open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

Can you just change the extension .log to .txt (.log extensions are not allowed on this forum), or better send a zip with the 8 PE00007-0000*.log files please?

Chi-Tuan

Hi Chi-Tuan,

Attached is a zip file contain the PE000007-*.log files. Note, the PE00007-00000.log file isn't produced.

Cheers

Hi Glen,

There is no error in your log files. The segmentation fault may comes from the modification of the bief module.
As the bief module have been changed, did you recompile the whole telemac system?

I tried your case too, and I didn't have this error. But I have another one:
WARNING!!! THE POINT WITH X/Y IN THE MESH = 19527837.335698456 -4371866.5294836313
CONVERTED TO LONGITUDE/LATITUDE = ****************
SEEMS TO BE OUTSIDE OF THE EXTENT OF THE TIDAL SOLUTION
WHOSE MIN/MAX FOR LATITUDE ARE: -90.08 90.08
CHECK THE KEYWORDS GEOGRAPHIC SYSTEM + ZONE NUMBER IN GEOGRAPHIC SYSTEM = 1 -1


@Chi-Tuan, this error comes from the initialization stage where read_mesh_coord.f transforms the lat/lon into the mercator for Telemac projection.
I think that is why I had previously modified the tpxo module in the projection conversion parts:but it may not be the best fix.

Regards,
Laurent

I suggest to consider my previous modifications in tpxo.f for the case GEOSYST==1, because the mesh coordinates has been previously converted to a mercator projection in read_mesh_coord.f

@Chi-Tuan,
I let you see if there can be a better fix.
I also enclose the telemac3d_init.f
A call to write_mesh was not modified (for T3DHYD).




@Glen,
With these additional modifications, your case runs.
I think you should also provide the latitude and longitude of origin of your mesh to ensure a correct mercator projection.
I also suggest you to start with these default settings:
TIDE GENERATING FORCE = NO
MINOR CONSTITUENTS INFERENCE = NO
COEFFICIENT TO CALIBRATE TIDAL RANGE = 1.
COEFFICIENT TO CALIBRATE TIDAL VELOCITIES = 999999.

I hope it helps.

Regards,
Laurent

Hello Laurent and Glen,

Laurent is right, you should follow his suggestions, in particular to start with the default values.
I started to work to your issue but dit not have so much time today. I will try later (not tomorrow).

Anyway, I would be interested in your configuration (OS + version of compiler) because I cannot reproduce your segmentation fault, but I can see that the results are not good.

I changed some subroutines a little bit to convert to longitude/latitude but I still have an issue, which may come from the values given to LONGITUDE OF ORIGIN POINT and LATITUDE OF ORIGIN POINT as Laurent raised.
I "usually" choose (because I do not model so many large models which need spherical coordinates) the longitudes and latitude around the center of the model, in your case, you can try LATITUDE OF ORIGIN POINT = -36. and LONGITUDE OF ORIGIN POINT = 175.4. The Mercator projection should be better.

Just to let you know (in particular Laurent), with write_mesh subroutibe, I have to change the call to it in telemac3d_init subroutine with 3D work arrays as the coordinates in 3D concern the number of 3D nodes. I could be filled in with the first layer, and then this layer will be copied on the other horizontal levels. The optional argument will eventually be removed I think, but I have to test it first and clean... when I have time.

I let you know.

Sorry for the delay.

Chi-Tuan

Hello Laurent and Glen,

Please find enclosed my fix for hour case.
For the steering files, the main differences between what you initially gave and what I suggest are:

GEOGRAPHIC SYSTEM = 5
LATITUDE OF ORIGIN POINT = -36.
LONGITUDE OF ORIGIN POINT = 175.4

rather than

GEOGRAPHIC SYSTEM = 1
LATITUDE OF ORIGIN POINT = 0

I have commented the 2 keywords:
- PARALLEL PROCESSORS (you can steer with the option --ncsize=8 after the telemac3d.py command),
- DEBUGGER = 1 as it works here.
I also used the tpxo7 global solution rather than PO because I have already had it here, not the PO, but the feature should be the same for long/lat writing.

It seems to work in parallel here (112 cores).

Please let me know if it is OK for you @Glen.

Chi-Tuan

Hi Chi-Tuan and Laurent,

Thank you, with the updated fortran files the model is now working. Interestingly, after Laurent said the model was working, I tested it with my on my Windows-10/GFORTRAN build and it still failed to run with MPI (Note: I have tested Telemac3D with MPI on example models). So, I run the model on our HPC Linux build and it worked fine. With Chi-Tuan latest Fortran code it is working now with both Windows and Linux builds.

My only comment is that the model doesn't seem to be reading the TPXO SATELLITE ALTIMETRY initial conditions. I can work around this by running a 2D simulation and using this as a restart for the 3D model.

Thank you both.

Regards