Identification and Optimization of Sharing Patterns for Scalable Shared-memory Multiprocessors PDF Download
Are you looking for read ebook online? Search for your book and save it on your Kindle device, PC, phones or tablets. Download Identification and Optimization of Sharing Patterns for Scalable Shared-memory Multiprocessors PDF full book. Access full book title Identification and Optimization of Sharing Patterns for Scalable Shared-memory Multiprocessors by Stefanos Kaxiras. Download full books in PDF and EPUB format.
Author: Stavros D Nikolopoulos Publisher: World Scientific ISBN: 9814493767 Category : Computers Languages : en Pages : 357
Book Description
This volume addresses the state-of-the-art and future directions of informatics. Several senior researchers and graduate students present their research and work here. The purpose of the book is to disseminate the latest scientific, engineering and technical information in various fields of informatics. It covers a wide range of subjects, from theoretical computer science, software engineering, systems and scientific computing to networking and applied research. The book can be used either as a reference for related scientific work or as educational material for advanced computer science courses.
Author: Daniel E. Lenoski Publisher: Elsevier ISBN: 1483296016 Category : Computers Languages : en Pages : 364
Book Description
Dr. Lenoski and Dr. Weber have experience with leading-edge research and practical issues involved in implementing large-scale parallel systems. They were key contributors to the architecture and design of the DASH multiprocessor. Currently, they are involved with commercializing scalable shared-memory technology.
Author: University of Wisconsin--Madison. Computer Sciences Dept Publisher: ISBN: Category : Computer architecture Languages : en Pages : 13
Book Description
Abstract: "The fine-grain nature of shared-memory multiprocessor communication introduces overheads that can be substantial. Using the Scalable Coherent Interface (SCI) as a base hardware platform and the SPLASH benchmark suite for applications, we analyze three techniques to reduce this overhead: (i) efficient synchronization primitives, and in particular a hardware primitive called QOLB; (ii) weakened memory ordering constraints; and (iii) optimization of the cache-coherence protocol for two nodes sharing data. We perform simulations both for current technology and technology that we anticipate will be available five years hence. We find that QOLB (of which this study performs the first detailed simulations) shows a large and consistent improvement, much larger than that predicted by Mellor-Crummey and Scott [19]. The relaxation of memory ordering constraints also provides a consistent performance improvement. In accordance with prior results, we show that a more aggressive memory model produces more substantial performance improvements. The optimization for two-node sharing shows mixed results, correlating unsurprisingly with the presence of that sharing pattern in an application. Our most important results are (i) that the overheads eliminated with these optimizations are largely orthogonal -- the performance gains from supporting multiple optimizations concurrently are for the most part additive -- and (ii) that technological improvements increase both these overheads and the success of the optimizations at reducing them."