Convey Computer Achieves Fast Implementation of Smith-Waterman Algorithm

Convey Computer Corp., a provider of hybrid-core computers integrating low cost and simple programming model of a commodity system with the performance of customized hardware architecture, has announced that company's implementation of the Smith-Waterman algorithm is 172x faster when compared to the conventional methods, and has reportedly been the fastest Smith-Waterman implementation executed till now. The implementation of Smith-Waterman, a well-known algorithm for performing local sequence alignment enabling determination of similar regions between two nucleotide or protein sequences, is widely made in life sciences applications for aligning DNA and protein sequences.

Life sciences applications need to analyze and process heavy amounts of data, where it's not a big deal to analyze 340 terabytes worth of data, and developing complete phylogenetic trees, a data intensive process, is a usual practice at the research lab. As the laboratory performance is increasing at such heavy rates leveraging Smith-Waterman and other algorithms, bioinformatics and computational biology researchers are having better analysis about genes to find enhanced ways to cure and manage diseases.

Leveraging its hybrid-core computing architecture that seamlessly integrates advanced computer architecture and compiler technology with commercial, off-the-shelf hardware, Convey Computer Corporation has been engaged in development of the computer processing productivity. Its architecture deploys an Intel Xeon processor and Xilinx (News - Alert) Field Programmable Gate Arrays or 'FPGAs' to offer enhanced performance and high-performance computing to the customers, at much lower energy costs. Further, the systems developed by the company are simple enough to be easily used by the programmers, and offer full support of an ANSI standard C, C++ and FORTRAN development environment.

The architecture is based upon Filed Programmable Gate Arrays or 'FPGAs', offering the developers comprehensive flexibility for advanced application development, and suitably fulfills the needs of the life sciences research institutions served by Convey. These laboratories include the University of South Carolina and Virginia Bioinformatics Institute or 'VBI' at Virginia Tech.

According to Dr. Jason D. Bakos, Assistant Professor, Department of Computer Science and Engineering at the University of South Carolina, his Heterogeneous and Reconfigurable Computing Group aims at deploying reconfigurable coprocessor technology to accelerate applications that have never been accelerated before. Bakos continued that these applications range from computational phylogenetics and sparse linear algebra to data mining and logic minimization. The research group also emphasizes upon devising new design methodologies for high-performance computing, ranging from development of new automatic partitioning tools to improvement of system architecture with multi-FPGA interconnects, Bakos added.

In January 2010, Convey Computer Corporation announced that HPCwire had selected Steve Wallach, the co-founder and chief scientist of the company for the publication's 2010 'People to Watch' award, which recognizes the individuals who contribute to the advancement of high-performance computing, HPC technology, and the HPC community at large.