基于消除冗余编译的Trace优化生成方法  

Optimization Method of Trace Generating Based on Eliminating Redundant Compilation

在线阅读下载全文

作  者:兰海燕[1] 吴艳霞[1] 刘海波[1] 

机构地区:[1]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001

出  处:《自动化技术与应用》2016年第7期32-35,53,共5页Techniques of Automation and Applications

基  金:黑龙江省青年自然基金(编号QC2014C067);黑龙江博士后科研启动基金(编号LBH-Q14056);哈尔滨市科技人才创新基金(编号2015RQQXJ072)

摘  要:Dalvik VM作为Android平台核心部分,决定了Android的整体性能表现。由于Dalvik虚拟机无法满足Android的日益增长,针对Dalvik JIT整体框架和解释器之间的交互模式,提出了基于消除冗余编译的Trace生成方法优化方案。该优化方法将Dalvik JIT划分为前端和后端两部分,主要包括基本块构造、控制流关系确定、识别及筛选循环、SSA形式转换和MIR转换为LIR、LIR转换为机器码,利用Linpack与Scimark2标准测试集,测试Dalvik JIT在不同阈值、不同优化模式下的性能表现。分析了Dalvik JIT及其优化对系统性能的影响、合理的阈值范围,实验表明经过消除冗余编译的Trace优化处理后的Dalvik虚拟机中的程序执行速度有了明显提高。As a core part of the Android platform, Dalvik VM determines the overall performance of Android. Since the Dalvik virtual machine can't meet the growing demand interactive mode between the overall framework of Dalvik JIT and interpretation. An optimization scheme of Trace Generation based on eliminate redundant compiled is proposed in this paper. The Dalvik JIT is divided into two parts by the optimized method, the front-end and back-end,including the basic block structure, the controlled flow relationship, identification and screening cycle, the form conversion of SSA, converted MIR into LIR, LIR converted into machine code. It tests Dalvik JIT's performance in threshold value and optimization model, and analyzes the Dalvik JIT and its optimization's effect for the performance of the system and the reasonable range of threshold by using the standard test set of Linpackand Scimark2. Experiments show that the elimination of redundant after compiling Trace optimized, Dalvik virtual machine execution speed of the program is significantly improved.

关 键 词:DALVIK虚拟机 JIT编译 冗余 TRACE 

分 类 号:TN911.4[电子电信—通信与信息系统]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象