Reading Grid Problem

[Dewu-Yang]

Dear all,
When I encrypt the grid, the following error occurs. I was confused. The grid works fine with sparse points, but once the number of grid points increases, this error appears. I would greatly appreciate it if you could provide me some advice!

_forrtl: severe (111): position number is outside of the valid range, unit -265, file /dssg/home/acct-saaxh/saaxh/YDW/horses3d/Solver/test/NavierStokes/Blade/RESULTS/blade_0000000000.hsol
Image PC Routine Line Source
libifcoremt.so.5 000014C9F26E0E79 for__io_return Unknown Unknown
libifcoremt.so.5 000014C9F274793C for_write_seq Unknown Unknown
horses3d.ns 0000000000861FEC Unknown Unknown Unknown
horses3d.ns 00000000007AC56D Unknown Unknown Unknown
horses3d.ns 000000000058AFAA Unknown Unknown Unknown
horses3d.ns 0000000000433A77 Unknown Unknown Unknown
horses3d.ns 0000000000404812 Unknown Unknown Unknown
libc-2.28.so 000014C9EC8ECCF3 _libc_start_main Unknown Unknown
horses3d.ns 000000000040471E Unknown Unknown Unknown

Best,
Dewu

[loganoz]

Hi @Dewu-Yang,

Sorry for the delay in the response, I have been out of office these weeks.
I think we have some issues with meshes with a very big number of elements. How many elements does your mesh have?

Best,
Gonzalo

[Dewu-Yang]

Dear Prof.Gonzalo,

Thank you very much for your response. The quantity of my h-type grid is around twenty thousand. The grid encounters errors when I set p≥5. Lowering the order is computationally feasible. It is expected that the computation termination occurs when high polynomial orders are used due to excessively small sizes at the finest grid locations.

Best,
Dewu

[loganoz]

I think that is something different then.
Can you compile and run in debug mode so we get more information?
@oscarmarino, @hmarbona, @Andres-MG, @stecolumbus any ideas?

[oscarmarino]

The problem here is for very high DOF, the hsol file cannot write anymore because the integer in fortran is overflow. Do you have active the save gradients option? If it is, by disabling it you can still save results (without gradients). We are aware of this problem and will provide a solution in the near future.

[oscarmarino]

This has been fixed now with the newest version.