In this paper, we describe the design of high-performance extended burrow wheeler transform based weighted sequence comparison algorithm for many core GPU s taking advantages of the full programmability offered by compute unified device architecture (CUDA) and its standard library thrust. Our extended burrow wheeler transform based weighted sequence comparison algorithm with thrust library implementation on CUDA is the fastest implementation of weighted sequence comparison algorithm than the

more »

... previous implementation of extended burrow wheeler transform based weighted sequence algorithm without using thrust library, and it is on average 56.3X times faster. Moreover, our present time implementation is also competitive with CPU implementations, being up to 2.9X times faster than comparable routine on 2.99 GHz Intel Pentium (R) 4 CPU with 3 GB RAM.