Author Topic: Numerical Geometry Optimization for CCSD(T)  (Read 8219 times)

mike

  • Jr. Member
  • **
  • Posts: 10
  • Karma: +0/-0
Numerical Geometry Optimization for CCSD(T)
« on: April 16, 2014, 07:58:48 PM »
Hi there,

As only CCSD(T) energies are available in TM, I'm wondering if it's possible to do a numerical optimization with CCSD(T), like it seems was done for MP2 in earlier versions of Turbomole. See: http://www.uni-koeln.de/themen/chemie/software/turbomole/DOK/html/node176.html

Does anyone already have such a code ?

Or is anyone familiar with the MP2OPT code mentioned in the above link ? Could it be easily modified to do numerical CCSD(T) optimizations ?

Many thanks,

Mike

antti_karttunen

  • Sr. Member
  • ****
  • Posts: 227
  • Karma: +1/-0
Re: Numerical Geometry Optimization for CCSD(T)
« Reply #1 on: April 16, 2014, 08:57:18 PM »
Hi Mike,

At least in TM 6.5 the "scripts" directory includes a script called NumGrad. It prints out a short documentation with NumGrad -help. It seems that it can also be combined with jobex (jobex -level NumGrad). I did a quick test at HF level and it seems to work OK (runs without crashing, that is). I have no idea why it isn't mentioned in the manual, maybe because it's a rather new addition.

Best,
Antti

uwe

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 558
  • Karma: +0/-0
Re: Numerical Geometry Optimization for CCSD(T)
« Reply #2 on: April 24, 2014, 05:15:38 PM »
Hi Mike, Hi Antti,

numerical gradients which can also be used in combination with jobex for geometry optimizations will officially be supported in TM version 6.6. It does work already in 6.5 but the usage is a bit complicated - Antti was on the right track...

Regards,

Uwe


Hauke

  • Full Member
  • ***
  • Posts: 37
  • Karma: +0/-0
Re: Numerical Geometry Optimization for CCSD(T)
« Reply #3 on: January 09, 2015, 07:41:04 PM »

This sounds interesting. Is there a reason why NumGrad is still not mentioned in the manual of TM 6.6?

christof.haettig

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 291
  • Karma: +0/-0
    • Hattig's Group at the RUB
Re: Numerical Geometry Optimization for CCSD(T)
« Reply #4 on: September 25, 2015, 02:20:24 PM »
Because things needed to be cleaned up.