From owner-nwchem-users Fri Mar 12 12:12:19 1999 Received: (from majordom@localhost) by odyssey.emsl.pnl.gov (8.8.8+Sun/8.8.5) id MAA13356 for nwchem-users-outgoing; Fri, 12 Mar 1999 12:02:55 -0800 (PST) Content-return: allowed Date: Fri, 12 Mar 1999 12:02:45 -0800 From: "Windus, Theresa L" Subject: Two problems with NWChem release 3.2.1 To: "'nwchem-users@emsl.pnl.gov'" Message-id: <816AB6738C0ED11196C600609720F7BE01E692B9@PNLMSE8.emsl.pnl.gov> MIME-version: 1.0 X-Mailer: Internet Mail Service (5.5.2232.9) Content-type: text/plain Sender: owner-nwchem-users@emsl.pnl.gov Precedence: bulk Hello, Two problems have recently been identified for NWChem release 3.2.1. The first is that ECPs with a local potential with a zero power of R are not computed with the appropriate integrals. An example of an ECP with this problem is the LANL2DZ ECP for Uranium: ECP U nelec 78 U ul 0 4.51627250 -0.22527100 1 368.68769520 -53.78057010 2 112.75767080 -516.30520530 2 30.20329150 -202.67382220 2 10.29783550 -62.56440490 2 3.49105150 -21.55895590 2 1.26683910 -1.53905450 U s 0 257.35294420 2.61188160 1 83.16054280 81.23566570 2 29.72005130 353.24653640 2 6.24821320 311.92922340 2 4.16346720 -423.55257950 2 3.00015110 293.21608360 2 1.56000000 -21.34425520 2 0.87000000 1.66925480 U p 0 288.72982500 1.84143880 1 76.78384200 75.70547550 2 24.82520030 288.46016740 2 5.12487350 250.72894290 2 3.76033000 -318.89680610 2 2.97617060 160.08449240 2 1.99190170 33.78018910 2 0.97388320 -2.06999970 U d 0 350.46994070 2.90820640 1 166.70913010 74.35776280 2 85.82061190 587.11403290 2 26.98422550 317.25436790 2 7.15655920 102.84221170 2 1.28408550 22.42142780 U f 0 251.26707540 3.93503390 1 108.01092180 64.94341200 2 114.80218950 265.18047760 2 39.91125450 389.42840900 2 11.90397840 153.47672210 2 3.38507720 12.08074440 2 1.26776070 -1.35101380 END If there is a zero in the first column of the ECP description, wrong results will occur. We are currently working on correcting this problem. The second issue concerns the calculation of numerical hessians (frequencies). The numerical step size between versions 3.2.0 and 3.2.1 was changed from 0.01 to 0.001. However, even though this helped a few users, it turns out to be too small in general cases to provide numerically accurate results. Therefore, users should change this value in calculations with version 3.2.1. This can be accomplished by adding the following line to the input deck before requesting the hessian or frequency: set stpr_gen:delta 0.01 This behavior will be changed in the next release. If you have additional questions concerning this information, please send an e-mail to mscf-consulting@emsl.pnl.gov. Sincerely, Theresa Windus MSCF Scientific Consulting Theresa Windus Environmental Molecular Science Laboratory Pacific Northwest National Laboratory tel: 509-372-6410 3335 Q. Ave., Mail Stop K1-83 fax: 509-375-6631 Richland, WA 99352 e-mail: Theresa.Windus@pnl.gov