An Acceleration Method of Data Table Parsing and Database Sorting Based on Mask Calculation

The rapid development of artificial intelligence and big data has put forward a higher requirements for database performance. As one of the basic operations of database, the sorting based on CPU operation cannot meet the requirements. Therefore, a heterogeneous sorting method combining CPU and FPGA has been developed. However, these methods only sort the fixed format data, and are not optimized for the storage format of the database, so it is difficult to apply to the database directly. Aiming at the above problems, this paper proposes an acceleration method of data table parsing and database sorting based on mask calculation. This method combines the parsing of the data table with the sorting operation by calculating the mask, which saves the time of parsing the data table, simplifies the control logic, and improves the sorting efficiency. Moreover, the dynamic reconfiguration of FPGA supports the multi-mode deployment of the kernel in space and time. Experimental results show that compared with software sorting method, the proposed method has a speedup of 25.1 for single-field sorting and 27.4 for multi-field sorting. In addition, the proposed method has a higher speedup and fewer hardware resource usage than the separate data table parsing and FPGA sorting method.