Hi [name removed]!
I am running Artemis model for a significantly larger node points. While running the simulation, I am facing the error attached below.
Also ,I am using 64 bit architecture workstation, but still it seems my code is compiled in 32-bit. Anyway, I can change my FORTRAN compilation to 64-bit so that I wont face this issue. Also, I am attaching my configuration file in this message.
Hi
not fully sure the switch to 64-bit will solve the problem as it’s related to integer kind in the program.
This probably needs more in-depth change to works…
By the way, this seems in relation with direct solver. To check, you could try another solver, even this step will probably be much longer…
Another option could be to install mumps and use it then you will be able to run in parallel on smaller subdomains…
Hope this helps
Hi.
Thank you for the reply.
Is there any procedure for how to install mumps and proceed further? Also my node numbers are more than 1.5 million. Will mumps solver accomodate it?
[name removed]
Hi
Mumps is a third party library so you should look to the dedicated website and install it accordingly.
After this step, you should modify your config file accordingly (see systel.edf.cfg for example)
hope this helps
As you seems to be on windows, you can take a look at wintel script to build MUMPS.
https://gitlab.pam-retd.fr/otm/wintel/-/blob/main/build-telemac.bat?ref_type=heads
