<div dir="ltr">Dear Florian and dear Samuel <o:p> </o:p> first of all thank you for your answer. The system on which I am focused is a coated metallic cluster of about one hundred atoms (600 electrons in total). I would like to obtain the electronic spectrum of such a system at b3lyp level and it is good to know that for such type of system the standard algorithm is faster than the linear scaling one. <o:p> </o:p> About my computational resources, at the moment, I have access to Dual Intel Haswell E5-2670 CPUs (2.3 Ghz) giving 24 cores per node with 56Gb/s FDR Infiniband interconnect. When running on 192 cores CP2K needs 560 sec for each real-time-propagation step (rtps), that is too high to obtain a spectrum in a reasonable time (for me reasonable would be 24-48 hours for each polarization direction of light). In this case I used the PADE approximant for the exponential but considering your answer I have planned new tests with the ARNOLDI algorithm. About the ADMM basis set I tested FIT11 for the metal and cpFIT3 for the other species. Reducing the basis sets to FIT3 leads to a reduction of the time for rtps to 460 sec, that is still too high and probably will give a spectrum that would not be accurate. Is there the possibility to reduce that time without increasing the number of cores? Which are the keyword with which I can 'play' to reduce the time for rtps? <o:p> </o:p> Finally my latest question concerns the segmentation fault that I obtain when I use the AUXILIARY_DENSITY_MATRIX_METHOD with the cam-b3lyp functional. In this case the problem comes out at the beginning of the scf procedure. Did you ever try that combination ? Did you ever run up against a similar issue? <o:p> </o:p> For Samuel: the INPUT that gave me the segmentation fault is : <o:p> </o:p> &FORCE_EVAL<o:p></o:p> METHOD QUICKSTEP<o:p></o:p> &DFT<o:p></o:p> &AUXILIARY_DENSITY_MATRIX_METHOD<o:p></o:p> METHOD BASIS_PROJECTION<o:p></o:p> ADMM_PURIFICATION_METHOD NONE<o:p></o:p> &END<o:p></o:p> &REAL_TIME_PROPAGATION<o:p></o:p> MAX_ITER 7<o:p></o:p> MAT_EXP BCH <o:p></o:p> PROPAGATOR ETRS<o:p></o:p> EXP_ACCURACY 1.0E-11<o:p></o:p> EPS_ITER 1.0E-2<o:p></o:p> INITIAL_WFN SCF_WFN<o:p></o:p> DENSITY_PROPAGATION .TRUE.<o:p></o:p> &END<o:p></o:p> BASIS_SET_FILE_NAME ../../../data/BASIS_MOLOPT<o:p></o:p> BASIS_SET_FILE_NAME ../../../data/BASIS_ADMM<o:p></o:p> POTENTIAL_FILE_NAME ../../../data/POTENTIAL<o:p></o:p> &MGRID<o:p></o:p> CUTOFF 100<o:p></o:p> &END MGRID<o:p></o:p> &QS<o:p></o:p> EPS_DEFAULT 1.0E-10<o:p></o:p> MAP_CONSISTENT<o:p></o:p> &END QS<o:p></o:p> &POISSON<o:p></o:p> PERIODIC NONE<o:p></o:p> PSOLVER MT<o:p></o:p> &END<o:p></o:p> <o:p> </o:p> &SCF<o:p></o:p> MAX_SCF 5<o:p></o:p> EPS_SCF 1.0E-6<o:p></o:p> SCF_GUESS ATOMIC<o:p></o:p> &OT ON<o:p></o:p> ROTATION<o:p></o:p> &END OT<o:p></o:p> &END SCF<o:p></o:p> &LS_SCF<o:p></o:p> EPS_FILTER 1.0E-10<o:p></o:p> EPS_LANCZOS 1.0E-8<o:p></o:p> MAX_ITER_LANCZOS 30<o:p></o:p> SIGN_SQRT_ORDER 3<o:p></o:p> &END LS_SCF<o:p></o:p> &XC<o:p></o:p> &XC_FUNCTIONAL NONE<o:p></o:p> &END XC_FUNCTIONAL<o:p></o:p> &HF<o:p></o:p> &SCREENING<o:p></o:p> EPS_SCHWARZ 1.0E-10<o:p></o:p> EPS_SCHWARZ_FORCES 1.0E-10<o:p></o:p> SCREEN_ON_INITIAL_P FALSE<o:p></o:p> &END<o:p></o:p> &INTERACTION_POTENTIAL<o:p></o:p> POTENTIAL_TYPE COULOMB<o:p></o:p> &END<o:p></o:p> &MEMORY<o:p></o:p> MAX_MEMORY 50<o:p></o:p> EPS_STORAGE_SCALING 0.1<o:p></o:p> TREAT_FORCES_IN_CORE TRUE<o:p></o:p> &END<o:p></o:p> &END<o:p></o:p> &END XC<o:p></o:p> &END DFT<o:p></o:p> &SUBSYS<o:p></o:p> <o:p> </o:p> &CELL<o:p></o:p> ABC 8.0 8.0 8.0<o:p></o:p> PERIODIC NONE<o:p></o:p> &END CELL<o:p></o:p> &COORD<o:p></o:p> O 0.000000 0.000000 -0.065587<o:p></o:p> H 0.000000 -0.757136 0.520545<o:p></o:p> H 0.000000 0.757136 0.520545<o:p></o:p> &END COORD<o:p></o:p> &KIND H<o:p></o:p> BASIS_SET DZVP-MOLOPT-GTH<o:p></o:p> POTENTIAL GTH-PBE-q1<o:p></o:p> AUX_FIT_BASIS_SET cFIT3<o:p></o:p> &END KIND<o:p></o:p> &KIND O<o:p></o:p> BASIS_SET DZVP-MOLOPT-GTH<o:p></o:p> POTENTIAL GTH-PBE-q6<o:p></o:p> AUX_FIT_BASIS_SET cFIT3<o:p></o:p> &END KIND<o:p></o:p> &END SUBSYS<o:p></o:p> &END FORCE_EVAL<o:p></o:p> &GLOBAL<o:p></o:p> PROJECT H2O-hybrid-EMD<o:p></o:p> RUN_TYPE EHRENFEST_DYN<o:p></o:p> PRINT_LEVEL MEDIUM<o:p></o:p> &END GLOBAL<o:p></o:p> &MOTION<o:p></o:p> &MD<o:p></o:p> ENSEMBLE NVE<o:p></o:p> STEPS 2<o:p></o:p> TIMESTEP [au_t] 0.25<o:p></o:p> TEMPERATURE 300.0<o:p></o:p> &END MD<o:p></o:p> &END MOTION<o:p></o:p> On Tuesday, May 5, 2015 at 12:10:11 PM UTC+2, Florian Schiffmann wrote:<blockquote class="gmail_quote" style="margin: 0;margin-left: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;"><div dir="ltr">Hi Luca, you are trying some serious stuff there ;-) I am afraid, the linear scaling version doesn't support hybrids yet. It's related to the problem that the sparsity of the exchange part does not match the sparsity of the overlap matrix which is used throughout CP2K (that might have caused the segfault). However, for medium sized systems, up to 1600 atoms the standard algorithm is anyway faster if you use the Arnoldi method for the matrix exponential. Here, it is possible to use hybrid functionals. However, ADMM is somewhat more tricky as the purified versions do not yield an analytic KS matrix which is required by the RTP method. Therefore ADMM/RTP does only work with PURIFICATION NONE. I only tested it on the spectrum of a water molecule and it seemed to work fine, however there is no absolute warranty that it will always work. So if you want to use it, handle with care. Cheers Flo </div></blockquote><div> </div></div>