Hello all,<div><br></div><div>I need to make use of range separated hybrid functionals for periodic systems, so I have pulled the latest development version of cp2k from github which adds this support to admm module (in hfx_admm_utils.F).</div><div><br></div><div>To the best of my knowledge, I have implemented it correctly, but the difference in final energy between PBE0 and MIX_CL_TRUNC is ~7 Ha, which seems rather large. When going from GGA -> PBE0 the energy difference <1 Ha. Obviously there should be some change when changing the interaction potential, but this has me considered.</div><div><br></div><div><b>GGA: </b> ENERGY| Total FORCE_EVAL ( QS ) energy (a.u.):             -983.840366364871670 <b>PBE0</b> : ENERGY| Total FORCE_EVAL ( QS ) energy (a.u.):             -983.680119471562080</div><div><b>MIX_CL_TRUNC</b>: ENERGY| Total FORCE_EVAL ( QS ) energy [a.u.]:    -974.290021294811368</div><div><br></div><div>Here is the XC section I am using to activate MIX_CL_TRUNC with ADMM. Can someone please sanity check and tell me if this is appropriate? </div><div><br></div><div><p>                &XC</p>
<p>                        &XC_FUNCTIONAL PBE</p>
<p>                                &PBE</p>
<p>                                        PARAMETRIZATION ORIG</p>
<p>                                        SCALE_C 1</p>
<p>                                        SCALE_X 0.665 # 1 - alpha - beta</p>
<p>                                &END PBE</p>
<p>                        &END XC_FUNCTIONAL</p>
<p>                        &HF</p>
<p>                                FRACTION 1</p>
<p>                                &SCREENING</p>
<p>                                        EPS_SCHWARZ 1e-07</p>
<p>                                        EPS_SCHWARZ_FORCES 1e-06</p>
<p>                                        SCREEN_ON_INITIAL_P True</p>
<p>                                        SCREEN_P_FORCES True</p>
<p>                                &END SCREENING</p>
<p>                                &INTERACTION_POTENTIAL</p>
<p>                                        POTENTIAL_TYPE MIX_CL_TRUNC</p>
<p>                                        CUTOFF_RADIUS 1.8</p>
<p>                                        T_C_G_DATA t_c_g.dat</p>
<p>                                        SCALE_LONGRANGE 0.085 # alpha</p>
<p>                                        SCALE_COULOMB 0.25 # beta</p>
<p>                                        OMEGA 0.94486</p>
<p>                                &END INTERACTION_POTENTIAL</p>
<p>                                &LOAD_BALANCE</p>
<p>                                        RANDOMIZE True</p>
<p>                                &END LOAD_BALANCE</p>
<p>                                &MEMORY</p>
<p>                                        EPS_STORAGE_SCALING 0.1</p>
<p>                                        MAX_MEMORY 2000</p>
<p>                                &END MEMORY</p>
<p>                        &END HF</p>
<p>                &END XC</p></div><div>Best,</div><div>Nicholas Winner</div>