Programming Club
Post-SCF, ab initio development.
(One of only two subgroups without plagarized html)
Info:
- Will attempt to meet every few weeks for presentation and/or discussion (check the calendar)
- Idea is to make everyone a better programmer and facilitate a better understanding of methods we use everyday by practical implementation
- Development will center around ab initio methods that build upon SCF
- Hopefully a group-oriented project will emerge (see Possible Future Development below)
For questions concerning the Programming Club, email
A good site for the referencing of C and C++ standard libraries, click here
Meeting Documentation
June 6, 2005: Similarities and differences in triples estimation methods, dpd library
Possible Future Development:
A dpd-type library capable of handling triples contributions
to allow the programming of a fast, efficient local version
of CCSDT in with the PSI3 ab initio software package.
Current Development:
I am currently interested in development of the CCSDT-n and
CCSDT models to include derivation of the spin-orbit equations,
creation of intermediates, and transformation of equations to spin-adapted form.
I have created handouts for the programming of various ab initio
problems. These include calculation of rotational constants, normal-coordinate
analysis, SCF (in-core and out-of-core algorithms), MP2, CISD, QCISD, CCD, and CCSD.
handouts
My current programming endevours rotate around the implementation
of integrals (one- and two-electron). I have completed and tested
so far overlap, kinetic, potential, and dipole integrals. Some were
coded using the recursion of Obara and Saika. At this time I am
implementing the two-electron repulsion integrals using recursion.
Future applications include the development of ab initio methods
based on my integral code.
integrals
I have developed tutorials on using shared libraries for C++ programming,
class and
library
My current programming project involves a string-based coupled cluster
code that allows arbitrary order excitations. I also have an interest in
developing multi-reference coupled cluster. I'm currently implementing
Brillouin-Wigner and Mukherjee's coupled cluster methods and working on a
spin-adapted UHF-CCSD code that takes advantage of spatial symmetry.
more info
I am working on programming MP2-R12 in a restricted open-shell (ROHF) formalism
to add to the RHF formalism already existent in PSI3.
I also want to explore ways of solving the problem of circumventing the
difficulties introduced by the 6-index integrals. The current solution is to use a resolution of the identity
(RI) to reduce these evil 6-index quantities to more friendly 4-index
quantities. The problem is that the atomic basis set is not
sufficiently complete for this RI, unless a very large
basis set is used. Auxillary basis has been investigated and I would like to give this a go
with a few different types of basis.
more info
I am currently designing a program, "The Termerator"
that generates equations for various electron correlation methods.
Currently I am only capable of generating .tex formatted outputs
of spin-orbit coupled cluster equations (energy and T amplitudes) for
any arbitrary set of T excitation operators. In the future I would
like to spin-adapt these equations and create intermediates as well
as extend this idea to other areas such as eom, lambda equations, etc.
"The Termerator"
Secondly, I am intereted in the extension of the current dpd library in
PSI3 to include the triples. To find
out more click here.
|
Login
