A Solution for the Multiple Writer Problem in Shared Memory Multiprocessors: Anticipated Intention to Commit in Adaptive Environment Confidential! Veljko.

Slides:



Advertisements
Similar presentations
L.N. Bhuyan Adapted from Patterson’s slides
Advertisements

Multi-core systems System Architecture COMP25212 Daniel Goodman Advanced Processor Technologies Group.
Thoughts on Shared Caches Jeff Odom University of Maryland.
IT Systems Multiprocessor System EN230-1 Justin Champion C208 –
Chapter Hardwired vs Microprogrammed Control Multithreading
1 Efficient Multithreading Implementation of H.264 Encoder on Intel Hyper- Threading Architectures Steven Ge, Xinmin Tian, and Yen-Kuang Chen IEEE Pacific-Rim.
How Multi-threading can increase on-chip parallelism
A. Frank - P. Weisberg Operating Systems Introduction to Tasks/Threads.
SyNAR: Systems Networking and Architecture Group Symbiotic Jobscheduling for a Simultaneous Multithreading Processor Presenter: Alexandra Fedorova Simon.
Introduction to Embedded Systems
1 Processes and Threads Chapter Processes 2.2 Threads 2.3 Interprocess communication 2.4 Classical IPC problems 2.5 Scheduling.
Operating Systems Lecture 09: Threads (Chapter 4)
Discovering and Understanding Performance Bottlenecks in Transactional Applications Ferad Zyulkyarov 1,2, Srdjan Stipic 1,2, Tim Harris 3, Osman S. Unsal.
Operating System Review September 10, 2012Introduction to Computer Security ©2004 Matt Bishop Slide #1-1.
Multiple Processor Systems. Multiprocessor Systems Continuous need for faster and powerful computers –shared memory model ( access nsec) –message passing.
Threads in Java. History  Process is a program in execution  Has stack/heap memory  Has a program counter  Multiuser operating systems since the sixties.
Computer Architecture Lecture10: Input/output devices Piotr Bilski.
MODERN OPERATING SYSTEMS Third Edition ANDREW S. TANENBAUM Chapter 2 Processes and Threads Tanenbaum, Modern Operating Systems 3 e, (c) 2008 Prentice-Hall,
CPE 731: Advanced Computer Architecture Research Report and Presentation 1.
Operating System Concepts and Techniques Lecture 2 Multiprogramming, Multithreading, Multiprocessing, and Multitasking M. Naghibzadeh Reference M. Naghibzadeh,
1 Multiprocessor and Real-Time Scheduling Chapter 10 Real-Time scheduling will be covered in SYSC3303.
© Janice Regan, CMPT 300, May CMPT 300 Introduction to Operating Systems Introduction to Concurrency.
Introduction to AMBA Veljko Milutinovic
CMP-MSI Feb. 11 th 2007 Core to Memory Interconnection Implications for Forthcoming On-Chip Multiprocessors Carmelo Acosta 1 Francisco J. Cazorla 2 Alex.
CENTRAL PROCESSING UNIT. CPU Does the actual processing in the computer. A single chip called a microprocessor. Composed of an arithmetic and logic unit.
THREAD Subject:T0934 / Multimedia Programming Foundation Session:3 Tahun:2009 Versi:1/0.
13-1 Chapter 13 Concurrency Topics Introduction Introduction to Subprogram-Level Concurrency Semaphores Monitors Message Passing Java Threads C# Threads.
MBA in the Nutshell Veljko Milutinovic
S&T Methodologies Veljko Milutinovic, Fellow of the IEEE
OS part I, , Part I Computer System Structures Department of Computer Engineering, PSUWannarat Suntiamorntut.
Multithreading The objectives of this chapter are: To understand the purpose of multithreading To describe Java's multithreading mechanism.
Cooperative Task Management without Manual Stack management Hanyun Tao EECS 582 – W161.
Chapter 3 RTOS Concepts And Definitions Department of Computer Science Hsu Hao Chen Professor Hsung-Pin Chang.
Where Testing Fails …. Problem Areas Stack Overflow Race Conditions Deadlock Timing Reentrancy.
7/9/ Realizing Concurrency using Posix Threads (pthreads) B. Ramamurthy.
Processes and threads.
Chapter 1: Introduction
Operating System Concepts and Techniques Lecture 2
Scheduler activations
Waltham, Massachusetts, USA Wichita State University (WSU), USA
Simultaneous Multithreading
The University of Adelaide, School of Computer Science
The University of Adelaide, School of Computer Science
April 6, 2001 Gary Kimura Lecture #6 April 6, 2001
Unit OS4: Scheduling and Dispatch
Effective Data-Race Detection for the Kernel
Steven Ge, Xinmin Tian, and Yen-Kuang Chen
12.4 Memory Organization in Multiprocessor Systems
Lecture 7 Processes and Threads.
The University of Adelaide, School of Computer Science
MODERN OPERATING SYSTEMS Third Edition ANDREW S
The University of Adelaide, School of Computer Science
QNX Technology Overview
Symmetric Multiprocessing (SMP)
I/O Systems I/O Hardware Application I/O Interface
Operating systems Process scheduling.
Massachusetts Institute of Technology
What is Concurrent Programming?
Lecture 4- Threads, SMP, and Microkernels
Outline Scheduling algorithms Multi-processor scheduling
Chapter 10 Multiprocessor and Real-Time Scheduling
The University of Adelaide, School of Computer Science
Lecture 17 Multiprocessors and Thread-Level Parallelism
Linux Process State Scheduling information Identifiers
Lecture 17 Multiprocessors and Thread-Level Parallelism
The University of Adelaide, School of Computer Science
Concurrency: Threads, Address Spaces, and Processes
Chapter 3: Process Management
Threads CSE 2431: Introduction to Operating Systems
Lecture 17 Multiprocessors and Thread-Level Parallelism
Presentation transcript:

A Solution for the Multiple Writer Problem in Shared Memory Multiprocessors: Anticipated Intention to Commit in Adaptive Environment Confidential! Veljko Milutinovic, Fellow of the IEEE

The Best Method to Present Research Results 1. Introduction 2. Problem Statement 3. Existing Solutions 4. Essence of the Proposed Solution 5. Conditions and Assumptions 6. Details 7. Modelling 8. Analysis 9. Conclusions 10. References

1. Introduction

2. Problem Statement: MRMW

3. Existing Solutions The BEST current MRMW solution comes from Stanford professor Kunle Olokotun: Two-Phase Commit The BEST current TMS/SMT solution comes from Barcelona professor Mateo Valero: Microsoft-Intel UPC

4. Essence of the Proposed Approach Major improvements of our approach: a. Includes a timer b. Includes an anticipatory action (Intention to Commit can be executed before the moment when a process is actually ready to commit) c. Imlies the usage of SMT (Simultaneous Multi Threading), so the process can keep busy, while waiting for the commit activity to complete. Monitors and adaptivity

5. Conditions and Assumptions Conditions: a. Size of MCM unit b. Processors vs processes and SMT c. Worst case delays in SMP vs DSM Assumptions: a. False sharing b. Add-on to a directory scheme c. System limits

6. Details Intention to Commit a. Vector of starting MCM unit addresses b. Vector of shared data involved c. Timer activated Counter Interrupt a. No formal conflict b. Yes formal conflict; no essential conflict c. Both formal and essential conflict Adaptive Priorities Adaptive Abort vs Resume Adaptive SMT

7. Analysis System Modelling, Performance Analysis, Complexity Analysis M5 GeneralSimulators vs UPC MareNostrum Microsoft vs Intel

8. Conclusion and References