ConSil Jeff Chase Duke University. Collaborators Justin Moore –received PhD in April, en route to Google. Did this research. Wrote this paper. Named the.

Slides:



Advertisements
Similar presentations
1 Using ANN to solve the Navier Stoke Equations Motivation: Solving the complete Navier Stokes equations using direct numerical simulation is computationally.
Advertisements

Support Vector Machines
SEKE 2014, Hyatt Regency, Vancouver, Canada
Visualization of hidden node activity in a feed forward neural network Adam Arvay.
Machine Learning Neural Networks
Multiple Criteria for Evaluating Land Cover Classification Algorithms Summary of a paper by R.S. DeFries and Jonathan Cheung-Wai Chan April, 2000 Remote.
Introduction to Neural Networks John Paxton Montana State University Summer 2003.
1 Automated Feature Abstraction of the fMRI Signal using Neural Network Clustering Techniques Stefan Niculescu and Tom Mitchell Siemens Medical Solutions,
October 14, 2010Neural Networks Lecture 12: Backpropagation Examples 1 Example I: Predicting the Weather We decide (or experimentally determine) to use.
CHAPTER 11 Back-Propagation Ming-Feng Yeh.
October 7, 2010Neural Networks Lecture 10: Setting Backpropagation Parameters 1 Creating Data Representations On the other hand, sets of orthogonal vectors.
CSC2535: 2013 Advanced Machine Learning Lecture 3a: The Origin of Variational Bayes Geoffrey Hinton.
CoolAir Temperature- and Variation-Aware Management for Free-Cooled Datacenters Íñigo Goiri, Thu D. Nguyen, and Ricardo Bianchini 1.
1 Neural plug-in motor coil thermal modeling Mo-Yuen Chow; Tipsuwan Y Industrial Electronics Society, IECON 26th Annual Conference of the IEEE, Volume:
Choosing Your Primary Research Method What do you need to find out that your literature did not provide?
Ranga Rodrigo April 5, 2014 Most of the sides are from the Matlab tutorial. 1.
Optimal Fan Speed Control for Thermal Management of Servers UMass-Amherst Green Computing Seminar September 21 st, 2009.
Thermal Aware Resource Management Framework Xi He, Gregor von Laszewski, Lizhe Wang Golisano College of Computing and Information Sciences Rochester Institute.
CHAPTER 12 ADVANCED INTELLIGENT SYSTEMS © 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang.
Chapter 8 Prediction Algorithms for Smart Environments
XI HE Computing and Information Science Rochester Institute of Technology Rochester, NY USA Rochester Institute of Technology Service.
U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 2011 Predicting Solar Generation from Weather Forecasts Using Machine Learning Navin.
A Shaft Sensorless Control for PMSM Using Direct Neural Network Adaptive Observer Authors: Guo Qingding Luo Ruifu Wang Limei IEEE IECON 22 nd International.
Sensor-Based Fast Thermal Evaluation Model For Energy Efficient High-Performance Datacenters Q. Tang, T. Mukherjee, Sandeep K. S. Gupta Department of Computer.
December 5, 2012Introduction to Artificial Intelligence Lecture 20: Neural Network Application Design III 1 Example I: Predicting the Weather Since the.
Random Sampling, Point Estimation and Maximum Likelihood.
Heterogeneity and Dynamicity of Clouds at Scale: Google Trace Analysis [1] 4/24/2014 Presented by: Rakesh Kumar [1 ]
IE 585 Introduction to Neural Networks. 2 Modeling Continuum Unarticulated Wisdom Articulated Qualitative Models Theoretic (First Principles) Models Empirical.
Chapter 9 Neural Network.
© Copyright 2004 ECE, UM-Rolla. All rights reserved A Brief Overview of Neural Networks By Rohit Dua, Samuel A. Mulder, Steve E. Watkins, and Donald C.
Chapter 3 Neural Network Xiu-jun GONG (Ph. D) School of Computer Science and Technology, Tianjin University
NEURAL NETWORKS FOR DATA MINING
Summer Report Xi He Golisano College of Computing and Information Sciences Rochester Institute of Technology Rochester, NY
Thermal-aware Issues in Computers IMPACT Lab. Part A Overview of Thermal-related Technologies.
Copyright © 2011, Performance Evaluation of a Green Scheduling Algorithm for Energy Savings in Cloud Computing Truong Vinh Truong Duy; Sato,
Artificial Neural Networks (ANN) modeling of the pulsed heat load during ITER CS magnet operation L. Savoldi Richard 1, R. Bonifetto 1, S. Carli 1, A.
IE 585 Competitive Network – Learning Vector Quantization & Counterpropagation.
Active Sampling for Accelerated Learning of Performance Models Piyush Shivam, Shivnath Babu, Jeff Chase Duke University.
Thermal Aware Data Management in Cloud based Data Centers Ling Liu College of Computing Georgia Institute of Technology NSF SEEDM workshop, May 2-3, 2011.
Learning Application Models for Utility Resource Planning Piyush Shivam, Shivnath Babu, Jeff Chase Duke University IEEE International Conference on Autonomic.
Confidence Interval Estimation For statistical inference in decision making:
An Artificial Neural Network Approach to Surface Waviness Prediction in Surface Finishing Process by Chi Ngo ECE/ME 539 Class Project.
CS621 : Artificial Intelligence
PARALLELIZATION OF ARTIFICIAL NEURAL NETWORKS Joe Bradish CS5802 Fall 2015.
1 Thermal Management of Datacenter Qinghui Tang. 2 Preliminaries What is data center What is thermal management Why does Intel Care Why Computer Science.
XI HE Computing and Information Science Rochester Institute of Technology Rochester, NY USA Rochester Institute of Technology Service.
Modeling Electricity Demand: A Neural Network Approach Christian Crowley GWU Department of Economics 28 th Annual IAEE International Conference June 6,
Neural Networks Presented by M. Abbasi Course lecturer: Dr.Tohidkhah.
1 1 Slide Simulation Professor Ahmadi. 2 2 Slide Simulation Chapter Outline n Computer Simulation n Simulation Modeling n Random Variables and Pseudo-Random.
OPERATING SYSTEMS CS 3530 Summer 2014 Systems and Models Chapter 03.
Xi He Golisano College of Computing and Information Sciences Rochester Institute of Technology Rochester, NY THERMAL-AWARE RESOURCE.
BME 353 – BIOMEDICAL MEASUREMENTS AND INSTRUMENTATION MEASUREMENT PRINCIPLES.
Measurements Measurements and errors : - Here, the goal is to have some understanding of the operation and behavior of electrical test instruments. Also,
Joe Bradish Parallel Neural Networks. Background  Deep Neural Networks (DNNs) have become one of the leading technologies in artificial intelligence.
Operation and Control Strategy of PV/WTG/EU Hybrid Electric Power System Using Neural Networks Faculty of Engineering, Elminia University, Elminia, Egypt.
Automated Cost-Aware Data Center Management (part 1) Justin Moore Advisor: Jeff Chase Committee Parthasarathy Ranganathan, Carla Ellis, Alvin Lebeck, Jun.
Neural network based hybrid computing model for wind speed prediction K. Gnana Sheela, S.N. Deepa Neurocomputing Volume 122, 25 December 2013, Pages 425–429.
Learning: Neural Networks Artificial Intelligence CMSC February 3, 2005.
Pattern Recognition Lecture 20: Neural Networks 3 Dr. Richard Spillman Pacific Lutheran University.
Learning with Neural Networks Artificial Intelligence CMSC February 19, 2002.
Energy Aware Network Operations
Machine Learning Supervised Learning Classification and Regression
Big data classification using neural network
Forecasting with Cyber-physical Interactions in Data Centers (part 3)
Neural Network Architecture Session 2
I.Panapakidis, A.Dagoumas Energy & Environmental Policy laboratory,
Deep Feedforward Networks
Automated Cost-Aware Data Center Management (part 4)
Luís Filipe Martinsª, Fernando Netoª,b. 
CS 4501: Introduction to Computer Vision Training Neural Networks II
Presentation transcript:

ConSil Jeff Chase Duke University

Collaborators Justin Moore –received PhD in April, en route to Google. Did this research. Wrote this paper. Named the system. –Something to do with “Get Smart” (?) Did not send me slides… Partha Ranganathan (HP) has led this work.

Context: Dynamic Thermal Management for Data Centers CRAC Rack Temperature Scale (C) Heat build-ups

Goals ConSil is part of a larger system to analyze data center thermals and manage heat proactively. –Temperature-aware workload placement –“Smart cooling” Preliminary conclusion: it is practical to reduce total energy by about 15% under “typical” conditions. –Your mileage may vary. Other goals: –Reduce capital cost with “common case” cooling system. Allow cluster to “burst”, but stop short of meltdown. –Improve long-term reliability and availability –Better data center design

“Green” Workload Placement Making Scheduling "Cool": Temperature-Aware Resource Assignment in Data Centers by Justin Moore, J. Chase, P. Ranganathan, and R. Sharma. In the 2005 USENIX Annual Technical Conference, April USENIX Annual Technical Conference Place workload intelligently to promote an even temperature distribution, given the “thermal topology” of the data center.

The Subproblem that Consil Solves How hot is point (x, y, z) in your data center? –Placement policies need a thermal map Option 1: install new instrumentation –Tradeoff $$$ vs. granularity Option 2: use built-in sensors –But: how to derive the inlet temperatures? If we can do that, then we can obtain a precise and accurate thermal map with low instrumentation cost.

Thermal Instrumentation Inlet Heat (Q inlet ) Heat Sources (Q workload ) Temperature Sensors (Q observed ) Observed: ▲= f(▲, ▲) Learn: ▲= g(▲, ▲)

ConSil in Context Workload measures

Learning a Model Learn statistical model for Y from m samples of YmYm s mn..s m2 s m1 smsm Y2Y2 s 2n..s 22 s 21 s2s2 Y1Y1 s 1n..s 12 s 11 s1s1 YXnXn..X2X2 X1X1 Attributes Samples

First Cut: Neural Nets Infer ambient temperature from an input sample: –Last N workload measure samples (epoch E) –Internal temperature sensor readings Use off-the-shelf FANN library Some static (SWAG) structural choices: –Four layers of neurons Input  hidden  hidden  output –Neurons use FANN sigmoid transform function –Train the net using FANN back-propagation to set input weights on each neuron.

Experiments with Consil Collected data for 12 servers in a data center. Pick servers whose inlet temperatures are known –i.e., they have a sensor near them 45 hours of data collected under active/varying load –Two server models (HP DL360 G3, Dell 1425) –CPU data: 1 second granularities –temperature data: 5 or 30 second granularities CPU utilization only –CPU uses 80% of power (225/275 watts peak) 266 Lines of FANN code

Methodology FFCV –Divide observations into fifths –Train on one fifth, test on four –Do it for each fifth –Compute SSE Output: CDFs of errors Sensitivity study –Training time –Accuracy

ConSil: Accuracy Accurate inference using workload and onboard data –75% of inferred values are within 1C of actual value

Sensitivity Time-to-train –Most significant: FFCV sub-experiment –Training time is highly data-dependent –Epoch length –Number of sensor/workload epochs Accuracy (SSE) –Most significant: FFCV sub-experiment –Indicates not enough variation in behavior –Coarse granularity (more history) improves

ConSil in Context Workload measures

Predicting Thermal Effects Model relationship using machine learning –Inputs: Workload data, AC settings, fan speeds –Output: Predicted thermal map –Learns from observations during normal operation –FANN neural net library –Active “burn in” may speed learning Weatherman: Automated, Online, and Predictive Thermal Mapping and Management for Data Centers by Justin Moore, J. Chase, and P. Ranganathan. Third IEEE International Conference on Autonomic Computing, June 2006.Third IEEE International Conference on Autonomic Computing

Accurate inferences using workload and AC data –Data from validated Flovent CFD models –92% of predicted values are within 1.0C of actual value Weatherman: Accuracy

Summary/Conclusion Machine learning is a useful tool for “autonomic” self- optimization. –Sense and respond –Optimizing control loops based on learned models Neural nets don’t always suck. –Initial results suggest they work well here. –Maybe we can do better. Need good baseline datasets for training/validation. –Variance –History

Why “ConSil”? Cone of Silence –“Mask out” unwanted signals

The maximum number of training iterations was set to $10^5$. Each neural net contained one input, one output, and two hidden layers. Each hidden layer contained twice the number of neurons as the input layer; varying the number of recent epochs we use as input, we vary the number of workload epochs~---~parameter $B$~---~and internal sensor epochs~---~parameter $C$~---~independently. Using general full factorial design analysis, we can identify which parameters have a significant effect when changed, and for which parameters we can simply select a ``reasonable'' value.