Research On The Real-time Performance Evaluation And Optimizing Method Of MIPS Architecture CNC System

Supported by Important National Science&Technology Specific ProjectsConcerning High-end CNC Machine Tools and Basic Manufacturing Equipment,Shenyang Institute of Computing Technology, National Engineering Research Centerfor High–End CNC has successfully developed the high-end NC system based onLoongson-2F, the MIPS framework processor. It’s the first time that Loongson seriesare applied in industrial field, which leads the domestic research. In the present there israrely research on the application of Loongson processor in industrial area, particularlythe high-end NC system with expertized requirements of complexity, real-timeperformance and stability. Thus, a method to estimate the stability and real-timeperformance of NC system based on MIPS framework is imperiously in need. Not onlyprovides it a reliable, stable and secure runtime environment for the system, but alsothe consultant of further development of Loongson series in real-time application.The methods of evaluating and optimizing the stability and real-time performanceof NC system based on the high-end machines with MIPS framework have beenstudied. First of all, integrated the characteristics, both software structure andestimation of real-time performance that widely using in embedded system, of NCsystem, aimed at Periodic Jitter and Event Latency, the key operations that affect theperformance and stability are indicated as evaluative factors on the basis of evaluationmodel of NC system. The time factors’ selection, the data acquisition of key indicatorsand mathematical statistics on experimental information are all involved. Primarily ananalytical method CNC-KDE(CNC key Department Evaluation),which is used for theestimation of key indicators including real-time performance and stability in PC-CNCsystem based on MIPS, is proposed. Then the real world-bask workload in NC systemis evaluated with CNC-KDE. Moreover, based on the experimental data and analyzedconsequence, relevant optimizing schemes to the shortages on software and hardwarearchitecture are also proposed.