From owner-chemistry@ccl.net Thu Oct 26 04:51:00 2006 From: "nokophala]_[aim.com" To: CCL Subject: CCL: intrinsic rates of spin transitions Message-Id: <-32869-061026025258-16120-aJ3b1Kzn/JWkWMhslnYQwA#server.ccl.net> X-Original-From: nokophala]_[aim.com Content-Type: multipart/alternative; boundary="--------MB_8C8C6DD68CA6BD5_E8C_6927_webmail-da05.sysops.aol.com" Date: Thu, 26 Oct 2006 02:16:29 -0400 MIME-Version: 1.0 Sent to CCL by: nokophala _ aim.com ----------MB_8C8C6DD68CA6BD5_E8C_6927_webmail-da05.sysops.aol.com Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="us-ascii" Hi Community, I would like to ask about the meaning of the pre-exponential factor for spin crossing or electronic transition. To calculate the rate for a normal reaction, the pre-exponential factor is typically kT/h, and the physical basis is clear (collision frequency). What is teh equivalent expression for an electronic transition?I also imagine the value is typiclly lower, hence the existance of spin forbidden reactions because hoppoing from PES of one state to that of another one is an electron spin change. Can some one please share their experiences on this issue (how to determine the pre-exponential factor in calculating rate constant of a spin transition from first principles, or alternatively, what does the pre-exponential factor really mean, vs kT/h). Kind regards, Noko Phala ________________________________________________________________________ Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. ----------MB_8C8C6DD68CA6BD5_E8C_6927_webmail-da05.sysops.aol.com Content-Transfer-Encoding: 7bit Content-Type: text/html; charset="us-ascii"

Hi Community,

I would like to ask about the meaning of the pre-exponential factor for spin crossing or electronic transition. To calculate the rate for a normal reaction, the pre-exponential factor is typically kT/h, and the physical basis is clear (collision frequency). What is teh equivalent expression for an electronic transition?I also imagine the value is typiclly lower, hence the existance of spin forbidden reactions because hoppoing from PES of one state to that of another one is an electron spin change. Can some one please share their experiences on this issue (how to determine the pre-exponential factor in calculating rate constant of a spin transition from first principles, or alternatively, what does the pre-exponential factor really mean, vs kT/h).

Kind regards,

Noko Phala

Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection.

----------MB_8C8C6DD68CA6BD5_E8C_6927_webmail-da05.sysops.aol.com-- From owner-chemistry@ccl.net Thu Oct 26 09:14:00 2006 From: "Robert E Duke rduke]-[email.unc.edu" To: CCL Subject: CCL: Require statics about GROMACS & AMBER MD simulation on cluster Message-Id: <-32870-061026090051-17219-4AyahxnjTYweg7qP/3y44w*server.ccl.net> X-Original-From: "Robert E Duke" Date: Thu, 26 Oct 2006 09:00:50 -0400 Sent to CCL by: "Robert E Duke" [rduke/./email.unc.edu] Sambit - The amber website, amber.scripps.edu, has a ton of benchmarks for both amber 8 and amber 9. A lot of the amber 9 benchmarks are for pmemd, rather than sander, as pmemd is our high-scaling md application for the most-commonly used md procedures, and sander instead supports the full range of functionality developed by the amber team. I am the pmemd developer, and I have made a reasonable effort to develop an application that does well on performance across the gamut of machines from workstations through linux clusters and medium-sized unix smp's on up to hundreds of processors at large supercomputer installations. GROMACS may indeed be faster on a small number of processors when using a single precision build; pmemd does not support a single precision build because single precision does not give, in my opinion, adequate trajectory reproducibility when run in parallel. Because of network indeterminacy which affects the order of operations in things like force summations, any md code run in parallel starts producing slightly different results after some number of steps due to differences in the rounding errors. With pmemd and a system with reasonable energetics (ie., one that is not blowing up), trajectories are typically reproducible for the first 300 to 500 steps with pmemd. I once tried single precision gromacs, and it was not reproducible for more than a few steps in single precision. While we understand the cause of this lack of reproducibility, I have chosen to not live with this situation in pmemd as a developer, because it then becomes very difficult to spot errors in code development. Just my opinion on using single precision in md (I know the real accuracy in md is pretty darn low due to all the guesstimates involved in forcefields and what have you - so the real issue is not greater actual accuracy). In my opinion, for a 5 node cluster, either gromacs in double precision or amber would probably work for you in terms of performance; it is possible that only one of these two systems has support for something you need - be it forcefields or some md functionality (and then there is always namd, but it has been more targetted at larger installations in my opinion). Best Regards - Bob Duke rduke- -email.unc.edu From owner-chemistry@ccl.net Thu Oct 26 10:43:00 2006 From: "Andrew Emerson a.emerson-$-cineca.it" To: CCL Subject: CCL: Require statics about GROMACS & AMBER MD simulation on cluster Message-Id: <-32871-061026100544-13440-VbnhnGc0j8wzZqoCpsrDvA]^[server.ccl.net> X-Original-From: "Andrew Emerson" Date: Thu, 26 Oct 2006 10:05:44 -0400 Sent to CCL by: "Andrew Emerson" [a.emerson]|[cineca.it] Dear Sambit I think if you can you need to try the system you want to simulate on your cluster with both Gromacs and Amber and then choose the best. I have had some experience with Gromacs and a simulation of a protein in water scales well upto about 16 processors on an IBM SP machine, but scales poorly on a Linux/Xeon cluster (not quite sure why there is such a difference). But I have also discovered that the parallel scaling in Gromacs depends on the number of solvent molecules so I really recommend trying the system you want to study on your hardware, rather than relying on published benchmarks. In terms of parallelism, NAMD is generally considered to scale the best but on only 5 nodes this may be less important than usability, available force-fields, etc. regards Andy Emerson From owner-chemistry@ccl.net Thu Oct 26 13:49:00 2006 From: "Jim Kress ccl_nospam[A]kressworks.com" To: CCL Subject: CCL: Require statics about GROMACS & AMBER MD simulation on cluster Message-Id: <-32872-061026134525-10156-MpDYpB807YZMTSyx56nJ+g~!~server.ccl.net> X-Original-From: "Jim Kress" Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="us-ascii" Date: Thu, 26 Oct 2006 13:44:47 -0400 MIME-Version: 1.0 Sent to CCL by: "Jim Kress" [ccl_nospam]^[kressworks.com] Gromacs scales poorly on Linux/PC clusters due to the latency and transmission speed of the interconnects between the PCs, especially if you use 100 Mbit or Gbit Ethernet. If you use Myrinet (or an equivalent high speed/ low latency) PC to PC interconnect Gromacs scales much better. That also explains the superior performance for the IBM SP since there is no low performance interconnect between the cpus and cpus to memory. Jim > -----Original Message----- > From: Andrew Emerson a.emerson-$-cineca.it > [mailto:owner-chemistry^ccl.net] > Sent: Thursday, October 26, 2006 10:45 AM > To: Kress, Jim > Subject: CCL: Require statics about GROMACS & AMBER MD > simulation on cluster > > Sent to CCL by: "Andrew Emerson" [a.emerson]|[cineca.it] Dear Sambit > > I think if you can you need to try the system you want to > simulate on your cluster with both Gromacs and Amber and then > choose the best. I have had some experience with Gromacs and > a simulation of a protein in water scales well upto about > 16 processors on an IBM SP machine, but scales poorly on a > Linux/Xeon cluster (not quite sure why there is such a > difference). But I have also discovered that the parallel > scaling in Gromacs depends on the number of solvent molecules > so I really recommend trying the system you want to study on > your hardware, rather than relying on published benchmarks. > > In terms of parallelism, NAMD is generally considered to > scale the best but on only 5 nodes this may be less important > than usability, available force-fields, etc. > > > regards > Andy Emerson > > > > -= This is automatically added to each message by the mailing > script =- To recover the email address of the author of the > message, please change the strange characters on the top line > to the ^ sign. You can also look up the X-Original-From: line > in the mail header.> Conferences: > http://server.ccl.net/chemistry/announcements/conferences/ > > Search Messages: http://www.ccl.net/htdig (login: ccl, > Password: search)> > -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > -+-+-+-+-+ > > > > > From owner-chemistry@ccl.net Thu Oct 26 18:48:01 2006 From: "Osman Guner oguner ~~ turquoisecons.com" To: CCL Subject: CCL: Call for papers... ACS - Chicago Message-Id: <-32873-061026184603-14460-IsH6p8pmmCic0hgdCNXfpw^server.ccl.net> X-Original-From: "Osman Guner" Content-Type: multipart/alternative; boundary="----=_NextPart_000_0008_01C6F915.D6981B80" Date: Thu, 26 Oct 2006 15:46:00 -0700 MIME-Version: 1.0 Sent to CCL by: "Osman Guner" [oguner _ turquoisecons.com] This is a multi-part message in MIME format. ------=_NextPart_000_0008_01C6F915.D6981B80 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Advances in Pharmacophores and 3D Screening =20 Organized by Chemical Information Division (CINF) Co-sponsored by Computers in Chemistry (COMP), and Medicinal Chemistry (MEDI) Divisions =20 At 233rd ACS National Meeting, Chicago, IL, March 25-29, 2007 =20 Recent developments in three-dimensional screening and data-mining techniques will be covered with specific focus in pharmacophores and 3D-searching. =20 If you are interested in being a speaker for this symposium, please = submit an abstract via the ACS OASYS system. A direct link to the CINF sessions = is below: =20 http://oasys.acs.org/acs/233nm/cinf/papers/index.cgi=20 =20 The deadline for abstract submission is November 17. =20 Feel free to contact me with any questions. =20 --- Osman F, G=FCner, PhD Principal, Turquoise Consulting oguner[]turquoisecons.com=20 http://www.turquoisecons.com=20 Skype: turquoisecons =20 =20 ------=_NextPart_000_0008_01C6F915.D6981B80 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable

Advances in Pharmacophores and 3D = Screening

Organized by Chemical Information Division = (CINF)

Co-sponsored by Computers in Chemistry (COMP), and = Medicinal Chemistry (MEDI) Divisions

=A0

At 233rd ACS National Meeting, Chicago, IL, March 25-29, 2007

Recent developments in three-dimensional screening = and data-mining techniques will be covered with specific focus in = pharmacophores and 3D-searching.

If you are interested in being a speaker for this = symposium, please submit an abstract via the ACS OASYS system. A direct link to the = CINF sessions is below:

http://oasy= s.acs.org/acs/233nm/cinf/papers/index.cgi

The deadline for abstract submission is November = 17.

Feel free to contact me with any = questions.

---

Osman F, G=FCner, PhD

Principal, Turquoise = Consulting

oguner[]turquoisecons.com&nbs= p;

http://www.turquoisecons.com&nb= sp;

Skype: turquoisecons

------=_NextPart_000_0008_01C6F915.D6981B80-- From owner-chemistry@ccl.net Thu Oct 26 21:45:00 2006 From: "Gonzalo Jimenez Oses gonzalo.jimenez###dq.unirioja.es" To: CCL Subject: CCL: Measurement of the volume of MO lobes Message-Id: <-32874-061026214026-31437-JU78qJ+Oc2NfKQXR8nUM1Q!^!server.ccl.net> X-Original-From: "Gonzalo Jimenez Oses" Date: Thu, 26 Oct 2006 21:40:25 -0400 Sent to CCL by: "Gonzalo Jimenez Oses" [gonzalo.jimenez*dq.unirioja.es] Dear CCL'ers Maybe my question is somewhat naive, but in these days I been trying to calculate the extension of each lobe in a given molecular orbital or, specifically, a NBO orbital. I realize that in some cases these lobes are unsymmetrical (because I can see it on their plots) but I wonder if there is some trivial way to quantify their size in terms of volume units or something related. I have found some methods to do something similar through the EFOE model (by Tomoda et al.), but I am afraid that it is not of easily available in standard QM packages. Any idea about that? Thank you very much in advance. Best regards, Gonzalo From owner-chemistry@ccl.net Thu Oct 26 23:41:00 2006 From: "Brock Askew chemguy1985*o*hotmail.com" To: CCL Subject: CCL: HyperChem Message-Id: <-32875-061026121805-28861-W3OhfT3XQrZDNicueX0SHw _ server.ccl.net> X-Original-From: "Brock Askew" Date: Thu, 26 Oct 2006 12:18:04 -0400 Sent to CCL by: "Brock Askew" [chemguy1985 .. hotmail.com] I am thinking of buying HyperChem for the Mac. Does any one have an openion about this product or any of Hypercubes other products or services before I pruchase the software? Thank you for your advice. Brock