Systems and Architecture
2-5 PM, February 16
In today’s rapidly evolving technological landscape, the convergence of systems and architecture plays a pivotal role in shaping the future. This student conference session aims to delve into the intricacies and significance of this intersection, offering insights into how these two fields intertwine to create innovative solutions and pave the way for groundbreaking advancements. Some of the learning outcomes include: 1. Comprehend the interdependencies between systems and architecture. 2. Gain insights into designing scalable, adaptable, and efficient systems. 3. Explore real-world applications and their impact on various industries. 4. Identify challenges and emerging trends in this dynamic field.
Keynote Speaker – Dr Akshitha Sriraman, Carnegie Mellon University
“Introducing Equitable Systems and Architectures”
Time: 2-2:50 PM, February 16
Abstract: The modern systems and architectures we build today are mostly either in the hands of the privileged or on the ground as e-waste in under-developed communities, since we, in the systems community, often do not consider the ethical implications of the systems we build. In this rather non-traditional talk that is more of a call to action, I will outline how you, as a systems and architecture researcher, can make equitable computing as a first-order systems consideration to enable efficient systems that are also equitable.
Biography: Akshitha Sriraman is an Assistant Professor in the departments of Electrical and Computer Engineering and Computer Science at Carnegie Mellon University. Her research bridges computer architecture and software systems, demonstrating the importance of that bridge in enabling efficient, sustainable, and equitable web services via solutions that span the compute stack. Her research has been recognized with several awards, including the NSF CAREER Award, the 2023 Intel Rising Star Award, the 2022 ACM SIGARCH/IEEE CS TCCA Outstanding Dissertation Award Honorable Mention, the 2022 ACM SIGOPS Dennis M. Ritchie Doctoral Dissertation Award Honorable Mention, and an IEEE Micro Top Picks distinction.
Keynote Speaker Q&A: 2:50-3 PM, February 16
Invited Student Speaker – Joseph Zuckerman, Columbia University
“Democratizing System-on-Chip Research with Open ESP“
Time: 3:10-3:40 PM, February 16
Abstract: Modern computing systems are relying more and more upon heterogeneous system-on-chip (SoC) architectures, which combine general-purpose processors with dedicated hardware accelerators. Due to their size, complexity, and heterogeneity, these systems are difficult to design and verify, especially for small teams. In this talk, I will present ESP, an open-source platform for SoC design, developed over the course of a decade of research and teaching at Columbia University, to address this challenge. I will discuss two important use cases of ESP. First, I will demonstrate ESP’s productivity gains in SoC design, as evidenced by the ASIC implementation of two domain-specific SoCs in the last two years. The two SoCs have increasing complexity, yet were both designed by less than 10 engineers in the span of just a few months. Second, I will show how ESP can be used as a valuable tool for computer architecture research to enable full-system evaluation of novel ideas. This potential is demonstrated by our recent work Cohmeleon, in which we applied reinforcement learning to dynamically select the best coherence mode for accelerators at runtime based on the status of the SoC.
Biography: Joseph Zuckerman is a Computer Science PhD candidate at Columbia University, working in the System-Level Design group, advised by Professor Luca Carloni. Joseph’s research interests include agile design methodologies, novel architectures, and runtime optimization of heterogeneous systems-on-chip. He currently leads the development of ESP, an open-source research platform for SoC design, in use throughout academia and industry. He is the recipient of a NSF graduate research fellowship. He previously completed his bachelor’s degree in Electrical Engineering from Harvard University in 2019.
Invited Student Speaker Q&A: 3:40-3:50 PM, February 16
Student Speakers
Time: 4-5 PM, February 16
Haoran Qiu
“Power-aware Deep Learning Model Serving with µ-Serve“
Yiqiu Sun
“MapReduce Framework for Processing-Using-Memory“
Benjamin Reidys
“RackBlox: A Software-Defined Rack-Scale Storage System with Network-Storage Co-Design“
Yuqi Xue
“Hardware-Assisted NPU Multi-tenancy for Improved Resource Utilization and Fairness with V10“