Speaker
Description
Justification for delivering demo and technical requirements (ONLY for demonstrations)
This is a project the results of which is most suitable for a demo presentation.
The audience will be able to interactively define quantum chemistry calculations and submit jobs
within the graphical interface and see in real time how the job is completed on the grid.
The technical requirements is:
large monitor
high-bandwidth internet connection
Keywords
Computational chemistry
Detailed analysis
The Extensible Computational Chemistry Environment (ECCE) allows users to build molecules in a graphical interface, atom by atom, or with molecular building blocks from a structure library. ECCE translates the generated strucures to input files in the format of different quantum chemstry softwares (e.g. Gaussian, NWChem) and submits calculations to predefined computational resources. It hides many details from the user, e.g. the handling of input files, the communication with different clusters with different batch systems. New interfaces are developed that enable the use of ECCE software package on EGEE grid resources. This is implemented from the view of the grid itself as a batch queue, with commands for submitting jobs and commands for retrieving jobs when they are finished. The results of the calculations are defined in a local database which can be analyzed in a common environment independently of where the jobs were actually run.
URL for further information
http://www.pdc.kth.se/projects/pdcprojects/ecce.html
Conclusions and Future Work
With ordinary batch environments it is possible to monitor the state of the calculations within ECCE while the jobs are running. This is currently not possible with glite, and it would be desirable, in collaboration with glite developers, to achieve this functionality.
Impact
The most sucessful grid applications have been those where the grid itself is invisible to the user. In this situation a large number of experienced as well as inexperienced users are attracted who can carry out work on new resources with minimal startup time.