CCL: QM/MM computational cost?

 Sent to CCL by: "Mgr. Lubos Vrbka" [lubos.vrbka[a]uochb.cas.cz]
 dear nathan,
 > Greetings CCLers.  Can anyone point me to a reference discussing the
 > computational cost of QM/MM simulations and how it scales with size of the
 > components?  I have some experience with quantum chemistry simulations, as
 > well as molecular dynamics simulations, but I'm new to QM/MM and am being
 > asked as part of a proposal I'm writing to discuss the computational cost
 o=
 > f
 > the calculations I want to do.  Right now the plan is to solvate a simple 6
 > atom quantum mechanical system with approximately 1000 TIP4P water
 > molecules.
 i don't have any reference, but it's not that big problem to estimate
 the result.
 the computational cost of the MD algorithm per 4000 atom and one
 timestep will be in any case much shorter then the cost of any decent QM
 method. only if you use some really cheap semiempirical method as QM and
 very large number of atoms in the MM region (but certainly more than
 4000), it *might* become comparable.
 btw, for 1000 water molecules, you can expect to get several nanoseconds
 per day from MD simulation. i don't know what QM method you need to use.
 of course you'll use wavefunction from the previous step as a guess for
 the current step, but still...
 if you do QM/MM optimization then depending on the setup and system
 (very large total number of atoms, small QM region, cheap QM), the
 diagonalization of the hessian might be the most time demanding step,
 but it doesn't seem to be your case.
 to conclude, you can almost certainly expect that the QM calculation
 will be the time determining step.
 regards,
 lubos
 --
 .....................................................
 Mgr. Lubos Vrbka
 Center for Biomolecules and Complex Molecular Systems
 Institute of Organic Chemistry and Biochemistry
 Academy of Sciences of the Czech Republic
 Prague, Czech Republic
 http://www.molecular.cz/~vrbka
 .....................................................