1 INTRODUCTION: COHERENT STATES FOR CHEMICAL SIMULATIONS The final challenge in the field of quantum chemis- try is to achieve direct, real-time simulations of chemical reactions. However, even with the present computer technology, full quantum mechanics de- scriptions of large molecular systems remain unfea- sible and recurrences to less costly classical mechan- ics treatments are inescapable. A realistic treatment of large molecules should employ a generalized hy- brid quantum/classical approach (Warshel & Levitt 1976) where some molecular degrees of freedom and/or some molecular regions are partitioned into quantum and classical descriptions. Degrees of free- dom for which a quantum treatment is less critical [e.g. translational, rotational (Morales, Deumens et al. 1999) and vibrational (Morales, Diz et al. 1995) motions] and/or molecular regions not undergoing chemical reactions (e.g. the bulk solvent surround- ing a solute) can be treated via classical mechanics with added quantum corrections on top (Miller 2001). Conversely, regions where quantum phenom- ena occur (e.g. tunneling) must be described quan- tum-mechanically. An ideal Q/C methodology should permit that the transition from quantum to classical mechanics be realized at any desired level of accuracy and in a continuous way. Such flexibil- ity can be obtained by exploiting the properties of coherent states (CS) functions (Klauder & Skager- stam 1985). Conceptually speaking, CS are sets of genuine quantum states that permits expressing full quantum dynamical equations in a classical format in terms of generalized positions i q and generalized conjugate momenta i p (Kramer & Saraceno 1981). More formally, the states i z , depending upon the complex parameters i i i z q ip = + , make a set of CS if they satisfy the following two conditions (Klauder & Skagerstam 1985): (1) continuity with respect to i z , and (2) resolution to unity: 1 = ( ) , i i i i d z z z z µ ∗ ∫ with positive measure ( ) , 0 i i d z z µ ∗ ≥ and ' 0 i i z z ≠ (i.e. a non-orthogonal and over-complete basis set). Some CS are also quasi-classical (Morales, Deumens et al. 1999) in the sense that there exists a Hamiltonian ˆ H so that the CS average positions ˆ i x = () i qt = ( ) i z t ( ) ˆ i i x z t and momenta ˆ i p = () i p t = ( ) i z t ( ) ˆ i i p z t evolve in time with ˆ H ac- cording to Hamilton classical equations: i q =  ( ) , i i Hq p ∂ / i p ∂ and i p H = −∂  ( ) , i i q p / i q ∂ , with H ( ) , i i q p = ( ) i z t () ˆ i Hz t . A CS-expressed dy- namics is fully quantum but in the closest possible form to classical mechanics; if the CS is also quasi- classical then a classical dynamics with a quantum state is obtained. The most common way to con- struct CS is by following the group-related Perelo- mov prescription (PP). Illustrating with the quantum harmonic oscillator quasi-classical Glauber CS, PP involves (Klauder & Skagerstam 1985): (1) finding adequate sets of Lie-algebra generators [here, the Weyl algebra creation (annihilation) † a ( a ) opera- tors] and then (2) applying an associated Lie-group irreducible unitary representation ( ) ˆ , U zz ∗ onto a fi- ducial state 0 to generate the CS: z = ˆ (, )0 Uzz ∗ = ( ) † exp za za ∗ − 0 . PP is simple in outline but its application to construct CS for chemical problems is far from being trivial. Furthermore, PP does not nec- Concurring Engineering Research and Applications: Next Generation Concurrent Engineering, 469-472 (2005), M. Sobolewski and P. Ghodous (eds.), ISPE, inc., ISBN 0-9768246-0-4 Grid implementation of the electron nuclear dynamics theory: a coherent states chemistry K. Tsereteli, S. Addepalli, J. Perez & J. A. Morales Department of Chemistry and Biochemistry & High Performance Computer Center, Texas Tech University PO Box 41061, Lubbock, TX 79409-1061, USA. ABSTRACT: We report our ongoing implementation of the electron nuclear dynamics (END) theory as a multi domain grid service involving Globus toolkit, Avaki compute grid, and Avaki data grid. END provides a genuine quantum dynamics but recast in a generalized classical Hamiltonian format via coherent states (CS) parameterization. END simulations of chemical reactions demand several trajectory calculations in quantum phase space and naturally call for a grid solution. The current grid implementation is based on the suite of programs CSTechG, which has been tested for various architectures under different conditions involving file staging and several queuing techniques. Different components and services of the application and sample cal- culations for the + H CH CH + ≡ reaction system are also presented.