Qinling: A Parametric Model in Speculative Multithreading

نویسندگان

Yuxiang Li

Yinliang Zhao

Bin Liu

چکیده

Speculative multithreading (SpMT) is a thread-level automatic parallelization technique that can accelerate sequential programs, especially for irregular applications that are hard to be parallelized by conventional approaches. Thread partition plays a critical role in SpMT. Conventional machine learning-based thread partition approaches applied machine learning to offline guide partition, but could not explicitly explore the law between partition and performance. In this paper, we build a parametric model (Qinling) with a multiple regression method to discover the inherent law between thread partition and performance. The paper firstly extracts unpredictable parameters that determine the performance of thread partition in SpMT; secondly, we build a parametric model Qinling with extracted parameters and speedups, and train Qinling offline, as well as apply it to predict the theoretical speedups of unseen applications. Finally, validation is done. Prophet, which consists of an automatic parallelization compiler and a multi-core simulator, is used to obtain real speedups of the input programs. Olden and SPEC2000 benchmarks are used to train and validate the parametric model. Experiments show that Qinling delivers a good performance to predict speedups of unseen programs, and provides feedback guidance for Prophet to obtain the optimal partition parameters.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Speculative Multithreading: An Object-Driven Approach

Speculative multithreading (SpMT) is a parallelizing execution model for single-threaded programs on multi-core architectures. In this paper, we introduce a new SpMT model, Object-Driven Speculative Multithreading, which exploits the structure and semantics of object-oriented programs to generate speculative parallelism. Within our technique, individual program objects take the responsibility t...

متن کامل

Towards a new model of speculative bubbles: nonparametric test with an application to the Tunisian Stock Index

Bubbles in asset prices have fascinated researchers in finance. Identify asset bubbles, by circumstances, on the stock market has been a growing number of research theoretical and empirical. On a theoretical level, it was assumed that the price dynamics reflect irrational behavior of economic agents and, therefore, should be excluded from a deal with the truly rational economic agents Burmeiste...

متن کامل

When All Else Fails, Guess: The Use of Speculative Multithreading for High-Performance Computing

| Fundamental physical limits are being encountered in the design of integrated circuits that will limit future increases in processor clock rates. As a result, computer architects are developing aggressive new mechanisms to execute instructions speculatively, that is, before it is known whether or not they should actually be executed, and even before the input values needed by the instructions...

متن کامل

Supporting Speculative Multithreading on Simultaneous Multithreaded Processors

Speculative multithreading is a technique that has been used to improve single thread performance. Speculative multithreading architectures for Chip multiprocessors (CMPs) have been extensively studied. But there have been relatively few studies on the design of speculative multithreading for simultaneous multithreading (SMT) processors. The current SMT based designs IMT [9] and DMT [2] use loa...

متن کامل