From owner-chemistry@ccl.net Mon Jun 5 08:19:00 2006 From: "Tirath Ramdas tirath(_)tpg.com.au" To: CCL Subject: CCL:G: AO angular momentum statistical distribution Message-Id: <-31896-060605075209-24949-hiR/g/PxkWBPXW6a1qx+Gg{:}server.ccl.net> X-Original-From: "Tirath Ramdas" Content-Transfer-Encoding: 7bit Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=original Date: Mon, 5 Jun 2006 20:26:03 +1000 MIME-Version: 1.0 Sent to CCL by: "Tirath Ramdas" [tirath{=}tpg.com.au] Hi Jeff, Thanks for your reply, but unless I misunderstand you, I think we both have different things in mind. Let me elaborate/clarify a bit, using the Rys quadrature method [1] as an example. One of the cost-determining factors for the Rys quadrature are the angular momentum xyz triplets for each of the four basis functions, eg for Gaussian primitive i, \lambda_i=nx+ny+nz, where \lambda_i is "closely related to the total angular momentum quantum number" [1]. These values determine: 1) The number of roots required to compute the quadrature -- NROOTS=(lambda_i+lambda_j+lambda_k+lambda_l)/2, as explained in [1]. 2) The depth of the recurrence that needs to be computed to get the 'I factors' -- as described in [1]. I have a handle on that. What I don't have a handle on is the nature of the basis functions that "typically" arrive at the input of the integral routine (specifically, the angular momenta associated with basis functions, which determine the nx,ny,nz). To put it another way; from recent observations of large workloads (i.e. big molecules, big basis sets), are most basis functions that are input into the quadrature routine low angular momentum or high angular momentum? Clearly an important factor is the choice of basis set, but I wonder if other things like integral screening also skew things in a manner related to angular momentum. It seems to me that an understanding of the ERI routine (or even the entire SCF procedure) in isolation is not sufficient to steer optimisation efforts because the performance is workload dependant. Ideally I would like a picture of the distribution describing workload characteristics, specifically angular momentum associated with basis functions. Bottom line > from my point of view is the old saying "make the common case fast"... maybe oversight on my part, but I've not been able to identify the common case. If there is such a thing, I'm sure someone on this list would know it :-) I'd even be very happy with domain specific workload intuition; eg "in transition metals we typically find that ..." Thanks again! -tirath Monash University [1] Rys, Dupuis, King, "Computation of Electron Repulsion Integrals Using the Rys Quadrature Method", J. Comp. Chem, p 154, vol4, 2, 1983 ----- Original Message ----- > From: "jhammond-x-uchicago.edu" To: "Ramdas, Tirath " Sent: Monday, June 05, 2006 2:46 AM Subject: CCL: AO angular momentum statistical distribution > Sent to CCL by: [jhammond=uchicago.edu] > GAMESS own documentation > (http://www.spec.org/hpc96/docs/RelatedPublications/gamess/re > fs.html is one version online) has the citations which give > the integral procedures in detail, from which you should be > able to figure out all you need. The computational cost of > SCF procedures is one of the most thoroughly studied aspects > of quantum chemistry and understanding the computational > cost of these procedures is far from intractible if you are > willing to read a few papers. > > After checking out the GAMESS-related literature you might > check out > > International Journal of Quantum Chemistry > Volume 40, Issue 6 , Pages 753 - 772 > The prism algorithm for two-electron integrals > Peter M. W. Gill, John A. Pople > > as one of but many papers on the subject which detail > integral algorithms and their cost. > > Jeff Hammond > University of Chicago From owner-chemistry@ccl.net Mon Jun 5 11:28:01 2006 From: "Manish Sud msud===san.rr.com" To: CCL Subject: CCL: New release of MayaChemTools package... Message-Id: <-31897-060605112606-6452-6FvkVR/wHLqeXOdVEXnL3w,,server.ccl.net> X-Original-From: "Manish Sud" Date: Mon, 5 Jun 2006 11:26:00 -0400 Sent to CCL by: "Manish Sud" [msud*san.rr.com] A new release of MayaChemTools, a growing collection of Perl scripts to support computational discovery needs, is available as free software; you can redistribute it and/or modify it under the terms of the GNU LGPL. New scripts: o InfoPeriodicTableElements.pl o InfoAminoAcids.pl o EelementalAnalysis.pl o ElementalAnalysisSDFiles.pl o ElementalAnalysisTextFiles.pl And enhancements to existing scripts. For further details about the scripts and to download the package, please visit www.MayaChemTools.org. "Interesting scripts", to paraphrase an old saying, "lie ahead." And I'm not done scripting yet. In the mean time, enjoy the current scripts. Your feedback is welcome. Manish Sud msud^^san.rr.com From owner-chemistry@ccl.net Mon Jun 5 15:32:01 2006 From: "Gustavo Mercier gamercier+*+yahoo.com" To: CCL Subject: CCL:G: AO angular momentum statistical distribution Message-Id: <-31898-060605144410-21050-wCzzRfwZkowG12YxJa6UTQ.@.server.ccl.net> X-Original-From: Gustavo Mercier Content-Transfer-Encoding: 8bit Content-Type: multipart/alternative; boundary="0-2056697415-1149533041=:81249" Date: Mon, 5 Jun 2006 11:44:01 -0700 (PDT) MIME-Version: 1.0 Sent to CCL by: Gustavo Mercier [gamercier,+,yahoo.com] --0-2056697415-1149533041=:81249 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Hi! It seems to me that what you are looking for is the distribution of the angular momentum (ni's) in the basis sets that are commonly used in quantum chemistry. This are my qualitative observations: 1) This has changed as computational power has increased. 2) For organic molecules it is common to have s,p,d functions (ie. n=0,1,2). It is unusual to go to f functions. 3) For inorganic molecules or organometallics, f and g functions may be added (n=3,4). I doubt that you see anyone using beyond g functions. 3) A common basis set for organic molecules is 6-31G(p,d) 4) You may be able to compute a rough frequency by looking at the type of atoms in a database (something simple like the Merck Test Collection in the CCL) and assigning values of n. For example, Carbon would have 1s,2s,2p,3d etc. I hope this helps. Take care! GM "Tirath Ramdas tirath(_)tpg.com.au" wrote: Sent to CCL by: "Tirath Ramdas" [tirath{=}tpg.com.au] Hi Jeff, Thanks for your reply, but unless I misunderstand you, I think we both have different things in mind. Let me elaborate/clarify a bit, using the Rys quadrature method [1] as an example. One of the cost-determining factors for the Rys quadrature are the angular momentum xyz triplets for each of the four basis functions, eg for Gaussian primitive i, \lambda_i=nx+ny+nz, where \lambda_i is "closely related to the total angular momentum quantum number" [1]. These values determine: 1) The number of roots required to compute the quadrature -- NROOTS=(lambda_i+lambda_j+lambda_k+lambda_l)/2, as explained in [1]. 2) The depth of the recurrence that needs to be computed to get the 'I factors' -- as described in [1]. I have a handle on that. What I don't have a handle on is the nature of the basis functions that "typically" arrive at the input of the integral routine (specifically, the angular momenta associated with basis functions, which determine the nx,ny,nz). To put it another way; from recent observations of large workloads (i.e. big molecules, big basis sets), are most basis functions that are input into the quadrature routine low angular momentum or high angular momentum? Clearly an important factor is the choice of basis set, but I wonder if other things like integral screening also skew things in a manner related to angular momentum. It seems to me that an understanding of the ERI routine (or even the entire SCF procedure) in isolation is not sufficient to steer optimisation efforts because the performance is workload dependant. Ideally I would like a picture of the distribution describing workload characteristics, specifically angular momentum associated with basis functions. Bottom line > from my point of view is the old saying "make the common case fast"... maybe oversight on my part, but I've not been able to identify the common case. If there is such a thing, I'm sure someone on this list would know it :-) I'd even be very happy with domain specific workload intuition; eg "in transition metals we typically find that ..." Thanks again! -tirath Monash University [1] Rys, Dupuis, King, "Computation of Electron Repulsion Integrals Using the Rys Quadrature Method", J. Comp. Chem, p 154, vol4, 2, 1983 ----- Original Message ----- > From: "jhammond-x-uchicago.edu" To: "Ramdas, Tirath " Sent: Monday, June 05, 2006 2:46 AM Subject: CCL: AO angular momentum statistical distribution > Sent to CCL by: [jhammond=uchicago.edu] > GAMESS own documentation > (http://www.spec.org/hpc96/docs/RelatedPublications/gamess/re > fs.html is one version online) has the citations which give > the integral procedures in detail, from which you should be > able to figure out all you need. The computational cost of > SCF procedures is one of the most thoroughly studied aspects > of quantum chemistry and understanding the computational > cost of these procedures is far from intractible if you are > willing to read a few papers. > > After checking out the GAMESS-related literature you might > check out > > International Journal of Quantum Chemistry > Volume 40, Issue 6 , Pages 753 - 772 > The prism algorithm for two-electron integrals > Peter M. W. Gill, John A. Pople > > as one of but many papers on the subject which detail > integral algorithms and their cost. > > Jeff Hammond > University of Chicagohttp://www.ccl.net/cgi-bin/ccl/send_ccl_messagehttp://www.ccl.net/chemistry/sub_unsub.shtmlhttp://www.ccl.net/spammers.txt-- Gustavo A. Mercier, Jr. MD,PhD Baylor University Medical Center Radiology American Radiology Associates 712 N. Washington, Suite 101 Dallas, TX 75246 214-826-8822 214-826-9792 fax gamercier]![yahoo.com --0-2056697415-1149533041=:81249 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: 8bit
Hi!
 
It seems to me that what you are looking for is the distribution of the angular momentum (ni's) in the basis sets that are commonly used in quantum chemistry.
 
This are my qualitative observations:
 
1) This has  changed as computational power has increased.
2) For organic molecules it is common to have s,p,d functions (ie. n=0,1,2). It is unusual to go to f functions.
3) For inorganic molecules or organometallics, f and g functions may be added (n=3,4). I doubt that you see anyone using beyond g functions.
3) A common basis set for organic molecules is 6-31G(p,d)
4) You may be able to compute a rough frequency by looking at the type of atoms in a database (something simple like the Merck Test Collection in the CCL) and assigning values of n. For example, Carbon would have 1s,2s,2p,3d etc.
 
I hope this helps.
 
Take care!
GM
 


"Tirath Ramdas tirath(_)tpg.com.au" <owner-chemistry]![ccl.net> wrote:
Sent to CCL by: "Tirath Ramdas" [tirath{=}tpg.com.au]
Hi Jeff,

Thanks for your reply, but unless I misunderstand you, I think we both have
different things in mind. Let me elaborate/clarify a bit, using the Rys
quadrature method [1] as an example.

One of the cost-determining factors for the Rys quadrature are the angular
momentum xyz triplets for each of the four basis functions, eg for Gaussian
primitive i, \lambda_i=nx+ny+nz, where \lambda_i is "closely related to the
total angular momentum quantum number" [1]. These values determine:

1) The number of roots required to compute the quadrature --
NROOTS=(lambda_i+lambda_j+lambda_k+lambda_l)/2, as explained in [1].
2) The depth of the recurrence that needs to be computed to get the 'I
factors' -- as described in [1].

I have a handle on that.

What I don't have a handle on is the nature of the basis functions that
"typically" arrive at the input of the integral routine (specifically, the
angular momenta associated with basis functions, which determine the
nx,ny,nz). To put it another way; from recent observations of large
workloads (i.e. big molecules, big basis sets), are most basis functions
that are input into the quadrature routine low angular momentum or high
angular momentum? Clearly an important factor is the choice of basis set,
but I wonder if other things like integral screening also skew things in a
manner related to angular momentum.

It seems to me that an understanding of the ERI routine (or even the entire
SCF procedure) in isolation is not sufficient to steer optimisation efforts
because the performance is workload dependant. Ideally I would like a
picture of the distribution describing workload characteristics,
specifically angular momentum associated with basis functions. Bottom line
> from my point of view is the old saying "make the common case fast"... maybe
oversight on my part, but I've not been able to identify the common case. If
there is such a thing, I'm sure someone on this list would know it :-)

I'd even be very happy with domain specific workload intuition; eg "in
transition metals we typically find that ..."

Thanks again!

-tirath
Monash University

[1] Rys, Dupuis, King, "Computation of Electron Repulsion Integrals Using
the Rys Quadrature Method", J. Comp. Chem, p 154, vol4, 2, 1983


----- Original Message -----
> From: "jhammond-x-uchicago.edu"
To: "Ramdas, Tirath "
Sent: Monday, June 05, 2006 2:46 AM
Subject: CCL: AO angular momentum statistical distribution


> Sent to CCL by: [jhammond=uchicago.edu]
> GAMESS own documentation
> (http://www.spec.org/hpc96/docs/RelatedPublications/gamess/re
> fs.html is one version online) has the citations which give
> the integral procedures in detail, from which you should be
> able to figure out all you need. The computational cost of
> SCF procedures is one of the most thoroughly studied aspects
> of quantum chemistry and understanding the computational
> cost of these procedures is far from intractible if you are
> willing to read a few papers.
>
> After checking out the GAMESS-related literature you might
> check out
>
> International Journal of Quantum Chemistry
> Volume 40, Issue 6 , Pages 753 - 772
> The prism algorithm for two-electron integrals
> Peter M. W. Gill, John A. Pople
>
> as one of but many papers on the subject which detail
> integral algorithms and their cost.
>
> Jeff Hammond
> University of Chicago


http://www.ccl.net/cgi-bin/ccl/send_ccl_message
http://www.ccl.net/cgi-bin/ccl/send_ccl_message
http://www.ccl.net/chemistry/sub_unsub.shtml
http://www.ccl.net/spammers.txt






--
Gustavo A. Mercier, Jr. MD,PhD
Baylor University Medical Center
Radiology
American Radiology Associates
712 N. Washington, Suite 101
Dallas, TX 75246
214-826-8822
214-826-9792 fax
gamercier]![yahoo.com 
--0-2056697415-1149533041=:81249-- From owner-chemistry@ccl.net Mon Jun 5 16:07:01 2006 From: "Victor Manuel Rosas-Garcia quimico69-,-yahoo.com" To: CCL Subject: CCL: Amber parameters Message-Id: <-31899-060605142559-19171-5CbJghTY59lLI2EKBWp8YQ^-^server.ccl.net> X-Original-From: Victor Manuel Rosas-Garcia Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=iso-8859-1 Date: Mon, 5 Jun 2006 10:25:51 -0700 (PDT) MIME-Version: 1.0 Sent to CCL by: Victor Manuel Rosas-Garcia [quimico69]^[yahoo.com] Hi Markus, If I remember properly, the numbers are: 1 (for highest quality parameters), 2 (for tentative values) and 3 (for generalized, and thus lowest quality, values). --- "Markus Weingarth m.weingarth*web.de" wrote: > Sent to CCL by: "Markus Weingarth" [m.weingarth,,web.de] > Hello, > > I have a little problem with amber. I would like to specify some new dihedral > angles, but I am not really familiar with all the parameter. > > After the given dihedral angle stands always a 2 or a 3. What is ment with > this number? > > > .would be nice if anyone could help me. > > Markus> > > > Victor M. Rosas García, PhD Coordinador del Posgrado en Ciencias Facultad de Ciencias Quimicas, UANL e-mail: quimico69]_[yahoo.com Tel: (81) 8329-4010 ext. 6253 Fax: (81) 8376-5375 __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From owner-chemistry@ccl.net Mon Jun 5 17:27:00 2006 From: "Elaine Meng meng!^!cgl.ucsf.edu" To: CCL Subject: CCL: Amber parameters Message-Id: <-31900-060605172516-12911-zzDak6aa25b6U8C2ByPdvw]_[server.ccl.net> X-Original-From: "Elaine Meng" Date: Mon, 5 Jun 2006 17:25:09 -0400 Sent to CCL by: "Elaine Meng" [meng ~~ cgl.ucsf.edu] Dear Markus et al., The number after the phase is the periodicity. See the description of "card 6" in the "force field parameter file specification" in http://amber.scripps.edu/formats.html I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. meng\a/cgl.ucsf.edu UCSF Computer Graphics Lab and Babbitt Lab Department of Pharmaceutical Chemistry University of California, San Francisco http://www.cgl.ucsf.edu/home/meng/index.html --- "Markus Weingarth m.weingarth*web.de" wrote: Sent to CCL by: "Markus Weingarth" [m.weingarth,,web.de] Hello, I have a little problem with amber. I would like to specify some new dihedral angles, but I am not really familiar with all the parameter. After the given dihedral angle stands always a 2 or a 3. What is ment with this number? .would be nice if anyone could help me. Markus> From owner-chemistry@ccl.net Mon Jun 5 18:02:00 2006 From: "Young Leh youngleh!^!gmail.com" To: CCL Subject: CCL: X-window Software with OpenGL Message-Id: <-31901-060605171731-12033-98BATifLAWO6+3brC2uz7g*|*server.ccl.net> X-Original-From: "Young Leh" Date: Mon, 5 Jun 2006 17:17:31 -0400 Sent to CCL by: "Young Leh" [youngleh!^!gmail.com] Dear CCLer, Could somebody please recommend any FREE X-win software that supports OpenGL? Thanks and have a nice day. Regards, Young Leh From owner-chemistry@ccl.net Mon Jun 5 19:21:00 2006 From: "Ross Walker ross===rosswalker.co.uk" To: CCL Subject: CCL: Amber parameters Message-Id: <-31902-060605173808-26580-KPuDWQ04RD8QeQ3DB1Bb4Q{}server.ccl.net> X-Original-From: "Ross Walker" Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="US-ASCII" Date: Mon, 5 Jun 2006 14:38:01 -0700 MIME-Version: 1.0 Sent to CCL by: "Ross Walker" [ross%rosswalker.co.uk] > If I remember properly, the numbers are: 1 (for highest > quality parameters), 2 > (for tentative values) and 3 (for generalized, and thus > lowest quality, > values). I don't know where you got this information from but it is certainly not correct. Here is the format of dihedral parameters as specified in the amber parm99.dat file: X -CT-CT-X 9 1.40 0.0 3. JCC,7,(1986),230 HC-CT-CT-Cl 1 0.25 0.0 1. Junmei et al, 1999 HC-CT-CT-Br 1 0.000 0.0 -3. JCC,7,(1986),230 HC-CT-CT-Br 1 0.55 0.0 1. Junmei et al, 1999 The equation is: Edihedral = sum(0.5*Vn(1+cos[n*theta-gamma]) Where the first column of the above lines defines the atom types making up the dihedral (an X is a wild card character). The second column specifies the divider to be used on the value of vn. Typically if you don't use wildcards (it is recommended that you always specify your dihedrals explicitly) then this value would be 1. For the X-CT-CT-X case this is 9 because there are 9 different combinations that come from this wilcard - e.g. ethane would have 3 H's on each end. The third column is the barrier height (Vn) in KCal/mol. The 4th column is the phase offset (gamma) in degrees, the fifth column is the periodicity (n) and the final column is the reference for this parameter. The negative sign on the periodicity is used when there is more than 1 dihedral defining the torsion between 4 atoms. The +ve values of n are used in the calculation of 1-4 non-bond interactions and so all 'duplicate', i.e. dihedrals with different values of n, should have a -'ve sign in front of the value of n. I hope this helps. Note for questions related to amber you are often much better off posting to the amber mailing list (http://amber.scripps.edu/#reflector) and also searching the archive of this mailing list (http://amber.ch.ic.ac.uk/archive/). All the best Ross /\ \/ |\oss Walker | HPC Consultant and Staff Scientist | | San Diego Supercomputer Center | | Tel: +1 858 822 0854 | EMail:- ross * rosswalker.co.uk | | http://www.rosswalker.co.uk | PGP Key available on request | Note: Electronic Mail is not secure, has no guarantee of delivery, may not be read every day, and should not be used for urgent or sensitive issues.