The functions of system LSI become more and more complicated



Similar documents
Optimizing Configuration and Application Mapping for MPSoC Architectures

Making Multicore Work and Measuring its Benefits. Markus Levy, president EEMBC and Multicore Association

PikeOS: Multi-Core RTOS for IMA. Dr. Sergey Tverdyshev SYSGO AG , Moscow

Data Centric Systems (DCS)

Agenda. Michele Taliercio, Il circuito Integrato, Novembre 2001

Client/Server Computing Distributed Processing, Client/Server, and Clusters

Introduction to Embedded Systems. Software Update Problem

Principles and characteristics of distributed systems and environments

CHAPTER 1 INTRODUCTION

A Hardware and Software Monitor for High-Level System-on-Chip Verification

Seven Challenges of Embedded Software Development

ICRI-CI Retreat Architecture track

Universal Flash Storage: Mobilize Your Data

Hardware in the Loop (HIL) Testing VU 2.0, , WS 2008/09

NanopowerCommunications: Enabling the Internet of Things OBJECTS TALK

Speeding up MATLAB and Simulink Applications

Real-Time Operating Systems for MPSoCs

Codesign: The World Of Practice

CycurHSM An Automotive-qualified Software Stack for Hardware Security Modules

Driving force. What future software needs. Potential research topics

on-chip and Embedded Software Perspectives and Needs

EECS 750: Advanced Operating Systems. 01/28 /2015 Heechul Yun

Applications to Computational Financial and GPU Computing. May 16th. Dr. Daniel Egloff

22S:295 Seminar in Applied Statistics High Performance Computing in Statistics

Design and Implementation of the Heterogeneous Multikernel Operating System

Advanced Encryption Standard (AES) User's Guide

ESE566 REPORT3. Design Methodologies for Core-based System-on-Chip HUA TANG OVIDIU CARNU

Cluster, Grid, Cloud Concepts

Multi-objective Design Space Exploration based on UML

Electronic system-level development: Finding the right mix of solutions for the right mix of engineers.

Parallel Computing. Benson Muite. benson.

Introduction to Cloud Computing

Cortex-A9 MPCore Software Development

Weighted Total Mark. Weighted Exam Mark

Part I Courses Syllabus

BY STEVE BROWN, CADENCE DESIGN SYSTEMS AND MICHEL GENARD, VIRTUTECH

Chapter 6, The Operating System Machine Level

GEDAE TM - A Graphical Programming and Autocode Generation Tool for Signal Processor Applications

numascale White Paper The Numascale Solution: Extreme BIG DATA Computing Hardware Accellerated Data Intensive Computing By: Einar Rustad ABSTRACT

White Paper The Numascale Solution: Extreme BIG DATA Computing

White Paper The Numascale Solution: Affordable BIG DATA Computing

System Software Integration: An Expansive View. Overview

Distributed Systems. REK s adaptation of Prof. Claypool s adaptation of Tanenbaum s Distributed Systems Chapter 1

Embedded Parallel Computing

How To Write A Windows Operating System (Windows) (For Linux) (Windows 2) (Programming) (Operating System) (Permanent) (Powerbook) (Unix) (Amd64) (Win2) (X

Unified Batch & Stream Processing Platform

Network for Sustainable Ultrascale Computing (NESUS)

HPC with Multicore and GPUs

Beyond Virtualization: A Novel Software Architecture for Multi-Core SoCs. Jim Ready September 18, 2012

Multi-core Curriculum Development at Georgia Tech: Experience and Future Steps

Chapter 12: Multiprocessor Architectures. Lesson 01: Performance characteristics of Multiprocessor Architectures and Speedup

Chapter 1: Introduction. What is an Operating System?

OpenMosix Presented by Dr. Moshe Bar and MAASK [01]

So#ware Tools and Techniques for HPC, Clouds, and Server- Class SoCs Ron Brightwell

FPGA area allocation for parallel C applications

Chapter 7: Distributed Systems: Warehouse-Scale Computing. Fall 2011 Jussi Kangasharju

Multicore Parallel Computing with OpenMP

VP/GM, Data Center Processing Group. Copyright 2014 Cavium Inc.

Chapter 1 Computer System Overview

Scalability and Programmability in the Manycore Era

Learning Objectives. Chapter 1: Networking with Microsoft Windows 2000 Server. Basic Network Concepts. Learning Objectives (continued)

Principles of Operating Systems CS 446/646

Tools Page 1 of 13 ON PROGRAM TRANSLATION. A priori, we have two translation mechanisms available:

Low-Overhead Hard Real-time Aware Interconnect Network Router

Client/Server and Distributed Computing

End-user Tools for Application Performance Analysis Using Hardware Counters

OpenACC Programming and Best Practices Guide

Designing and Building Applications for Extreme Scale Systems CS598 William Gropp

RTAI. Antonio Barbalace (modified by M.Moro 2011) RTAI

ELEC 5260/6260/6266 Embedded Computing Systems

DB2 for i5/os: Tuning for Performance

AP Computer Science AB Syllabus 1

What is a System on a Chip?

LSI SAS inside 60% of servers. 21 million LSI SAS & MegaRAID solutions shipped over last 3 years. 9 out of 10 top server vendors use MegaRAID

AMD Opteron Quad-Core

Chapter 16 Distributed Processing, Client/Server, and Clusters

High Performance Computing. Course Notes HPC Fundamentals

Parallel Programming Survey

High Performance or Cycle Accuracy?

Lecture 23: Multiprocessors

be architected pool of servers reliability and

Hardware Acceleration for Just-In-Time Compilation on Heterogeneous Embedded Systems

Why study Operating Systems? CS 372. Why study. Why study. Introduction to Operating Systems

CS521 CSE IITG 11/23/2012


SHARED HASH TABLES IN PARALLEL MODEL CHECKING

HPC and Grid Concepts

Big Data and Cloud Computing for GHRSST

Software and the Concurrency Revolution

Transcription:

The functions of system LSI become more and more complicated Current requirement Data processing Compliant to new formats Further expand requirement Innovations of the user interface Recognizing outside Requirement of processing performance Trend 2007 ITRS page 10, Figure SYSD7 SOC Consumer Portable Processing Performance Trends System LSI Ex. Smile detection technology

New functions require increase of # of processors New functions TAT: Turn Around Time SW: Software

AMP: Asynchronous Multi-Processor Processor Control Message Sub system Single-Core AMP

SMP: Symmetrical Multi-Processor Symmetric Simple SMP (ex.1) Symmetric AMP Simple SMP (ex.2)

Complex SMP and Many-Core homogenous heterogynous Simple SMP Complex SMP Many-Core

It will happen when complex SMP is introduced into Current mainstream Requires new paradigm Paradigm change AMP Simple SMP Complex SMP Many-Core

Problem area: design, implementation & debug Design Implementation Debug Performance Reusability Productivity

Replace to parallelizable Increase max parallel degree Increase self-propelled periods Optimize hierarchy memory access

Increase of the fault due to the programming difficulty Parallel movement Disturbance Communication between SMP

Increase of the difficulty of the debugging work itself Behavior of multiple processors Divergence between source code & object Multiple programming languages

Implementation & debugging issues are at the fore front Communication between SMP becomes important

The challenge of the productivity is late considerably Parallel algorithm, API for parallel computing, automatic parallelization etc.. Design Implementation Debug

Backward compatibility covered the time lag System LSI HW development AMP Simple SMP Tool chain development Brush up Tools for AMP Optimized for simple SMP Software production Brush up Reuse assets for AMP For simple SMP HW: Hardware Backward compatibility

HW, SW & tools have to tackle it in a cooperative way! System LSI HW development Tool chain development Software production Simple SMP Ready to change For simple SMP Ready to change For simple SMP Ready to change Complex SMP Optimized for complex SMP Optimized for complex SMP Cooperative action!

Shift to the Many-Core depends on the SW productivity It requires cooperation among HW, SW & tools AMP Simple SMP Complex SMP Many-Core SW productivity Cooperation of HW/SW/Tools