Compiler Autotuning Based on Hot Function for SHENWEI Processor.

In the field of HPC (High Performance Computing), tons of compiling optimizing technologies have been proposed. However, users are usually confused about which optimizing should be enabled to speed up their applications on the target processor. In this paper, an iterative compiling framework is proposed to probe the compiler options automatically to optimize programs. Tightly coupling with the system software of the SHENWEI family processors, this framework provides flexible iterative supports based on hot functions of many-core programs, which dramatically improve user-friendliness of the computer systems built on SHENWEI family processors. The evaluation shows that with the help of autotuning, the performance of the SPEC2000 cases and the SPEC2006 cases can be improved about 14% and 25% respectively on average compared with the -O3 optimizing level.