AbstractMulti-task parallel algorithm is applied to accelerate the program of DNA Sequence Reads Compressor (DSRC), which is specialized for compressing DNA sequencing data file - FASTQ format file. The compression process is first divided into two parallel tasks: data process task and data input task. Then, the data process task is further divided into two parts: Title data process part and other data process part. Three high parallelism models for the Title data processing are developed: multi-CPU model, CPU+GPU model and CPU+MIC model. All three models are tested, and the result shows that nearly 3x acceleration can be achieved. The maximum throughput observed among all tested cases is 144MB/S, and the average throughput is 120MB/S. Research finds that the acceleration ratio depends on the types of sequencing platforms. The program performs better on data from the ILLUMINA and SOLiD sequencing platforms.