Predicting LLM Inference Latency: A Roofline-Driven ML Method
- Saki Imai
- Rina Nakazawa
- et al.
- 2024
- NeurIPS 2024
Rina Nakazawa
Title
Staff Research Scientist
Bio
I joined IBM in 2014 after receiving M.S. degree of Computer Science from Ochanomizu University in Japan.
- Graduated Ochanomizu University in 2012
- Received Master degree of Computer Science from Ochanomizu University in 2014
- Received PhD degree of Computer Science from Ochanomizu University in 2020
My current research interests include performance analysis, container technologies, microservices, and visualization.
Publications
Best-Effort Power Model Serving for Energy Quantification of Cloud Instances
- 2024
- MASCOTS 2024
Optimizing GPU Multiplexing for Efficient and Cost-Effective Access to Diverse Large Language Models in GPU Clusters
- Yue Zhu
- Chen Wang
- et al.
- 2024
- MASCOTS 2024
Process-based Efficient Power Level Exporter
- Marcelo Amaral
- Huamin Chen
- et al.
- 2024
- CLOUD 2024
Kepler: A Framework to Calculate the Energy Consumption of Containerized Applications
- Marcelo Amaral
- Huamin Chen
- et al.
- 2023
- CLOUD 2023
ImageJockey: A framework for container performance engineering
- 2020
- CLOUD 2020
ConfAdvisor: A performance-centric configuration tuning framework for containers on kubernetes
- Tatsuhiro Chiba
- Rina Nakazawa
- et al.
- 2019
- IC2E 2019
Visualization Tool for Designing Microservices with the Monolith-First Approach
- Rina Nakazawa
- Takanori Ueda
- et al.
- 2018
- VISSOFT 2018
Taming Performance Degradation of Containers in the Case of Extreme Memory Overcommitment
- 2017
- CLOUD 2017
Patents
- GB
- 2612727
- US
- 11423025
- US
- 11265264
- US
- 10664378
- US
- 10430250
- US
- 10142197
Projects
Top collaborators
TC
Tatsuhiro Chiba
STSM, Manager, Hybrid Cloud
SC
Sunyanan Choochotkaew
Researcher, Hybrid Cloud
MC
Marcelo Carneiro do Amaral
Staff Research Scientist
ST
Scott Trent
Staff Research Scientist & Certified Linux Geek