PEARL
Performance, Efficiency, Autonomy, and Reliability in Computing Systems
The main goal of our Research Lab is to derive unique insights on and improve the performance, efficiency, and reliability of large-scale computing systems. Our research questions and revisits long-standing concepts and abstractions in system software, in the context of technology trends and workload evolution. We use experimental and theoretical approaches to enable autonomic management of system resources. Our current emphasis is on the automated synthesis of memory management algorithms and data structures, new abstractions and architectures for Cloud Computing and taming extreme heterogeneity in High Performance Computing.
RESEARCH
HETEROGENEOUS SCALABLE MEMORY SYSTEMS
Managing inherent and architected heterogeneity in current and future memory systems. Revisiting virtual memory management abstractions and algorithms in light of new memory technologies and emerging workloads.
SERVERLESS COMPUTING
Introducing new abstractions and capabilities for Functions-as-a-Service to add hardware diversity and enable higher performance. Supported by NSF
DEMOCRATIZING MACHINE LEARNING TRAINING
Research on enabling training of ML models on any device, anywhere.
SYNTHESIS OF HIGH-PERFORMANCE DATA STRUCTURES
Rethinking how data structures for high-performance computing are designed and explore new data structure synthesis methodologies.
PROJECTS
HardLambda (generously supported by NSF)
VirtEdgeAI (generously supported by Cisco)
OUR TEAM
DIMITRIOS NIKOLOPOULOS
Lab Director
Professor of Computer Science and by courtesy Electrical and Computer Engineering, working in system software and high-performance computing. Current interests include memory systems and new approaches to Cloud Computing.
MOUSTAFA KAHLA
PhD Student
Identifying and automating missed compiler optimizations in parallel programming models.
EMAD ABDRABOU
PhD Student
Containerization and autoscaling technologies for network slicing.
JACQUELINE BRUEN
PhD Student
Serverless computing and its interplay with emerging heterogeneous memory technologies
DANIEL MOYER
MSc (2021)
Designed and implemented the first capability for ephemeral communication and stateful sharing in serverless computing, prototyped on AWS Lambda.
LALITHA KUPPA
Undergraduate Researcher
Rethinking page migration in Linux in an era of extreme heterogeneity and non-volatility in server memory systems.
MRUNAL MANISH KHINVASARA
MSc Student
Working on algorithmic and system improvement for large-scale graph mining systems.
MEL CAMERON
MSc Student
Working on new educational tools and visualization technologies for teaching computing systems concepts in CS curricula.
ISHHAN GULATI
MSc Student
Hierarchical and Scalable Consensus Protocols for Edge Computing
MATTHEW JACKSON
MSc Student
Federated Learning using Heterogeneous Edge Computing Resources
OPEN POSITIONS
RESEARCH ASSISTANTS (ALL LEVELS)
We have a mission to develop young talent in Computing Systems Research and we are looking for new students to join our team at all levels (PhD, MSc, or Undergraduate). You will have excellent opportunities and state of the art equipment to address challenging research problems and shape the future of large-scale computing systems. We offer an open, collaborative and conducive research environment. We have trained more than fifty (50) PostDoctoral Fellows, PhD Students, MSc students, and Undergraduates, all of whom are now in senior and leadership positions in academia and industry. Our researchers benefit from our strong network of research partners in government and industry and have extensive opportunities for internships and joint research.