Running High-Resolution Basilisk Simulation on Supercomputer Without Root Access

101 views
Skip to first unread message

Nivesh Mishra

unread,
Jul 30, 2025, 8:48:54 AMJul 30
to basilisk-fr

Hello Basilisk users,

I’m currently running a simulation using Basilisk that requires a high grid resolution, and therefore demands significant computational resources. I would like to run this simulation on a supercomputer; however, I do not have root access on the system.

Is there a recommended way to run Basilisk in such an environment without requiring administrative privileges? Any guidance on setting up and executing high-resolution simulations under these constraints would be greatly appreciated.

Thank you in advance for your support!

Edoardo Cipriano

unread,
Jul 30, 2025, 9:25:49 AMJul 30
to basilisk-fr
Hello,

I think nobody has root privileges on supercomputers, unless they are the admins.

Check out the following link for running simulations on a supercomputer. This method works even if Basilisk is not installed in the system - you just need mpi.


The logic is:
1. On your local machine, you compile the Basilisk C code into a plain C99 code using the -source option of qcc.
2. Copy the resulting file to the supercomputer.
3. Write a run.sh script to be submitted to the queue system of the supercomputer. The file compiles the C99 source code and runs the simulation.

Otherwise, if you can install Basilisk on the supercomputer (it should not require root privileges) you can perform step 1 directly on the remote machine. 

Best,
Edoardo

Conor Olive

unread,
Jul 30, 2025, 9:36:12 AMJul 30
to basilisk-fr
Supercomputer users are virtually never given root access as it would not be practical for many reasons. There is a page somewhere on basilisk.fr explaining how to compile into portable C code using the -source option on qcc. I think that is all you should probably need to know from the basilisk side of things that would be different from your local machine. Other than that, anything particular to site packages (e.g. loading compilers, user programs, shared libraries) and job scheduling (e.g. slurm) will be particular to your cluster and there should be a help page for it somewhere giving you a good place to start learning more. Probably wise to have some basic knowledge about MPI as well.

Nivesh Mishra

unread,
Jul 30, 2025, 11:08:24 AMJul 30
to Conor Olive, basilisk-fr

Thank you, Edoardo and Conor. I will try using your suggestions. I hope it works well.


--
You received this message because you are subscribed to the Google Groups "basilisk-fr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to basilisk-fr...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/basilisk-fr/fe10b87f-4a8d-4064-b25d-2a566ee75715n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages