About

Andrew A. Chien is the William Eckhardt Distinguished Service Professor of Computer Science at the University of Chicago. The page does not provide specific details about his research focus, background, or key contributions.

Research topics

• Computer Science
• Artificial Intelligence
• Machine Learning
• Environmental science
• Algorithm
• Environmental economics
• Economics
• Waste management
• Engineering
• Environmental engineering

Selected publications

• The Emerging Challenge of Variable Capacity Scheduling

  2026-02-06

  book-chapter

  This book focuses on managing and allocating resources in the new age of renewable power generation. High-performance computing platforms, data centers and cloud computing environments all experience fluctuating capacity over time due to external factors like weather patterns or grid conditions. From a computing perspective, the goal is to optimize resource utilization and minimize disruptions by dynamically adjusting job scheduling and resource allocation strategies in response to these capacity variations. From a societal perspective, the opportunity is to enhance sustainability efforts and to minimize the environmental impact of computing.

  PublisherDOI

• From High-performance Computing Datacenters to Grid-integrated AI Datacenters

  2026-02-06

  book-chapter1st authorCorresponding

  To provide understanding of the opportunities and challenges of datacenters, we describe the basic elements of a datacenter – including power distribution, cooling, and of course computing. We elucidate how these systems create opportunities for dynamic flexibility in capacity, the key element required for grid-integration, and the driving motivation for variable capacity scheduling. We discuss grid dynamics and how growing datacenter demand challenges the decarbonization of the grid with variable renewable generation. This leads to discussion of the important role that datacenters can play in both increasing grid power capacity and accelerating grid decarbonization, and what datacenters (and computation workloads) need to do to help society forward into a more sustainable future, combined with the growth of artificial intelligence. Finally, we discuss some implications for scheduling problems, and some interesting research directions.

  PublisherDOI

• Scheduling Algorithms and Techniques with Variable Capacity Resources

  2026-02-06

  book-chapter

  This chapter surveys the impact of resource variability onto scheduling algorithms and metrics. First, we survey recent work related to energy-aware scheduling. Next, we discuss characteristics of current and future HPC and data centers, with a focus on power sources and their variation. Then, we investigate several scheduling techniques that can be deployed for batch schedulers to cope with, or even benefit from, changes in the number of computing resources. The chapter continues with a broad overview of the challenges that face scheduling techniques to account for variability, from new workload types up to combined optimization metrics. We conclude by sketching two case-studies, one for applications with variable running times, and one for risk-aware mapping jobs on dynamically changing platforms.

  PublisherDOI

• Scheduling Variable Capacity Resources for Sustainability

  2026-02-06

  book
  PublisherDOI

• DCGen 1.1 Technical Report: Generating Datacenter Configurations (including IT, Power, Cooling)

  arXiv (Cornell University) · 2026-03-15

  preprintOpen accessSenior author

  Diversification of digital applications and workloads has driven the development of diverse datacenter architectures on ever-larger scales. These datacenters consist of complex IT, power, and cooling systems with interdependencies that influence configuration and performance. As datacenters scale and power density increase, designing realistic models becomes more difficult, particularly for research, because it requires understanding all layers of the datacenter and how they interact. Consequently, many studies rely on outdated or unrealistic designs. To support research in datacenter hardware design principles, operational dynamics, cooling mechanisms, and interactions of these facilities with the electrical grid, we have designed DCGen, a tool which can generate a variety of datacenter configurations (including IT hardware, cooling and power distribution infrastructures) at various electrical power, compute capability, and area targets.The tool captures power and space characteristics of IT, cooling, and power infrastructures at both the rack and datacenter levels, enabling modeling of power, energy, and space. DCGen leverages specific use cases such as AI training, AI inference, and cloud services, to select reference and canonical IT hardware configurations, producing realistic mixes of server types. It can target datacenter scale in terms of both power (e.g., 10 MW, 100 MW, 1 GW) and compute capability. For cooling and power distribution infrastructures, DCGen chooses components from a production equipment catalog that optimizes for space or power efficiency while meeting the datacenter capacity requirements. This tool supports research using realistic datacenter designs through ``what-if'' scenario exploration, including studies of power density evolution over time, grid interconnection capacity planning, datacenter-grid interactions, and space management.

  PublisherDOI

• UpDown: A Supercomputer Co-Designed for Scalable Graph Processing

  IEEE Transactions on Parallel and Distributed Systems · 2026-04-06

  article1st authorCorresponding

  Traditional supercomputers have focused on dense computation performance as exemplified by HPL. Graph processing applications differ with extreme irregularity (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$10^{9}$</tex-math></inline-formula> imbalance in skewed, real-world graphs) that produces unpredictable work, parallelism, memory access, and communication. Together, these make scalable performance and programming difficult. We describe the UpDown system architecture, co-designed for irregular graph computations. UpDown provides efficient fine-grained thread invocations (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\sim$</tex-math></inline-formula>10 instructions), direct messaging (no network interface card) for scalable local and global messaging, and split-transaction memory operations that enable extremely high memory bandwidth. Combined with architectural support for global addressing and an aggressive network design, these UpDown features enable direct exploitation of edge and vertex parallelism, using it to deliver breakthrough graph processing performance and programmability. We evaluate the performance of the UpDown system using a challenging suite of graph applications (Pagerank, Breadth-first Search, Triangle Counting, Partial Match, etc). For a single-node, results show 100-fold performance advantage over multicore CPUs. Compared to today's fastest scalable parallel computers UpDown achieves 1000-fold performance increases. UpDown delivers these levels of performance with high-level programmability, these programs directly express vertex-edge parallelism which UpDown exploits directly in hardware.

  PublisherDOI

• DCGen 1.1 Technical Report: Generating Datacenter Configurations (including IT, Power, Cooling)

  arXiv (Cornell University) · 2026-03-15

  articleOpen accessSenior author

  Diversification of digital applications and workloads has driven the development of diverse datacenter architectures on ever-larger scales. These datacenters consist of complex IT, power, and cooling systems with interdependencies that influence configuration and performance. As datacenters scale and power density increase, designing realistic models becomes more difficult, particularly for research, because it requires understanding all layers of the datacenter and how they interact. Consequently, many studies rely on outdated or unrealistic designs. To support research in datacenter hardware design principles, operational dynamics, cooling mechanisms, and interactions of these facilities with the electrical grid, we have designed DCGen, a tool which can generate a variety of datacenter configurations (including IT hardware, cooling and power distribution infrastructures) at various electrical power, compute capability, and area targets.The tool captures power and space characteristics of IT, cooling, and power infrastructures at both the rack and datacenter levels, enabling modeling of power, energy, and space. DCGen leverages specific use cases such as AI training, AI inference, and cloud services, to select reference and canonical IT hardware configurations, producing realistic mixes of server types. It can target datacenter scale in terms of both power (e.g., 10 MW, 100 MW, 1 GW) and compute capability. For cooling and power distribution infrastructures, DCGen chooses components from a production equipment catalog that optimizes for space or power efficiency while meeting the datacenter capacity requirements. This tool supports research using realistic datacenter designs through ``what-if'' scenario exploration, including studies of power density evolution over time, grid interconnection capacity planning, datacenter-grid interactions, and space management.

  PublisherOA PDF

• Distribution and Management of Datacenter Load Decoupling

  ArXiv.org · 2025-11-12

  preprintOpen accessSenior author

  The exploding power consumption of AI and cloud datacenters (DCs) intensifies the long-standing concerns about their carbon footprint, especially because DCs' need for constant power clashes with volatile renewable generation needed for grid decarbonization. DC flexibility (a.k.a. load adaptation) is a key to reducing DC carbon emissions by improving grid renewable absorption. DC flexibility can be created, without disturbing datacenter capacity by decoupling a datacenter's power capacity and grid load with a collection of energy resources. Because decoupling can be costly, we study how to best distribute and manage decoupling to maximize benefits for all. Key considerations include site variation and datacenter-grid cooperation. We first define and compute the power and energy needs of datacenter load decoupling, and then we evaluate designed distribution and management approaches. Evaluation shows that optimized distribution can deliver &gt;98% of the potential grid carbon reduction with 70% of the total decoupling need. For management, DC-grid cooperation (2-way sharing and control vs. 1-way info sharing) enables 1.4x grid carbon reduction. Finally, we show that decoupling may be economically viable, as on average datacenters can get power cost and carbon emissions benefits greater than their local costs of decoupling. However, skew across sites suggests grid intervention may be required.

  PublisherOA PDFDOI

• How Fast Can Graph Computations Go on Fine-grained Parallel Architectures

  ArXiv.org · 2025-07-01

  preprintOpen accessSenior author

  Large-scale graph problems are of critical and growing importance and historically parallel architectures have provided little support. In the spirit of co-design, we explore the question, How fast can graph computing go on a fine-grained architecture? We explore the possibilities of an architecture optimized for fine-grained parallelism, natural programming, and the irregularity and skew found in real-world graphs. Using two graph benchmarks, PageRank (PR) and Breadth-First Search (BFS), we evaluate a Fine-Grained Graph architecture, UpDown, to explore what performance codesign can achieve. To demonstrate programmability, we wrote five variants of these algorithms. Simulations of up to 256 nodes (524,288 lanes) and projections to 16,384 nodes (33M lanes) show the UpDown system can achieve 637K GTEPS PR and 989K GTEPS BFS on RMAT, exceeding the best prior results by 5x and 100x respectively.

  PublisherOA PDFDOI

• Welcome: Sustainability and Computing Special Section

  Communications of the ACM · 2025-06-27 · 14 citations

  articleOpen access1st authorCorresponding
  PublisherOA PDFDOI

Recent grants

• CRISP 2.0 Type 2: Collaborative Research: Exploiting Interdependencies Between Computing and Electrical Power Infrastructures to Maximize Resilience and Flexibility

  NSF · $1.1M · 2018–2023

• Project/Proposal Title: EAGER: Creating a New Paradigm for Computer Architecture and Implementation: The 10 X 10 Idea

  NSF · $300k · 2010–2012

• High-Performance, Adaptive Routing in Multiprocessor Networks

  NSF · $282k · 1993–1997

• II-New: RIVER: A Research Infrastructure to Explore Volatility, Energy-Efficiency, and Resilience

  NSF · $1.0M · 2014–2019

• Project/Proposal Title: EAGER: Creating a New Paradigm for Computer Architecture and Implementation: The 10 X 10 Idea

  NSF · $255k · 2011–2013

Frequent coauthors

• Liuzixuan Lin

  University of Chicago

  65 shared

• Rajini Wijayawardana

  University of Chicago

  55 shared

• Varsha Rao

  University of Chicago

  54 shared

• Wedan Emmanuel Gnibga

  Institut de Recherche en Informatique et Systèmes Aléatoires

  54 shared

• Hai Thanh Nguyen

  Can Tho University

  54 shared

• Martin Holcmann

  Comprehensive Cancer Center Vienna

  36 shared

• Barbara Drobits-Handl

  36 shared

• Ana Korosec

  Medical University of Vienna

  36 shared

Labs

• CERES Center for Unstoppable ComputingPI

Education

• Ph.D., Computer Science

  University of California, Berkeley

  1991

• M.S., Computer Science

  University of California, Berkeley

  1987

• B.S., Electrical Engineering and Computer Science

  University of California, Berkeley

  1985

Awards & honors

• Distinguished service professor (2015)
• Elected AAAS fellow (2008)
• Elected IEEE fellow (2006)
• Best paper award, HPSSS (2013)
• Best student paper award, IEEE Symposium on Cloud Computing…