International Roadmap for Devices and Systems (IRDS™) 2022 Edition
2022 IRDS Executive Documents
Executive Summary
Citation: IEEE International Roadmap for Devices and Systems, “Executive Summary 2022.” Institute of Electrical and Electronics Engineers, 2022, doi: 10.60627/C13Z-V363.
https://doi.org/10.60627/c13z-v363
The IEEE International Roadmap for Devices and Systems (IRDS) is the continuation and extension of the National Technology Roadmap for Semiconductors/International Technology Roadmap for Semiconductors (NTRS/ITRS) with a record of uninterrupted publications spanning from 1992 to 2015. The IRDS has followed several of the successful methodologies demonstrated by the NTRS/ITRS with additional expansion to system integration, data centers and Internet communications requirements as well as quantum computing and quantum communications. The IRDS has been previously published in 2016, 2017, 2018, 2020 and 2021.
2022 IRDS International Focus Teams (IFTs) Roadmap Reports
Systems & Architectures
Citation: IEEE International Roadmap for Devices and Systems, “Systems and Architectures.” Institute of Electrical and Electronics Engineers, 2023, doi: 10.60627/JC06-2E45.
https://doi.org/10.60627/jc06-2e45
The Systems and Architectures section of the roadmap serves as a bridge between application benchmarks and component technologies. The systems analyzed in this section cover a broad range of applications of computing, electronics, and photonics. In this chapter, we identify key challenges and technology requirements by observing the patterns in play in the systems and architectures anticipated during the period under study
Outside System Connectivity
Citation: IEEE International Roadmap for Devices and Systems, “Outside System Connectivity.” Institute of Electrical and Electronics Engineers, 2023, doi: 10.60627/P687-BE68.
https://doi.org/10.60627/p687-be68
With the rapid growth of artificial intelligence (AI) applications, high data rate internet applications and IoT, communication rates in data centers need to grow to support high speed access and provide high speed low latency communication between servers, special AI processors and memory or other servers. With increased use of high resolution video, virtual reality, and augmented reality applications, ever higher data rate communication with lower latency is required in data centers. Communication between racks and switches is carried by optical interconnects; however, the capacity of switches is increasing faster than the capacity of fiber interconnects. Also, the power consumed by the switches in data centers is ~approximately 30% of the power consumed in the data center. To reduce switching power and overcome the fiber capacity gap, companies are working to integrate silicon photonic I/O into the switch package. There are also efforts to extend fiber into the server into packages and thus reduce power. Currently, short range (<100m) fiber communication uses VCSELs with multi-mode fiber, which has a significant cost and power advantage over single mode communication. There is an interest in employing single mode fiber communication throughout data centers. However, there are significant technical challenges with developing and implementing single mode fiber with low loss and cost for these applications. These challenges will be discussed in this chapter.
Read the full chapter…
More Moore
The More Moore International Focus Team (IFT) of the International Roadmap of Devices and Systems (IRDS) provides physical, electrical, and reliability requirements for logic and memory technologies to sustain power, performance, area, cost (PPAC) scaling for edge and cloud applications. This is done over a time horizon of 15 years for mainstream/high-volume manufacturing (HVM). Read the full chapter…
Yield Enhancement
The Yield Enhancement Chapter of IRDS is dedicated to ensuring that semiconductor wafer processing environment is best equipped for identifying, reducing, and avoiding yield-relevant defects and contamination. Read the full chapter…
Factory Integration
The Factory Integration (FI) chapter of the IRDS is dedicated to ensuring that the microelectronics manufacturing infrastructure contains the necessary components to produce items at affordable cost and high volume. Realizing the potential of Moore’s Law requires taking full advantage of device feature size reductions, new materials, yield improvement to near 100%, wafer size increases, and other manufacturing productivity improvements. This in turn requires a factory system that can fully integrate additional factory components and utilize these components collectively to deliver items that meet specifications determined by other IRDS international focus teams (IFTs) as well as cost, volume, and yield targets. Preserving the decades-long trend of 30% per year reduction in cost per function also requires capturing all possible cost reduction opportunities. These include opportunities in front-end as well as back end production, facilities, yield management and improvement, increased system integration such as up and down the supply chain and improving environmental health and safety. FI challenges play a key role realizing these opportunities and many FI technology challenges are becoming limiters to achieving major technology milestones.
Factory Integration Tables (XLSX)
Lithography
Historically, improvements in lithography have been key enablers to make improved chip technologies. The International Roadmap for Devices and Systems (IRDS) Lithography roadmap predicts how current patterning capability can support future chip generations and where challenges and improvements are needed. It gives a roadmap of expected key lithography requirements. It is intended to be used by semiconductor industry participants, by industry analysts, and by researchers who want or need to know how the industry will evolve in the future and what challenges need to be addressed. Read the full chapter…
Beyond CMOS & Emerging Materials Integration
Dimensional and functional scaling1 of CMOS is driving information processing technology into a broadening spectrum of new applications. Scaling has enabled many of these applications through increased performance and complexity. As dimensional scaling of CMOS will eventually approach fundamental limits, new information processing devices and microarchitectures for both existing and new functions are being explored. This is driving interest in new devices for information processing and memory, new technologies for heterogeneous integration of multiple functions, and new paradigms for system architecture. This chapter, therefore, provides an IRDS perspective on emerging research device technologies and serves as a bridge between conventional CMOS and the realm of nanoelectronics beyond the end of CMOS scaling. Read the full chapter…
Cryogenic Electronics & Quantum Information Processing
The goal of this International Roadmap for Devices and Systems (IRDS) chapter is to survey, catalog, and assess the status of technologies in the areas of cryogenic electronics and quantum information processing. Application drivers are identified for sufficiently developed technologies and application needs are mapped as a function of time against projected capabilities to identify challenges requiring research and development effort.
Cryogenic Electronics & Quantum Information Processing tables (XLSX)
Metrology
The Metrology Chapter identifies emerging measurement challenges and knowledge gaps in devices, systems, and integration in the semiconductor industry and describes research and development pathways for meeting them. This includes, but is not limited to, measurement needs for CMOS transistors, novel devices beyond CMOS, technologies for interconnects and chiplets architectures, and materials characterization for process development in integrated devices and systems. This also includes advances in metrology required for research and development, process control and device / system performance assurance in manufacturing environments.
Application Benchmarking
The Applications Benchmarking chapter of the IRDS was started as a “top end” to the roadmap. Its mission was to identify key application drivers and to track and roadmap their performance. This activity resulted in several editions of the chapter since the inception of IRDS in 2016. However, this present edition is the last edition of the chapter in this form. Included here is the introduction from the 2020 edition of the AB chapter with some edits to account for the changes as outlined in this preamble.
2022 IRDS White Papers
Autonomous Machine Computing
Autonomous machines, e.g. autonomous vehicles, are extremely complex systems that integrate many pieces of technologies [1]. For autonomous machines to become an integral part of our daily life, we are still facing many technical challenges. In this white paper, we categorize the technical challenges that are related to devices and systems, introduce the current status and roadblocks, as well as potential research directions.
Mass Digital Storage & Novolatile Memory
Digital storage technology is a key ingredient in all computer systems, from computing in large and small data centers, at the edge of networks and in endpoint processors used in civic infrastructure, factories, agriculture, homes and various wearable and embedded devices. Digital storage is where data lives. The amount of digital data generated is growing exponentially, driven by the increasing use of various sensors, including video cameras, the use of large training sets for various types of artificial intelligence training and once trained, for the use of AI models for inference from the data center to the edge and finally connected endpoint.
Although most of this data is transitory and not kept longer term, its use generates data that is retained for longer or shorter periods of time, driving the demand for more digital storage. As with other computing technologies, in order to enable this greater storage of data, digital storage is often distributed across a hierarchy of technologies that trade off performance (both bandwidth and latency) with the cost of storage. In addition, endurance and the expected life of the storage media are important characteristics that disadvantage or favor mass digital storage technologies, particularly for high use and archival applications.
There are many types of digital storage technology, including new non-volatile solid state storage technologies. These technologies continue to evolve, generally lowering the price to storage the growing volume of digital content. Many of these technologies compete with each other for various applications (e.g HDDs and SSDs). This roadmap will discuss the drivers, enablers and challenges for various digital storage technologies including NAND flash memory and solid state drives (SSDs) as well as other SSDs built on other non-volatile memory technologies, hard disk drives (HDDs), magnetic tape and various types of optical storage and optical disc recording technology. We also intend to include DNA storage for the first time.This roadmap will look at the same factors for various non-volatile solid-state memories that are in more limited commercial production and in use as stand-alone memory as well as memory built into embedded devices such as Systems on Chip (SoCs), including magnetic random-access memory (MRAM), resistive random-access memory (RRAM), ferroelectric random access memory (FRAM) and phase change memory (PCM) and other solid-state non volatile memory technologies that are not yet in commercial production. These factors will be used to create a roadmap table for the evolution of these various mass storage and non-volatile memory technologies (built upon expected changes in important specifications for each technology) for the next ten years. Read the full chapter…
More Than Moore
Since its inception in 1998, the objective of the International Technology Roadmap for Semiconductors (ITRS) has been to identify the technical challenges that had to be addressed in order to ensure that microelectronics would be able to remain a driver for innovation in a wide range of applications. This has resulted in an industrial/academic agenda for pre-competitive research, which is continuously being updated to take into account new trends.
Over the years, the scope of the ITRS was enlarged to include not only the CMOS-based digital domain for memory and microprocessor devices (driven by miniaturization, as described by Moore’s Law), but also heterogeneous integration of multi-functional analog and mixed-signal technologies for smart system applications (More than Moore). At the same time, the perspective of the roadmap shifted from being mostly technology driven to being increasingly determined by application requirements. In line with this, the ITRS changed into the International Roadmap for Devices and Systems (IRDS™).
The roadmapping effort has given rise to new insights in innovation methodology and strategy. This is in particular the case for More than Moore, which requires a highly multidisciplinary R&D environment. It has become clear that progress in highly complex technology fields can only be achieved by cooperation along the complete innovation chain, which implies that multiple fields of expertise can be combined for the development of generic technology modules, which can be made available on open technology platforms. This trend is clearly demonstrated in the present developments in, e.g., the automotive industry and the medical domain.
The objective of the present More than Moore white paper is to provide an overview of a number of technology/application areas that are representative for the More than Moore domain in the sense that they require multifunctional heterogeneous system solutions, rather than miniaturization of devices only.