[hitzhwan]

I always has errors like this, what is the problem?

Node 0: Process 31532: Received signal SIGSEGV.

================================================== ============================
MPI Application rank 0 exited before MPI_Finalize() with status 2

[LuckyTran]

sigsegv is a segmentation fault which most commonly means you tried to access a memory location that doesn't exist.


Either it happens sporadically because your have issues with your RAM stability. Or it happens very repeatedly because you're using a UDF.

[hitzhwan]

Originally Posted by LuckyTran

sigsegv is a segmentation fault which most commonly means you tried to access a memory location that doesn't exist.


Either it happens sporadically because your have issues with your RAM stability. Or it happens very repeatedly because you're using a UDF.

It always takes place, but this case does not have a udf ,can you tell me what is the reason ? is that the software's problem? The picture is error sketch.


360??20190109072934995.jpg

[LuckyTran]

Can you confirm whether it is numerical and specific to your case? If you run the case again does it crash at the same time?


Or is it hardware and indiscriminate? If you run a different case, a tutorial maybe, it also crashes?


It always crashes at the same time or randomly?

[hitzhwan]

Originally Posted by LuckyTran

Can you confirm whether it is numerical and specific to your case? If you run the case again does it crash at the same time?


Or is it hardware and indiscriminate? If you run a different case, a tutorial maybe, it also crashes?


It always crashes at the same time or randomly?

You ask a right question, it crashes randomly, I try several cases, sometimes it will happen randomly , but sometimes it won't, I do not know why ,and my case is simple, I thick there is no mistakes in my case.

[LuckyTran]

Originally Posted by hitzhwan

I always has errors like this, what is the problem?

Node 0: Process 31532: Received signal SIGSEGV.

================================================== ============================
MPI Application rank 0 exited before MPI_Finalize() with status 2

These two together means one of the parallel nodes couldn't communicate back with the head node.


So what's your hardware configuration look like? Single machine? Multiple machine? A simple network timeout or one machine going to sleep can cause this.

[Jiricbeng]

Generally Fluent is very unstable and abrupt crashes are not anything exceptional. I experienced several times that simulation even diverged suddenly and when running again (everything the same) it converged well. Not even talking about SIGSEGV error, it is casual.


Fluent/ANSYS developers should rather focus on stability improvement rather than changing icons, trying to give a new coat to an old code.

[hitzhwan]

Originally Posted by LuckyTran

These two together means one of the parallel nodes couldn't communicate back with the head node.


So what's your hardware configuration look like? Single machine? Multiple machine? A simple network timeout or one machine going to sleep can cause this.

Mine is a single machine,I have set the machine not go to sleep，but the problem exists ,too.

[jan.kracik]

Hello, has anyone succeeded in sorting out this issues? I have been experiencing this on and off but only in a parallel setup. It is very annoying because I am not able to fully utilise the power of my station. I have been thinking about starting with OpenFOAM. Ansys Fluent is very unstable, I have also experienced unjustified shutdowns for no reason. Thanks!

[vlad]

At my 17.2 Fluent version this would regularly happen after interpolation and restarting the simulation.
I found out that if instead dual precision and parallel if I just initially start the simulation as single precision serial it would work then stop and restart it as dp parallel and everything works.