open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear Jean-Michel,

I am still battling with old problem, on trying to prescribe tracers values at a near closed wall boundary points. I have coded in bord3d as per your last post on this matter. It compiles but at the end I am getting the Segmentation fault error. I have tried both: serial and parallel with same results. Please advise if there is anything else to try. I appreciate your help. Please see attached fortarn and result with error message files from Serial and Parallel run. Thank you in advance.

Kind Regards!

Violeta

Hello Violeta,

A segmentation fault (in bord3d.f which is in you Fortran file) is due to a programming error, so we should find it. Maybe if you compile a version with full checkings and traceback of your compiler it will tell you at what line in the Fortran you have this error.

I see e.g. in your Fortran a test IF(IPLAN.GT.29.D0), it probably works but IPLAN.GT.29 would be safer.

Concentrate on serial run, as you use hardcoded boundary points it is not supposed to work in parallel like this.

With best regards,

Jean-Michel Hervouet

Thank you very much Jean-Michel. I will do so. Thank you for your helpful advice, as always.

Kind Regards!

Violeta

Dear Jean-Michel,

Segmentation Fault error.

I have been thinking and trying for few days but it turned out to be not so easy on my computer, as I thought. I could not recompile the full version of Telemac with -fcheck=all command, if that what you meant. Also I have tried to use valgrind for checking but libc6-dbg is missing on my pc and even if I uploaded this lib it is still not working.
Is there another way to check the fortran script? After running -fcheck=all with gcc I got some output, but not what I expected. Please see attached. I am still digging on how I can make the checking to work on my computer. I have tried on cluster too, but my authority is very limited there. It should not be a problem to do the -fcheck but it might be something with my gfortran compiler or I am just entering command not in right order, even if I searched all internet on this. Sorry to be a pain. Please advise. Thanks in advance.

Kind Regards!

Violeta

Hello Violeta,

So it would be simpler if I test your case on my computer, if it is not too heavy to send.

Regards,

JMH

Dear Jean-Michel,

Thank you very much for your help. This is the first thing I will do Monday morning. Unfortunately all my cases are saved on my computer in work. Thanks again.

Have a nice week-end!

Kind Regards!

Violeta

Dear Jean-Michel,

Thank you very much for your reply. Here are the files I am using to run this case. Please advise what can I do to improve the case. Thank you very much, your help is much appreciated.

Kind Regards!

Violeta

I have changed extension of cas file as it was not going through. Please see attached. Thank you.

Kind Regards!

Violeta

Hello Violeta,

Well received, thank you. The crash is in your modified bord3d.f, line 588. You are using here an undefined ITRAC, a loop from ITRAC=1 to NTRAC would be necessary around these lines with ITRAC.

Regards,

JMH

So quick! Thank you very much Jean-Michel. I will do so. Fingers crossed:). Thanks.

Kind Regards!

Violeta