Performance models for master/slave parallel programs

Size: px
Start display at page:

Download "Performance models for master/slave parallel programs"

Transcription

1 Performance models for master/slave parallel programs PASM 2004 Lucas Baldo Leonardo Brenner Luiz G. Fernandes Paulo Fernandes Afonso Sales {lbaldo, lbrenner, gustavo, paulof, PUCRS - FACIN - PPGCC Porto Alegre, , Brazil HP Brazil - CNPq/Brazil PASM 2004: Performance models for master/slave parallel programs p.1/22

2 Motivation Foresee implementation behavior in early stages of the parallel program development Identification of performance indices Provide an easy-to-use (or at least feasible) stochastic modeling Any modeling technique could be used (PEPA, SPN, etc) Very dependable on programmer previous experience PASM 2004: Performance models for master/slave parallel programs p.2/22

3 Outline Stochastic Automata Networks (SAN) Case Study: Propagation Algorithm Proposed models Asynchronous Synchronous Performance Indices Conclusions PASM 2004: Performance models for master/slave parallel programs p.3/22

4 SAN (Concept) Proposed by B. Plateau Each subsystem is represented by a stochastic automaton Local events change the local state of only one automaton Synchronizing events change the local state of several automata simultaneously Interaction among automata: Synchronizing events Functional rates and/or functional probabilities PASM 2004: Performance models for master/slave parallel programs p.4/22

5 SAN (Example) A (1) A (2) 0 (1) e 1 e 2 1 (1) 0 (2) e 4 e 2 (π 1 ) e 2 (π 2 ) e 5 2 (2) 1 (2) Type Event Rate loc e 1 f syn e 2 µ loc e 3 σ loc e 4 δ loc e 5 τ f = [( st A (2) == 0 (2)) λ ] + [( st A (2) == 2 (2)) γ ] e 3 Reachability =! [( st A (1) == 1 (1)) && ( st A (2) == 1 (2))] PASM 2004: Performance models for master/slave parallel programs p.5/22

6 SAN (Markov) λ 0 (1) 0 (2) 1 (1) 0 (2) δ τ µπ 2 µπ 1 δ τ 0 (1) 2 (2) σ 0 (1) 1 (2) 1 (1) 2 (2) σ 1 (1) 1 (2) γ PASM 2004: Performance models for master/slave parallel programs p.6/22

7 Master/Slave Paradigm One master node generates and allocates work (tasks) to N slave nodes Each slave node receives tasks, computes them, and sends the result back Master node receives and analyses the computed results by slaves (summation) PASM 2004: Performance models for master/slave parallel programs p.7/22

8 Master/Slave Generic SAN Model Master T x c s r 1..r N Rx up IT x up begins the program Slave (i) s c p i r i execution ends the parallel application (returns to the initial situation) sends tasks to the slaves performs the summation finishes a task by the i th slave sends the result from i th slave back to the master r i (π) I P r T x up s p i r i (1 π) PASM 2004: Performance models for master/slave parallel programs p.8/22

9 Case Study Image interpolation application method to create smooth and realistic virtual views starts with two source images Propagation Algorithm input: seed pairs based on a region growing technique goal: match the largest possible region PASM 2004: Performance models for master/slave parallel programs p.9/22

10 Asynchronous Implementation One master, one buffer and N slaves The results sent by the slaves are stored on a repository (buffer) Master does not need to synchronize the reception of results PASM 2004: Performance models for master/slave parallel programs p.10/22

11 SAN Model of Asynchronous Implementation Master Buffer Slave (1) Slave (i) Slave (N) T x K I I I c(g2) c(g1) Rx s 1..s N r 1..r N r 1..r N K 1 c c P r up s 1 r 1 (1 π) up s i... P r r i (1 π)... P r up s N r N (1 π) up r 1..r N. c r 1 (π) p 1 r i (π) p i r N (π) p N IT x 0 T x T x T x g 1 = (nb [Slave (1)..Slave (N) ] I == 0) g 2 = (nb [Slave (1)..Slave (N) ] I > 0) Type Event Rate Type Event Rate Type Event Rate syn up λ syn µ syn c σ syn s i δ syn r i α loc p i γ PASM 2004: Performance models for master/slave parallel programs p.11/22

12 Parameterization for Asynchronous Implementation Input values: BL slaves buffer length (5.500 points) P S percentage of slices extension over its neighborhoods (0.5) N P number of points ( ) N S RN P N B number of slices (granularity) [ ] real number of points 2 (1 + P S) + (NS 2) (1 + 2P S) number of buffer for each slice RNP BL NS NP NS PASM 2004: Performance models for master/slave parallel programs p.12/22

13 Parameterization for Asynchronous Implementation Rates: s i δ = 1 T T up λ = 1 T T N T T send a new task N number of slaves µ = 1000 (insignificant time) p i γ = 1 CM c σ = 1 ER r i α = 1 T P CM compute final matches ER evaluate results T P send a result pack PASM 2004: Performance models for master/slave parallel programs p.13/22

14 Synchronous Implementation One master and N slaves The results are sending to the master Master waits all slaves have finished their tasks before it starts a new task distribution PASM 2004: Performance models for master/slave parallel programs p.14/22

15 SAN Model of Synchronous Implementation Master Slave (1) Slave (i) Slave (N) T x I I I c s up s up s up s r 1..r N Rx P r r 1 (1 π)... P r r i (1 π)... P r r N (1 π) up r 1 (π) p 1 r i (π) p i r N (π) p N IT x T x T x T x f c = (nb [Slave (1)..Slave (N) ] I == N) σ Type Event Rate Type Event Rate Type Event Rate syn up λ syn µ loc c f c syn s δ syn r i α loc p i γ PASM 2004: Performance models for master/slave parallel programs p.15/22

16 Parameterization for Synchronous Implementation Rates: s δ = 1 T T up λ = 1 T T N T T send a new task N number of slaves µ = 1000 (insignificant time) p i γ = 1 CM c σ = 1 ER r i α = 1 T P N CM compute final matches ER evaluate results T P send a result pack PASM 2004: Performance models for master/slave parallel programs p.16/22

17 Performance Indices Stationary analysis Local state probabilities: Slave in T x state master unavailable Master in Rx state master unavailable Master in T x state slaves unavailable PASM 2004: Performance models for master/slave parallel programs p.17/22

18 Synchronous vs. Asynchronous Synchronous Model Asynchronous Model 0.4 coarse grain fine grain 0.4 coarse grain fine grain Probability of slave in Tx state Probability of slave in Tx state number of slave nodes number of slave nodes N Coarse Fine Synchronous N Coarse Fine Asynchronous PASM 2004: Performance models for master/slave parallel programs p.18/22

19 Buffer Size Asynchronous Implementation with coarse grain Asynchronous Implementation with fine grain slaves 3 slaves 4 slaves 5 slaves 6 slaves 7 slaves slaves 3 slaves 4 slaves 5 slaves 6 slaves 7 slaves Probability of slave in Tx state Probability of slave in Tx state Buffer size Buffer size Buffer size 2 Slaves 3 Slaves 4 Slaves 5 Slaves 6 Slaves 7 Slaves Asynchronous (Coarse) Buffer size 2 Slaves 3 Slaves 4 Slaves 5 Slaves 6 Slaves 7 Slaves Asynchronous (Fine) PASM 2004: Performance models for master/slave parallel programs p.19/22

20 Granularity Asynchronous Model Asynchronous Model 0.98 coarse grain fine grain 0.14 coarse grain fine grain Probability of master in Rx state Probability of master in Tx state number of slave nodes number of slave nodes N Coarse Fine Receiving N Coarse Fine Transmitting PASM 2004: Performance models for master/slave parallel programs p.20/22

21 Conclusions We can: Choose the paradigm Choose the buffer size Choose the granularity Find possible bottlenecks before implementation We cannot (future work): Estimate execution time Transient analysis Provide an easy-to-use stochastic modeling Automatic modeling tool PASM 2004: Performance models for master/slave parallel programs p.21/22

22 Contact Performance Evaluation Group PASM 2004: Performance models for master/slave parallel programs p.22/22

Performance models for master/slave parallel programs 1

Performance models for master/slave parallel programs 1 PASM 2004 Preliminary Version Performance models for master/slave parallel programs 1 Lucas Baldo 2 Leonardo Brenner 3 Luiz Gustavo Fernandes 4 Paulo Fernandes 5 Afonso Sales 6 Faculdade de Informática

More information

Representing Software Usage Models with Stochastic Automata Networks

Representing Software Usage Models with Stochastic Automata Networks epresenting Software Usage Models with Stochastic Automata Networks André G. Farina CPTS-PPGCC-PUCS FACIN Av. Ipiranga, 6681 Porto Alegre BAZIL farina@cpts.pucrs.br Paulo Fernandes PPGCC-PUCS FACIN Av.

More information

Auto-Negotiation for 10GBASE-T

Auto-Negotiation for 10GBASE-T Auto-Negotiation for 10GBASE-T Eric Lynskey Outline Overview of 10/100 auto-negotiation Auto-negotiation additions for 1000BASE-T Possible extensions for 10GBASE-T Motivation for auto-negotiation Handshaking

More information

Uniprocessor Scheduling. Basic Concepts Scheduling Criteria Scheduling Algorithms. Three level scheduling

Uniprocessor Scheduling. Basic Concepts Scheduling Criteria Scheduling Algorithms. Three level scheduling Uniprocessor Scheduling Basic Concepts Scheduling Criteria Scheduling Algorithms Three level scheduling 2 1 Types of Scheduling 3 Long- and Medium-Term Schedulers Long-term scheduler Determines which programs

More information

High Performance Computing

High Performance Computing Master Degree Program in Computer Science and Networking, 2014-15 High Performance Computing 1 st appello January 13, 2015 Write your name, surname, student identification number (numero di matricola),

More information

Agenda. 1 Business Processes. 2 Modeling Techniques Used in this Work. 3 Automated Conversion from Business Process Models to SAN

Agenda. 1 Business Processes. 2 Modeling Techniques Used in this Work. 3 Automated Conversion from Business Process Models to SAN IME - USP Modeling Techniques for Business Process Performance Analysis Kelly Rosa Braghetto Advisor: João Eduardo Ferreira 1 Co-advisor: Jean-Marc Vincent 2 1 Dept. de Ciência da Computação 2 Laboratoire

More information

Parametric Real Time System Feasibility Analysis Using Parametric Timed Automata

Parametric Real Time System Feasibility Analysis Using Parametric Timed Automata Parametric Real Time System Feasibility Analysis Using Parametric Timed Automata PhD Dissertation Yusi Ramadian Advisor : Luigi Palopoli Co advisor : Alessandro Cimatti 1 Real Time System Applications

More information

Parallel Hybrid Monte Carlo Algorithms for Matrix Computations

Parallel Hybrid Monte Carlo Algorithms for Matrix Computations Parallel Hybrid Monte Carlo Algorithms for Matrix Computations V. Alexandrov 1, E. Atanassov 2, I. Dimov 2, S.Branford 1, A. Thandavan 1 and C. Weihrauch 1 1 Department of Computer Science, University

More information

A new caching policy for cloud assisted Peer-to-Peer video on-demand services

A new caching policy for cloud assisted Peer-to-Peer video on-demand services A new caching policy for cloud assisted Peer-to-Peer video on-demand services Franco Robledo, Pablo Rodríguez-Bocca, Pablo Romero and Claudia Rostagnol Facultad de Ingeniería, Universidad de la República.

More information

Resource Constrained Project Scheduling. Reservations and Timetabling

Resource Constrained Project Scheduling. Reservations and Timetabling DM87 SCHEDULING, TIMETABLING AND ROUTING Lecture 13 Resource Constrained Project Scheduling. Reservations and Timetabling Marco Chiarandini DM87 Scheduling, Timetabling and Routing 2 Preprocessing: Temporal

More information

intro, applications MRF, labeling... how it can be computed at all? Applications in segmentation: GraphCut, GrabCut, demos

intro, applications MRF, labeling... how it can be computed at all? Applications in segmentation: GraphCut, GrabCut, demos Image as Markov Random Field and Applications 1 Tomáš Svoboda, svoboda@cmp.felk.cvut.cz Czech Technical University in Prague, Center for Machine Perception http://cmp.felk.cvut.cz Talk Outline Last update:

More information

CELLULAR automata (CA) are mathematical models for

CELLULAR automata (CA) are mathematical models for 1 Cellular Learning Automata with Multiple Learning Automata in Each Cell and its Applications Hamid Beigy and M R Meybodi Abstract The cellular learning automata, which is a combination of cellular automata

More information

Partitioned Fixed-Priority Scheduling of Parallel Tasks Without Preemptions

Partitioned Fixed-Priority Scheduling of Parallel Tasks Without Preemptions Partitioned Fixed-Priority Scheduling of Parallel Tasks Without Preemptions *, Alessandro Biondi *, Geoffrey Nelissen, and Giorgio Buttazzo * * ReTiS Lab, Scuola Superiore Sant Anna, Pisa, Italy CISTER,

More information

Performance Testing from UML Models with Resource Descriptions *

Performance Testing from UML Models with Resource Descriptions * Performance Testing from UML Models with Resource Descriptions * Flávio M. de Oliveira 1, Rômulo da S. Menna 1, Hugo V. Vieira 1, Duncan D.A. Ruiz 1 1 Faculdade de Informática Pontifícia Universidade Católica

More information

Tracing Lineage Beyond Relational Operators

Tracing Lineage Beyond Relational Operators Tracing Lineage Beyond Relational Operators Mingwu Zhang 1 Xiangyu Zhang 1 Xiang Zhang 2 Sunil Prabhakar 1 1 Computer Science 2 Bindley Bioscience Center Purdue University Introduction Lineage (Data Provenance)

More information

High Performance Printing: Increasing Personalized Documents Rendering through PPML Jobs Profiling and Scheduling

High Performance Printing: Increasing Personalized Documents Rendering through PPML Jobs Profiling and Scheduling 2009 International Conference on Computational Science and Engineering High Performance Printing: Increasing Personalized Documents Rendering through PPML Jobs Profiling and Scheduling Thiago Nunes, Mateus

More information

The Finite Element Method for the Analysis of Non-Linear and Dynamic Systems. Prof. Dr. Eleni Chatzi, J.P. Escallo n Lecture December, 2013

The Finite Element Method for the Analysis of Non-Linear and Dynamic Systems. Prof. Dr. Eleni Chatzi, J.P. Escallo n Lecture December, 2013 The Finite Element Method for the Analysis of Non-Linear and Dynamic Systems Prof. Dr. Eleni Chatzi, J.P. Escallo n Lecture 11-17 December, 2013 Institute of Structural Engineering Method of Finite Elements

More information

15-744: Computer Networking. Overview. Queuing Disciplines. TCP & Routers. L-6 TCP & Routers

15-744: Computer Networking. Overview. Queuing Disciplines. TCP & Routers. L-6 TCP & Routers TCP & Routers 15-744: Computer Networking RED XCP Assigned reading [FJ93] Random Early Detection Gateways for Congestion Avoidance [KHR02] Congestion Control for High Bandwidth-Delay Product Networks L-6

More information

Design and Implementation of Measurement-Based Resource Allocation Schemes Using the Realtime Traffic Flow Measurement Architecture

Design and Implementation of Measurement-Based Resource Allocation Schemes Using the Realtime Traffic Flow Measurement Architecture Design and Implementation of Measurement-Based Resource Allocation Schemes Using the Realtime Traffic Flow Measurement Architecture Robert D. Callaway, Michael Devetsikiotis, and Chao Kan Department of

More information

Some aspects of parallel program design. R. Bader (LRZ) G. Hager (RRZE)

Some aspects of parallel program design. R. Bader (LRZ) G. Hager (RRZE) Some aspects of parallel program design R. Bader (LRZ) G. Hager (RRZE) Finding exploitable concurrency Problem analysis 1. Decompose into subproblems perhaps even hierarchy of subproblems that can simultaneously

More information

Workloads Programmierung Paralleler und Verteilter Systeme (PPV)

Workloads Programmierung Paralleler und Verteilter Systeme (PPV) Workloads Programmierung Paralleler und Verteilter Systeme (PPV) Sommer 2015 Frank Feinbube, M.Sc., Felix Eberhardt, M.Sc., Prof. Dr. Andreas Polze Workloads 2 Hardware / software execution environment

More information

A Novel Design Framework for the Design of Reconfigurable Systems based on NoCs

A Novel Design Framework for the Design of Reconfigurable Systems based on NoCs Politecnico di Milano & EPFL A Novel Design Framework for the Design of Reconfigurable Systems based on NoCs Vincenzo Rana, Ivan Beretta, Donatella Sciuto Donatella Sciuto sciuto@elet.polimi.it Introduction

More information

Quantitative Synthesis for Concurrent Programs

Quantitative Synthesis for Concurrent Programs Quantitative Synthesis for Concurrent Programs Pavol Černý, Krishnendu Chatterjee, Thomas A. Henzinger, Arjun Radhakrishna, and Rohit Singh IST Austria IIT Bombay Abstract. We present an algorithmic method

More information

Analysis of Different Reference Plane Setups for the Calibration of a Mobile Laser Scanning System

Analysis of Different Reference Plane Setups for the Calibration of a Mobile Laser Scanning System Analysis of Different Reference Plane Setups for the Calibration of a Mobile Laser Scanning System 18. Internationaler Ingenieurvermessungskurs Graz, Austria, 25-29 th April 2017 Erik Heinz, Christian

More information

Vision 3D articielle Disparity maps, correlation

Vision 3D articielle Disparity maps, correlation Vision 3D articielle Disparity maps, correlation Pascal Monasse monasse@imagine.enpc.fr IMAGINE, École des Ponts ParisTech http://imagine.enpc.fr/~monasse/stereo/ Contents Triangulation Epipolar rectication

More information

CS 268: Computer Networking

CS 268: Computer Networking CS 268: Computer Networking L-6 Router Congestion Control TCP & Routers RED XCP Assigned reading [FJ93] Random Early Detection Gateways for Congestion Avoidance [KHR02] Congestion Control for High Bandwidth-Delay

More information

Remote and Partial Reconfiguration of FPGAs: Tools and Trends

Remote and Partial Reconfiguration of FPGAs: Tools and Trends Remote and Partial Reconfiguration of FPGAs: Tools and Trends Daniel Mesquita, Fernando Moraes, José palma, Leandro Moller, Ney Calazans Laboratoire de Informatique, de Robotique et de Microéletronique

More information

Multiprocessor Scheduling Using Parallel Genetic Algorithm

Multiprocessor Scheduling Using Parallel Genetic Algorithm www.ijcsi.org 260 Multiprocessor Scheduling Using Parallel Genetic Algorithm Nourah Al-Angari 1, Abdullatif ALAbdullatif 2 1,2 Computer Science Department, College of Computer & Information Sciences, King

More information

Model-based Analysis of Event-driven Distributed Real-time Embedded Systems

Model-based Analysis of Event-driven Distributed Real-time Embedded Systems Model-based Analysis of Event-driven Distributed Real-time Embedded Systems Gabor Madl Committee Chancellor s Professor Nikil Dutt (Chair) Professor Tony Givargis Professor Ian Harris University of California,

More information

UML2SAN: TOWARD A NEW SOFTWARE PERFORMANCE ENGINEERING APPROACH

UML2SAN: TOWARD A NEW SOFTWARE PERFORMANCE ENGINEERING APPROACH UML2SAN: TOWARD A NEW SOFTWARE PERFORMANCE ENGINEERING APPROACH Ihab Sbeity 1, Mohamed Dbouk 1 and Ilfat Ghamlouche 2 1 Faculty of Sciences, Lebanese University, Lebanon 2 Faculty of Economical Sciences

More information

Parallel Programming Patterns

Parallel Programming Patterns Parallel Programming Patterns Moreno Marzolla Dip. di Informatica Scienza e Ingegneria (DISI) Università di Bologna http://www.moreno.marzolla.name/ Copyright 2013, 2017, 2018 Moreno Marzolla, Università

More information

Design Improvement and Implementation of 3D Gauss-Markov Mobility Model

Design Improvement and Implementation of 3D Gauss-Markov Mobility Model Design Improvement and Implementation of 3D Gauss-Markov Mobility Model Mohammed J.F. Alenazi, Cenk Sahin, and James P.G. Sterbenz Department of Electrical Engineering & Computer Science Information &

More information

Optimal Routing and Scheduling in Multihop Wireless Renewable Energy Networks

Optimal Routing and Scheduling in Multihop Wireless Renewable Energy Networks Optimal Routing and Scheduling in Multihop Wireless Renewable Energy Networks ITA 11, San Diego CA, February 2011 MHR. Khouzani, Saswati Sarkar, Koushik Kar UPenn, UPenn, RPI March 23, 2011 Khouzani, Sarkar,

More information

Parallel Computing. Slides credit: M. Quinn book (chapter 3 slides), A Grama book (chapter 3 slides)

Parallel Computing. Slides credit: M. Quinn book (chapter 3 slides), A Grama book (chapter 3 slides) Parallel Computing 2012 Slides credit: M. Quinn book (chapter 3 slides), A Grama book (chapter 3 slides) Parallel Algorithm Design Outline Computational Model Design Methodology Partitioning Communication

More information

DDSS: Dynamic Dedicated Servers Scheduling for Multi Priority Level Classes in Cloud Computing

DDSS: Dynamic Dedicated Servers Scheduling for Multi Priority Level Classes in Cloud Computing DDSS: Dynamic Dedicated Servers Scheduling for Multi Priority Level Classes in Cloud Computing Husnu Saner Narman Md. Shohrab Hossain Mohammed Atiquzzaman School of Computer Science University of Oklahoma,

More information

OUTLINE. SPI Theory SPI Implementation STM32F0 SPI Resources System Overview Registers SPI Application Initialization Interface Examples

OUTLINE. SPI Theory SPI Implementation STM32F0 SPI Resources System Overview Registers SPI Application Initialization Interface Examples SERIAL PERIPHERAL INTERFACE (SPI) George E Hadley, Timothy Rogers, and David G Meyer 2018, Images Property of their Respective Owners OUTLINE SPI Theory SPI Implementation STM32F0 SPI Resources System

More information

ANT/OR. An Optimisation Model for Staff Planning in a Home Care Organisation

ANT/OR. An Optimisation Model for Staff Planning in a Home Care Organisation University of Antwerp Operations Research Group ANT/OR An Optimisation Model for Staff Planning in a Home Care Organisation P.A. Maya Duque 1,2 M. Castro 1 P. Goos 1 K. Sörensen 1 1 Faculty of Applied

More information

Machine-Learning-Based Flow scheduling in OTSSenabled

Machine-Learning-Based Flow scheduling in OTSSenabled Machine-Learning-Based Flow scheduling in OTSSenabled Datacenters Speaker: Lin Wang Research Advisor: Biswanath Mukherjee Motivation Traffic demand increasing in datacenter networks Cloud-service, parallel-computing,

More information

Programming, numerics and optimization

Programming, numerics and optimization Programming, numerics and optimization Lecture C-4: Constrained optimization Łukasz Jankowski ljank@ippt.pan.pl Institute of Fundamental Technological Research Room 4.32, Phone +22.8261281 ext. 428 June

More information

Analysis of High Speed TCP Variants and TCP over Cognitive Radio

Analysis of High Speed TCP Variants and TCP over Cognitive Radio Analysis of High Speed TCP Variants and TCP over Cognitive Radio Sudheer Poojary Advisor: Vinod Sharma IISc, Bangalore November 30, 2015 Sudheer Poojary Analysis of High Speed TCP Variants and TCP over

More information

Diffusing Your Mobile Apps: Extending In-Network Function Virtualisation to Mobile Function Offloading

Diffusing Your Mobile Apps: Extending In-Network Function Virtualisation to Mobile Function Offloading Diffusing Your Mobile Apps: Extending In-Network Function Virtualisation to Mobile Function Offloading Mario Almeida, Liang Wang*, Jeremy Blackburn, Konstantina Papagiannaki, Jon Crowcroft* Telefonica

More information

Performance and Usability Evaluation of a Pattern-Oriented Parallel Programming Interface for Multi-Core Architectures

Performance and Usability Evaluation of a Pattern-Oriented Parallel Programming Interface for Multi-Core Architectures Performance and Usability Evaluation of a Pattern-Oriented Parallel Programming Interface for Multi-Core Architectures Dalvan Griebler, Daniel Adornes, Luiz Gustavo Fernandes Pontifícia Universidade Católica

More information

Numerical schemes for Hamilton-Jacobi equations, control problems and games

Numerical schemes for Hamilton-Jacobi equations, control problems and games Numerical schemes for Hamilton-Jacobi equations, control problems and games M. Falcone H. Zidani SADCO Spring School Applied and Numerical Optimal Control April 23-27, 2012, Paris Lecture 2/3 M. Falcone

More information

Outline. Petri nets. Introduction Examples Properties Analysis techniques. 1 EE249Fall04

Outline. Petri nets. Introduction Examples Properties Analysis techniques. 1 EE249Fall04 Outline Petri nets Introduction Examples Properties Analysis techniques 1 Petri Nets (PNs) Model introduced by C.A. Petri in 1962 Ph.D. Thesis: Communication with Automata Applications: distributed computing,

More information

Parallel Programming Concepts. Parallel Algorithms. Peter Tröger

Parallel Programming Concepts. Parallel Algorithms. Peter Tröger Parallel Programming Concepts Parallel Algorithms Peter Tröger Sources: Ian Foster. Designing and Building Parallel Programs. Addison-Wesley. 1995. Mattson, Timothy G.; S, Beverly A.; ers,; Massingill,

More information

Hierarchical Composition and Abstraction In Architecture Models

Hierarchical Composition and Abstraction In Architecture Models Hierarchical Composition and Abstraction In Architecture Models Pam Binns and Steve Vestal Honeywell Labs {pam.binns, steve.vestal}@honeywell.com Supported by the Air Force Office of Scientific Research

More information

Ch. 7: Benchmarks and Performance Tests

Ch. 7: Benchmarks and Performance Tests Ch. 7: Benchmarks and Performance Tests Kenneth Mitchell School of Computing & Engineering, University of Missouri-Kansas City, Kansas City, MO 64110 Kenneth Mitchell, CS & EE dept., SCE, UMKC p. 1/3 Introduction

More information

Before creating a new cluster or joining a node to an existing cluster, keep in mind the following restrictions:

Before creating a new cluster or joining a node to an existing cluster, keep in mind the following restrictions: Cluster Management Before starting Before creating a new cluster or joining a node to an existing cluster, keep in mind the following restrictions: It's strongly suggested to configure all Application

More information

HPPA. Online Security Assessment in Brazil. J. Jardim June 2-5,2013. Software & Information Systems

HPPA. Online Security Assessment in Brazil. J. Jardim June 2-5,2013. Software & Information Systems Online Security Assessment in Brazil J. Jardim June 2-5,2013 Software & Information Systems Topics Brazilian System Overview DSA Characteristics Application and Cases New Developments Brazilian System

More information

LDA for Big Data - Outline

LDA for Big Data - Outline LDA FOR BIG DATA 1 LDA for Big Data - Outline Quick review of LDA model clustering words-in-context Parallel LDA ~= IPM Fast sampling tricks for LDA Sparsified sampler Alias table Fenwick trees LDA for

More information

Regularization and Markov Random Fields (MRF) CS 664 Spring 2008

Regularization and Markov Random Fields (MRF) CS 664 Spring 2008 Regularization and Markov Random Fields (MRF) CS 664 Spring 2008 Regularization in Low Level Vision Low level vision problems concerned with estimating some quantity at each pixel Visual motion (u(x,y),v(x,y))

More information

298 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 15, NO. 1, JANUARY 2016

298 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 15, NO. 1, JANUARY 2016 298 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 5, NO., JANUARY 206 Hybrid Network Coding for Unbalanced Slotted ALOHA Relay Networks Shijun Lin, Member, IEEE, Liqun Fu, Member, IEEE, Jianmin Xie,

More information

Segmentation of 3D Materials Image Data

Segmentation of 3D Materials Image Data Enhanced Image Modeling for EM/MPM Segmentation of 3D Materials Image Data Dae Woo Kim, Mary L. Comer School of Electrical and Computer Engineering Purdue University Presentation Outline - Motivation -

More information

Region-based Segmentation

Region-based Segmentation Region-based Segmentation Image Segmentation Group similar components (such as, pixels in an image, image frames in a video) to obtain a compact representation. Applications: Finding tumors, veins, etc.

More information

Impact of Clustering on Epidemics in Random Networks

Impact of Clustering on Epidemics in Random Networks Impact of Clustering on Epidemics in Random Networks Joint work with Marc Lelarge INRIA-ENS 8 March 2012 Coupechoux - Lelarge (INRIA-ENS) Epidemics in Random Networks 8 March 2012 1 / 19 Outline 1 Introduction

More information

Static Pruning of Terms In Inverted Files

Static Pruning of Terms In Inverted Files In Inverted Files Roi Blanco and Álvaro Barreiro IRLab University of A Corunna, Spain 29th European Conference on Information Retrieval, Rome, 2007 Motivation : to reduce inverted files size with lossy

More information

Database Architectures

Database Architectures B0B36DBS, BD6B36DBS: Database Systems h p://www.ksi.m.cuni.cz/~svoboda/courses/172-b0b36dbs/ Lecture 11 Database Architectures Authors: Tomáš Skopal, Irena Holubová Lecturer: Mar n Svoboda, mar n.svoboda@fel.cvut.cz

More information

Worst-case Ethernet Network Latency for Shaped Sources

Worst-case Ethernet Network Latency for Shaped Sources Worst-case Ethernet Network Latency for Shaped Sources Max Azarov, SMSC 7th October 2005 Contents For 802.3 ResE study group 1 Worst-case latency theorem 1 1.1 Assumptions.............................

More information

Performance Optimization and Analysis of Blade Designs under Delay Variability

Performance Optimization and Analysis of Blade Designs under Delay Variability Performance Optimization and Analysis of Blade Designs under Delay Variability Dylan Hand, Hsin-Ho Huang, Benmao Chang, Yang Zhang *, Matheus Trevisan Moreira, Melvin Breuer, Ney Laert Vilar Calazans,

More information

REDUCED ORDER MODELING IN MULTISPECTRAL PHOTOACOUSTIC TOMOGRAPHY

REDUCED ORDER MODELING IN MULTISPECTRAL PHOTOACOUSTIC TOMOGRAPHY REDUCED ORDER MODELING IN MULTISPECTRAL PHOTOACOUSTIC TOMOGRAPHY Arvind Saibaba Sarah Vallélian Statistical and Applied Mathematical Sciences Institute & North Carolina State University May 26, 2016 OUTLINE

More information

SRCMap: Energy Proportional Storage using Dynamic Consolidation

SRCMap: Energy Proportional Storage using Dynamic Consolidation SRCMap: Energy Proportional Storage using Dynamic Consolidation By: Akshat Verma, Ricardo Koller, Luis Useche, Raju Rangaswami Presented by: James Larkby-Lahet Motivation storage consumes 10-25% of datacenter

More information

Sensor Tasking and Control

Sensor Tasking and Control Sensor Tasking and Control Outline Task-Driven Sensing Roles of Sensor Nodes and Utilities Information-Based Sensor Tasking Joint Routing and Information Aggregation Summary Introduction To efficiently

More information

Module 20: Multi-core Computing Multi-processor Scheduling Lecture 39: Multi-processor Scheduling. The Lecture Contains: User Control.

Module 20: Multi-core Computing Multi-processor Scheduling Lecture 39: Multi-processor Scheduling. The Lecture Contains: User Control. The Lecture Contains: User Control Reliability Requirements of RT Multi-processor Scheduling Introduction Issues With Multi-processor Computations Granularity Fine Grain Parallelism Design Issues A Possible

More information

Random Walks and Universal Sequences

Random Walks and Universal Sequences Random Walks and Universal Sequences Xiaochen Qi February 28, 2013 Abstract A random walk is a chance process studied in probability, which plays an important role in probability theory and its applications.

More information

ECE519 Advanced Operating Systems

ECE519 Advanced Operating Systems IT 540 Operating Systems ECE519 Advanced Operating Systems Prof. Dr. Hasan Hüseyin BALIK (10 th Week) (Advanced) Operating Systems 10. Multiprocessor, Multicore and Real-Time Scheduling 10. Outline Multiprocessor

More information

Dependable and Secure Systems Dependability

Dependable and Secure Systems Dependability Dependable and Secure Systems Dependability Master of Science in Embedded Computing Systems Quantitative Dependability Analysis with Stochastic Activity Networks: the Möbius Tool Andrea Domenici DII, Università

More information

The Capacity of Wireless Networks

The Capacity of Wireless Networks The Capacity of Wireless Networks Piyush Gupta & P.R. Kumar Rahul Tandra --- EE228 Presentation Introduction We consider wireless networks without any centralized control. Try to analyze the capacity of

More information

Parameterization of triangular meshes

Parameterization of triangular meshes Parameterization of triangular meshes Michael S. Floater November 10, 2009 Triangular meshes are often used to represent surfaces, at least initially, one reason being that meshes are relatively easy to

More information

arxiv: v1 [cs.pl] 21 Apr 2011

arxiv: v1 [cs.pl] 21 Apr 2011 Quantitative Synthesis for Concurrent Programs, Pavol Černý1, Krishnendu Chatterjee 1, Thomas A. Henzinger 1, Arjun Radhakrishna 1, and Rohit Singh 2 arxiv:1104.4306v1 [cs.pl] 21 Apr 2011 1 IST Austria

More information

Formal Modeling and Analysis of Stream Processing Systems

Formal Modeling and Analysis of Stream Processing Systems Formal Modeling and Analysis of Stream Processing Systems Linh T.X. Phan March 2009 Computer and Information Science University of Pennsylvania 1 High performance Highly optimized requirements systems

More information

Towards a Wireless Lexicon. Philip Levis Computer Systems Lab Stanford University 20.viii.2007

Towards a Wireless Lexicon. Philip Levis Computer Systems Lab Stanford University 20.viii.2007 Towards a Wireless Lexicon Philip Levis Computer Systems Lab Stanford University 20.viii.2007 Low Power Wireless Low cost, numerous devices Wireless sensornets Personal area networks (PANs) Ad-hoc networks

More information

Automated Formal Methods for Embedded Systems

Automated Formal Methods for Embedded Systems Automated Formal Methods for Embedded Systems Bernd Finkbeiner Universität des Saarlandes Reactive Systems Group 2011/02/03 Bernd Finkbeiner (UdS) Embedded Systems 2011/02/03 1 / 48 Automated Formal Methods

More information

Image Segmentation Using Iterated Graph Cuts Based on Multi-scale Smoothing

Image Segmentation Using Iterated Graph Cuts Based on Multi-scale Smoothing Image Segmentation Using Iterated Graph Cuts Based on Multi-scale Smoothing Tomoyuki Nagahashi 1, Hironobu Fujiyoshi 1, and Takeo Kanade 2 1 Dept. of Computer Science, Chubu University. Matsumoto 1200,

More information

CS534: Introduction to Computer Vision Edges and Contours. Ahmed Elgammal Dept. of Computer Science Rutgers University

CS534: Introduction to Computer Vision Edges and Contours. Ahmed Elgammal Dept. of Computer Science Rutgers University CS534: Introduction to Computer Vision Edges and Contours Ahmed Elgammal Dept. of Computer Science Rutgers University Outlines What makes an edge? Gradient-based edge detection Edge Operators Laplacian

More information

Subdivision. Outline. Key Questions. Subdivision Surfaces. Advanced Computer Graphics (Spring 2013) Video: Geri s Game (outside link)

Subdivision. Outline. Key Questions. Subdivision Surfaces. Advanced Computer Graphics (Spring 2013) Video: Geri s Game (outside link) Advanced Computer Graphics (Spring 03) CS 83, Lecture 7: Subdivision Ravi Ramamoorthi http://inst.eecs.berkeley.edu/~cs83/sp3 Slides courtesy of Szymon Rusinkiewicz, James O Brien with material from Denis

More information

Graph Coloring via Constraint Programming-based Column Generation

Graph Coloring via Constraint Programming-based Column Generation Graph Coloring via Constraint Programming-based Column Generation Stefano Gualandi Federico Malucelli Dipartimento di Elettronica e Informatica, Politecnico di Milano Viale Ponzio 24/A, 20133, Milan, Italy

More information

<March 2011> doc.: IEEE thz. Submission. Sebastian Priebe, TU Braunschweig

<March 2011> doc.: IEEE thz. Submission. Sebastian Priebe, TU Braunschweig Slide Spatial and Temporal Dispersion in THz Indoor Propagation Channels Sebastian Priebe, Martin Jacob, Thomas Kürner Institut für Nachrichtentechnik, Technische Universität Braunschweig, Germany Slide

More information

COMP/CS 605: Introduction to Parallel Computing Topic: Parallel Computing Overview/Introduction

COMP/CS 605: Introduction to Parallel Computing Topic: Parallel Computing Overview/Introduction COMP/CS 605: Introduction to Parallel Computing Topic: Parallel Computing Overview/Introduction Mary Thomas Department of Computer Science Computational Science Research Center (CSRC) San Diego State University

More information

Scalable Algorithmic Techniques Decompositions & Mapping. Alexandre David

Scalable Algorithmic Techniques Decompositions & Mapping. Alexandre David Scalable Algorithmic Techniques Decompositions & Mapping Alexandre David 1.2.05 adavid@cs.aau.dk Introduction Focus on data parallelism, scale with size. Task parallelism limited. Notion of scalability

More information

Today s class. Scheduling. Informationsteknologi. Tuesday, October 9, 2007 Computer Systems/Operating Systems - Class 14 1

Today s class. Scheduling. Informationsteknologi. Tuesday, October 9, 2007 Computer Systems/Operating Systems - Class 14 1 Today s class Scheduling Tuesday, October 9, 2007 Computer Systems/Operating Systems - Class 14 1 Aim of Scheduling Assign processes to be executed by the processor(s) Need to meet system objectives regarding:

More information

Stochastic Processing Networks: What, Why and How? Ruth J. Williams University of California, San Diego

Stochastic Processing Networks: What, Why and How? Ruth J. Williams University of California, San Diego Stochastic Processing Networks: What, Why and How? Ruth J. Williams University of California, San Diego http://www.math.ucsd.edu/~williams 1 OUTLINE! What is a Stochastic Processing Network?! Applications!

More information

This formula shows that partitioning the network decreases the total traffic if 1 N R (1 + p) < N R p < 1, i.e., if not all the packets have to go

This formula shows that partitioning the network decreases the total traffic if 1 N R (1 + p) < N R p < 1, i.e., if not all the packets have to go Chapter 3 Problem 2 In Figure 3.43 of the text every node transmits R bps, and we assume that both network partitions consist of 1 N nodes. So the total traffic generated by the nodes 2 of each Ethernet

More information

CSC630/COS781: Parallel & Distributed Computing

CSC630/COS781: Parallel & Distributed Computing CSC630/COS781: Parallel & Distributed Computing Algorithm Design Chapter 3 (3.1-3.3) 1 Contents Preliminaries of parallel algorithm design Decomposition Task dependency Task dependency graph Granularity

More information

Basics of UART Communication

Basics of UART Communication Basics of UART Communication From: Circuit Basics UART stands for Universal Asynchronous Receiver/Transmitter. It s not a communication protocol like SPI and I2C, but a physical circuit in a microcontroller,

More information

Branch-price-and-cut for vehicle routing. Guy Desaulniers

Branch-price-and-cut for vehicle routing. Guy Desaulniers Guy Desaulniers Professor, Polytechnique Montréal, Canada Director, GERAD, Canada VeRoLog PhD School 2018 Cagliari, Italy, June 2, 2018 Outline 1 VRPTW definition 2 Mathematical formulations Arc-flow formulation

More information

COMMUNICATION AND I/O ARCHITECTURES FOR HIGHLY INTEGRATED MPSoC PLATFORMS OUTLINE

COMMUNICATION AND I/O ARCHITECTURES FOR HIGHLY INTEGRATED MPSoC PLATFORMS OUTLINE COMMUNICATION AND I/O ARCHITECTURES FOR HIGHLY INTEGRATED MPSoC PLATFORMS Martino Ruggiero Luca Benini University of Bologna Simone Medardoni Davide Bertozzi University of Ferrara In cooperation with STMicroelectronics

More information

Chapter 3 Complexity of Classical Planning

Chapter 3 Complexity of Classical Planning Lecture slides for Automated Planning: Theory and Practice Chapter 3 Complexity of Classical Planning Dana S. Nau CMSC 722, AI Planning University of Maryland, Spring 2008 Licensed under the Creative Commons

More information

Towards Robust and Flexible Low-Power Wireless Networking

Towards Robust and Flexible Low-Power Wireless Networking Towards Robust and Flexible Low-Power Wireless Networking Philip Levis (joint work with Leonidas Guibas) Computer Systems Lab Stanford University 3.vii.2007 Low Power Wireless Low cost, numerous devices

More information

Spatial Outlier Detection

Spatial Outlier Detection Spatial Outlier Detection Chang-Tien Lu Department of Computer Science Northern Virginia Center Virginia Tech Joint work with Dechang Chen, Yufeng Kou, Jiang Zhao 1 Spatial Outlier A spatial data point

More information

PROVIDING SERVICE DIFFERENTIATION IN OBS NETWORKS THROUGH PROBABILISTIC PREEMPTION. YANG LIHONG (B.ENG(Hons.), NTU)

PROVIDING SERVICE DIFFERENTIATION IN OBS NETWORKS THROUGH PROBABILISTIC PREEMPTION. YANG LIHONG (B.ENG(Hons.), NTU) PROVIDING SERVICE DIFFERENTIATION IN OBS NETWORKS THROUGH PROBABILISTIC PREEMPTION YANG LIHONG (B.ENG(Hons.), NTU) A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF ENGINEERING DEPARTMENT OF ELECTRICAL &

More information

Scalable Bayes Clustering for Outlier Detection Under Informative Sampling

Scalable Bayes Clustering for Outlier Detection Under Informative Sampling Scalable Bayes Clustering for Outlier Detection Under Informative Sampling Based on JMLR paper of T. D. Savitsky Terrance D. Savitsky Office of Survey Methods Research FCSM - 2018 March 7-9, 2018 1 / 21

More information

Multiprocessor Support

Multiprocessor Support CSC 256/456: Operating Systems Multiprocessor Support John Criswell University of Rochester 1 Outline Multiprocessor hardware Types of multi-processor workloads Operating system issues Where to run the

More information

Efficient Scheduling of Scientific Workflows using Hot Metadata in a Multisite Cloud

Efficient Scheduling of Scientific Workflows using Hot Metadata in a Multisite Cloud Efficient Scheduling of Scientific Workflows using Hot Metadata in a Multisite Cloud Ji Liu 1,2,3, Luis Pineda 1,2,4, Esther Pacitti 1,2,3, Alexandru Costan 4, Patrick Valduriez 1,2,3, Gabriel Antoniu

More information

Virtualization in Wireless Networks

Virtualization in Wireless Networks ORCA-PROJECT.EU Virtualization in Wireless Networks Jonathan van de Belt, Yi Zhang CONNECT, TCD CROWNCOM 2018 TUTORIAL Gent 20-09-2018 ORCHESTRATION AND RECONFIGURATION CONTROL ARCHITECTURE Why use virtualization?

More information

Automated Design Flow for Coarse-Grained Reconfigurable Platforms: an RVC-CAL Multi-Standard Decoder Use-Case

Automated Design Flow for Coarse-Grained Reconfigurable Platforms: an RVC-CAL Multi-Standard Decoder Use-Case XIV International Conference on Embedded Computer and Systems: Architectures, MOdeling and Simulation SAMOS XIV - 2014 July 14 th - Samos Island (Greece) Carlo Sau, Luigi Raffo DIEE Università degli Studi

More information

Fine-Grain NBAR for Selective Applications

Fine-Grain NBAR for Selective Applications By default NBAR operates in the fine-grain mode, offering NBAR's full application recognition capabilities. Used when per-packet reporting is required, fine-grain mode offers a troubleshooting advantage.

More information

Model checking pushdown systems

Model checking pushdown systems Model checking pushdown systems R. Ramanujam Institute of Mathematical Sciences, Chennai jam@imsc.res.in Update Meeting, IIT-Guwahati, 4 July 2006 p. 1 Sources of unboundedness Data manipulation: integers,

More information

NONDETERMINISTIC MOORE

NONDETERMINISTIC MOORE NONDETERMINISTIC MOORE AUTOMATA AND BRZOZOWSKI'S ALGORITHM G. Castiglione, A. Restivo, M. Sciortino University of Palermo Workshop PRIN Varese, 5-7 Settembre 2011 SUMMARY A class of nondeterministic Moore

More information

OPTIMIZING COARSE- GRAINED UNITS IN FLOATING POINT HYBRID FPGA

OPTIMIZING COARSE- GRAINED UNITS IN FLOATING POINT HYBRID FPGA OPTIMIZING COARSE- GRAINED UNITS IN FLOATING POINT HYBRID FPGA Chi Wai Yu 1, Alastair M. Smith 2, Wayne Luk 1, Philip Leong 3, Steven J.E. Wilton 2 1 Dept of Computing Imperial College London, London {cyu,wl}@doc.ic.ac.uk

More information

Basic vs. Reliable Multicast

Basic vs. Reliable Multicast Basic vs. Reliable Multicast Basic multicast does not consider process crashes. Reliable multicast does. So far, we considered the basic versions of ordered multicasts. What about the reliable versions?

More information