渐进式智能回溯向量化代码调优方法

为了充分发挥高性能计算机的计算能力,缓解程序员设计和编写并行程序的压力,扩充可用软件集合,设计并实现了利用交互界面深入挖掘程序中的可向量化语句,优化生成代码中的向量化语句,提高生成代码的执行效率.该方法对充分发挥高性能计算机的计算能力,增强系统可用性和扩展应用范围具有重要的意义,同时能够提供有效的辅助手段和工具支持.渐进式智能回溯向量化代码调优架构通过对用户提交的串行程序进行程序分析和变换,采用串行程序分析、数据依赖分析、向量化分析等技术手段,根据分析结果对程序进行变换和优化,自动生成最终的向量化代码.该方法通过分析串行程序中潜在的并行性,将其自动变换为等价的向量化代码形式,大大简化了程序员的工作.     

Data dependence and program restructuring

Data dependence concepts are reviewed, concentrating on and extending previous work on direction vectors. A bit vector representation of direction vectors is discussed. Various program restructuring transformations, such as loop circulation (a form of loop interchanging), reversal, skewing, sectioning (strip mining), combing, and rotation, are discussed in terms of their effects on the execution of the program, the required dependence tests for legality, and the effects of each transformation on the dependence graph. The bit vector representation of direction vectors is used to develop simple and efficient bit vector operations for the dependence tests and to generate the modified direction vector for each transformation. Finally, a simple method to interchange complex convex loop limits is given, which is useful when several loop restructuring operations are being applied in sequence.This work was supported by NSF Grant CCR-8906909 and DARPA Grant MDA972-88-J-1004.     

程序并行化中数据收集代码自动生成算法研究

着重论述了串行程序并行化过程中的数据收集部分代码的自动生成。提出利用等价类的方法获取数据的最后写关系,并建立包括计算划分、循环迭代和数据最后写关系的不等式限制系统,最后利用FME消元法对不等式限制系统进行消元处理,最终实现数据收集代码的自动生成。     

XDP: A compiler intermediate language extension for the representation and optimization of data movement

The ability to represent, manipulate, and optimize data placement and movement between processors in a distributed address space machine is crucial in allowing compilers to generate efficient code. Data placement is embodied in the concept of dataownership. Data movement can include not just the transfer of data values but the transfer of ownership as well. However, most existing compilers for distributed address space machines either represent these notions in a language-or machine-dependent manner, or represent data or ownership transfer implicitly. In this paper we describe XDP, a set of intermediate language extensions for representing and manipulating data and ownership transfers explicitly in a compller. XDP is supported by a set of per-processor structures that can be used to implement ownership testing and manipulation at run-time, XDP provides a uniform framework for translating and optimizing sequential, data parallel, and message-passing programs to a distributed address space machine. We describe analysis and optimization techniques for this explicit representation. Finally, we compare the intermediate languages of some current distributed address space compilers with XDP.     

分布式系统中数据分解的研究

数据分解对消息传递并行机下的并行编译器取得高性能至关重要。根据编译器自动得出的数据分解（映射数据到处理机）信息，C语言版本的发送／接收消息循环嵌套可产生出来，从而在处理机之间实现分布数据。不仅一个已被证明且功能强大的数学模型用于产生数据分解代码，而且一个形式化的算法及其实现也已给出。初步实验结果显示该算法能显著提高性能。     

分支嵌套循环的自动并行化研究

GCC编译器是一种受广大研究者青睐的开源优化编译器,但它仅仅能够对完美嵌套循环进行依赖分析。为了更好地挖掘嵌套循环粗粒度的并行,深入研究了GCC5.1数据依赖分析过程,提出了一种能够处理分支嵌套循环的依赖测试方法。首先识别出分支嵌套循环,然后分析数组下标与分支嵌套循环外层索引变量的关系,最后计算出外层循环索引变量的距离向量,并通过检测距离向量判断循环是否存在依赖。实验结果表明,该方法能够正确、有效地分析出分支嵌套循环的依赖关系。     

Two-Modulus Codes with Summation of On-Data Bits for Technical Diagnostics of Discrete Systems

A fundamentally new approach to building a code with summation of on-data bits based on the selection and separate check of subsets of bits of the data vector is presented. The properties of the proposed code are analyzed in comparison with the classic and modified Berger codes. The advantages and disadvantages of new codes with summation of on-data bits are noted. The basic properties of the proposed codes with summation that should be taken into account in solving problems of technical diagnostics are established. The results of experimental applications of the developed codes to the organization of concurrent error detection systems of combinational benchmarks from LGSynth`89 are given.     

Evaluation of fortran vector compilers and preprocessors

Many scientific codes can achieve significant performance improvement when executed on a computer equipped with a vector processor. Vector constructs in source code should be recognized by a vectorizing compiler or preprocessor. This paper discusses, from a general point of view, how a vectorizing compiler/preprocessor can be evaluated. The areas discussed include data dependence analysis, IF loop analysis, nested loops, loop interchanging, loop collapsing, indirect addressing, use of temporary storage, and order of arithmetic. The ideas presented are based on vectorization of over a million lines of production codes and an extensive test suite developed to evaluate preprocessors under varying degrees of code complexity. Areas for future research are also discussed.