Gyeongseo Park
Research Interests
- Energy-Efficient Computing
- Systems for Network
- Virtualization
- Systems for Machine Learning
Education
- Ph.D. in Electrical Engineering and Computer Science, DGIST, Daegu, Mar, 2017 – 2024
- B.S. in Electrical and Computer Engineering (Magna Cum Laude), Ajou University, Suwon, Feb, 2016
Experience
Research Assistant, March 2017 to August 2024
- ML for Systems & Systems for ML
- Participating in industry-university cooperation to develop PIM (Processing In Memory) for distributed computing (e.g., Spark, Hadoop) [2019 – 2020]
- Participating in industry-university cooperation to develop PIM for DLRM (Deep Learning Recommendation Model) [2020 – 2021]
- Participating in industry-university cooperation to develop distributed computing framework providing tiered-memory management (e.g., CXL Memory, Persistent Memory, etc.) [2022 – 2024]
- Developing reinforcement learning-based dynamic core allocation for latency-critical services [Work in Progress]
- Energy-Efficient Systems for Latency-Critical Services
- Developing network load-aware power management for latency-critical services (CAL ’20, MICRO ’21)
- Developing network interrupt rate management for improving the energy efficiency of latency-critical services (ICCD ’20, IEEE Access ‘20)
- Developing latency-critical container-aware power management (ISET ’18)
- Developing dynamic core allocation in virtualization (i.e., KVM) for improving the performance and energy efficiency of latency-critical services [Work in Progress]
Skiils and Techniques
- Machine Learning Algorithms: DLRM, DQN, PPO, etc.
- Programing Language: C, Python, C++
- Frameworks/Packages: Pytorch, Pandas, eBPF
Publications
- (Conference) Co-UP: Comprehensive Core and Uncore Power Management for Latency-Critical Workloads
- (Conference) vSPACE: Supporting Parallel Network Packet Processing in Virtualized Environments through Dynamic Core Management
- (Thesis) Dynamic Core Allocation for Latency-Critical Workloads in Data-Center Servers
- (Journal) NoHammer: Preventing Row Hammer With Last-Level Cache Management
- (Journal) CoreNap: Energy Efficient Core Allocation for Latency-Critical Workloads
- (Conference) NMAP: Power Management Based on Network Packet Processing Mode Transition for Latency-Critical Workloads
- (Conference) Improving the Efficiency of Power Management via Dynamic Interrupt Management
- (Journal) Co-Adjusting Voltage/Frequency State and Interrupt Rate for Improving Energy-Efficiency of Latency-Critical Applications
- (Journal) Network Packet Processing Mode-Aware Power Management for Data Center Servers
