PowerPrep : A power management technique for user-facing datacenter workloads
Open Access
Author:
Govindaraj, Vineetha
Graduate Program:
Computer Science and Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
October 28, 2020
Committee Members:
Mahmut Taylan Kandemir, Thesis Advisor/Co-Advisor John Morgan Sampson, Thesis Advisor/Co-Advisor Chitaranjan Das, Program Head/Chair Vijaykrishnan Narayanan, Thesis Advisor/Co-Advisor
Keywords:
Datacenter workloads computer architecture
Abstract:
Modern data center applications are user facing/latency critical. We analyze and
generalize characteristics of such applications i.e., high idleness, unpredictable CPU usage,
and high sensitivity to CPU performance. In spite of such execution characteristics,
datacenter operators disable sleep states to optimize performance. Deep-sleep states hurt
performance mainly due to: a) high wake-latency and b) cache warm-up after exiting
deep-sleep. To address these challenges, we quantify three necessary characteristics
required to realize deep-sleep states in datacenter applications: a) low wake-latency, b)
low resident power, and c) selective retention of cache-state. Using these observations,
we show how emerging technological advances can be leveraged to improve the energy
efficiency of latency-critical datacenter workloads.
