OptimML: Joint Control of Inference Latency and Server Power Consumption for ML Performance Optimization
G Chen, X Wang - ACM Transactions on Autonomous and Adaptive …, 2024 - dl.acm.org
Power capping is an important technique for high-density servers to safely oversubscribe the
power infrastructure in a data center. However, power capping is commonly accomplished …
power infrastructure in a data center. However, power capping is commonly accomplished …
Performance optimization of machine learning inference under latency and server power constraints
G Chen, X Wang - 2022 IEEE 42nd International Conference on …, 2022 - ieeexplore.ieee.org
Power capping is an important technique for high-density servers to safely oversubscribe the
power infrastructure in a data center. However, power capping is commonly accomplished …
power infrastructure in a data center. However, power capping is commonly accomplished …
Efficient Processing of Machine Learning Workloads on Data Center Servers
G Chen - 2024 - rave.ohiolink.edu
Data centers often need to run lots of Machine Learning (ML) applications with stringent
Service-Level Objective (SLO) requirements, such as inference latency. With bursty Machine …
Service-Level Objective (SLO) requirements, such as inference latency. With bursty Machine …
Adaptive power management through thermal aware workload balancing in internet data centers
J Yao, H Guan, J Luo, L Rao… - IEEE Transactions on …, 2014 - ieeexplore.ieee.org
The past decade witnessed the tremendous growth of online services and applications.
Together with the increase of cloud computing, more and more computation are hosted by …
Together with the increase of cloud computing, more and more computation are hosted by …
DeepPower: Deep Reinforcement Learning based Power Management for Latency Critical Applications in Multi-core Systems
Latency-critical (LC) applications are widely deployed in modern datacenters. Effective
power management for LC applications can yield significant cost savings. However, it poses …
power management for LC applications can yield significant cost savings. However, it poses …
A reinforcement learning approach for performance-aware reduction in power consumption of data center compute nodes
As Exascale computing becomes a reality, the energy needs of compute nodes in cloud data
centers will continue to grow. A common approach to reducing this energy demand is to limit …
centers will continue to grow. A common approach to reducing this energy demand is to limit …
Pack & cap: adaptive dvfs and thread packing under power caps
The ability to cap peak power consumption is a desirable feature in modern data centers for
energy budgeting, cost management, and efficient power delivery. Dynamic voltage and …
energy budgeting, cost management, and efficient power delivery. Dynamic voltage and …
Network packet processing mode-aware power management for data center servers
In data center servers, power management (PM) exploiting Dynamic Voltage and Frequency
Scaling (DVFS) for processors can play a crucial role to improve energy efficiency. However …
Scaling (DVFS) for processors can play a crucial role to improve energy efficiency. However …
Sleepscale: Runtime joint speed scaling and sleep states management for power efficient data centers
Power consumption in data centers has been growing significantly in recent years. To
reduce power, servers are beingequipped with increasingly sophisticated power …
reduce power, servers are beingequipped with increasingly sophisticated power …
Apparate: Rethinking Early Exits to Tame Latency-Throughput Tensions in ML Serving
Machine learning (ML) inference platforms are tasked with balancing two competing goals:
ensuring high throughput given many requests, and delivering low-latency responses to …
ensuring high throughput given many requests, and delivering low-latency responses to …