive2p2c2d --> Wall time 6107 sec.
I observe such behaviour on both versions 2.12 and 3.0.0.

Can you please post the complete output for the run with 4 cores? Plus the 
initial vtk files? And, pretty please with sugar on top, the number of physical 
cores on your machine?

Bernd

Kind regards,
Lorenzo


_______________________________________________
Dumux mailing list
Dumux@listserv.uni-stuttgart.de
https://listserv.uni-stuttgart.de/mailman/listinfo/dumux

Reply via email to