Can we make the new 60K good until June 30? ----------------------------------------- Ray Bair Argonne National Laboratory and the University of Chicago On 3/20/17, 4:12 PM, "[email protected] on behalf of Bair, Raymond A." <[email protected] on behalf of [email protected]> wrote: This 60K request looks OK to me. Ray ----------------------------------------- Ray Bair Argonne National Laboratory and the University of Chicago On 3/19/17, 12:58 PM, "[email protected] on behalf of [email protected]" <[email protected] on behalf of [email protected]> wrote: Hello, A change in allocation has been requested: Requester: mpfeifle (Mark Pfeifle) Project: pozdyn Title: Dissociation Dynamics of Primary Ozonides in Alkene Ozonolysis Description: The trajectory calculations will be carried out using a self-made driver program by one of our group members (http://tcg.cse.anl.gov/papr/codes/dint.html). It interfaces either the Gaussian 09 or Molpro 2012/2015 program packages to do the actual electronic structure calculations. One typical trajectory run (ca. 1000 time steps using 4 gradient evaluations per iteration) takes about two days on a single processor with density functional theory methods and a moderate-sized one-electron basis set (Gaussian 09). We expect to carry out around 200-300 trajectory runs to achieve statistically significant results. This has to be done for two different transition states that are kinetically relevant. The DFT calculations can be run in parallel in principle. However, according to initial tests, the parallelization efficiency is not optimal for these kind of jobs. Therefore, the quantum-chemical calculations themselves will be run as single-core jobs. Parallelization will be simply achieved by running N separate trajectory jobs on one node (e.g. N=16 for the blues machines). The computational burden within the driver routine is negligible compared to the cost of the electronic structure calculations themselves. In addition to the DFT runs, CASPT2 calculations using Molpro are planned for at least one of the transition states. These are computationally more expensive than DFT calculations, but otherwise the same considerations about parallization apply in this case, too. Presumably, only one member will work on this project. Current: undetermined amount Justification: Requested: 60000 A specific reason has been given: The calculations for one of the transition states involved are twice as expensive as initially anticipated because a slightly different electronic structure method has to be used (UHF vs. RHF). Furthermore, we would like to study the temperature dependence of the dynamics because the initial results at room temperature are very interesting and and promising. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System _______________________________________________ allocations-admins mailing list [email protected] https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins _______________________________________________ allocations-admins mailing list [email protected] https://lists.lcrc.anl.gov/mailman/listinfo/allocations-admins