Research And Practice Of FPGA Online Experimental System

With the rapid development of Internet technology,network technology and computer control technology,online teaching has become an important educational method.Due to the limitation of time and space,the traditional experimental methods can no longer meet the needs of the current experiments,nor can they meet the needs of online education.The FPGA(Field Programmable Gate Array)online experimental system researched and designed in this paper can better solve the online development problems of various experiments based on the FPGA platform.Most of the existing remote experiment systems are mainly embedded systems,which have the advantages of small size and low cost,but have the disadvantages of weak processing capacity,high deployment difficulty,and inability to serve multiple users at the same time.And the existing remote FPGA experiment system usually needs to install huge and expensive EDA software on the user's computer to complete the design,synthesis,simulation and so on,and finally configure it to a remote hardware platform for hardware testing.This method has higher requirements on the user's hardware and system platform.In response to the above problems,the FPGA online experiment system studied in this paper uses the B/S model.According to the characteristics of FPGA experiments and the idea of microservers,a network architecture based on independent compilation server clusters and experimental server clusters is designed.The compilation server cluster provides users with online editing,logic synthesis,constraint establishment,structure synthesis,and timing analysis services.The experimental server cluster provides users with FPGA remote configuration,remote signaling,telemetry and remote control of experimental equipment,and video observation of experimental phenomena.The system structure has the advantages of convenient deployment,easy development,strong scalability,support for high concurrency,and cross-platform.Aiming at the control requirements of remote experimental equipment,this paper designs a general hardware control module for remote experimental equipment.This module takes MCU and FPGA as the core,and can provide up to 40 channels of switch control or relay control.Realized DDS-based high-resolution arbitrary clock source,digitally controlled voltage source,digitally controlled potentiometer;equal-accuracy frequency,pulse width and duty cycle measurement,high-precision analog voltage measurement and so on;realized serial port and network communication.The universal hardware control module can be used to perform switching control,signal generation,acquisition and measurement of the experimental equipment,and remotely interact with the user through the experimental server.Another key to FPGA online experiment is the remote configuration of FPGA.The dissertation researches the FPGA remote configuration scheme,and uses the experimental server to directly invoke the official quartus_pgm programming tool provided by Intel from the command line to realize the remote configuration of the FPGA.The observation of the experimental phenomenon is implemented by a web camera,which is simple and convenient.The thesis builds a compile server using SSH(Struts2,Hibernate4,Spring4)framework,and combines the FCKeditor editor to realize the online design input,online synthesis,constraint establishment,adaptation,timing analysis of FPGA experiments.Similarly,the experimental server is built using the SSH framework,and the remote configuration of the FPGA is realized.The remote measurement and control of the experimental equipment is realized through the hardware control module,so that the user can complete the experimental operation of the FPGA experimental platform on the browser.In order to optimize the compilation server cluster and the experiment server cluster,the system designed a compilation scheduling service system and an experimental platform scheduling service system.In order to maximize the efficiency of the system,the paper uses queuing theory to mathematically model these two subsystems.In the compilation and scheduling service system,according to the characteristics of FPGA online compilation,the minimum connection number and weighted polling algorithm were improved,making the compilation server cluster run more efficiently.In the experimental platform scheduling service system,the M/M/n/m model of the queuing theory is used to set the key parameters in the experimental platform scheduling service system.Combined with the task request parameters,the relationship between the task request and the experimental server utilization rate is studied.The optimal relationship laid the foundation for the large-scale deployment of the system.