Parallel Speedup.

Slides:



Advertisements
Similar presentations
Optimal PRAM algorithms: Efficiency of concurrent writing “Computer science is no more about computers than astronomy is about telescopes.” Edsger Dijkstra.
Advertisements

Potential for parallel computers/parallel programming
CIS December '99 Introduction to Parallel Architectures Dr. Laurence Boxer Niagara University.
Parallel Processing & Distributed Systems Thoai Nam Chapter 2.
Advanced Topics in Algorithms and Data Structures Lecture pg 1 Recursion.
Parallel System Performance CS 524 – High-Performance Computing.
11Sahalu JunaiduICS 573: High Performance Computing5.1 Analytical Modeling of Parallel Programs Sources of Overhead in Parallel Programs Performance Metrics.
1 Lecture 4 Analytical Modeling of Parallel Programs Parallel Computing Fall 2008.
Arquitectura de Sistemas Paralelos e Distribuídos Paulo Marques Dep. Eng. Informática – Universidade de Coimbra Ago/ Quantitative.
Performance Metrics Parallel Computing - Theory and Practice (2/e) Section 3.6 Michael J. Quinn mcGraw-Hill, Inc., 1994.
On-line adaptive parallel prefix computation Jean-Louis Roch, Daouda Traoré and Julien Bernard Presented by Andreas Söderström, ITN.
Parallel System Performance CS 524 – High-Performance Computing.
The Group Runtime Optimization for High-Performance Computing An Install-Time System for Automatic Generation of Optimized Parallel Sorting Algorithms.
Computer Science 320 Measuring Speedup. What Is Running Time? T(N, K) says that the running time T is a function of the problem size N and the number.
Performance Evaluation of Parallel Processing. Why Performance?
1 Interconnects Shared address space and message passing computers can be constructed by connecting processors and memory unit using a variety of interconnection.
1 Introduction to Parallel Computing. 2 Multiprocessor Architectures Message-Passing Architectures –Separate address space for each processor. –Processors.
“elbowing out” Processors used Speedup Efficiency timeexecution Parallel Processors timeexecution Sequential Efficiency   
CS453 Lecture 3.  A sequential algorithm is evaluated by its runtime (in general, asymptotic runtime as a function of input size).  The asymptotic runtime.
CS4231 Parallel and Distributed Algorithms AY 2006/2007 Semester 2 Lecture 3 (26/01/2006) Instructor: Haifeng YU.
Parallel Suffix Array Construction by Accelerated Sampling Matthew Felice Pace University of Warwick Joint work with Alexander Tiskin University of Warwick.
Mathematical Proofs And how they are applied to Computer Olympiads.
Dean Tullsen UCSD.  The parallelism crisis has the feel of a relatively new problem ◦ Results from a huge technology shift ◦ Has suddenly become pervasive.
Parallel Processing & Distributed Systems Thoai Nam And Vu Le Hung.
Parallel Programming with MPI and OpenMP
Limits to Computation How do you analyze a new algorithm? –Put it in the form of existing algorithms that you know the analysis. –For example, given 2.
A System Performance Model Distributed Process Scheduling.
3/12/2013Computer Engg, IIT(BHU)1 INTRODUCTION-1.
Computer Science 320 Measuring Sizeup. Speedup vs Sizeup If we add more processors, we should be able to solve a problem of a given size faster If we.
1 Potential for Parallel Computation Chapter 2 – Part 2 Jordan & Alaghband.
1 5. Abstract Data Structures & Algorithms 5.6 Algorithm Evaluation.
DCS/1 CENG Distributed Computing Systems Measures of Performance.
Partial Orderings: Selected Exercises
Potential for parallel computers/parallel programming
Parallel Computing and Parallel Computers
PERFORMANCE EVALUATIONS
Performance Evaluation Frédéric Desprez INRIA
COMPUTATIONAL MODELS.
SCALABILITY ANALYSIS.
What Exactly is Parallel Processing?
Introduction to parallel algorithms
Sorting Quiz questions
Multi-Node Broadcasting in Hypercube and Star Graphs
April 30th – Scheduling / parallel
Chapter 3: Principles of Scalable Performance
Complexity Measures for Parallel Computation
STUDY AND IMPLEMENTATION
Deterministic PDAs - DPDAs
Introduction to parallel algorithms
Numerical Algorithms Quiz questions
Parallelismo.
Professor Ioana Banicescu CSE 8843
University of Pittsburgh
Coarse Grained Parallel Selection
University of Pittsburgh
Analytical Modeling of Parallel Systems
Multithreaded Programming in Cilk Lecture 1
PERFORMANCE MEASURES. COMPUTATIONAL MODELS Equal Duration Model:  It is assumed that a given task can be divided into n equal subtasks, each of which.
Parallel Algorithms A Simple Model for Parallel Processing
Parallel Computing and Parallel Computers
Potential for parallel computers/parallel programming
Potential for parallel computers/parallel programming
Complexity Measures for Parallel Computation
Analytical Modeling of Parallel Systems
Potential for parallel computers/parallel programming
Chapter 5 Parallel Lines and Related Figures
CSE 332: Parallel Algorithms
Potential for parallel computers/parallel programming
Introduction to parallel algorithms
Presentation transcript:

Parallel Speedup

Performance of a Parallel Algorithm n : problem size (e.g.: sort n numbers) p : number of processors Tp : parallel time Ts : sequential time (optimal sequ. alg.) S = Ts / Tp : speedup (1Sp) S S=p super-linear linear sub-linear p

Speedup linear speedup S = p optimal super linear speedup S > p : impossible Proof. Assume that parallel algorithm A has a speedup S > p for processors, i.e. S = Ts / Tp > p. Hence: Ts > T·p. Simulate A on a sequential, single processor machine. Then Tp(1) = Tp · p < Ts. Hence, Ts was not optimal. Contradiction.

S p

Scaled Speedup Ts may be unknown (in fact, for most real experiments this is the case) Scaled speedup S’ = Tp(1) / Tp(p) S’  S

Efficiency e = S / p efficiency (0e1) optimal linear speedup S = p  e = 1 e’ = S’ / p scaled efficiency