Dear Wien2k Communoty, i 'm doing calculation for large nanowire structure ( ~ 350 atoms) cpu time calculation for lapw0 is 4 * less fast than lapw1? there is a mean to reduce it ? may be with appropriate .machines (how) because there is no parallelisation for lapw0? i use 2 k points and the .machines file 1:compute-1 1:compute-1 1:compute-1 1:compute-1 1:compute-1 1:compute-1 1:compute-1 1:compute-1 granularity:1 extrafine:1 start (Tue Nov 5 09:33:46 CET 2013) with lapw0 (40/99 to go) cycle 1 (Tue Nov 5 09:33:46 CET 2013) (40/99 to go) > lapw0 (09:33:46) 76361.156u 871.920s 21:38:34.84 99.1% 0+0k > 0+0io 151000pf+0w > lapw1 -c (07:12:21) 20041.500u 69.663s 5:35:49.96 99.8% 0+0k > 0+0io 163pf+0w > lapw2 -c (12:48:12) WARNING: EF not accurate, new > emin,emax,NE-min,NE-max -0.440243565733419 -0.440243559860867 935.999989735308 936.000005624997 3777.368u 426.339s 1:10:07.17 99.9% 0+0k 0+0io 10pf+0w > lcore (13:58:19) 10.514u 4.378s 0:15.62 95.2% 0+0k 0+0io 6pf+0w > mixer (13:59:18) 111.761u 108.394s 3:56.40 93.1% 0+0k 0+0io > 14pf+0w :ENERGY convergence: 0 0.0001 1.9910072550000000 :CHARGE convergence: 0 0.0000 2.4688953 ec cc and fc_conv 0 1 1 cycle 2 (Wed Nov 6 14:03:15 CET 2013) (39/98 to go) > lapw0 (14:03:15) 64170.045u 564.585s 18:08:09.24 99.1% 0+0k > 0+0io 146601pf+0w > lapw1 -c (08:11:25) 20268.697u 65.214s 5:39:54.68 99.7% 0+0k > 0+0io 163pf+0w > lapw2 -c (13:51:20) 3780.375u 406.236s 1:09:51.34 99.8% 0+0k > 0+0io 11pf+0w > lcore (15:01:11) 10.297u 4.136s 0:16.32 88.3% 0+0k 0+0io 6pf+0w > mixer (15:02:22) 112.917u 106.817s 3:57.76 92.4% 0+0k 0+0io > 12pf+0w :ENERGY convergence: 0 0.0001 .8686897250000000 :CHARGE convergence: 0 0.0000 2.2008227 ec cc and fc_conv 0 1 1 Regards
_______________________________________________ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html