Stream Processor Architecture

Media processing applications, such as three-dimensional graphics, video compression, and image processing, currently demand 10-100 billion operations per second of sustained computation. Fortunately, hundreds of arithmetic units can easily fit on a modestly sized 1cm2 chip in modern VLSI. The challenge is to provide these arithmetic units with enough data to enable them to meet the computation demands of media processing applications. Conventional storage hierarchies, which frequently include caches, are unable to bridge the data bandwidth gap between modern DRAM and tens to hundreds of arithmetic units. A data bandwidth hierarchy, however, can bridge this gap by scaling the provided bandwidth across the levels of the storage hierarchy. The stream programming model enables media processing applications to exploit a data bandwidth hierarchy effectively. Media processing applications can naturally be expressed as a sequence of computation kernels that operate on data streams. This programming model exposes the locality and concurrency inherent in these applications and enables them to be mapped efficiently to the data bandwidth hierarchy. Stream programs are able to utilize inexperience local data bandwidth when possible and consume expensive global data bandwidth only when necessary. Stream Processor Architecture presents the architecture of the Imagine streaming media processor, which delivers a peak performance of 20 billion floating-point operations per second. Imagine efficiently supports 48 arithmetic units with a three-tiered data bandwidth hierarchy. At the base of the hierarchy, the streaming memory system employs memory access scheduling to maximize the sustained bandwidth of external DRAM. At the center of the hierarchy, the global stream register file enables streams of data to be recirculated directly from one computation kernel to the next without returning data to memory. Finally, local distributed register files that directly feed the arithmetic units enable temporary data to be stored locally so that it does not need to consume costly global register bandwidth. The bandwidth hierarchy enables Imagine to achieve up to 96% of the performance of a stream processor with infinite bandwidth from memory and the global register file.

Get started with Apache Flink, the open source framework that powers some of the world’s largest stream processing applications. With this practical book, you’ll explore the fundamental concepts of parallel stream processing and discover how this technology differs from traditional batch data processing. Longtime Apache Flink committers Fabian Hueske and Vasia Kalavri show you how to implement scalable streaming applications with Flink’s DataStream API and continuously run and maintain these applications in operational environments. Stream processing is ideal for many use cases, including low-latency ETL, streaming analytics, and real-time dashboards as well as fraud detection, anomaly detection, and alerting. You can process continuous data of any kind, including user interactions, financial transactions, and IoT data, as soon as you generate them. Learn concepts and challenges of distributed stateful stream processing Explore Flink’s system architecture, including its event-time processing mode and fault-tolerance model Understand the fundamentals and building blocks of the DataStream API, including its time-based and statefuloperators Read data from and write data to external systems with exactly-once consistency Deploy and configure Flink clusters Operate continuously running streaming applications

Stream processing is a novel distributed computing paradigm that supports the gathering, processing and analysis of high-volume, heterogeneous, continuous data streams, to extract insights and actionable results in real time. This comprehensive, hands-on guide combining the fundamental building blocks and emerging research in stream processing is ideal for application designers, system builders, analytic developers, as well as students and researchers in the field. This book introduces the key components of the stream computing paradigm, including the distributed system infrastructure, the programming model, design patterns and streaming analytics. The explanation of the underlying theoretical principles, illustrative examples and implementations using the IBM InfoSphere Streams SPL language and real-world case studies provide students and practitioners with a comprehensive understanding of such applications and the middleware that supports them.

This book constitutes the refereed proceedings of the 7th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2007, held in Hangzhou, China in June 2007. Focusing on two broad areas of parallel and distributed computing, the papers are organized in topical sections on parallel algorithms, parallel architecture, grid computing, peer-to-peer technologies, and advanced network technologies.

Get started with Apache Flink, the open source framework that powers some of the world’s largest stream processing applications. With this practical book, you’ll explore the fundamental concepts of parallel stream processing and discover how this technology differs from traditional batch data processing. Longtime Apache Flink committers Fabian Hueske and Vasia Kalavri show you how to implement scalable streaming applications with Flink’s DataStream API and continuously run and maintain these applications in operational environments. Stream processing is ideal for many use cases, including low-latency ETL, streaming analytics, and real-time dashboards as well as fraud detection, anomaly detection, and alerting. You can process continuous data of any kind, including user interactions, financial transactions, and IoT data, as soon as you generate them. Learn concepts and challenges of distributed stateful stream processing Explore Flink’s system architecture, including its event-time processing mode and fault-tolerance model Understand the fundamentals and building blocks of the DataStream API, including its time-based and statefuloperators Read data from and write data to external systems with exactly-once consistency Deploy and configure Flink clusters Operate continuously running streaming applications

Multicore Processors and Systems provides a comprehensive overview of emerging multicore processors and systems. It covers technology trends affecting multicores, multicore architecture innovations, multicore software innovations, and case studies of state-of-the-art commercial multicore systems. A cross-cutting theme of the book is the challenges associated with scaling up multicore systems to hundreds of cores. The book provides an overview of significant developments in the architectures for multicore processors and systems. It includes chapters on fundamental requirements for multicore systems, including processing, memory systems, and interconnect. It also includes several case studies on commercial multicore systems that have recently been developed and deployed across multiple application domains. The architecture chapters focus on innovative multicore execution models as well as infrastructure for multicores, including memory systems and on-chip interconnections. The case studies examine multicore implementations across different application domains, including general purpose, server, media/broadband, network processing, and signal processing. Multicore Processors and Systems is the first book that focuses solely on multicore processors and systems, and in particular on the unique technology implications, architectures, and implementations. The book has contributing authors that are from both the academic and industrial communities.

Ascend AI Processor Architecture and Programming: Principles and Applications of CANN offers in-depth AI applications using Huawei’s Ascend chip, presenting and analyzing the unique performance and attributes of this processor. The title introduces the fundamental theory of AI, the software and hardware architecture of the Ascend AI processor, related tools and programming technology, and typical application cases. It demonstrates internal software and hardware design principles, system tools and programming techniques for the processor, laying out the elements of AI programming technology needed by researchers developing AI applications. Chapters cover the theoretical fundamentals of AI and deep learning, the state of the industry, including the current state of Neural Network Processors, deep learning frameworks, and a deep learning compilation framework, the hardware architecture of the Ascend AI processor, programming methods and practices for developing the processor, and finally, detailed case studies on data and algorithms for AI. Presents the performance and attributes of the Huawei Ascend AI processor Describes the software and hardware architecture of the Ascend processor Lays out the elements of AI theory, processor architecture, and AI applications Provides detailed case studies on data and algorithms for AI Offers insights into processor architecture and programming to spark new AI applications

The refereed proceedings of the 12th Asia-Pacific Computer Systems Architecture Conference are presented in this volume. Twenty-six full papers are presented together with two keynote and eight invited lectures. Collectively, they represent some of the most important developments in computer systems architecture. The papers emphasize hardware and software techniques for state-of-the-art, multi-core and multi-threaded architectures.