V.: An Adaptive Data Prefetching Scheme for Biosequence Database Search on Reconfigurable Platforms

Searching on DNA and protein databases using sequence comparison algorithms has become one of the most powerful techniques to better understand the functionality of particular biological sequences. However, the requirements to process the biological data exceed the ability of general-purpose processor. The core of sequence alignment algorithm was implemented as fine-grained parallel architecture that was running on a commercial-off-the-shelf (COTS) FPGA board, where supercomputer performance has been achieved. However, reconfigurable computing platforms have utilized a PCI bus as the communications channel, limiting the communication speed between the host processor and the FPGA. This communication bottleneck often offsets the application speedup enabled by FPGA. In this paper we present an adaptive data prefetching scheme to avoid reconfigurable coprocessor stalls due to data unavailability through profiling techniques and quantitative analysis. Experimental results satisfied time constraints with various query sequences and show that we can effectively eliminate a major portion of data access penalty