Dice icon indicating copy to clipboard operation
Dice copied to clipboard

Bus errors with PySCF/slurm/SHCI

Open jackweber1296 opened this issue 2 years ago • 1 comments

Hi,

Has anyone had issues with seemingly random (but fairly common) bus errors and out of range errors when running many Dice calculations (interfaced with PySCF)? We are having ~50% of our calculations crash when running Dice/SHCI and wanted to know if anyone else has seen similar behavior or if this is only an issue with our cluster.

Thanks, Jack

jackweber1296 avatar Jan 17 '23 20:01 jackweber1296

Hi Jack, Bus errors often happen when you recompile the program, causing the executable on the hard disk to change, while it is being used in a calculation. But if that is not what is happening in your case, can you send us an example input file that is causing the error.

Sandeep.

On Tue, Jan 17, 2023 at 12:40 PM Jack Weber @.***> wrote:

Hi,

Has anyone had issues with seemingly random (but fairly common) bus errors and out of range errors when running many Dice calculations (interfaced with PySCF)? We are having ~50% of our calculations crash when running Dice/SHCI and wanted to know if anyone else has seen similar behavior or if this is only an issue with our cluster.

Thanks, Jack

— Reply to this email directly, view it on GitHub https://github.com/sanshar/Dice/issues/9, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABVW4EP6F2RLSUDZK5O2TLWS37THANCNFSM6AAAAAAT6JT7D4 . You are receiving this because you are subscribed to this thread.Message ID: @.***>

sanshar avatar Jan 17 '23 21:01 sanshar