In this paper we present new approaches to high performance protein database scanning on two novel massively parallel architectures to gain supercomputer power at low cost. The first architecture is built around a Beowulf PCcluster linked by a high-speed network and fine-grained parallel Systola 1024 processor boards connected to each node. The second architecture is the Fuzion 150, a new parallel computer with a linear SIMD array of 1536 processing elements on a single chip. We present the design of a database scanning application based on the SmithWaterman algorithm in order to derive efficient mappings onto these architectures. The implementations lead to significant runtime savings for large-scale database scanning. This result shows that both architectures provide highthroughput sequence similarity analysis solutions at a good price/performance ratio.

Protein sequences with unknown functionality are often compared to a set of known sequences to detect functional similarities. Efficient dynamic-programming algorithms exist for solving this problem, however current solutions still require significant scan times. These scan time requirements are likely to become even more severe due to exponential database growth. In this paper we present a new approach to bio-sequence database scanning using re-configurable FPGA-based hardware platforms to gain high performance at low cost. Efficient mappings of the Smith-Waterman algorithm using fine-grained parallel processing elements (PEs) that are tailored towards the parameters of a query have been designed. We use customization opportunities available at run-time to dynamically hyper customize the systolic array to make better use of available resource. Our FPGA implementation achieves a speedup of approximately 170 for linear gap penalties and 125 for affine gap penalties as compared to a standard desktop computing platform. We show how hyper-customization at run-time can be used to further improve the performance.

Molecular biologists frequently compare an unknown protein sequence with a set of other known sequences (a database scan) to detect functional similarities. Even though efficient dynamic programming algorithms exist for the problem, the required scanning time is still very high, and because of the rapid database growth finding fast solutions is of highest importance to research in this area. In this paper we present a new approach to biosequence database scanning on reconfigurable hardware platforms to gain high performance at low cost. To derive an efficient mapping onto this type of architecture, we have designed fine-grained parallel processing elements (PEs). Since our solution is based on reconfigurable hardware, we can design PEs that are tailored towards the parameters of a query. This results in an implementation with significant runtime savings on a standard off-the-shelf FPGA. 1.

With the unstoppable growth in data collection, data mining is playing an important role in the way massive data sets are analyzed. Trends clearly indicate that future decision making systems would weigh on even quicker and more reliable models used for data analysis. In order to achieve this, current algorithms and computing systems have to be optimized and tuned to effectively process the large volumes of raw data to be seen in future. In this paper, we present a brief overview of the current approaches and challenges faced in system design. The paper starts out by highlighting the uniqueness of data mining applications, which actually makes current “generic ” system designs unsuitable for mining large data. Subsequently, we summarize the current innovations and efforts made by researchers to design systems to efficiently process data mining workloads. 1

The UCSC Kestrel parallel processor is a single-board linear array processor with 512 8-bit processing elements. In the process of building the machine, we have touched nearly all aspects of computer engineering, from VLSI layout to board design and debugging, and from device drivers to new algorithm development. The programmable array is primarily designed for several core algorithms from computational biology, on which Kestrel can outperform a workstation by a factor of 20. We have also considered a variety of other algorithms, including graph coloring, computational chemistry, and neural network evaluation.