<div dir="ltr">Hi All,<div><br></div><div>I have a large system (about 5K atoms, protein) that I am trying to run in LS-SCF. I have found that when I set MGRID too high (over about 300) I get this strange crash:</div><div><br></div><div><div>*******************************************************************************</div><div> *   ___                                                                       *</div><div> *  /   \                                                                      *</div><div> * [ABORT]                                                                     *</div><div> *  \___/                     Index to radix array not found.                  *</div><div> *    |                                                                        *</div><div> *  O/|                                                                        *</div><div> * /| |                                                                        *</div><div> * / \                                                      pw/fft_tools.F:292 *</div><div> *******************************************************************************</div><div><br></div><div><br></div><div> ===== Routine Calling Stack =====</div><div><br></div><div>            6 pw_grid_setup</div><div>            5 pw_env_rebuild</div><div>            4 qs_env_rebuild_pw_env</div><div>            3 qs_env_setup</div><div>            2 qs_init_subsys</div><div>            1 CP2K</div><div><br></div></div><div>I am on OLCF Summit, running 7.0 with DBCSR and GPU support. Also using threads and MPI. Here I am using about 24-54 nodes. </div><div>When I set the grid a bit coarser, I have no problems converging an SCF in under 5 minutes.</div><div><br></div><div>Thanks for any help!</div><div><br></div><div><br></div></div>