| The construction of the multi-core embedded operating system experimental platform can facilitate the accurate and accurate inspection of the efficiency and performance of the new algorithm under the operating system,and is of great significance to the research of the multi-core operating system.Based on the operating system experiment platform and its performance evaluation method,this paper improves the existing multi-core simulation platform,establishes the operating system performance evaluation system,and uses the combination of simulation experiment and evaluation model to continuously correct the relevant mechanisms and algorithms.The main contents of this article are as follows:First of all,for the overall architecture design of the experimental platform of the microkernel operating system of heterogeneous multi-core processors,this article based on Simics,using the Simics platform as a digital modeling platform model,designed a microkernel operating system for heterogeneous multi-core processors Simulation experiment platform CC-GLL.Using the processor supported by Simics as a reference model,by connecting the HLA adapter and Verilog PLI through the inter-core communication mechanism of task classification,the deployment of HDL simulators is assisted,so that the simulation platform can simulate the real environment in which the processor runs,and inherits The error detection function of the simulator makes the CC-GLL simulation experiment platform more extensible and portable;for the design of the operating system performance evaluation system,this paper is based on the existing expert evaluation method and safety checklist method and other evaluation models An evaluation model suitable for microkernel embedded operating system was constructed,and a new set of performance evaluation system for embedded operating system was designed.Secondly,for the core module design in the overall architecture of the microkernel operating system experimental platform for heterogeneous multi-core processors,in the optimization of the simulation platform model ICP simulation module,in order to solve the particle swarm algorithm when searching for the optimal result,the comparison is influenced by the parameter settings.The big problem,this paper improves the particle swarm algorithm.The optimized algorithm improves the efficiency of finding the optimal result and adjusting the test performance,and accelerates the simulation rate of the platform.In terms of the design of the memory mechanism of the platform,this paper designs a mach's virtual memory management mechanism.This memory management mechanism is implemented through the slab allocator when the platform's memory is reasonably and efficiently allocated,so that the time spent in the allocation and release of memory during the simulation experiment platform is more Short;in terms of inter-process communication design,this paper designs a communication strategy based on shared memory.The basic principle of this strategy is to represent all the registers in the shared area of the memory and establish a linked list,which can easily put all kinds of Data information is read and written into the shared area of the memory,which improves the communication rate.In the process/thread design,this paper proposes a dual-stack design model.This model reduces the system overhead generated by the address space during the switching process.Accelerate the operating efficiency of the program,thereby improving the performance of the CC-GLL simulation experiment platform.Finally,in the same test environment,the test efficiency program designed in this article is placed in the ?C/OS-II operating system,and then the test program of this article is run in the CC-GLL simulation experimental platform and Visual Studio 2010 built in this article.The communication rate and the allocation and release of memory are tested and tested.Finally,a comparative analysis of the operation result data proves that the CC-GLL platform designed in this paper has higher efficiency during operation. |
PDF Full Text Request