Abstract: The bioinformatics specific supercomputer is the key project supported by the Knowledge Innovation Program of CAS. The challenge is to accelerate the important core algorithms of bioinformation processing based on the modern computer system architecture and reconfigurable computing chip, and reach tens of times improvement. In this paper, the three typical algorithms, BLAST, dynamic programming, and Zuker-RNA secondary structure prediction, are optimized with hiding I/O latency, fine granularity paralleling, and pipeline paralleling, and C simulators are developed to evaluate the performance of those algorithms. The results show the performance can be improved efficiently by the special machine.