Hello, A change in allocation has been requested: Requester: jellinek (Julius Jellinek) Project: Comp_Nanocatalysis Title: Theoretical/Computational Studies of the Fundamentals of One-Component and Alloy Metal Based Nanocatalysis Relevant to Alternative Sources of Energy Description: The subject of this project is structural, electronic, and chemical reactivity properties of one-component and alloy metal nanocatalysts, both gas phase and deposited on oxide supports. The complexity of these systems stems from the fact that even for a one-component catalytic cluster or nanoparticle the number of different possible structural forms (isomers) is very large. For the case of alloys particles, the complexity increases even further as each isomeric form gives rise to a large number of secondary structures (homotops) that differ by the placement of the different types of atoms between the sites of that isomer. As a consequence, the configuration space that has to be searched is immense. We are using the parallelized version of the NWChem package installed on Fusion. The current work focuses on pure Pt and Mo and mixed Pt/Mo clusters. So far we have covered the size range from 2 to 13 atoms and made some test runs on a 55 atom cluster. The cases of l ow or no symmetry are particularly problematic. Even for a system as small as a 13-atom Pt cluster we repeatedly encountered references to “segmentation error”. The largest number of processors we have used so far is 128. The reason for this is not lack of scalability in performance, but simply the desire to save the very limited resources we have been allocated so far. Our performance tests indicate that we could use profitably hundreds or even thousands of processors in runs on systems with about 50-100 atoms. Our plans are to extend our studies to this size range and beyond. In addition, we plan to add in the computations also the atomistic representation of the supports. The expected number of project members is 4. Current: undetermined amount Justification: The allocation request is based on estimates made on Fusion and on Carver at NERSC. Tests on Pt_12Mo cluster with the Stuttgart pseudopotential and a basis set of 38 contracted functions per atom performed with 32, 64, and 128 processors indicate full parallel scaling efficiency. Tests on a 55 atom cluster with 128 and 200 processors give a comparable scaling. Requested: 621000 A specific reason has been given: We were "holding back the horses" in the first half of FY11, yet we ran out of time (allocation) by Feb. 18, 2011. The work planned for the second half of FY11 is substantially more demanding computationally, because we will focus on pure and mixed nanocatalysts of larger sizes (50 atoms and more) and we will expand the studies to include their chemical reactivity with various molecules. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System