Memory problem in SAPT calculation

Dear PSI4 community,

I run a SAPT0/jun-cc-VDZ calculation for a dimer system containing approximately 100 atoms. Specifying one node, 4 processors, 200 GB memory in the input file. I found the job restart automatically without completing and changing to the other node. Our HPC cluster has 64 RAM, 24 processors per node. Please kindly help me if anyone solved such a problem.

Thanks and regards,
Kisan

I think that Psi4 cannot access memory across multiple nodes. If the nodes in your HPC cluster have 64 GB of memory, you cannot run a Psi4 calculation that requires more than 64 GB, even if your job has access to more memory than that on other nodes.

Thank you @ccavender for your kind suggestion