From owner-chemistry@ccl.net Wed Nov 13 15:05:01 2013 From: "Laidig, Bill laidig.wd]=[pg.com" To: CCL Subject: CCL: Recommendation for QMD Review Message-Id: <-49337-131113150335-14412-oP42rbygZO+bgjQ1bIlZ/A#,#server.ccl.net> X-Original-From: "Laidig, Bill" Content-Language: en-US Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset="utf-8" Date: Wed, 13 Nov 2013 20:03:29 +0000 MIME-Version: 1.0 Sent to CCL by: "Laidig, Bill" [laidig.wd-.-pg.com] Hi, I am currently trying to decide what QMD program(s) to advocate for our modeling group and am looking for a review(s) of the current state-of-the-art in terms of: 1) Theory (what methodologies are in use, best practices, frontiers, areas of agreement/disagreement); 2) Programs (speed, scaling behavior in CPUs/GPUs, ease of use, analysis abilities, cost...); 3) Simulations (practical limits in box size, atoms number, element types, addition of constraints/pathways...) Most of what I have found is either too specific or biased. By the way, what I mean by QMD is Born-Oppenheimer dynamics (classical nuclei with QM electrons), either CP or PAW based. I am leaning toward Abinit (scales well, free) or VASP (fast, not free), but am pretty open to other suggestions. I am a traditional quantum chemist by training and have a lot of classical protein simulation experience as well, but not very knowledgable about plane wave simulations. I really want to find out what folks are using now and the computational cost in CPU hrs. for systems as a function of box size and simulation length. Thanks, Bill Bill Laidig BRTC-BRB 8611 Becket Rd West Chester, OH 45069-7053 From owner-chemistry@ccl.net Wed Nov 13 21:13:00 2013 From: "Michael Sullivan (IHPC) Michael#,#ihpc.a-star.edu.sg" To: CCL Subject: CCL: Recommendation for QMD Review Message-Id: <-49338-131113205108-31226-bLHr8ftGhfwSK+wGYgxfQQ__server.ccl.net> X-Original-From: "Michael Sullivan (IHPC)" Content-Language: en-US Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset="us-ascii" Date: Thu, 14 Nov 2013 09:50:56 +0800 MIME-Version: 1.0 Sent to CCL by: "Michael Sullivan (IHPC)" [Michael-.-ihpc.a-star.edu.sg] Bill, I would also consider Quantum ESPRESSO (http://www.quantum-espresso.org/). It can run on GPUs and has a number of good packages for post-processing. The reference for QE is here: P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo, A. Dal Corso, S. Fabris, G. Fratesi, S. de Gironcoli, R. Gebauer, U. Gerstmann, C. Gougoussis, A. Kokalj, M. Lazzeri, L. Martin-Samos, N. Marzari, F. Mauri, R. Mazzarello, S. Paolini, A. Pasquarello, L. Paulatto, C. Sbraccia, S. Scandolo, G. Sclauzero, A. P. Seitsonen, A. Smogunov, P. Umari, R. M. Wentzcovitch, J.Phys.:Condens.Matter 21, 395502 (2009), http://arxiv.org/abs/0906.2569 http://dx.doi.org/10.1088/0953-8984/21/39/395502 I find QE pretty easy to use and scales quite well. Mike michael]_[ihpc.a-star.edu.sg http://www.ihpc.a-star.edu.sg/ On 14 Nov, 2013, at 4:03 am, Laidig, Bill laidig.wd]=[pg.com wrote: > > Sent to CCL by: "Laidig, Bill" [laidig.wd-.-pg.com] > Hi, > > I am currently trying to decide what QMD program(s) to advocate for our modeling group and am looking for a review(s) of the current state-of-the-art in terms of: > > 1) Theory (what methodologies are in use, best practices, frontiers, areas of agreement/disagreement); > 2) Programs (speed, scaling behavior in CPUs/GPUs, ease of use, analysis abilities, cost...); > 3) Simulations (practical limits in box size, atoms number, element types, addition of constraints/pathways...) > > Most of what I have found is either too specific or biased. > > By the way, what I mean by QMD is Born-Oppenheimer dynamics (classical nuclei with QM electrons), either CP or PAW based. I am leaning toward Abinit (scales well, free) or VASP (fast, not free), but am pretty open to other suggestions. I am a traditional quantum chemist by training and have a lot of classical protein simulation experience as well, but not very knowledgable about plane wave simulations. I really want to find out what folks are using now and the computational cost in CPU hrs. for systems as a function of box size and simulation length. > > Thanks, > Bill > > Bill Laidig > BRTC-BRB > 8611 Becket Rd > West Chester, OH > 45069-7053> > IHPC Values :: Impact :: Honesty :: Performance :: Co-operation This email is confidential and may be privileged. If you are not the intended recipient, please delete it and notify us immediately. Please do not copy or use it for any purpose, or disclose its contents to any other person. Thank you. From owner-chemistry@ccl.net Wed Nov 13 23:30:01 2013 From: "Bradley Welch bwelch5[-]slu.edu" To: CCL Subject: CCL:G: Difference between Gaussian 09's SVP,TZVP and def2-svp, and def2-tzvp? Message-Id: <-49339-131113223236-563-zx+w+ZUe5URv62Rqd1oWAg:_:server.ccl.net> X-Original-From: "Bradley Welch" Date: Wed, 13 Nov 2013 22:32:35 -0500 Sent to CCL by: "Bradley Welch" [bwelch5[a]slu.edu] Dear all, I have some organometallic complexes I'd like to do an optimization on before I do some additional calculations (mainly energetics). From what I understand the SVP and TZVP basis sets in Gaussian 09 are older iterations of the aldrich basis sets. Barring age (and different exponents), are there any other differences in terms of extra functions on the newer ones? Before doing the optimization, I wanted to get an idea of time by doing a single point on the system with both sets of basis sets. I ran one with TZVP (the built in one in gaussian) on C,H,N,O and Br and used the def2-tzvp&pseudo potential on platinum. I reran the same molecule but with the updated basis sets on C,N,H,O,Br and def2-tzvp&pseudopotential on Pt. For the built in one I got a total of 2224 basis functions. With the def2 ones from the EMSL I got 3208 basis functions. The calculation using the def2 ones also took significantly longer. I'm including in this email dropbox links to my input files. Maybe my issue is I forgot a blank space or something like that. (using def2 basis sets from EMSL) https://www.dropbox.com/s/0x0wfs1ju8vxitw/dimerdef2pseudo.inp (using built in TZVP on C,H,N,O,Br, def2&pseudopotential on Pt) https://www.dropbox.com/s/5rs6nnbpnkg7in4/dimercomp3pipi.inp