Next week we will be welcoming four teams to Santiago de Compostela, Galicia, Spain, to participate in CESGHACK4. Alongside our expert mentors from Appentra, CESGA and RWTH Aachen University, the teams will be working on improving performance of codes across a wide range of scientific applications.
MestreTeam
MestreTeam from Mestrelab Research, Santiago de Compostela, Spain, working on developing chemistry software. Global Spectral Deconvolution (GSD) is a Mestrelab’s proprietary algorithm for the fast and fully automatic detection and deconvolution of all spectral signals present in a frequency domain NMR spectrum. Its main goal is to extract all resonances from the 1D-NMR spectrum in the presence of noise, baseline distortions, spikes and signals caused by factors other than molecules in the sample. Resulting peaks are characterized by their fundamental NMR properties, including chemical shift (i.e. resonance frequency), amplitude, line width at half height and line shape.
Because the algorithm complexity depends linearly on the number of peaks, it is expected that its parallelization can result in significant boost of its computational performance.
dRPMD
The dRPMD team comes again to CESGAHACK in order to keep on accelerating their dRPMD code. They come from the Consejo Superior de Investigaciones Científicas (CSIC, the Spanish National Research Council) and Universidad Autónoma de Madrid, Spain. They are developing a direct approach of the Ring Polymer Molecular Dynamics (dRPMD) method to calculate reaction rates at low temperatures (between 10 and 100K) of organic molecules. This is of great importance to astrochemistry, the study of molecules in the Universe. In the particular case under study, methanol + OH, these studies will also serve to validate recent experimental results. The team plans to improve the parallel efficiency of their code so that they can apply for time on a PRACE Tier-0 supercomputer.
IRLab
IRLab team, from University of A Coruña, is a research group that is working on the development of new models and techniques to improve the effectiveness and efficiency of information retrieval systems. The team also works on computational biology problems such as protein structure prediction. This problem is known as the “holy grail” of computational biology for its potential impact on drug discovery.
EIATA
EIATA team, from European Institute for Aviation Training and Accreditation of Universidad Rey Juan Carlos, Madrid, are developing new strategies for the next generation of solvers in computational fluid and continuum mechanics. Their project explores alternative formulations of well-known algorithms to reach new solutions to classical problems.
Leave a Reply