Topic: State-of-the-art Many Body Computational Chemistry Codes
Date Received: September 16, 2009
We are served well with the exception that we can't do any calculation that we really want to do. We have no computational chemistry codes that are efficient enough to do the size of calculations that we need to do to compare to experiments. This has been the situation since the other Larry dragged me into this computational chemistry game decades ago. I have been grasping at straws ever since. We are thinking about encouraging Hpcmo to grasp at another. The idea is to start from scratch to generate many body computational chemistry codes that leverage the next generation parallel computer architecture. One suggestion for the wave(let) of the future that has been suggested is producing a general multi-body code based on wavelet methods using multi-scaling techniques. The fellow that we have talked to about this is Dr. Harrison of Oak Ridge who has some fledgling results.
We recommended this to Dr. Kedzoria who is trying to host Dr. Harrison (MADNESS) of Oak Ridge here for a seminar as part of his HPC education program. This kind of effort may merit a significant HPC push (or not). If we can be made to believe in the touted advantages of wavelet methods and multi-scaling for computational chemistry problems we may try to push your BC team to change your baseline.
We will let you know.
BC Team Feedback
Reply Date: September 24, 2009
Thank you kindly for your recent input to the Baseline Configuration (BC) team. The development of new state-of-the-art computational codes that leverage many-core parallel computer systems should be referred to Dr. Doug Post, HPCMP Chief Scientist, who oversees the HPCMP CREATE Program. Please kindly send a note to Dr. Post (email@example.com) with your request for state-of-the-art many body computational chemistry codes. We recommend that you copy Mr. Andy Mark (firstname.lastname@example.org) and Mr. Guy Hammer (email@example.com) in your note to Dr. Post since both Andy and Guy address a diversity of software needs across the Program.