June 16, 2020, 16:17
|
identifying a bug efficiently
|
#1
|
Senior Member
Louis Gagnon
Join Date: Mar 2009
Location: Stuttgart, Germany
Posts: 306
Rep Power: 15
|
Hi,
I found a bug, but I'm not sure how to pinpoint where it comes from because the error is not consistently the same. I'm quite sure it's a bug because it happens on two different computers and with OF1906 and OF1912... It always occur during an iteration of the wall distance yWall using the (GAMG) advectionDiffusion (Poisson) method. If I use meshWave and leave everything else untouched the error does not occur.
It's always a segmentation fault, but with a different "description" which seems to depend on the state of the simulation. It also always occur after more of the the same runtime (in this case the clocktime is ~3 hours on 55 procs so not so quick to reproduce). I'm under the impression that some pointer exceeds its allowed dimensions, but only after a while and for a quite large simulation...
Can someone hint me as to the most efficient procedure to identify the source of the error?
|
|
|