Researchers at IBM’s laboratories in Zurich have developed a new algorithm that can sort, correlate, and analyze millions of random data sets in minutes. Without the algorithm, the analysis would have taken days for supercomputers to process, says IBM researcher Costas Bekas. He says the algorithm could be used to analyze data measuring electricity usage and air or water pollution levels. The algorithm also could be used to break down data from global financial markets. The algorithm combines models of data calibration and statistical analysis that can assess measurement models and relationships between data sets. Bekas says the algorithm, which can analyze nine terabytes of data in less than 20 minutes, makes data analysis more cost and energy efficient because it reduces the load on supercomputers.
View Full Article
For More Information Visit: http://www.cpccci.com

The Partnership for Advanced Computing in Europe’s work package for future petaflop computer technologies beyond 2010 recently assessed 12 prototypes for next-generation computer architectures. The evaluation included full systems, system components, software prototypes, and several research activities. For example, CINES and LRZ jointly evaluated a hybrid system containing both thin and fat nodes and compute accelerators within a shared file system. NCF assessed a system of ClearSpeed/PetaPath accelerator boards together with the ClearSpeed programming language. CEA looked at the performance of graphics processing units (GPUs) using CAPS hybrid multicore parallel programming. The CSC studied the maturity of OpenCL and performance improvements for multi-GPU programming on NVIDIA Tesla and AMD Firestream cards. CSCS evaluated the PGAS programming model using the Cray Compiler Environment for UPC and CAF. And EPCC evaluated the HARWEST Compiling Environment for developing programs on the FPGA-based Maxwell supercomputer.
View Full Article
For More Information Visit: http://www.cpccci.com

Researchers at the Warsaw University of Technology in Poland have studied the characteristics of voice over Internet Protocol (VoIP) calls in an attempt to gain a better understanding of ordinary traffic. Security remains an issue for VoIP calls, which can be hijacked and used to send confidential information over the Internet. Wojciech Mazurczyk and colleagues decided to study ordinary VoIP calls so experts would have a way to compare and contrast regular calls with those that have been embedded with stolen data. VoIP calls can be compromised by changing the order in which the digital packets are sent, or by deliberately delaying certain packets that have embedded data, a technique known as Lost Audio Packet Steganography (LACK). The team’s research shows that packets are not normally re-ordered in a way for hiding data, so attacks that re-order data are not a real threat. However, LACK attacks would be difficult to spot because of the routine loss of data packets.

View Full Article

For More Information Visit: http://www.cpccci.com