ITR: Collaborative Research--Ascertaining Runtime Branch Characteristics through Algebraic Analysis of Programs

ITR：协作研究--通过程序的代数分析确定运行时分支特征

• 批准号: 0082671
• 负责人: Kevin Skadron
• 金额: $ 11万
• 依托单位: University of Virginia Main Campus
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2000
• 资助国家: 美国
• 起止时间: 2000-09-15 至 2003-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0082671&HistoricalAwards=false
• 关键词: ITR; Collaborative; Research; Ascertaining; Runtime

The goal of this research is to reduce the cost and improve theperformance of observation-based branch characterization mechanisms inthe compiler and hardware. Often, correlation discovered at great costor missed entirely through execution can be determined in a simplealgebraic fashion at compile time. Relationships between programstructures can be inferred from these algebraic expressions andsubsequently conveyed to compiler optimizations and to the hardwarethrough appropriate mechanisms to be developed by this research. Onceemployed, these relationships can refocus the efforts expended byobservation-based mechanisms, or can eliminate the need for themaltogether.

本研究的目标是在编译器和硬件中降低基于观测的分支特征化机制的成本并提高其性能。 通常，在编译时，可以用简单的代数方法确定在大成本上发现的或在执行过程中完全丢失的相关性。 程序结构之间的关系可以从这些代数表达式中推断出来，并随后通过本研究开发的适当机制传达给编译器优化和硬件。 一旦使用，这些关系可以重新集中基于观察的机制所花费的努力，或者可以完全消除对它们的需要。