|Posted on December 14, 2012 at 12:55 PM|
The use of parallel architectures opens the possibility to address challenges that would be impossible otherwise. The biggest supercomputer in Catalonia is Mare Nostrum, which has been recently extended with Mare Nostrum 3 and belongs to the Barcelona Supercomputer Center (BSC). I have benefited from BSC resources in the past and currently I'm involved into two projects granted by BSC:
- Benchmark Calculations on Few-Electron Harmonium Atoms.
- Transition metal chemistry with PNOF5.
The first deals with the continuation of the benchmark calculation for harmonium we started with Jerzy Cioslowski during my postdoc in Szczecin. Now thanks to Krzysztof Strasburger we have a method which not only provides highly accurate energetics but also wavefunctions. The time granted in this period we partially cover our computational needs to deal with four-electron harmonium. These results will be used to calibrate and develop new DFT functionals. Indeed, the three electron results have already been used to elaborate the Starting grant project proposal DFTCorr.
The second project is the continuation of the applicability of PNOF5 code of Mario Piris (recently parallelized by me) to molecular relevant systems. First we focus on small biological systems now we turn our sight to transition metal chemistry. This work is tightly related to the Marie-Curie Career-Integration-Grant Project that I was awarded in 2011. Other people in the IQCC has been granted computational time in BSC.
In addition, lately I had the opportunity to attend a free course on parallel computing. In the course they gave the basics of OpenMP, MPI and many debugging tools to enhance the performance of a parallel code. Some of these tools will be soon installed in our own cluster and they will be certainly used to tune the programs behind the aforementioned projects.