This book delves into the abstract mathematical universe of read-modify-write (RMW) synchronization operations as a foundation for understanding efficient data processing in massively parallel computing systems. The author establishes a new and broad framework for representing RMW operations, a family of memory access operations used in computer architectures. The book extends a technique known as memory request combining, originally introduced in the NYU Ultracomputer, to arbitrary RMW operations. The author provides a formal correctness proof for this combining mechanism and analyzes its practicality across a wide range of memory access operations. This includes a variety of synchronization primitives, such as test-and-set, fetch-and-add, and more general data-level synchronization methods. The work also explores the close relationship between the combining mechanism and parallel prefix evaluation, a well-known computational problem. The book provides a firm foundation for understanding a family of efficient parallel processing techniques foundational to the construction of modern parallel computer architectures.
"synopsis" may belong to another edition of this title.
Larry Rudolph, Ph.D., received his doctorate from NYU's Courant Institute. He is currently a Principal Research Scientist at MIT and a co-faculty member at the New England Complex Science Institute. His most recent research has been in the field of mobile computing.
Snir is Senior Manager, Scalable Parallel Systems, IBM Research Division.
"About this title" may belong to another edition of this title.
US$ 52.43 shipping from Germany to U.S.A.
Destination, rates & speedsFREE shipping from United Kingdom to U.S.A.
Destination, rates & speedsSeller: Forgotten Books, London, United Kingdom
Paperback. Condition: New. Print on Demand. This book delves into the abstract mathematical universe of read-modify-write (RMW) synchronization operations as a foundation for understanding efficient data processing in massively parallel computing systems. The author establishes a new and broad framework for representing RMW operations, a family of memory access operations used in computer architectures. The book extends a technique known as memory request combining, originally introduced in the NYU Ultracomputer, to arbitrary RMW operations. The author provides a formal correctness proof for this combining mechanism and analyzes its practicality across a wide range of memory access operations. This includes a variety of synchronization primitives, such as test-and-set, fetch-and-add, and more general data-level synchronization methods. The work also explores the close relationship between the combining mechanism and parallel prefix evaluation, a well-known computational problem. The book provides a firm foundation for understanding a family of efficient parallel processing techniques foundational to the construction of modern parallel computer architectures. This book is a reproduction of an important historical work, digitally reconstructed using state-of-the-art technology to preserve the original format. In rare cases, an imperfection in the original, such as a blemish or missing page, may be replicated in the book. print-on-demand item. Seller Inventory # 9781332088713_0
Quantity: Over 20 available
Seller: PBShop.store US, Wood Dale, IL, U.S.A.
PAP. Condition: New. New Book. Shipped from UK. Established seller since 2000. Seller Inventory # LW-9781332088713
Quantity: 15 available
Seller: PBShop.store UK, Fairford, GLOS, United Kingdom
PAP. Condition: New. New Book. Shipped from UK. Established seller since 2000. Seller Inventory # LW-9781332088713
Quantity: 15 available
Seller: Buchpark, Trebbin, Germany
Condition: Sehr gut. Zustand: Sehr gut | Sprache: Englisch | Produktart: Bücher. Seller Inventory # 26049600/2
Quantity: 1 available