cp2k cannot exit from queue when calculates large system being interfaced with ASE

Dear cp2k users and experts,

I want to use ASE as a job submission tool on supercomputer and further I plan to use ASE to perform Genetic Algorithm global optimization. However, I find for relatively large system containing more than 200 atoms (generated by GA, so it is always a “bad” structure), calculations always get stuck after 6 converged scf cycles regardless of calculation method (classical DFT or semi-empirical method GFN-xTB). Every time when I meet this problem, I check job status on queue, it is still “running” but actually there is no action can be seen.

I also find that for a system has about 20 atoms, cp2k can exit from queue successfully.

I attach all my input files and a detailed description in the following, any solutions and suggestions will be appreciated.

Thanks in advance 🙂

Yike

 

Read more here: Source link