摘要
针对传统测试程序在并行测试中效率提升不明显的问题,结合测试流程提出了基于任务、数据、数据流的3种分解方法,以解决测试中的可并行性分解问题.将大幅增加了并行能力的多核(MCP)技术引入并行测试平台,使得任务并行度的实时性进一步得以提升,进而提出了基于工作量的并行测试任务调度策略,优化了MCP平台的测试流程与任务调度,解决了单核测试程序对MCP技术利用率不高的问题.实验结果表明,基于MCP的测试平台较之单核平台的测试效率可提升30%-50%.
The efficiency of the parallel test is limited with the traditional test program. Three decomposed methods, which are based on missions, data and data flows, are brought forward to solve the problem of the parallel decomposition in test processes. The multi-core technology is introduced into the field of the parallel test, and it improves the real-time capability of test task with its parallel ability. The schedule strategy based on the task amount is advanced to optimize the test flow and the task schedule of MCP system, and solves the problem that single-core program couldn't utilize MCP technology effectively. The experiment results show that the performance of the plat based on MCP is enhanced about 30%-50% against the single-core plat.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2008年第6期683-687,共5页
Journal of Xi'an Jiaotong University
基金
总装"十一五"重点预研资助项目(51317030103)
关键词
多核
并行测试
测试程序
并行调度策略
multi-core
parallel test
test process
parallel schedule strategy