Towards Better Understanding of Black-box Auto-Tuning: A Comparative Analysis for Storage Systems
|
|
|
- Alannah Jenkins
- 5 years ago
- Views:
Transcription
1 Towards Better Understanding of Black-box Auto-Tuning: A Comparative Analysis for Storage Systems 2018 USENIX Annual Technical Conference Zhen Cao 1, Vasily Tarasov 2, Sachin Tiwari 1, and Erez Zadok 1 1 Stony Brook University; 2 IBM Research Almaden; Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 1
2 Outline l Introduction l Background l Experiment Settings l Evaluations l Related Work l Conclusions & Future Work Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 2
3 Motivation l Why tuning storage systems? uslow storage impacts I/O bound workloads udefault settings are sub-optimal utuning can provide significant gains 9 [FAST 10] l Manual tuning is intractable l Auto-tuning storage systems ublack-box optimization is promising ulack of comparison of techniques ulack of understanding Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 3
4 Contributions l First comparative study on auto-tuning storage systems u 5 techniques l Various aspects u Cumulative & instantaneous throughput u Impacts of hyper-parameters l Explanations on evaluation results u From storage perspective l Future Goal: complete solution to tune storage systems Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 4
5 Outline l Introduction l Background l Experiment Settings l Evaluations l Related Work l Conclusions & Future Work Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 5
6 Concepts l Storage system u File system, underlying storage hardware and any layers between them l Parameters u Configurable options u E.g., file-system block size l Parameter values u E.g., 1K, 2K, 4K (Ext4 block size) l Configuration u Combination of parameter values u E.g., [Ext4, 4K, data=ordered] l Parameter space u All possible configurations l Hyper-parameter Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 6
7 Challenges l Vast parameter space uext4: 59 parameters, configs udevices, Layers udistributed, large-scale l Discrete and non-numeric ulinux I/O scheduler: noop, cfq, deadline l Non-linearity l Sensitivity to environment uhardware & workloads Manual Tuning Inefficient Gradient Unavailable Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 7
8 Inapplicable Methods l Control Theory uunstable in controlling non-linear systems l Supervised Machine Learning ulong training phase uhigh-quality training data l Inapplicable or inefficient to serve as the core auto-tuning algorithm ucould be helpful as a supplement Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 8
9 Black-box Optimization l Successfully applied in auto-tuning system configurations l Examples u Genetic Algorithms (GA) u Simulated Annealing (SA) u Bayesian Optimization (BO) l Obliviousness to system s internals Configuration evaluate select Evaluation Results Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 9
10 Key Factors l Fitness: optimization objective(s) uthroughput, latency, energy, ucomplex cost functions l Exploration usearch the unvisited area (e.g., randomly) l Exploitation uutilize neighborhood or history l History uhow much past data is kept and used for exploration/exploitation Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 10
11 Applied Methods l Simulated Annealing (SA) l Genetic Algorithms (GA) l Deep Q-Network (DQN) l Bayesian Optimization (BO) l Random Search (RS) urandom selection without replacement Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 11
12 Genetic Algorithms l Inspired by natural evolution l Concepts ugene: file system, block size, uallele: Ext4, XFS, Btrfs, uchromosome: configuration upopulation: set of configurations l Selection l Genetic operators ucrossover umutation History Exploitation vs. Exploration Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 12
13 Outline l Introduction l Background l Experiment Settings l Evaluations l Related Work l Conclusions & Future Work Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 13
14 l Hardware Experimental Setup u M1: 2 Intel Xeon single-core 2.8GHz CPU, 2G RAM, 73GB Seagate SCSI drive u M2: 1 Intel Xeon quad-core 2.4GHz CPU, 24G RAM, 4 drives (SAS-HDD 500GB, SAS-HDD 146GB, 1 SATA-HDD, SSD) l Filebench u Macro-workloads: fileserver, mailserver, webserver, dbserver u Default working set size Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 14
15 Experiment Setup (cont.) l Search spaces ustorage V1 File system, inode size, block size, block group, journal options, mount options, special options ustorage V2 V1 + I/O scheduler 6,222 configurations l Methodology uexhaustive Search Storage V2: 4 workloads 4 devices 3+ runs for each configuration Collected over 2+ years usimulate auto-tuning algorithms Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 15
16 Outline l Introduction l Background l Experiment Settings l Evaluations l Related Work l Conclusions & Future Work Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 16
17 Best Throughput Best Throughput (kops/s) M2-Mailserver-HDD3 GA SA BO DQN RS Time (hrs) Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 17
18 Success rate for finding near-optimal configurations Near-optimal configuration: one with throughput higher than 99% of the global optimal value. Percentage of Runs 100% 80% 60% 40% 20% GA SA BO DQN RS M2-Fileserver-HDD Time (hrs) Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 18
19 Instant Throughput Throughput (kops/s) M2-Mailserver-HDD Time (hrs) RS SA GA DQN BO Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 19
20 Genetic Algorithm (GA) Diversity Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 20
21 Correlation Analysis l Correlation analysis u Ordinary Least Squares (OLS) u Example: block size and journal option are the most correlated Ext4 parameter (Fileserver, SSD) l Explanations on evaluation results u GA and BO can identify important parameters through history u SA keeps no history ; thus performs poorly u DQN spends too much time on exploration u Random Search Near-optimal configurations take up 4.5% of the whole search space (M2, Mailserver, HDD). Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 21
22 Outline l Introduction l Background l Experiment Settings l Evaluations l Related Work l Conclusions & Future Work Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 22
23 Related Work l Auto-tuning storage u Storage system design (bin-packing heuristics) [Alvarez et al.] u Data recovery scheduling (GA) [Keeton et al.] u HDF5 optimization (GA) [Behzad et al.] u Lustre optimization (DQN) [Li et al.] l Auto-tuning other systems u Database [Alipourfard et al.] u Cloud VMs [Aken et al.] Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 23
24 Outline l Introduction l Background l Experiment Settings l Evaluations l Related Work l Conclusions & Future Work Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 24
25 Conclusions & Contributions l First comparative analysis on 5 techniques on auto-tuning storage systems u Efficiency on finding near-optimal configurations u Instant throughput l Provide insights from storage perspective u Importance of parameters E.g., impact of mutation rates on convergence l Valuable datasets u Will release to public Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 25
26 Future Work l More complex workloads and search spaces l Hyper-parameter tuning l More sophisticated auto-tuning u E.g., penalty functions to cope with costly parameter changes Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 26
27 Towards Better Understanding of Black-box Auto-Tuning: A Comparative Analysis for Storage Systems Zhen Cao, Vasily Tarasov, Sachin Tiwari, and Erez Zadok Towards Better Understanding of Black-box Auto-Tuning (ATC 18) 27
VMware VMmark V1.1 Results
Vendor and Hardware Platform: IBM System x3950 M2 Virtualization Platform: VMware ESX 3.5.0 U2 Build 110181 Performance VMware VMmark V1.1 Results Tested By: IBM Inc., RTP, NC Test Date: 2008-09-20 Performance
One Optimized I/O Configuration per HPC Application
One Optimized I/O Configuration per HPC Application Leveraging I/O Configurability of Amazon EC2 Cloud Mingliang Liu, Jidong Zhai, Yan Zhai Tsinghua University Xiaosong Ma North Carolina State University
A Tale of Two Erasure Codes in HDFS
A Tale of Two Erasure Codes in HDFS Dynamo Mingyuan Xia *, Mohit Saxena +, Mario Blaum +, and David A. Pease + * McGill University, + IBM Research Almaden FAST 15 何军权 2015-04-30 1 Outline Introduction
NEC PerforCache. Influence on M-Series Disk Array Behavior and Performance. Version 1.0
NEC PerforCache Influence on M-Series Disk Array Behavior and Performance. Version 1.0 Preface This document describes L2 (Level 2) Cache Technology which is a feature of NEC M-Series Disk Array implemented
Minimization of Energy Loss using Integrated Evolutionary Approaches
Minimization of Energy Loss using Integrated Evolutionary Approaches Attia A. El-Fergany, Member, IEEE, Mahdi El-Arini, Senior Member, IEEE Paper Number: 1569614661 Presentation's Outline Aim of this work,
Weather Research and Forecasting (WRF) Performance Benchmark and Profiling. July 2012
Weather Research and Forecasting (WRF) Performance Benchmark and Profiling July 2012 Note The following research was performed under the HPC Advisory Council activities Participating vendors: Intel, Dell,
ArcGIS Deployment Pattern. Azlina Mahad
ArcGIS Deployment Pattern Azlina Mahad Agenda Deployment Options Cloud Portal ArcGIS Server Data Publication Mobile System Management Desktop Web Device ArcGIS An Integrated Web GIS Platform Portal Providing
Advanced Machine Learning Practical 4b Solution: Regression (BLR, GPR & Gradient Boosting)
Advanced Machine Learning Practical 4b Solution: Regression (BLR, GPR & Gradient Boosting) Professor: Aude Billard Assistants: Nadia Figueroa, Ilaria Lauzana and Brice Platerrier E-mails: aude.billard@epfl.ch,
Oil Field Production using Machine Learning. CS 229 Project Report
Oil Field Production using Machine Learning CS 229 Project Report Sumeet Trehan, Energy Resources Engineering, Stanford University 1 Introduction Effective management of reservoirs motivates oil and gas
A heuristic algorithm for the Aircraft Landing Problem
22nd International Congress on Modelling and Simulation, Hobart, Tasmania, Australia, 3 to 8 December 2017 mssanz.org.au/modsim2017 A heuristic algorithm for the Aircraft Landing Problem Amir Salehipour
CPU Consolidation versus Dynamic Voltage and Frequency Scaling in a Virtualized Multi-Core Server: Which is More Effective and When
1 CPU Consolidation versus Dynamic Voltage and Frequency Scaling in a Virtualized Multi-Core Server: Which is More Effective and When Inkwon Hwang, Student Member and Massoud Pedram, Fellow, IEEE Abstract
WRF performance tuning for the Intel Woodcrest Processor
WRF performance tuning for the Intel Woodcrest Processor A. Semenov, T. Kashevarova, P. Mankevich, D. Shkurko, K. Arturov, N. Panov Intel Corp., pr. ak. Lavrentieva 6/1, Novosibirsk, Russia, 630090 {alexander.l.semenov,tamara.p.kashevarova,pavel.v.mankevich,
S0214 : GPU Based Stacking Sequence Generation For Composite Skins Using GA
S0214 : GPU Based Stacking Sequence Generation For Composite Skins Using GA Date: 16th May 2012 Wed, 3pm to 3.25pm(Adv. Session) Sathyanarayana K., Manish Banga, and Ravi Kumar G. V. V. Engineering Services,
StreamSVM Linear SVMs and Logistic Regression When Data Does Not Fit In Memory
StreamSVM Linear SVMs and Logistic Regression When Data Does Not Fit In Memory S.V. N. (vishy) Vishwanathan Purdue University and Microsoft vishy@purdue.edu October 9, 2012 S.V. N. Vishwanathan (Purdue,
Free Search in Multidimensional Space
Free Search in Multidimensional Space Technology School, Maritime and Technology Faculty Southampton Solent University Content Aims Test Problems Bump test function Michalewicz test function Norwegian
Quantum Artificial Intelligence and Machine Learning: The Path to Enterprise Deployments. Randall Correll. +1 (703) Palo Alto, CA
Quantum Artificial Intelligence and Machine : The Path to Enterprise Deployments Randall Correll randall.correll@qcware.com +1 (703) 867-2395 Palo Alto, CA 1 Bundled software and services Professional
ASTRA BASED SWARM OPTIMIZATIONS OF THE BERLinPro INJECTOR
ASTRA BASED SWARM OPTIMIZATIONS OF THE BERLinPro INJECTOR M. Abo-Bakr FRAAC4 Michael Abo-Bakr 2012 ICAP Warnemünde 1 Content: Introduction ERL s Introduction BERLinPro Motivation Computational Aspects
A parallel metaheuristics for the single machine total weighted tardiness problem with sequence-dependent setup times
A parallel metaheuristics for the single machine total weighted tardiness problem with sequence-dependent setup times Wojciech Bożejko Wroc law University of Technology Institute of Computer Engineering,
Environment (Parallelizing Query Optimization)
Advanced d Query Optimization i i Techniques in a Parallel Computing Environment (Parallelizing Query Optimization) Wook-Shin Han*, Wooseong Kwak, Jinsoo Lee Guy M. Lohman, Volker Markl Kyungpook National
Anomaly Detection and Categorization Using Unsupervised Deep Learning
Anomaly Detection and Categorization Using Unsupervised S6340 Thursday 7 th April 2016 GPU Technology Conference A. Stephen McGough, Noura Al Moubayed, Jonathan Cumming, Eduardo Cabrera, Peter Matthews,
Genetic Algorithm. Outline
Genetic Algorithm 056: 166 Production Systems Shital Shah SPRING 2004 Outline Genetic Algorithm (GA) Applications Search space Step-by-step GA Mechanism Examples GA performance Other GA examples 1 Genetic
Quantitative Estimation of the Performance Delay with Propagation Effects in Disk Power Savings
Quantitative Estimation of the Performance Delay with Propagation Effects in Disk Power Savings Feng Yan 1, Xenia Mountrouidou 1, Alma Riska 2, and Evgenia Smirni 1 1 College of William and Mary, Williamsburg,
Energy-aware scheduling for GreenIT in large-scale distributed systems
Energy-aware scheduling for GreenIT in large-scale distributed systems 1 PASCAL BOUVRY UNIVERSITY OF LUXEMBOURG GreenIT CORE/FNR project Context and Motivation Outline 2 Problem Description Proposed Solution
CONTEMPORARY ANALYTICAL ECOSYSTEM PATRICK HALL, SAS INSTITUTE
CONTEMPORARY ANALYTICAL ECOSYSTEM PATRICK HALL, SAS INSTITUTE Copyright 2013, SAS Institute Inc. All rights reserved. Agenda (Optional) History Lesson 2015 Buzzwords Machine Learning for X Citizen Data
FPGA Implementation of a HOG-based Pedestrian Recognition System
MPC Workshop Karlsruhe 10/7/2009 FPGA Implementation of a HOG-based Pedestrian Recognition System Sebastian Bauer sebastian.bauer@fh-aschaffenburg.de Laboratory for Pattern Recognition and Computational
Integer weight training by differential evolution algorithms
Integer weight training by differential evolution algorithms V.P. Plagianakos, D.G. Sotiropoulos, and M.N. Vrahatis University of Patras, Department of Mathematics, GR-265 00, Patras, Greece. e-mail: vpp
Sparse and Robust Optimization and Applications
Sparse and and Statistical Learning Workshop Les Houches, 2013 Robust Laurent El Ghaoui with Mert Pilanci, Anh Pham EECS Dept., UC Berkeley January 7, 2013 1 / 36 Outline Sparse Sparse Sparse Probability
Che-Wei Chang Department of Computer Science and Information Engineering, Chang Gung University
Che-Wei Chang chewei@mail.cgu.edu.tw Department of Computer Science and Information Engineering, Chang Gung University } 2017/11/15 Midterm } 2017/11/22 Final Project Announcement 2 1. Introduction 2.
Reinforcement Learning
Reinforcement Learning Model-Based Reinforcement Learning Model-based, PAC-MDP, sample complexity, exploration/exploitation, RMAX, E3, Bayes-optimal, Bayesian RL, model learning Vien Ngo MLR, University
Segment-Fixed Priority Scheduling for Self-Suspending Real-Time Tasks
Segment-Fixed Priority Scheduling for Self-Suspending Real-Time Tasks Junsung Kim, Björn Andersson, Dionisio de Niz, and Raj Rajkumar Carnegie Mellon University 2/31 Motion Planning on Self-driving Parallel
Doing Right By Massive Data: How To Bring Probability Modeling To The Analysis Of Huge Datasets Without Taking Over The Datacenter
Doing Right By Massive Data: How To Bring Probability Modeling To The Analysis Of Huge Datasets Without Taking Over The Datacenter Alexander W Blocker Pavlos Protopapas Xiao-Li Meng 9 February, 2010 Outline
REINFORCEMENT LEARNING
REINFORCEMENT LEARNING Larry Page: Where s Google going next? DeepMind's DQN playing Breakout Contents Introduction to Reinforcement Learning Deep Q-Learning INTRODUCTION TO REINFORCEMENT LEARNING Contents
Multi-Approximate-Keyword Routing Query
Bin Yao 1, Mingwang Tang 2, Feifei Li 2 1 Department of Computer Science and Engineering Shanghai Jiao Tong University, P. R. China 2 School of Computing University of Utah, USA Outline 1 Introduction
ArcGIS GeoAnalytics Server: An Introduction. Sarah Ambrose and Ravi Narayanan
ArcGIS GeoAnalytics Server: An Introduction Sarah Ambrose and Ravi Narayanan Overview Introduction Demos Analysis Concepts using GeoAnalytics Server GeoAnalytics Data Sources GeoAnalytics Server Administration
Parameter estimation using simulated annealing for S- system models of biochemical networks. Orland Gonzalez
Parameter estimation using simulated annealing for S- system models of biochemical networks Orland Gonzalez Outline S-systems quick review Definition of the problem Simulated annealing Perturbation function
Large Scale Semi-supervised Linear SVMs. University of Chicago
Large Scale Semi-supervised Linear SVMs Vikas Sindhwani and Sathiya Keerthi University of Chicago SIGIR 2006 Semi-supervised Learning (SSL) Motivation Setting Categorize x-billion documents into commercial/non-commercial.
Deep Learning. Convolutional Neural Networks Applications
Deep Learning Using a Convolutional Neural Network Dr. Ing. Morris Riedel Adjunct Associated Professor School of Engineering and Natural Sciences, University of Iceland Research Group Leader, Juelich Supercomputing
Basics of reinforcement learning
Basics of reinforcement learning Lucian Buşoniu TMLSS, 20 July 2018 Main idea of reinforcement learning (RL) Learn a sequential decision policy to optimize the cumulative performance of an unknown system
Local Search and Optimization
Local Search and Optimization Outline Local search techniques and optimization Hill-climbing Gradient methods Simulated annealing Genetic algorithms Issues with local search Local search and optimization
CAEFEM v9.5 Information
CAEFEM v9.5 Information Concurrent Analysis Corporation, 50 Via Ricardo, Thousand Oaks, CA 91320 USA Tel. (805) 375 1060, Fax (805) 375 1061 email: info@caefem.com or support@caefem.com Web: http://www.caefem.com
JOINT WMO TECHNICAL PROGRESS REPORT ON THE GLOBAL DATA PROCESSING AND FORECASTING SYSTEM AND NUMERICAL WEATHER PREDICTION RESEARCH ACTIVITIES FOR 2006
JOINT WMO TECHNICAL PROGRESS REPORT ON THE GLOBAL DATA PROCESSING AND FORECASTING SYSTEM AND NUMERICAL WEATHER PREDICTION RESEARCH ACTIVITIES FOR 2006 [TURKEY/Turkish State Meteorological Service] 1. Summary
Recurrent Latent Variable Networks for Session-Based Recommendation
Recurrent Latent Variable Networks for Session-Based Recommendation Panayiotis Christodoulou Cyprus University of Technology paa.christodoulou@edu.cut.ac.cy 27/8/2017 Panayiotis Christodoulou (C.U.T.)
Vibro-acoustic Analysis for Noise Reduction of Electric Machines
Vibro-acoustic Analysis for Noise Reduction of Electric Machines From Flux to OptiStruct Example : Synchronous Machine Flux 2D coupling to OptiStruct Patrick LOMBARD Application Team Manager Patrick.Lombard@cedrat.com
Analysis of Software Artifacts
Analysis of Software Artifacts System Performance I Shu-Ngai Yeung (with edits by Jeannette Wing) Department of Statistics Carnegie Mellon University Pittsburgh, PA 15213 2001 by Carnegie Mellon University
RAID+: Deterministic and Balanced Data Distribution for Large Disk Enclosures
RAID+: Deterministic and Balanced Data Distribution for Large Disk Enclosures Guangyan Zhang, Zican Huang, Xiaosong Ma SonglinYang, Zhufan Wang, Weimin Zheng Tsinghua University Qatar Computing Research
MSC HPC Infrastructure Update. Alain St-Denis Canadian Meteorological Centre Meteorological Service of Canada
MSC HPC Infrastructure Update Alain St-Denis Canadian Meteorological Centre Meteorological Service of Canada Outline HPC Infrastructure Overview Supercomputer Configuration Scientific Direction 2 IT Infrastructure
Energy-efficient Mapping of Big Data Workflows under Deadline Constraints
Energy-efficient Mapping of Big Data Workflows under Deadline Constraints Presenter: Tong Shu Authors: Tong Shu and Prof. Chase Q. Wu Big Data Center Department of Computer Science New Jersey Institute
Reducing Noisy-Neighbor Impact with a Fuzzy Affinity- Aware Scheduler
Reducing Noisy-Neighbor Impact with a Fuzzy Affinity- Aware Scheduler L U I S T O M Á S A N D J O H A N T O R D S S O N D E PA R T M E N T O F C O M P U T I N G S C I E N C E U M E Å U N I V E R S I T
Zebo Peng Embedded Systems Laboratory IDA, Linköping University
TDTS 01 Lecture 8 Optimization Heuristics for Synthesis Zebo Peng Embedded Systems Laboratory IDA, Linköping University Lecture 8 Optimization problems Heuristic techniques Simulated annealing Genetic
Stochastic Modelling of Electron Transport on different HPC architectures
Stochastic Modelling of Electron Transport on different HPC architectures www.hp-see.eu E. Atanassov, T. Gurov, A. Karaivan ova Institute of Information and Communication Technologies Bulgarian Academy
Event-based formulations for the RCPSP with production and consumption of resources
Event-based formulations for the RCPSP with production and consumption of resources Oumar KONE 1, Christian ARTIGUES 1, Pierre LOPEZ 1, and Marcel MONGEAU 2 May 2010 1: CNRS ; LAAS ; 7 avenue du colonel
A Spatial Data Infrastructure for Landslides and Floods in Italy
V Convegno Nazionale del Gruppo GIT Grottaminarda 14 16 giugno 2010 A Spatial Data Infrastructure for Landslides and Floods in Italy Ivan Marchesini, Vinicio Balducci, Gabriele Tonelli, Mauro Rossi, Fausto
arxiv: v2 [cs.ne] 22 Feb 2013
![arxiv: v2 [cs.ne] 22 Feb 2013](/thumbs/76/74046790.jpg "arxiv: v2 [cs.ne] 22 Feb 2013")
Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint arxiv:1301.3533v2 [cs.ne] 22 Feb 2013 Xanadu C. Halkias DYNI, LSIS, Universitè du Sud, Avenue de l Université - BP20132, 83957 LA
SOLVING SPARSE REPRESENTATIONS FOR OBJECT CLASSIFICATION USING QUANTUM D-WAVE 2X MACHINE
SOLVING SPARSE REPRESENTATIONS FOR OBJECT CLASSIFICATION USING QUANTUM D-WAVE 2X MACHINE! Nga Nguyen, Amy Larson, Carleton Coffrin, John Perry, Gary Salazar, and Garrett Kenyon Los Alamos National Laboratory
Location Prediction of Moving Target
Location of Moving Target Department of Radiation Oncology Stanford University AAPM 2009 Outline of Topics 1 Outline of Topics 1 2 Outline of Topics 1 2 3 Estimation of Stochastic Process Stochastic Regression:
Enumeration and generation of all constitutional alkane isomers of methane to icosane using recursive generation and a modified Morgan s algorithm
The C n H 2n+2 challenge Zürich, November 21st 2015 Enumeration and generation of all constitutional alkane isomers of methane to icosane using recursive generation and a modified Morgan s algorithm Andreas
Tennis player segmentation for semantic behavior analysis
Proposta di Tennis player segmentation for semantic behavior analysis Architettura Software per Robot Mobili Vito Renò, Nicola Mosca, Massimiliano Nitti, Tiziana D Orazio, Donato Campagnoli, Andrea Prati,
Sparse solver 64 bit and out-of-core addition
Sparse solver 64 bit and out-of-core addition Prepared By: Richard Link Brian Yuen Martec Limited 1888 Brunswick Street, Suite 400 Halifax, Nova Scotia B3J 3J8 PWGSC Contract Number: W7707-145679 Contract
Performance, Power & Energy. ELEC8106/ELEC6102 Spring 2010 Hayden Kwok-Hay So
Performance, Power & Energy ELEC8106/ELEC6102 Spring 2010 Hayden Kwok-Hay So Recall: Goal of this class Performance Reconfiguration Power/ Energy H. So, Sp10 Lecture 3 - ELEC8106/6102 2 PERFORMANCE EVALUATION
Exam Machine Learning for the Quantified Self with answers :00-14:45
Exam Machine Learning for the Quantified Self with answers 21. 06. 2017 12:00-14:45 NOTES: 1. YOUR NAME MUST BE WRITTEN ON EACH SHEET IN CAPITALS. 2. Answer the questions in Dutch or English. 3. Points
Member Level Forecasting Using SAS Enterprise Guide and SAS Forecast Studio
Paper 2240 Member Level Forecasting Using SAS Enterprise Guide and SAS Forecast Studio Prudhvidhar R Perati, Leigh McCormack, BlueCross BlueShield of Tennessee, Inc., an Independent Company of BlueCross
I/O Devices. Device. Lecture Notes Week 8
I/O Devices CPU PC ALU System bus Memory bus Bus interface I/O bridge Main memory USB Graphics adapter I/O bus Disk other devices such as network adapters Mouse Keyboard Disk hello executable stored on
REPORT, RE0813, MIL, ENV, 810G, TEMP, IN-HOUSE, , PASS
REPORT, RE0813, MIL, ENV, 810G, TEMP, IN-HOUSE, REV. A C r y s t a l G r o u p I n c 8 5 0 K a c e n a R d., H i a w a t h a, I A P h o n e : 8 7 7-2 7 9-7 8 6 3 F a x : 3 1 9-3 9 3-2 3 3 8 8 / 1 5 / 2
Optimization Techniques for Virtual Machine Placement and Migration
http://www.iaras.org/iaras/ournals/imcm Optimization Techniques for Virtual Machine Placement and Migration SATORU OHTA Department of Electrical and Computer Engineering, Toyama Prefectural University,
JOINT WMO TECHNICAL PROGRESS REPORT ON THE GLOBAL DATA PROCESSING AND FORECASTING SYSTEM AND NUMERICAL WEATHER PREDICTION RESEARCH ACTIVITIES FOR 2007
JOINT WMO TECHNICAL PROGRESS REPORT ON THE GLOBAL DATA PROCESSING AND FORECASTING SYSTEM AND NUMERICAL WEATHER PREDICTION RESEARCH ACTIVITIES FOR 2007 [TURKEY/Turkish State Meteorological Service] 1. Summary
The Green Index (TGI): A Metric for Evalua:ng Energy Efficiency in HPC Systems
The Green Index (TGI): A Metric for Evalua:ng Energy Efficiency in HPC Systems Wu Feng and Balaji Subramaniam Metrics for Energy Efficiency Energy- Delay Product (EDP) Used primarily in circuit design
Optimization strategy for MASNUM surface wave model
Hillsboro, September 27, 2018 Optimization strategy for MASNUM surface wave model Zhenya Song *, + * First Institute of Oceanography (FIO), State Oceanic Administrative (SOA), China + Intel Parallel Computing
The treatment of uncertainty in uniform workload distribution problems
The treatment of uncertainty in uniform workload distribution problems tefan PE KO, Roman HAJTMANEK University of šilina, Slovakia 34 th International Conference Mathematical Methods in Economics Liberec,
Indicator Constraints in Mixed-Integer Programming
Indicator Constraints in Mixed-Integer Programming Andrea Lodi University of Bologna, Italy - andrea.lodi@unibo.it Amaya Nogales-Gómez, Universidad de Sevilla, Spain Pietro Belotti, FICO, UK Matteo Fischetti,
Introduction to ArcGIS GeoAnalytics Server. Sarah Ambrose & Noah Slocum
Introduction to ArcGIS GeoAnalytics Server Sarah Ambrose & Noah Slocum Agenda Overview Analysis Capabilities + Demo Deployment and Configuration Questions ArcGIS GeoAnalytics Server uses the power of distributed
Generating Urban Mobility Data Sets Using Scalable GANs
Generating Urban Mobility Data Sets Using Scalable GANs Abhinav Jauhri & John Paul Shen ECE Department Carnegie Mellon University {ajauhri, jpshen}@cmu.edu Objective Generate city-scale human mobility
Energy-Efficient Real-Time Task Scheduling in Multiprocessor DVS Systems
Energy-Efficient Real-Time Task Scheduling in Multiprocessor DVS Systems Jian-Jia Chen *, Chuan Yue Yang, Tei-Wei Kuo, and Chi-Sheng Shih Embedded Systems and Wireless Networking Lab. Department of Computer
Application of Optimization Methods and Edge AI
and Hai-Liang Zhao hliangzhao97@gmail.com November 17, 2018 This slide can be downloaded at Link. Outline 1 How to Design Novel Models with Methods Embedded Naturally? Outline 1 How to Design Novel Models
Development of a Deep Recurrent Neural Network Controller for Flight Applications
Development of a Deep Recurrent Neural Network Controller for Flight Applications American Control Conference (ACC) May 26, 2017 Scott A. Nivison Pramod P. Khargonekar Department of Electrical and Computer
Geog 469 GIS Workshop. Managing Enterprise GIS Geodatabases
Geog 469 GIS Workshop Managing Enterprise GIS Geodatabases Outline 1. Why is a geodatabase important for GIS? 2. What is the architecture of a geodatabase? 3. How can we compare and contrast three types
Artificial Intelligence for HVAC Systems
Artificial Intelligence for HVAC Systems V. Ziebart, ACSS IAMP ZHAW Motivation More than 80% of the energy consumed in Swiss households is used for room heating & cooling and domestic hot water.* The increasing
Parallel PIPS-SBB Multi-level parallelism for 2-stage SMIPS. Lluís-Miquel Munguia, Geoffrey M. Oxberry, Deepak Rajan, Yuji Shinano
Parallel PIPS-SBB Multi-level parallelism for 2-stage SMIPS Lluís-Miquel Munguia, Geoffrey M. Oxberry, Deepak Rajan, Yuji Shinano ... Our contribution PIPS-PSBB*: Multi-level parallelism for Stochastic
Deduction and Induction. Stephan Schulz. The Inference Engine. Machine Learning. A Match Made in Heaven
Deduction and Induction A Match Made in Heaven Stephan Schulz The Inference Engine Machine Learning Deduction and Induction A Match Made in Heaven Stephan Schulz or a Deal with the Devil? The Inference
An Experimental Evaluation of Passage-Based Process Discovery
An Experimental Evaluation of Passage-Based Process Discovery H.M.W. Verbeek and W.M.P. van der Aalst Technische Universiteit Eindhoven Department of Mathematics and Computer Science P.O. Box 513, 5600
UTPlaceF 3.0: A Parallelization Framework for Modern FPGA Global Placement
UTPlaceF 3.0: A Parallelization Framework for Modern FPGA Global Placement Wuxi Li, Meng Li, Jiajun Wang, and David Z. Pan University of Texas at Austin wuxili@utexas.edu November 14, 2017 UT DA Wuxi Li
Continuous heat flow analysis. Time-variant heat sources. Embedded Systems Laboratory (ESL) Institute of EE, Faculty of Engineering
Thermal Modeling, Analysis and Management of 2D Multi-Processor System-on-Chip Prof David Atienza Alonso Embedded Systems Laboratory (ESL) Institute of EE, Falty of Engineering Outline MPSoC thermal modeling
FPGA Implementation of a Predictive Controller
FPGA Implementation of a Predictive Controller SIAM Conference on Optimization 2011, Darmstadt, Germany Minisymposium on embedded optimization Juan L. Jerez, George A. Constantinides and Eric C. Kerrigan
URP 4273 Section 3058 Survey Of Planning Information Systems (3 Credits) Spring 2017
URP 4273 Section 3058 Survey Of Planning Information Systems (3 Credits) Spring 2017 Instructor: Office Periods: Stanley Latimer 466 ARCH Phone: 352 294-1493 e-mail: latimer@geoplan.ufl.edu Monday Thursday,
Extreme Sparse Multinomial Logistic Regression: A Fast and Robust Framework for Hyperspectral Image Classification
Article Extreme Sparse Multinomial Logistic Regression: A Fast and Robust Framework for Hyperspectral Image Classification Faxian Cao 1, Zhijing Yang 1, *, Jinchang Ren 2, Wing-Kuen Ling 1, Huimin Zhao
Classification of Hand-Written Digits Using Scattering Convolutional Network
Mid-year Progress Report Classification of Hand-Written Digits Using Scattering Convolutional Network Dongmian Zou Advisor: Professor Radu Balan Co-Advisor: Dr. Maneesh Singh (SRI) Background Overview
URP 4273 Section 8233 Introduction to Planning Information Systems (3 Credits) Fall 2017
URP 4273 Section 8233 Introduction to Planning Information Systems (3 Credits) Fall 2017 Instructor: Office Periods: Stanley Latimer 466 ARCH Phone: 352 294-1493 e-mail: latimer@geoplan.ufl.edu Monday
Table S1 shows the SOM parameters that were used in the main manuscript. These correspond to the default set of options of the SOM Toolbox.
Supplemental file 1: SOM training parameters and sensitivity analysis. Table S1 shows the SOM parameters that were used in the main manuscript. These correspond to the default set of options of the SOM
Faster Machine Learning via Low-Precision Communication & Computation. Dan Alistarh (IST Austria & ETH Zurich), Hantian Zhang (ETH Zurich)
Faster Machine Learning via Low-Precision Communication & Computation Dan Alistarh (IST Austria & ETH Zurich), Hantian Zhang (ETH Zurich) 2 How many bits do you need to represent a single number in machine
Justin Appleby CS 229 Machine Learning Project Report 12/15/17 Kevin Chalhoub Building Electricity Load Forecasting
Justin Appleby CS 229 Machine Learning Project Report 12/15/17 Kevin Chalhoub Building Electricity Load Forecasting with ARIMA and Sequential Linear Regression Abstract Load forecasting is an essential
LC Commissioning, Operations and Availability
International Technology Recommendation Panel X-Band Linear Collider Path to the Future LC Commissioning, Operations and Availability Tom Himel Stanford Linear Accelerator Center April 26-27, 2004 Integrating
CS162 Operating Systems and Systems Programming Lecture 17. Performance Storage Devices, Queueing Theory
CS162 Operating Systems and Systems Programming Lecture 17 Performance Storage Devices, Queueing Theory October 25, 2017 Prof. Ion Stoica http://cs162.eecs.berkeley.edu Review: Basic Performance Concepts
Reducing The Computational Cost of Bayesian Indoor Positioning Systems
Reducing The Computational Cost of Bayesian Indoor Positioning Systems Konstantinos Kleisouris, Richard P. Martin Computer Science Department Rutgers University WINLAB Research Review May 15 th, 2006 Motivation
A Quantum Chemistry Domain-Specific Language for Heterogeneous Clusters
A Quantum Chemistry Domain-Specific Language for Heterogeneous Clusters ANTONINO TUMEO, ORESTE VILLA Collaborators: Karol Kowalski, Sriram Krishnamoorthy, Wenjing Ma, Simone Secchi May 15, 2012 1 Outline!
Developing IoT-Based Factory Automation Using F-RAM
Developing IoT-Based Factory Automation Using F-RAM Douglas Mitchell Flash Memory Summit 2018 Santa Clara, CA 1 INDUSTRIAL SYSTEMS TODAY Industry 4.0 The smart factory, Cyber-physical systems monitor factory
Regression, Curve Fitting and Optimisation
Supervised by Elena Zanini STOR-i, University of Lancaster 4 September 2015 1 Introduction Root Finding 2 3 Simulated Annealing 4 5 The Rosenbrock Banana Function 6 7 Given a set of data, what is the optimum
Online Algorithms for a Generalized Parallel Machine Scheduling Problem
PARALLEL MACHINE SCHEDULING PROBLEM 1 Online Algorithms for a Generalized Parallel Machine Scheduling Problem István SZALKAI 1,2, György DÓSA 1,3 1 Department of Mathematics, University of Pannonia, Veszprém,
先進的計算基盤システムシンポジウム SACSIS2012 Symposium on Advanced Computing Systems and Infrastructures SACSIS /5/18 VM 1 VM VM 0-1 1) 18% 50% 2) 3) Matching
VM 1 VM VM 0-1 1) 18% 50% 2) 3) Matching-based Virtual Machine Packing Algorithm for Lower Power Consumption SATOSHI TAKAHASHI, ATSUKO TAKEFUSA, MAIKO SHIGENO, HIDEMOTO NAKADA, TOMOHIRO KUDOH and AKIKO
Saving Power in the Data Center. Marios Papaefthymiou
Saving Power in the Data Center Marios Papaefthymiou Energy Use in Data Centers Several MW for large facilities Responsible for sizable fraction of US electricity consumption 2006: 1.5% or 61 billion kwh
What s Cooking? Predicting Cuisines from Recipe Ingredients
What s Cooking? Predicting Cuisines from Recipe Ingredients Kevin K. Do Department of Computer Science Duke University Durham, NC 27708 kevin.kydat.do@gmail.com Abstract Kaggle is an online platform for
Jake Diebolt, GIS Technician/Coordinator
Jake Diebolt, GIS Technician/Coordinator Population: 13 090 (2006 Census) 7 Organized Municipalities, two Incorporated Towns and two unorganized Townships. Population doubles during the summer tourist