code_saturne User's Forum

Hello Everybody,

I am using Code Saturne V3.0.1 and the humid atmospherics module. I have a problem when setting the sedimentation flag to 1 (routine atini1). When I do a mono processor run, my calculation ends correctly but crash at the second iteration when I do a parallel run (no matter the number of procs).

It seems that there is a mpi issue when sedimentation is active. Maybe there is a problem with my system?

I am using cohard and al for nucleation, Bouzereau for subgrid model and the gaussian distribution model.

Also I would like to have a confirmation of the output variables for this module :

LiqWater : droplets transported in kg/kg ?
Tot Water : water vapor + liquid water (droplets transported) in kg/kg ?
TotDrop : number of droplets ?

I red and the Bouzereau paper that there is a variable qr for the rainwater. Is this variable included in LiqWater ?

Thank you for your help

Regards

Hi again,

I have moved on the version 3.2.2 of code saturne. I re built my case with the same parameters. I have now an error message before calculation stage (whereas it is the same calculation than v3.0.1) :

SIGFPE signal (floating point exception) intercepted!

Call stack:
1: 0x2ba03c2f55bb <+0x26c5bb> (libsaturne.so.0)
2: 0x2ba03c2ff1dd <+0x2761dd> (libsaturne.so.0)
3: 0x2ba03c2ffa04 <cs_gradient_scalar+0x604> (libsaturne.so.0)
4: 0x2ba03c153df0 <cs_field_gradient_potential+0x90> (libsaturne.so.0)
5: 0x2ba03c154136 <cs_f_field_gradient_potential+0xe6> (libsaturne.so.0)
6: 0x2ba03c20b5e3 <predvv_+0xb03> (libsaturne.so.0)
7: 0x2ba03c1fd32f <navstv_+0x1e22> (libsaturne.so.0)
8: 0x2ba03c228205 <tridim_+0x46c5> (libsaturne.so.0)
9: 0x2ba03c10a77e <caltri_+0x34ae> (libsaturne.so.0)
10: 0x2ba03c0ec125 <cs_run+0x3f5> (libsaturne.so.0)
11: 0x2ba03c0ec2ed <main+0x16d> (libsaturne.so.0)
12: 0x2ba04001cd5d <__libc_start_main+0xfd> (libc.so.6)
13: 0x401649 <> (cs_solver)
End of stack

It seems that there is a problem with "HUMID" atmosphere (run goes normally when atmosphere is set to constant density).

In attached my listing of calculation, xml, meteo and user routines files in data.zip.

Thank you for your help

regards

Hello,

Version 3.2 is not maintained anymore, and 3.3 won't be for long, and does not have all bug fixes.

Version 4.0 beta will go into "release candidate" mode very soon (probably tomorrow, so check for updates), and as it is a validated version, it has many more bugfixes.

So you'll probably waste less much time rebuilding your data (again) with 4.0 than staying with 3.2 (sorry).

Actually, the automatic XML update you get by reloading your 3.2 XML file with 4.0 should mostly work, so you'll only need to check/fix your setup, not rebuild it completely.
You will need to rebuild your user subroutines, but in the case of boundary conditions, it should be mostly copy/paste (the number of arguments has changed, but the icodcl/rcodcl logic is still the same).

Regards,

Yvan

Hello Yvan,

Thank you for your indications.

Regards