Skip to main content

Über dieses Buch

This book provides a comprehensive overview of key technologies being used to address challenges raised by continued device scaling and the extending gap between memory and central processing unit performance. Authors discuss in detail what are known commonly as “More than Moore” (MtM), technologies, which add value to devices by incorporating functionalities that do not necessarily scale according to “Moore's Law”. Coverage focuses on three key technologies needed for efficient power management and cost per performance: novel memories, 3D integration and photonic on-chip interconnect.



Chapter 1. Impact of TSV and Device Scaling on the Quality of 3D ICs

TSVs have negative effects such as area, delay, and power overhead because of non-negligible TSV area and capacitance. Therefore, obtaining benefits such as wirelength reduction and performance improvement from 3D integration is highly dependent on the TSV size and capacitance. To reduce the negative effects, TSVs have been downscaled and sub-micron TSVs are expected to be commercially available in the near future. Meanwhile, devices have also been downscaled beyond 32 and 22 nm, so future 3D ICs will very likely be built with sub-micron TSVs and advanced device technologies. In this chapter, the impact of TSVs on the quality of today and future 3D ICs is investigated based on GDSII-level layouts.
Dae Hyun Kim, Sung Kyu Lim

Chapter 2. 3D Integration Technology

The emerging three-dimensional (3D) chip architectures, with their intrinsic capability of reducing the wire length, is one of the promising solutions to mitigate the interconnect problem in modern microprocessor designs. To leverage the benefits of fast latency, high bandwidth, and heterogeneous integration capability that are offered by 3D technology, new design methodologies should be developed targeting the unique feature of 3D integration. In this chapter, various approaches to model 3D electrical behavior, handle 3D thermal reliability problems, and design future 3D microprocessors are surveyed.
Yuan Xie, Qiaosha Zou

Chapter 3. Design and Optimization of Spin-Transfer Torque MRAMs

In this chapter, reviews the basics and modeling of spin-transfer torque magnetic RAM (STT-MRAM) for circuit-level failure analysis. A methodology for analyzing failures in STT-MRAM bit-cells is also presented. The optimization of STT-MRAM bit-cells using the presented framework is then discussed, along with several circuit and array architecture-level failure mitigation techniques. We will show that despite the relatively high write energy in STT-MRAM, large capacity last level caches based on STT-MRAM can be more energy efficient than their SRAM counterparts due to the unique characteristics of STT-MRAM.
Xuanyao Fong, Sri Harsha Choday, Kaushik Roy

Chapter 4. Embedded STT-MRAM: Device and Design

Spin-transfer-torque magnetoresistive random access memory (STT-MRAM) is made of a combination of semiconductor integrated circuits (IC) and a dense array of nanometer-scale magnetic tunnel junctions (MTJ). This emerging memory is of growing technological interest due to its potential to bring disruptive device innovation to the world of electronics. STT-MRAM is capable of providing high speed, unlimited endurance, and nonvolatility simultaneously, which is often recognized as a unique advantage over conventional and other emerging memories. While the technology is at an early stage and evolving in multiple platforms, STT-MRAM is particularly compelling as an embedded memory for system-on-chip (SOC). STT-MRAM can be integrated into SOC without altering baseline logic platforms both in process and in design. This chapter overviews key device and circuit subjects from the perspective of co-designing logic and MTJ.
Seung H. Kang, Seong-Ook Jung

Chapter 5. A Thermal and Process Variation Aware MTJ Switching Model and Its Applications in Soft Error Analysis

Spin-transfer torque random access memory (STT-RAM) has recently gained increased attention from circuit design and architecture societies. Although STT-RAM offers a good combination of small cell size, nanosecond access time and non-volatility for embedded memory applications, the reliability of STT-RAM is severely impacted by device variations and environmental disturbances. In this work, we develop a compact switching model for magnetic tunneling junction (MTJ), which is the data storage device in STT-RAM cells. By leveraging the capability to simulate the impact of thermal and process variations on MTJ switching, our model is able to analyze the diverse mechanisms of STT-RAM write operation failures. Besides the impacts of thermal and process variation, the soft error induced by radiation striking on the access transistor is another important threat to the MTJ reliability. It can also be analyzed by using our model. The incurred computation cost of our model is much less than the conventional macro-magnetic model, and hence, enabling its applications in comprehensive STT-RAM reliability analysis and design optimizations.
Peiyuan Wang, Enes Eken, Wei Zhang, Rajiv Joshi, Rouwaida Kanj, Yiran Chen

Chapter 6. Main Memory Scaling: Challenges and Solution Directions

The memory system is a fundamental performance and energy bottleneck in almost all computing systems. Recent system design, application, and technology trends that require more capacity, bandwidth, efficiency, and predictability out of the memory system make it an even more important system bottleneck. At the same time, DRAM technology is experiencing difficult technology scaling challenges that make the maintenance and enhancement of its capacity, energy-efficiency, and reliability significantly more costly with conventional techniques.
In this chapter, after describing the demands and challenges faced by the memory system, we examine some promising research and design directions to overcome challenges posed by memory scaling. Specifically, we describe three major solution directions: (1) enabling new DRAM architectures, functions, interfaces, and better integration of the DRAM and the rest of the system (an approach we call system-DRAM co-design), (2) designing a memory system that employs emerging non-volatile memory technologies and takes advantage of multiple different technologies (i.e., hybrid memory systems), (3) providing predictable performance and QoS to applications sharing the memory system (i.e., QoS-aware memory systems). We also briefly describe our ongoing related work in combating scaling challenges of NAND flash memory.
Onur Mutlu

Chapter 7. Nano-Photonic Networks-on-Chip for Future Chip Multiprocessors

To meet energy-efficient performance demands, the computing industry has moved to parallel computer architectures, such as chip-multi-processors (CMPs), internally interconnected via networks-on-chip (NoC) to meet growing communication needs. Achieving scaling performance as core counts increase to the hundreds in future CMPs, however, will require high performance, yet energy-efficient interconnects. Silicon nanophotonics is a promising replacement for electronic on-chip interconnect due to its high bandwidth and low latency, however, prior techniques have required high static power for the laser and ring thermal tuning. We propose a novel nano-photonic NoC architecture, LumiNOC, optimized for high performance and power-efficiency. This paper makes three primary contributions: a novel, nanophotonic architecture which partitions the network into subnets for better efficiency; a purely photonic, in-band, distributed arbitration scheme; and a channel sharing arrangement utilizing the same waveguides and wavelengths for arbitration as data transmission. In a 64-node NoC under synthetic traffic, LumiNOC enjoys 50 % lower latency at low loads and ∼ 40 % higher throughput per Watt on synthetic traffic, versus other reported photonic NoCs. LumiNOC reduces latencies ∼ 40 % versus an electrical 2D mesh NoCs on the PARSEC shared-memory, multithreaded benchmark suite.
Cheng Li, Paul V. Gratz, Samuel Palermo

Chapter 8. Design Automation for On-Chip Nanophotonic Integration

Recent breakthroughs in silicon photonics technology are enabling the integration of optical devices into silicon-based semiconductor processes. Significant developments in silicon photonic manufacturing and integration are enabling investigations into applications beyond that of traditional telecom: sensing, filtering, signal processing, quantum technology—and even optical computing. In effect, we are now seeing a convergence of communications and computation, where the traditional roles and boundaries of optics and microelectronics are becoming blurred. As the applications for opto-electronic integrated circuits (OEICs) are developed, and manufacturing capabilities expand, design support is necessary to fully exploit the potential of this technology. Photonic design automation represents an opportunity to take OEIC design to a larger scale, facilitating design-space exploration, and laying the foundation for current and future optical applications—thus fully realizing the potential of this technology.
This chapter describes our work on design automation for integrated optic system design. Using a building-block model for optical devices, we provide an EDA-inspired design flow and methodologies for optical design automation. Underlying these flows and methodologies are new supporting techniques in behavioral and physical synthesis. We also provide modeling for optical devices, and determine optimization and constraint parameters that guide the automation techniques. Starting from a logic design model, we describe how conventional logic synthesis and physical design techniques (placement, global and detail routing) can be applied in a top-down fashion to engineer a fully automated design flow for integrated optical systems.
Christopher Condrat, Priyank Kalla, Steve Blair
Weitere Informationen