Welcome to Beyond Petaflops: Specialized Architectures for Power Efficient Scientific Computing SIAM Conference on Computational Science and Engineering.

Slides:



Advertisements
Similar presentations
Multi-core processors. 2 Processor development till 2004 Out-of-order Instruction scheduling Out-of-order Instruction scheduling.
Advertisements

The Brave New World of Software Adam Kemp Staff Software Engineer National Instruments.
Multi-core systems System Architecture COMP25212 Daniel Goodman Advanced Processor Technologies Group.
Computer Abstractions and Technology
1 A Self-Tuning Cache Architecture for Embedded Systems Chuanjun Zhang*, Frank Vahid**, and Roman Lysecky *Dept. of Electrical Engineering Dept. of Computer.
Power Reduction Techniques For Microprocessor Systems
Device Tradeoffs Greg Stitt ECE Department University of Florida.
A Feedback and Continuous Improvement Tool to ACE Audit Readiness Keith S. Joy Quality Manager September 14, 2004.
Moore's Law By: Brad Sparks, Noah Al-Atta, Brenden Leis, Carly Reese, Malaika Messi.
Lecture 2: Modern Trends 1. 2 Microprocessor Performance Only 7% improvement in memory performance every year! 50% improvement in microprocessor performance.
Chapter1 Fundamental of Computer Design Dr. Bernard Chen Ph.D. University of Central Arkansas.
Microsoft Technical Computing Modeling the world with greater fidelity Wolfgang Dreyer, TC - Microsoft Germany.
SYNAR Systems Networking and Architecture Group CMPT 886: Special Topics in Operating Systems and Computer Architecture Dr. Alexandra Fedorova School of.
Neutron Scattering Experiment Automation with Python RT2010 Conference, Lisbon, Portugal (PCM-26) Piotr Żołnierczuk, Rick Riedel Neutron Scattering Science.
Room: E-3-31 Phone: Dr Masri Ayob TK 2123 COMPUTER ORGANISATION & ARCHITECTURE Lecture 4: Computer Performance.
Mercury Nozzle Status V.B. Graves Hg Jet Design Meeting Princeton University Nov 15, 2004.
Trevor Burton6/19/2015 Multiprocessors for DSP SYSC5603 Digital Signal Processing Microprocessors, Software and Applications.
Mercury Catcher Status V.B. Graves Hg Jet Design Meeting Princeton University Nov 15, 2004.
Future Computers CSCI 107, Spring When Moore’s law runs out of room When transistors become only tens of atoms thick –Quantum mechanics applies.
Single-Chip Multi-Processors (CMP) PRADEEP DANDAMUDI 1 ELEC , Fall 08.
Chapter1 Fundamental of Computer Design Dr. Bernard Chen Ph.D. University of Central Arkansas Fall 2010.
Introduction to Parallel Computing. Serial Computing.
Multi Core Processor Submitted by: Lizolen Pradhan
Lecture 03: Fundamentals of Computer Design - Trends and Performance Kai Bu
Last Time Performance Analysis It’s all relative
Multi-core Programming Introduction Topics. Topics General Ideas Moore’s Law Amdahl's Law Processes and Threads Concurrency vs. Parallelism.
Lecture 1: Performance EEN 312: Processors: Hardware, Software, and Interfacing Department of Electrical and Computer Engineering Spring 2013, Dr. Rozier.
1 Recap (from Previous Lecture). 2 Computer Architecture Computer Architecture involves 3 inter- related components – Instruction set architecture (ISA):
1 CS/EE 6810: Computer Architecture Class format:  Most lectures on YouTube *BEFORE* class  Use class time for discussions, clarifications, problem-solving,
MS108 Computer System I Lecture 2 Metrics Prof. Xiaoyao Liang 2014/2/28 1.
Grad Student Visit DayUniversity of Wisconsin-Madison Wisconsin Computer Architecture Guri SohiMark HillMikko LipastiDavid WoodKaru Sankaralingam Nam Sung.
Oak Ridge Regional Emergency Management Forum David Milan Manager, Emergency Preparedness Department Pollard Auditorium October 26, 2006 Oak Ridge National.
Graphical Expert System for Analyzing Nuclear Facility Vulnerability David Sulfredge Oak Ridge National Laboratory December 3, 2002 O AK R IDGE N ATIONAL.
Reminder Lab 0 Xilinx ISE tutorial Research Send me an if interested Looking for those interested in RC with skills in compilers/languages/synthesis,
Outline  Over view  Design  Performance  Advantages and disadvantages  Examples  Conclusion  Bibliography.
1 Application Scalability and High Productivity Computing Nicholas J Wright John Shalf Harvey Wasserman Advanced Technologies Group NERSC/LBNL.
Computational Sprinting on a Real System: Preliminary Results Arun Raghavan *, Marios Papaefthymiou +, Kevin P. Pipe +#, Thomas F. Wenisch +, Milo M. K.
1 Wisconsin Computer Architecture Guri SohiMark HillMikko LipastiDavid WoodKaru Sankaralingam Nam Sung Kim.
Morgan Kaufmann Publishers
W E L C O M E. T R I G A T E T R A N S I S T O R.
O AK R IDGE N ATIONAL L ABORATORY U. S. D EPARTMENT OF E NERGY 1 Ideas on a Framework and Methods for Estimating the Benefits of Government- Sponsored.
Multi-core processors. 2 Processor development till 2004 Out-of-order Instruction scheduling Out-of-order Instruction scheduling.
1 Spallation Neutron Source Data Analysis Jessica Travierso Research Alliance in Math and Science Program Austin Peay State University Mentor: Vickie E.
EUROHEAT & POWER CLIMESPACE PRESENTATION. Cooling – The New Market Dimension  Paris District Cooling Network.
O AK R IDGE N ATIONAL L ABORATORY U. S. D EPARTMENT OF E NERGY 1 Ten Steps Toward Understanding Export Controlled Information Dave Hamrin, ORNL 2010 ECCO.
Lx: A Technology Platform for Customizable VLIW Embedded Processing.
VLSI Design & Embedded Systems Conference January 2015 Bengaluru, India Few Good Frequencies for Power-Constrained Test Sindhu Gunasekar and Vishwani D.
Applications Kay Kasemir ORNL/SNS Using Information and pictures from Matthias Clausen, Jan Hatje, and Helge Rickens (DESY) October 2007.
DR. SIMING LIU SPRING 2016 COMPUTER SCIENCE AND ENGINEERING UNIVERSITY OF NEVADA, RENO Session 3 Computer Evolution.
O AK R IDGE N ATIONAL L ABORATORY U.S. D EPARTMENT OF E NERGY Data Requirements for Climate and Carbon Research John Drake, Climate Dynamics Group Computer.
Architectural Design of Large Hotel and Energy Used for Internal Space Thermal Control by Dr Bin Su School of Architecture and Landscape Architecture UNITEC,
Hardware Trends CSE451 Andrew Whitaker. Motivation Hardware moves quickly OS code tends to stick around for a while “System building” extends way beyond.
Hardware Trends CSE451 Andrew Whitaker. Motivation Hardware moves quickly OS code tends to stick around for a while “System building” extends way beyond.
CS203 – Advanced Computer Architecture
History of Computers and Performance David Monismith Jan. 14, 2015 Based on notes from Dr. Bill Siever and from the Patterson and Hennessy Text.
William Stallings Computer Organization and Architecture 6th Edition
High Performance Computer Architecture:
CS203 – Advanced Computer Architecture
Lynn Choi School of Electrical Engineering
Introduction Characteristics Advantages Limitations
Scaling Digital Computing Beyond Moore’s Law
Architecture & Organization 1
Architecture & Organization 1
CS/EE 6810: Computer Architecture
Adaptive Single-Chip Multiprocessing
Chapter 1 Introduction.
The University of Adelaide, School of Computer Science
Intel CPU for Desktop PC: Past, Present, Future
Presentation transcript:

Welcome to Beyond Petaflops: Specialized Architectures for Power Efficient Scientific Computing SIAM Conference on Computational Science and Engineering Monday, February 19, 2007 Costa Mesa California John Shalf LBNL/NERSC NERSC 5

What is Going on? New Constraints –Power limits clock rates –Cannot squeeze more performance from ILP (complex cores) either! But Moore’s Law continues! –What to do with all of those transistors if everything else is flat-lining? –Now, #cores per chip doubles every 18 months instead of clock frequency! Power Consumption is chief concern for system architects Power-Efficiency is the primary concern of consumers of computer systems! Figure courtesy of Kunle Olukotun, Lance Hammond, Herb Sutter, and Burton Smith

Source: sandpile.org Microprocessors: Up Against the Wall(s) Microprocessors are hitting a power wall –Higher clock rates and greater leakage increasing power consumption Reaching the limits of what non-heroic heat solutions can handle Newer technology becoming more difficult to produce, removing the previous trend of “free” power improvement From Joe Gebis: UCB/LBNL

Example: ORNL Power Cost Projection Immediate need to add 8 MW to prepare for 2007 installs of new systems NLCF petascale system could require an additional 10 MW by 2008 Need total of MW for projected systems by 2011 Numbers just for computers: add 75% for cooling Cooling will require 12,000 – 15,000 tons of chiller capacity Cost estimates based on $0.05 kW/hr $3M $17M $9M $23M $31M O AK R IDGE N ATIONAL L ABORATORY U. S. D EPARTMENT OF E NERGY Data taken from Energy Management System-4 (EMS4). EMS4 is the DOE corporate system for collecting energy information from the sites. EMS4 is a web-based system that collects energy consumption and cost information for all energy sources used at each DOE site. Information is entered into EMS4 by the site and reviewed at Headquarters for accuracy. Yikes!

Common Ground for MS02 Speakers When power costs more than capital procurement costs –It is cost-effective to explore more customized solutions –It is more power-efficient to specialize to the application –It is more power-efficient to employ massive parallelism See –Embedded computing market has been doing this for years! Start with Science Requirements and then design a machine to conform to requirements –This as opposed to building a petaflop machine and trying to find things that will run on it Using this approach, you can reach aggressive performance goals... –Far more rapidly than a conventional approach would allow –Using a reasonable power budget –Using a reasonable development budget