From owner-chemistry@ccl.net Fri Mar 23 08:47:00 2007 From: "Steve Gwaltney drg51^-^ra.msstate.edu" To: CCL Subject: CCL:G: ccsd(t) disc requirements Message-Id: <-33886-070322192205-17375-ThTYOxmeFBBtcSpGE8wULw{:}server.ccl.net> X-Original-From: Steve Gwaltney Content-transfer-encoding: 7BIT Content-type: text/plain; charset=ISO-8859-1; format=flowed Date: Thu, 22 Mar 2007 17:21:24 -0500 MIME-version: 1.0 Sent to CCL by: Steve Gwaltney [drg51:_:ra.msstate.edu] ysubboti:ucalgary.ca wrote: > Sent to CCL by: ysubboti/./ucalgary.ca > Well, the system is quite big - 73 atoms. Actually I didn't specify any > symmetry, so by default it is C1. > 644 basis functions, 1085 primitive gaussians, 674 cartesian basis > functions > 129 alpha electrons 129 beta electrons. A CCSD(T) calculation of this size represents quite a challenge. Have you estimated how long this calculation will take, even if you get enough disk space? A quick estimate, based on a 400 basis function CCSD(T) calculation I ran recently, give a calculation time of 24 years! Even if my estimate is off significantly, I doubt you can run this calculation on a single processor. I suspect that your only hope is to either get time on a large supercomputer cluster or to switch to a linear scaling CCSD(T) code. Steve Gwaltney -- Dr. Steven Gwaltney Phone: 662-325-7602 Assistant Professor Fax: 662-325-1618 Department of Chemistry, Mail: Box 9573 Center for Environmental Health Sciences, Mississippi State University and HPCC Center for Computational Sciences Mississippi State, MS 39762 From owner-chemistry@ccl.net Fri Mar 23 10:02:01 2007 From: "Green Power powergreen .. gmail.com" To: CCL Subject: CCL:G: How to force two objects parallel during geometry optimization Message-Id: <-33887-070323095110-11197-59KksVNCZNqc4cEZpXh4dg::server.ccl.net> X-Original-From: "Green Power" Content-Type: multipart/alternative; boundary="----=_Part_216462_20454001.1174657851278" Date: Fri, 23 Mar 2007 08:50:51 -0500 MIME-Version: 1.0 Sent to CCL by: "Green Power" [powergreen:-:gmail.com] ------=_Part_216462_20454001.1174657851278 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Hi, For example, force two ammonia molecules parallel to each other during geometry optimization using Gaussian. In other words, the two objects are related by translation symmetry. But I don't want to do periodic calculation, because I just want to force two fragments in a molecule to be parallel to each other. ModRedundant function seems not sufficient to do this. Any suggestion is appreciated. Tian ------=_Part_216462_20454001.1174657851278 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline
Hi,

For example, force two ammonia molecules parallel to each other during geometry optimization using Gaussian. In other words, the two objects are related by translation symmetry. But I don't want to do periodic calculation, because I just want to force two fragments in a molecule to be parallel to each other. ModRedundant function seems not sufficient to do this.
Any suggestion is appreciated.

Tian
------=_Part_216462_20454001.1174657851278-- From owner-chemistry@ccl.net Fri Mar 23 10:55:01 2007 From: "Orlin Blajiev Orlin.Blajiev[#]vub.ac.be" To: CCL Subject: CCL: ONIOM and basis set Message-Id: <-33888-070323103727-29134-yXG9zQQN3wnSOQ3j5uhxXw()server.ccl.net> X-Original-From: Orlin Blajiev Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Fri, 23 Mar 2007 15:36:29 +0100 MIME-Version: 1.0 Sent to CCL by: Orlin Blajiev [Orlin.Blajiev_._vub.ac.be] Hi anybody, I will appreciate some information if it is possible to use ONIOM and different basis set for the atoms put in the high level. If is possible how would the input look like. Thank you in advance. Best regards, Orlin Orlin Blajiev Department of Metallurgy, Electrochemistry and Materials Science Faculty of Applied Science Vrije Universiteit Brussel Pleinlaan 2, B-1050 Brussels Belgium http://www.vub.ac.be/META/ tel.: 32-(0)2-6293538 fax : 32-(0)2-6293200 From owner-chemistry@ccl.net Fri Mar 23 11:19:01 2007 From: "Gustavo Seabra gustavo.seabra]^[gmail.com" To: CCL Subject: CCL:G: ccsd(t) disc requirements Message-Id: <-33889-070323103326-27478-4lyz0IBpZeA9BL59XgYnyQ]![server.ccl.net> X-Original-From: Gustavo Seabra Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Fri, 23 Mar 2007 10:33:00 -0400 MIME-Version: 1.0 Sent to CCL by: Gustavo Seabra [gustavo.seabra*|*gmail.com] Steve Gwaltney drg51^-^ra.msstate.edu wrote: > > Sent to CCL by: Steve Gwaltney [drg51:_:ra.msstate.edu] > ysubboti:ucalgary.ca wrote: >> Sent to CCL by: ysubboti/./ucalgary.ca > >> Well, the system is quite big - 73 atoms. Actually I didn't specify any >> symmetry, so by default it is C1. >> 644 basis functions, 1085 primitive gaussians, 674 cartesian basis >> functions >> 129 alpha electrons 129 beta electrons. > > A CCSD(T) calculation of this size represents quite a challenge. Have > you estimated how long this calculation will take, even if you get > enough disk space? A quick estimate, based on a 400 basis function > CCSD(T) calculation I ran recently, give a calculation time of 24 years! > Even if my estimate is off significantly, I doubt you can run this > calculation on a single processor. I suspect that your only hope is > to either get time on a large supercomputer cluster or to switch to a > linear scaling CCSD(T) code. > > Steve Gwaltney > Hi Steve, I'm not a CCSD guy, but I found you numbers very interesting. Could you detail a bit more how you reached them? Thanks, Gustavo Seabra. From owner-chemistry@ccl.net Fri Mar 23 13:38:00 2007 From: "Steve Gwaltney drg51{}ra.msstate.edu" To: CCL Subject: CCL:G: ccsd(t) disc requirements Message-Id: <-33890-070323133119-3437-ms2tCdf7bPd03baA9xMP5g^server.ccl.net> X-Original-From: Steve Gwaltney Content-transfer-encoding: 7BIT Content-type: text/plain; charset=ISO-8859-1; format=flowed Date: Fri, 23 Mar 2007 12:30:07 -0500 MIME-version: 1.0 Sent to CCL by: Steve Gwaltney [drg51]-[ra.msstate.edu] Gustavo Seabra gustavo.seabra]^[gmail.com wrote: > > Sent to CCL by: Gustavo Seabra [gustavo.seabra*|*gmail.com] > Steve Gwaltney drg51^-^ra.msstate.edu wrote: >> >> Sent to CCL by: Steve Gwaltney [drg51:_:ra.msstate.edu] >> ysubboti:ucalgary.ca wrote: >>> Sent to CCL by: ysubboti/./ucalgary.ca >> >>> Well, the system is quite big - 73 atoms. Actually I didn't specify any >>> symmetry, so by default it is C1. >>> 644 basis functions, 1085 primitive gaussians, 674 cartesian basis >>> functions >>> 129 alpha electrons 129 beta electrons. >> >> A CCSD(T) calculation of this size represents quite a challenge. Have >> you estimated how long this calculation will take, even if you get >> enough disk space? A quick estimate, based on a 400 basis function >> CCSD(T) calculation I ran recently, give a calculation time of 24 years! >> Even if my estimate is off significantly, I doubt you can run this >> calculation on a single processor. I suspect that your only hope is >> to either get time on a large supercomputer cluster or to switch to a >> linear scaling CCSD(T) code. >> >> Steve Gwaltney >> > Hi Steve, > > I'm not a CCSD guy, but I found you numbers very interesting. Could you > detail a bit more how you reached them? > > Thanks, > > Gustavo Seabra. I recently ran a 64 electron, 396 basis function single point CCSD(T) energy calculation in Cs symmetry using ACES II on one processor of an IBM pSeries 630 with 1.2 GHz processors. Although faster CPUs are now available, this computer has the advantage of having an I/O system designed to run CC calculations, giving a CPU time that was 95% of the wall clock time. The total wall clock time to run the calculation was 719666 seconds. The time estimate was calculated using the asymptotic scaling formula for CCSD(T) of o^3v^4/h^2, where o is the number of occupied orbitals, v is the number of virtual orbitals, and h is the order of the point group. To get the time estimate, I took the ratio of occupied orbitals (129/32) cubed times the ratio of the virtual orbitals (644-129)/(396-32) to the fourth power, times 4 for the difference in point group symmetries. That gave a scaling factor of 1050. I multiplied this by the time for my calculation to get the final estimate of 24 years. Clearly, several major assumptions are built into that time estimate. First, I am assuming that the original author is performing a RHF based calculation, as I did. An unrestricted calculation would add an extra factor of 2 to the time needed. Second, I did not include any correction for dropping orbitals. I actually dropped 9 core orbitals. But if you assume that the ratio of dropped occupied orbitals to total occupied orbitals is about the same for the two calculations, the estimate still holds. Third, I assumed that the time for the CC calculation scaled as the cost of the triples correction, instead of separating out the CCSD part from the (T) part. The CCSD calculation scales as o^2v^4 times the number of iterations needed to converge the CCSD equations. However, since the number of iterations is often roughly equal to the number of occupied orbitals, the CCSD part and the (T) part often take close to the same amount of time, and it is valid to estimate the total time from the (T) scaling equation only. Finally, I used the total time of the calculation for my estimate. I probably should have used only the time spent in the CC code for this estimate, and then scaled the SCF and integral transformation times separately. Doing so drops the estimate down to 21 years for a single point energy. What this exercise shows is that using current computer technology, the only way to run a CCSD(T) calculation with more than around 400 to 500 basis functions requires either a very large supercomputer cluster or requires the use of a linear scaling algorithm (and these have their own set of issues). Steve Gwaltney -- Dr. Steven Gwaltney Phone: 662-325-7602 Assistant Professor Fax: 662-325-1618 Department of Chemistry, Mail: Box 9573 Center for Environmental Health Sciences, Mississippi State University and HPCC Center for Computational Sciences Mississippi State, MS 39762 From owner-chemistry@ccl.net Fri Mar 23 14:23:03 2007 From: "Yubo Fan yubofan{=}mail.chem.tamu.edu" To: CCL Subject: CCL: cyclic peptide Message-Id: <-33891-070323123809-30004-DKwUDmLUpsr5uDvFVFM7BQ]|[server.ccl.net> X-Original-From: "Yubo Fan" Content-Transfer-Encoding: 8bit Content-Type: text/plain;charset="gb2312";format="flowed" Date: Fri, 23 Mar 2007 10:37:37 -0500 MIME-Version: 1.0 Sent to CCL by: "Yubo Fan" [yubofan===mail.chem.tamu.edu] Hi, everyone, I have a problem about visualizing a cyclic peptide on VMD. Is there anyway to define the connection between N- and C-terminals in a pdf file? Any Advice? Thanks. Yubo Fan yubofan_._mail.chem.tamu.edu From owner-chemistry@ccl.net Fri Mar 23 14:50:01 2007 From: "=?ISO-8859-1?Q?=D6d=F6n?= Farkas farkas(_)chem.elte.hu" To: CCL Subject: CCL:G: ONIOM and basis set Message-Id: <-33892-070323134230-5924-W0q2WdzHLyJvxnhzMYTRRA^server.ccl.net> X-Original-From: =?ISO-8859-1?Q?=D6d=F6n?= Farkas Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=UTF-8 Date: Fri, 23 Mar 2007 18:40:29 +0100 Mime-Version: 1.0 Sent to CCL by: =?ISO-8859-1?Q?=D6d=F6n?= Farkas [farkas:+:chem.elte.hu] Hi Orlin, If there is no better idea then it can be done using the "External" keyword ( http://www.gaussian.com/g_ur/k_external.htm ). A little programming might be necessary to make it work for external Gaussian calls but this way any kind of special layer can be used in ONIOM. Best wishes, Ödön On Fri, 2007-03-23 at 15:36 +0100, Orlin Blajiev Orlin.Blajiev[#]vub.ac.be wrote: > Sent to CCL by: Orlin Blajiev [Orlin.Blajiev_._vub.ac.be] > Hi anybody, > > I will appreciate some information if it is possible to use ONIOM and > different basis set for the atoms put in the high level. If is possible > how would the input look like. Thank you in advance. > > Best regards, > Orlin > > Orlin Blajiev > Department of Metallurgy, Electrochemistry and Materials Science > Faculty of Applied Science > Vrije Universiteit Brussel > Pleinlaan 2, B-1050 Brussels > Belgium > > http://www.vub.ac.be/META/ > > tel.: 32-(0)2-6293538 > fax : 32-(0)2-6293200> > > -- Ödön Farkas Associate professor Deparment of Organic Chemistry and Laboratory of Chemical Informatics, Institute of Chemistry, Eötvös Loránd University, Budapest Address: 1/A Pázmány Péter sétány, H-1117 Budapest, Hungary Phone: +36-1-372-2570 Cell phone: +36-30-255-3111 Fax: +36-1-372-2620 URL: http://organ.elte.hu/farkas From owner-chemistry@ccl.net Fri Mar 23 16:10:01 2007 From: "John Stone johns()ks.uiuc.edu" To: CCL Subject: CCL: cyclic peptide Message-Id: <-33893-070323160016-20158-haKAV/2B9LYv4gwhJ92Rbw|*|server.ccl.net> X-Original-From: John Stone Content-Disposition: inline Content-Type: text/plain; charset=us-ascii Date: Fri, 23 Mar 2007 14:21:23 -0500 Mime-Version: 1.0 Sent to CCL by: John Stone [johns{=}ks.uiuc.edu] Yubo, I have a test build of VMD 1.8.6 you can use which should work for your structure. Contact me off-list with information on what platform you're running on and I can post a test build that includes the cyclic structure handling code for you. Cheers, John Stone vmd/./ks.uiuc.edu On Fri, Mar 23, 2007 at 10:37:37AM -0500, Yubo Fan yubofan{=}mail.chem.tamu.edu wrote: > > Sent to CCL by: "Yubo Fan" [yubofan===mail.chem.tamu.edu] > > Hi, everyone, > > I have a problem about visualizing a cyclic peptide on > VMD. Is there anyway to define the connection between N- > and C-terminals in a pdf file? > > Any Advice? Thanks. > > Yubo Fan > yubofan[-]mail.chem.tamu.edu> > -- NIH Resource for Macromolecular Modeling and Bioinformatics Beckman Institute for Advanced Science and Technology University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801 Email: johns/./ks.uiuc.edu Phone: 217-244-3349 WWW: http://www.ks.uiuc.edu/~johns/ Fax: 217-244-6078 From owner-chemistry@ccl.net Fri Mar 23 23:16:00 2007 From: "Robert Fiske rfiske_===hotmail.com" To: CCL Subject: CCL: 2d graphing Message-Id: <-33894-070323150058-20909-/kftxhU/CjB0ACoR5KYaOQ||server.ccl.net> X-Original-From: "Robert Fiske" Date: Fri, 23 Mar 2007 15:00:55 -0400 Sent to CCL by: "Robert Fiske" [rfiske_(!)hotmail.com] I am trying to construct 2-D graphs of WHAM data, but am running into trouble. I have been trying to use Matplotlib's contour function to draw the graphs. However, with my Z values any combination of coordinates that were not sampled results in a 0 in the array (which is then graphed as if it were real data) I have tried interpolating the data usign polin2 in Numerical recipes but hit the same problem. This seems like a simple issue, so I figuring I am probably missing something obvious as 2-D graphing and WHAM seem pretty common. Am I using the wrong tool for the job, or have I left out a crucial step? Thank you for your time and any assistance you can provide Robert Fiske rfiske_]=[hotmail.com Cleveland State University Gogonea Group