Link of the week - Salute to SALUTE
SALUTE-or Stochastic ALgorithms for Ultra-fast Transport in sEmiconductors-is a pilot grid application developed to solve computationally intensive problems in semiconductor physics.
Semiconductors-materials that can both conduct and insulate electricity-are invaluable to modern electronics. Their ability to conduct depends on the ability of "carriers"-electrons or phonons-to transport electrical charge.
This carrier transport occurs on a quantum level: the temporal scale is so short it is measured in femtoseconds, where one femtosecond is 10-15 seconds, or one millionth of a nanosecond.
Simulations on this sort of scale rely on complicated integral equations, solvable using SALUTE's novel and grid-enabled Monte Carlo algorithms.
Studying the quantum effects that occur at the nanometer and femtosecond scale has important scientific and social impact, since they can improve the process of designing nano-devices.
Scientists and engineers working in nanotechnology, including those from Austria, Bulgaria, Croatia, the UK and the U.S., have already expressed interest in SALUTE results.
SALUTE uses data from millions and even billions of trajectories. To manage all this data, we use the SEE-GRID infrastructure, which consists of more than 30 clusters, over 800 CPUs and more than 10 Terabytes of storage.
We also make use of new services available in EGEE's gLite middleware to enable a more efficient and robust implementation of SALUTE, especially in data management. We have submitted more then 15,000 jobs, creating a substantial peak load for the infrastructure.
SALUTE was developed at the Institute for Parallel Processing at the Bulgarian Academy of Sciences in Sofia, Bulgaria. It is funded by and uses grid infrastructure established as part of the SEE-GRID2 project.
SEE-GRID2 (South Eastern European Grid enabled e-Infrastructure Development) is co-funded by the European Commission and consists of thirteen partners: twelve contractors from SE European countries and CERN.
- The SALUTE team, Institute for Parallel Processing, Bulgarian Academy of Sciences