From owner-chemistry@ccl.net Wed Jan 18 14:17:00 2012 From: "Sebastian Kozuch kozuchs-,-yahoo.com" To: CCL Subject: CCL: G09: CCSD method and large basis Sets Message-Id: <-46162-120118120926-12555-AbueUFCfBgGVGrnFCU2s8Q(!)server.ccl.net> X-Original-From: Sebastian Kozuch Content-Type: multipart/alternative; boundary="-1822405001-1267926514-1326906556=:1240" Date: Wed, 18 Jan 2012 09:09:16 -0800 (PST) MIME-Version: 1.0 Sent to CCL by: Sebastian Kozuch [kozuchs|yahoo.com] ---1822405001-1267926514-1326906556=:1240 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable A good compromise between accuracy and timing is achieved with double hybri= d DFT. Usually much more accurate than hybrid functionals, and for sure muc= h faster than couple cluster. However, if you need chemical accuracy (~1 kc= al/mol) you should first compare them to CCSD(T)/CBS with a small model sys= tem. And always pray that you don't have big non-dynamic correlation, or ev= en couple cluster will be a failure!=0A=0A=A0=0Axxxxxxxxxxxxxxxxxxxxxxxxxxx= xxxxxxxxxx=0A..........Sebastian Kozuch...........=0Axxxxxxxxxxxxxxxxxxxxxx= xxxxxxxxxxxxxxx=0A....Weizmann Institute of Science....=0A...........Rehovo= t, Israel...........=0A.. sebastian.kozuch{=}weizmann.ac.il ..=0Ahttp://yfaat= .ch.huji.ac.il/kozuch.htm=0Axxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx=0A=0A=0A_= _______________________________=0A From: Roger Robinson scu98rkr%a%gmail.co= m =0ATo: "Kozuch, Sebastian " =0ASent: Wednesday, January 18, 2012 6:05 PM=0ASubject: CCL: G09: C= CSD method and large basis Sets=0A =0A=0ASent to CCL by: Roger Robinson [sc= u98rkr^-^gmail.com]=0ADear All,=0A=0AI want to accurately calculate barrier= heights for several sets of reactions.=0A=0AI am looking at molecules up t= o say 9-10 large atoms + Hydrogens. Currently I am using G09 for my calcula= tions=0A=0AFor most of my work I am looking at using the M06-2X functional = for calculating geometries, frequencies and zero point energies and for oth= er bits and bobs.=0A=0AI have just about managed to install GAMESS and I am= considering maybe the M08-SO functional for future work.=0A=0ABut a I also= need to calculate the barrier heights with as much precision as possible.= =0A=0AI have looked at a few papers such as=0A=0AKinetics of hydrogen-trans= fer isomerizations of butoxyl radicals by Zhen and Truhlar=0A=0AThey sugges= t CCSD(T)/CBS+CV+Rel.//M06-2X/MG3S as a very accurate method.=0A=0AIm looki= ng at slightly bigger molecules them and I probably dont have the same reso= urces available as them.=0A=0AI have managed to calculate barrier heights a= t the CCSD(T)/aug-cc-pVDZ//M06-2X/MG3S level so far.=0AI am currently tryin= g to do so at the CCSD(T)/aug-cc-pVTZ//M06-2X/MG3S level.=0A=0ASo far the R= WF files have got to ~330GB ! Im not too sure I have the resources to do th= is.=0A=0AThe Timings Comparisons on http://cccbdb.nist.gov/=0A=0ASeem to su= ggest VTZ will be 42 slower for a molecule with 3 large atoms, which does n= t sound right to me. This might be becuase the examples on http://cccbdb.ni= st.gov/ run quite quickly so there is alot of I/O activity ?=0A=0A42 times = longer would be well beyond my capabilities.=0A=0ADoes anyone have any thou= gh on=A0 CCSD(T)/maug-cc-pVTZ or CCSD(T)/cc-VTZ.=0A=0AWould CCSD(T)/aug-cc-= pVDZ//M06-2X/MG3S be about as accurate as Im going to get for a molecule wi= th 9 carbons ?=0AI can use 12 core machine shared memory machines but Im ho= ping the jobs last weeks not months.=0A=0AAny help appreciated,=0A=0ARoger= =0A=0A=0A=0A-=3D This is automatically added to each message by the mailing= script =3D-=0ATo recover the email address of the author of the message, p= lease change=0Athe strange characters on the top line to the {=} sign. You ca= n also=0A=0A=0AE-mail = to subscribers: CHEMISTRY{=}ccl.net or use:=0A=A0 =A0 http://www.ccl.net/cgi= -bin/ccl/send_ccl_message=0A=0AE-mail to administrators: CHEMISTRY-REQUEST{=}= ccl.net or use=0A=A0 =A0=0A= =0ASubscribe/Unsubscribe:=A0 =A0 =A0 http://www.ccl.net/chemistry/sub_unsub= .shtml=0A=0A=0A=0AJob= : http://www.ccl.net/jobs Conferences: http://server.ccl.net/chemistry/anno= uncements/conferences/=0A=0ASearch Messages: http://www.ccl.net/chemistry/s= earchccl/index.shtml=0A=0AIf your mail bounces from CCL with 5.7.1 error, c= heck:=0A=A0 =A0=0A=0ARTFI: http://www.ccl.= net/chemistry/aboutccl/instructions/ ---1822405001-1267926514-1326906556=:1240 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable
A good compromise bet= ween accuracy and timing is achieved with double hybrid DFT. Usually much m= ore accurate than hybrid functionals, and for sure much faster than couple = cluster. However, if you need chemical accuracy (~1 kcal/mol) you should fi= rst compare them to CCSD(T)/CBS with a small model system. And always pray = that you don't have big non-dynamic correlation, or even couple cluster wil= l be a failure!
 
xxxxxxxxxxxxxxxxxxxxxxxxxxxx= xxxxxxxxx
= ..........Sebastian Kozuch...........
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
....Weizmann Institute= of Science....
...........Rehovot, Israel...........
.. sebastian.kozuch{=}weizmann.ac.il ..
htt= p://yfaat.ch.huji.ac.il/kozuch.htm
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
<= div style=3D"font-family: arial, helvetica, sans-serif; font-size: 10pt;"> =
From: Roger Robinson s= cu98rkr%a%gmail.com <owner-chemistry{=}ccl.net>
To: "Kozuch, Sebastian " <kozuchs{=}= yahoo.com>
Sent: W= ednesday, January 18, 2012 6:05 PM
http://cccbdb.nist.gov/ run q= uite quickly so there is alot of I/O activity ?

42 times longer woul= d be well beyond my capabilities.

Does anyone have any though on&nbs= p; CCSD(T)/maug-cc-pVTZ or CCSD(T)/cc-VTZ.

Would CCSD(T)/aug-cc-pVDZ= //M06-2X/MG3S be about as accurate as Im going to get for a molecule with 9= carbons ?
I can use 12 core machine shared memory machines but Im hoping the jobs last weeks not months.

Any help appreciated,
Roger



-=3D This is automatically added to each message by t= he mailing script =3D-
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
E-mail to subscribers: CHEMISTRY{=}ccl.net or use:
   = ;

E-mail to adminis= trators: CHEMISTRY-REQUEST{=}ccl.net or use
  &nbs= p; http://www.ccl.net/cgi-bin/ccl/send_ccl_message

Subscribe= /Unsubscribe:     

Before posting, check = wait time at: http://www.ccl.net

Se= arch Messages: http://www.ccl.net/chemistry/searchccl/index.shtml

If= your mail bounces from CCL with 5.7.1 error, check:
    http= ://www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/chemistry/about= ccl/instructions/




---1822405001-1267926514-1326906556=:1240--