From owner-chemistry@ccl.net Thu Oct 9 10:37:01 2008 From: "zhendong zhao zzhao=olemiss.edu" To: CCL Subject: CCL:G: G03 on Barcelona's Message-Id: <-37879-081009101529-14058-gaOP29srW/7rZgIXnEq9eQ_+_server.ccl.net> X-Original-From: zhendong zhao Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-1 Date: Thu, 9 Oct 2008 08:37:05 -0500 Mime-Version: 1.0 Sent to CCL by: zhendong zhao [zzhao||olemiss.edu] Dear CCLers, This problem was discussed several month ago. I want to add some cents to this discussion. I tested Phenom (9850) on AMD 780G chipset and NVIDIA nforce 750a SLI chipset. Only X86_64bit g03 has weird problems on Phenom CPUs, surprisingly, X86_32bit g03 is stable on Phenom CPUs! ( Binary G03 E.01 shipped by Gassian Inc.) First, Phenom platform is stable using four cores. In my case, X86_32bit version g03 runs well, without any problems. Other known QC program suites, GAMESS, MPQC have no problems on Phenom CPUs. Second, Only X86_64bit version g03 has very weird problems on Phenom CPUs. Increasing CPU voltage does fix problems (credits to Dr. Michael Kullman), I have several successful running records, but the system still has randomly freezing problems, it is NOT good for long-time calculation jobs. Third, the system is stable with adding more CPU voltage when running X86_32bit g03, but the system will sooner shutdown automatically when running X86_64bit version g03. It seems that X86_64bit version takes some specific CPU instructions, and cause CPU overheating. Maybe replacing stock CPU fan with more powerful heatsink may fix this problem when feeding more CPU voltage. Good luck, ZZ On Wed, 8 Oct 2008 20:14:47 -0300 "Jones de Andrade johannesrs\a/gmail.com" wrote: > Well, given the fact that G03 also doesn't work on phenoms B2 but > seems to work on B3s, it's possibly exactly the bastard TLB bug that > somehow affects G03. >=20 > On Wed, Oct 8, 2008 at 1:06 PM, Mikolaj Feliks > mikolaj.feliks**pwr.wroc.pl < owner-chemistry a ccl.net> wrote: >=20 > > > > Sent to CCL by: Mikolaj Feliks [mikolaj.feliks:+:pwr.wroc.pl] > > > > How is it possible that B3 revision of Barcelona is compatible with > > Gaussian, whereas B2 is not? B2 and B3 are the same > > microarchitecture, the only difference between these two is > > a TLB bug fixed in B3. > > > > To learn more about a TLB bug, please see: > > http://en.wikipedia.org/wiki/AMD_Phenom#TLB_errata > > > > I am not very familiar with Barcellonas, but for desktop Phenom B2 > > processors there are various software patches and workarounds that > > disable the TLB. I can't see the reason why not to run Gaussian on > > B2 Barcellonas. > > > > > > Regards, mfx > > > > > > On Wed, Oct 08, 2008 at 10:16:59AM -0300, Jones de Andrade > > johannesrs%x% gmail.com wrote: > > > Better say: B2 Barcellonas are not compatible with G03. Kill your > > > suplier > > if > > > they sold you B2s instead of B3s. > > > > > > On Wed, Oct 8, 2008 at 3:17 AM, Jerome Kieffer jerome.Kieffer^ > > > terre-adelie.org wrote: > > > > > > > > > > > Sent to CCL by: Jerome Kieffer > > > > [jerome.Kieffer[*]terre-adelie.org] On Sat, 4 Oct 2008 16:39:26 > > > > -0400 "Gennady L Gutsev gennady.gutsev_+_famu.edu" > > > > > ccl.net> > > > > wrote: > > > > > > > > > we have installed a cluster of 50 Barcelona's but Gaussian > > > > > jobs are not running. If I submit, say, 10 jobs, 5 die in a > > > > > few seconds, 2 in a few minutes, 3 in hours to several days. > > > > > Have anybody experienced similar problems? I wonder what > > > > > could be wrong: or processors themselves, either G03 is not > > > > > well adapted to the Barcelona platform, either the problem > > > > > with the system support? > > > > > > > > Hi, the Barcelona is NOT compatible with Gaussian > > > > http://www.gaussian.com/g03_plat.htm > > > > > > > > -- > > > > J=E9r=F4me KIEFFER : http://www.terre-adelie.org > > > > ? v=E9lo, prendre une rue ? contre-sens est moins dangeureux > > > > que prendre un boulevard dans le sens l=E9gal. ? qui la faute ? > > > > > > > > > > > > > > > > - This is automatically added to each message by the mailing > > > > script -> > > > > > > > > > > > > -- > > > > -\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\- > > Mikolaj Feliks > > > > Molecular Modeling & Quantum Chemistry Group > > Institute of Physical & Theoretical Chemistry > > Wroclaw University of Technology, Poland > > > > http://ichfit.ch.pwr.wroc.pl/?q=3Duser/81 > > -\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\- > > > > > > > > -=3D This is automatically added to each message by the mailing > > script =3D-> > > > > From owner-chemistry@ccl.net Thu Oct 9 11:56:00 2008 From: "Veronique LEGRAND vflegrand^^^free.fr" To: CCL Subject: CCL:G: Re2:check the total number of molecular orbital Message-Id: <-37880-081009102303-14934-7wWqb9xkktPyXWqf4AqiQw|,|server.ccl.net> X-Original-From: "Veronique LEGRAND" Date: Thu, 9 Oct 2008 10:22:59 -0400 Sent to CCL by: "Veronique LEGRAND" [vflegrand(-)free.fr] Dear Ol Ga, I think we could consider my first problem solved. I understood the impossibility to have all the informations I wanted. For the second one, it's to say the number of orbitals, it's still confused. So if you are agree, I would like just checking different parts. -first, whatever the calculation .wfn writes only the non-zero occupation numbers. Right? -You told me that the occupation numbers must be : 0,1 with UHF and without density=current. Or 0,1 for UHF whatever the method? And, in the other hand all non-zero numbers with density=current. Right? But I'm in UMP2. I know that there is a big part claculated in HF and after just the addition of the correction for the electronic correlation. Does it change something to the value expected in occupation number? -I suppose I could read these occupation numbers on the output file of Gaussian, but under which name? I tried to find by myself, of course, and my hypothesis doesn't fit with the caracteristics you give to me. It's to say: 0,1 ..... I asked myself whether there is an inversion somewhere. Becuse you tell that with density=current the are all the MOs, but I observed the contrary. I have more orbitals wihtout this keyword. And an other point to finish. If I don't use density=current with my MP2 method, is there something realised in MP2? I feel not, but in what I am in this case, HF? Oh it's so complicated. I hope I will stop to bore you quickly. Veronique > "Ol Ga eurisco1+*+pochta.ru" wrote: > > Sent to CCL by: "Ol Ga" [eurisco1-#-pochta.ru] > Dear Veronique LEGRAND, > > I understood your _two_ problems after additional comments. > About your first point: > > As it is noted on the native site of NBO programs in section FAQ > > http://www.chem.wisc.edu/~nbo5/faq.htm > > Q10. I tried to go to a higher-level MP2 or CASSCF treatment, but suddenly there were no NBO orbital energies and no table of 2nd-order perturbative energies? What's wrong? > NBO evaluates "orbital energies" and 2nd-order stabilization energies only when there is a well-defined 1-electron effective Hamiltonian operator (e.g., Fock or Kohn-Sham operator). Such an operator is unavailable for correlated descriptions, except those of DFT type > > You can see it is methodological problem (not only a problem of NBO 3.1 and NBO 5). Obviously, if you want to use correlated methods with NBO you have only one choice it is DFT. > > About your second point: > > As it is noted on the native site of AIM (and .wfn) > > http://www.chemistry.mcmaster.ca/aimpac/notes/aimweb.txt > > Density=current is vital when running correlated jobs. Otherwise the SCF (HF) > density is used for popultion analysis, and more importantly, in the wavefunction > file. If all the orbital populations in your wavefunction file are 2.00 you can > be sure you have a Hartree-Fock wfn. If all the orbitals have energy=0.0 you have > a natural orbital description. For HF densities this is fine, but the individual > MOs are meaningless - try again without density=current > > And additional comments from me : > In the .wfnfile are printed orbitals which have non-zero occupation numbers. If you use the route section _without_ DENSITY=CURRENT > > #p UMP2(full)/6-31G** GFPrint GFinput Scf=VeryTight Output=WFN Punch=Archive iop(5/14=2) > # Maxdisk=500000000 Pop=(Full,NBO) Test > > 0, 2 are occupation numbers of RHF orbitals and 0, 1 are occupation numbers of UHF orbitals. You will get a list _alpha_orbital_1, , _alpha_orbital_n (it is highest occupied _alpha_orbital) > and list _beta_ orbital_(M+1) where M is number of atomic orbitals in your case in .wfnfile. > Please, note an unusual sequence in the numbered list (n<(M+1)) . > > If you use the route section _with_ DENSITY=CURRENT > > #p MP2(full)/6-31G** GFPrint density=current GFinput Scf=VeryTight Output=WFN Punch=Archive iop(5/14=2) > # Maxdisk=500000000 Pop=(Full,NBO) Test > > You will get a list of _all_ orbitals in .wfn because NBO analysis will be performed and _all_ orbitals have non-zero occupation numbers. > > So, you will get different numbers of orbitals in two cases. > > > I wish you good luck. > Sincerely, > > Ol Ga > > > > > > Sent to CCL by: "Veronique LEGRAND" [vflegrand*_*free.fr] > > First, I want thank you for your opinion on my problem, I feel less desesperate. > > Maybe I omitted certain details in the last message. I know that > > MP2 needs the density=current keywords, but when I used this syntax > > to my NBO calculation, I hadn't the analysis of Perturbation Theory > > part in the output. And I really need it to explain the electron > > tranfer occuring in my molecule. So I was obliged to choose a less > > optimum solution, it's to say omitted this keyword. If you have an other one, I take it. > > About the other point: the number of orbitals. I checked it again. > > And I realised thanks to you, that gaussian gives the same number of > > Molecular orbitals in both case. The difference is in the file .wfn, > > where in a case there are 127 and in the other 330. > > It's unbelievable. So I just run these 2 files one more time, and I tried one another. > > If I see something knew, or if it's like now I will tell you. > > Sincerely yours. > > Veronique > >> "Ol Ga eurisco1[]pochta.ru" wrote: > >> > >> Sent to CCL by: "Ol Ga" [eurisco1*o*pochta.ru] From owner-chemistry@ccl.net Thu Oct 9 16:36:01 2008 From: "David Hose Anthrax_brothers]|[hotmail.com" To: CCL Subject: CCL:G: serious BUGs in Gaussian03's O3LYP and IOP Message-Id: <-37881-081009163437-2841-/W6zokgh1+wZ8/FT8mSVJw!A!server.ccl.net> X-Original-From: "David Hose" Date: Thu, 9 Oct 2008 16:34:34 -0400 Sent to CCL by: "David Hose" [Anthrax_brothers^hotmail.com] Wenli, I can't comment about the O3LYP density functional as I have no personal experience of it. However, your comment the BUG in the IOp handling of OPT&Freq versus Opt then a separate Freq job, is I believe FEATURE of G03. If you check Donald Truhlar web pages, there is a section on the Minnesota density functionals developed by his group. See http://comp.chem.umn.edu/info/DFT.htm Each of his group's DFT's that can be run in Gaussian by adding IOp's commands to the route section, they make the explicit comment that the frequency calculations must be run as separate jobs. Thanks for making us aware of this issue and providing a fix. Regards, Dave. ------------ Sent to CCL by: "Wenli Zou" [zorkzou++gmail.com] Recently I performed O3LYP(LDA=VWN5)/cc-pVTZ test calculations for F- by using several quantum chemistry softwares. The energies are listed below. -99.698059008 PC-GAMESS -99.698055077 NWChem (*) -99.698015916 ORCA -99.811519454 Gaussian03 -100.262303658 PQS (*) O3LYP can be defined in NWChem by dft XC vwn_5 0.19 lyp 0.81 Hfexch 0.1161 slater 0.9262 OPTX 0.8133 end For OPTX functional, however, there is a bug in the source code $NWCHEM_TOP/src/nwdft/input_dft/xc_inp.F. To correct it, go to the line "xfac(2)=1.05151d0" (at about Line 288, depending on the version), replace it by "if (abs(xfac(2)).lt.1.d-4) xfac(2)=1.05151d0", and recompile it. We can see that the energies of Gaussian03 and PQS are very different from the others, which means their O3LYPs are WRONG! I noticed that some other research groups have also found the problem of Gaussian03's O3LYP. For example, see Phys. Chem. Chem. Phys., 2004, 6, 673 - 676, Tests of second-generation and third-generation density functionals for thermochemical kinetics Yan Zhao, Jingzhi Pu, Benjamin J. Lynch and Donald G. Truhlar In the paper, they could not reproduce the standard O3LYP values by using Gaussian03. Yet, according to the above tests, their final PQS-O3LYP results are also questionable. There is a temporary way to correct this BUG. In the Route Section, the parameters of the correlation functionals must be specified explicitly by IOP(3/77=0813309262), i.e., # cc-pvtz IOp(3/74=-24) IOP(3/77=0813309262) or (for Gaussian 03.c.01 and higher) # o3lyp/cc-pvtz IOP(3/77=0813309262) Then a correct energy of -99.6980592818 a.u. can be obtained. It implies that either the original parameters of Slater and/or OPTX functionals are not right, or the right parameters are changed somewhere by error. Its highly recommended to check the source code of Gaussian03, if someone has it. Whatever the case may be, the published O3LYP results by Gaussian03 are not reliable. For opt+freq computation, there is still another problem. If we define # o3lyp/cc-pvtz IOP(3/77=0813309262) opt freq we'll find that the keyword IOP(3/77=0813309262) can not be passed to the frequency calculation, so the O3LYP energy (as well as the properties) in the second step is still not right. This may be a BUG about IOP. To avoid it, the geometry optimization and frequency calculations must be performed separately, either in two jobs, or in a multi-step job. I asked some friends to help me do the tests by using several revisions of Gaussian03. For Gaussian03 b.x and c.x (and probably a.x), this correction is effective. But for Gaussian03 d.x (and probably e.x), the energies with and without IOP(3/77=0813309262) are still not right. It seems there is a new BUG in the IOP keyword since Gaussian03 d.x. B.T.W. The Gaussian03 calculations were performed on the PC-Cluster in Dept. Chem., Peking Univ. It is site licensed. From owner-chemistry@ccl.net Thu Oct 9 17:11:00 2008 From: "N. Sukumar nagams ~~ rpi.edu" To: CCL Subject: CCL:G: Re2:check the total number of molecular orbital Message-Id: <-37882-081009164725-8474-NOMRagMlHv97xt5EuExcvQ[]server.ccl.net> X-Original-From: "N. Sukumar" Content-Disposition: inline Content-Transfer-Encoding: binary Content-Type: text/plain Date: Thu, 09 Oct 2008 15:36:39 -0400 MIME-Version: 1.0 Sent to CCL by: "N. Sukumar" [nagams{}rpi.edu] ==============Original message text=============== On Thu, 09 Oct 2008 10:22:59 EDT "Veronique LEGRAND vflegrand^^^free.fr" wrote: Sent to CCL by: "Veronique LEGRAND" [vflegrand(-)free.fr] Dear Ol Ga, ... -first, whatever the calculation .wfn writes only the non-zero occupation numbers. Right? ... Veronique ===========End of original message text=========== To have g03 print ALL the orbitals to the .wfn file, include the option IOp(99/18=-1) Thus: # ... OUT=WFN IOp(99/18=-1) For details see http://www.gaussian.com/g_tech/overlay_9999.htm#99_18 N. Sukumar Associate Research Professor of Chemistry and Chemical Biology Center for Biotechnology and Interdisciplinary Studies Rensselaer Polytechnic Institute http://reccr.chem.rpi.edu/ From owner-chemistry@ccl.net Thu Oct 9 19:07:01 2008 From: "Guenter Grethe ggrethe[A]comcast.net" To: CCL Subject: CCL: Reminder - CINF Scholarships for Scientific Excellence Message-Id: <-37883-081009190058-8920-NoDD88GCudcgXK0JRFA7Pg]^[server.ccl.net> X-Original-From: "Guenter Grethe" Date: Thu, 9 Oct 2008 19:00:55 -0400 Sent to CCL by: "Guenter Grethe" [ggrethe(_)comcast.net] REMINDER REMINDER - REMINDER The deadline for applying for the scholarship and submitting an abstract to OASYS is October 20, 2008. CINF Scholarship for Scientific Excellence Sponsored by Symyx The scholarship program of the Division of Chemical Information (CINF) of the American Chemical Society (ACS) funded by Symyx is designed to reward graduate and postdoctoral students in chemical information and related sciences for scientific excellence and to foster their involvement in CINF. Up to five scholarships valued at $1,000 each will be presented at the 237th ACS National Meeting in Salt Lake City, March 22 26, 2009. Applicants must be enrolled at a certified college or university, and they will present a poster during the Welcoming Reception of the division on Sunday evening at the National Meeting. Additionally, they will have the option to also show their poster at the Sci-Mix session on Monday night. Abstracts for the poster must be submitted electronically through OASYS. To apply, please inform the Chair of the selection committee, Guenter Grethe at ggrethe(a)comcast.net, that you are applying for a scholarship. Submit your abstract at http://oasys.acs.org/acs/237nm/cinf/papers/index.cgi by clicking on CINF Scholarship for Scientific Excellence. To enter your abstract, just follow the instructions. The deadline for submitting an abstract to OASYS is October 20, 2008. Additionally, please send a 2,000-word abstract describing the work to be presented in electronic form to the Chair of the selection committee by January 15, 2008. Any questions related to applying for one of the scholarships should be directed to the same e-mail address. Winners will be chosen based on contents, presentation and relevance of the poster and they will be announced during the reception. The contents shall reflect upon the students work and describe research in the field of cheminformatics and related sciences. Winning posters will be marked Winner of Symyx-CINF Scholarship for Scientific Excellence at the poster session. Guenter Grethe About Symyx Symyx Technologies, Inc. is the scientific R&D integration partner to companies in the life sciences, chemicals, energy, electronics and consumer products industries. With scientific R&D under tremendous economic and technical pressure, we help companies reduce R&D risk and enhance R&D productivity to help them bring more and better products to market quickly and cost-effectively. Our integrated technology platform combines Symyx Software (electronic laboratory notebooks, content, laboratory logistics and analysis), Symyx Tools (software-driven integrated workflows) and Symyx Research (collaborative research and directed services) to support the entire R&D process. In October 2007, Symyx acquired MDL Information Systems, Inc., a leading provider of innovative informatics software, databases and services that accelerate successful scientific R&D by improving the speed and quality of scientists decision making. Information about Symyx, including reports and other information filed by Symyx with the Securities and Exchange Commission, is available at www.symyx.com. From owner-chemistry@ccl.net Thu Oct 9 23:04:00 2008 From: "javier sacr jsacristan++ictp.csic.es" To: CCL Subject: CCL: torsional/total energy MD negative? Message-Id: <-37884-081009061940-6475-OANW68yUcmxBgEcS8ZT1Uw/a\server.ccl.net> X-Original-From: "javier sacr" Date: Thu, 9 Oct 2008 06:19:36 -0400 Sent to CCL by: "javier sacr" [jsacristan ~ ictp.csic.es] Hi guys, I hope you can give me any idea about: After a long molecular dynamic simulation of a polymer under periodic boundary conditions, dihedral and total energies are lower than zero, (negative). do you know the physical meaning of that result? total energy obtained at the end of the MD should be positive? thanks for reading and for your help js From owner-chemistry@ccl.net Thu Oct 9 23:39:00 2008 From: "Giovanni Scalmani giovanni-.-gaussian.com" To: CCL Subject: CCL:G: Gaussian03 reproduces published O3LYP results. Message-Id: <-37885-081009170205-16979-oHIyc6HhGLZaiQ9YcCPTqQ#%#server.ccl.net> X-Original-From: Giovanni Scalmani Content-Transfer-Encoding: 7bit Content-Type: text/plain Date: Thu, 09 Oct 2008 16:08:14 -0400 Mime-Version: 1.0 Sent to CCL by: Giovanni Scalmani [giovanni%gaussian.com] Our reference to reproduce OLYP and O3LYP results has been the following paper: W.-M. Hoe, A.J. Cohen, N.H. Handy, Chem. Phys. Letters 341 (2001) 319-328 ("O3LYP paper"). There are some ambiguities in the description of O3LYP in this paper. If we interpret these ambiguities one way, we get the same results as other programs but cannot reproduce the results in the paper. If we interpret them the other way, we can reproduce the results in the paper. Gaussian 03 revision D and later have been fixed to agree with the paper. The full details are given below. According to the paper, the OPTX exchange functional is defined as Rho^(4/3)*(A1*Cx + A2*Y^2) where Y = (g*x^2)/(1+g*x^2) and Rho is the density (alpha or beta), Cx is (3/4)*(6/PI)^(1/3), g=0.006 and x is the reduced density gradient x = |Grad_Rho|/(Rho^(4/3)). The two parameters are defined as follows A1=1.05151 and A2=1.43169. These values of the parameters allowed us to reproduce the OLYP results. Later in the paper, the authors give the values of the parameters for the O3LYP functional: A=0.1161 (% of exact exchange), B=0.9262 and C=0.8133. However, assuming that B is A1 and C is A2, produces results different from those in the paper. Using the values A1=0.9262 and A2=0.8133*1.43169 produces results in agreement with the paper. In Gaussian03 Rev. C.01 both the keywords OLYP and O3LYP were made available, but only OLYP gave results in agreement with the O3LYP paper, while O3LYP give wrong results. Setting IOp(3/77=0813309262) did not fixed the problem, but rather led to A1=0.9262 and A2=0.8133. Starting from Rev. D.01 the value of A2=0.8133*1.43169 was used by default and we were finally able to reproduce the published O3LYP results. Following are the details of some calculations that can be easily carried out with Gaussian03 to reproduce the results from Table 2 of the O3LYP paper. The agreement is very good for the H and He atoms, while is less than perfect for heavier atoms. However, given the larger absolute values of the energy and the uncertainty in the details of the calculations that were carried out to produce the published results (basis set, integration grid, SCF convergence, etc.), we are still confident that our implementation is correct. Nota that, G03 Rev. C.01 using IOp(3/77=0813309262) - which apparently agrees with other implementations of O3LYP - is clearly not able to reproduce the published data with the same accuracy. Basis set: - O3LYP paper : "PVQZ" (possibly from the CADPAC library, no literature reference is given) - following calculations: cc-pVQZ (5d,7f) Integration grid: - O3LYP paper : "high accuracy grid" - following calculations: int(grid=499974) (499 radial and 974 angular points) SCF convergence: - O3LYP paper : unspecified - following calculations: scf(tight) (RMS DP<1.0d-08, Max DP< 1.0d-06, DE<1.0d-6) Table 2: "Errors is total atomic energies (in hartrees) (theory-experiment)" Exact energies: H = -0.5000, He = -2.9037, Li = -7.4781, Be = -14.6674, F = -99.7339, Ne = -128.9376. H atom ====== G03 Rev. G03 Table 2 E.01 BLYP +0.0022 +0.0022 E.01 OLYP +0.0016 +0.0015 E.01 B3LYP (with VWN5) +0.0010 +0.0010 E.01 O3LYP +0.0002 +0.0001 C.01 O3LYP+IOp(3/77=0813309262) +0.0072 E.01 B97-1 -0.0028 -0.0028 He atom ======= G03 Rev. G03 Table 2 E.01 BLYP -0.0031 -0.0031 E.01 OLYP -0.0035 -0.0036 E.01 B3LYP (with VWN5) -0.0042 -0.0042 E.01 O3LYP -0.0064 -0.0065 C.01 O3LYP+IOp(3/77=0813309262) +0.0173 E.01 B97-1 -0.0031 -0.0031 Li atom ======= G03 Rev. G03 Table 2 E.01 BLYP -0.0033 -0.0045 E.01 OLYP -0.0073 -0.0083 E.01 B3LYP (with VWN5) -0.0038 -0.0045 E.01 O3LYP -0.0095 -0.0102 C.01 O3LYP+IOp(3/77=0813309262) +0.0302 E.01 B97-1 -0.0070 -0.0078 Be atom ======= G03 Rev. G03 Table 2 E.01 BLYP +0.0070 +0.0059 E.01 OLYP +0.0008 -0.0004 E.01 B3LYP (with VWN5) +0.0087 +0.0080 E.01 O3LYP -0.0001 -0.0010 C.01 O3LYP+IOp(3/77=0813309262) +0.0569 E.01 B97-1 +0.0027 +0.0021 F atom ======= G03 Rev. G03 Table 2 E.01 BLYP -0.0296 -0.0307 E.01 OLYP -0.0108 -0.0123 E.01 B3LYP (with VWN5) -0.0048 -0.0055 E.01 O3LYP -0.0058 -0.0070 C.01 O3LYP+IOp(3/77=0813309262) +0.1138 E.01 B97-1 -0.0091 -0.0097 Ne atom ======= G03 Rev. G03 Table 2 E.01 BLYP -0.0290 -0.0303 E.01 OLYP -0.0120 -0.0138 E.01 B3LYP (with VWN5) +0.0002 -0.0006 E.01 O3LYP -0.0053 -0.0068 C.01 O3LYP+IOp(3/77=0813309262) +0.1221 E.01 B97-1 -0.0050 -0.0057