Bandwidth Rocks (1) Latency Lags Bandwidth (last ~20 years) Performance Milestones Disk: 3600, 5400, 7200, 10000, 15000 RPM.

Slides:



Advertisements
Similar presentations
IT253: Computer Organization
Advertisements

Communication-Avoiding Algorithms Jim Demmel EECS & Math Departments UC Berkeley.
CA 714CA Midterm Review. C5 Cache Optimization Reduce miss penalty –Hardware and software Reduce miss rate –Hardware and software Reduce hit time –Hardware.
Computer Architecture & Organization
Lecture 2: Modern Trends 1. 2 Microprocessor Performance Only 7% improvement in memory performance every year! 50% improvement in microprocessor performance.
CSE 490/590, Spring 2011 CSE 490/590 Computer Architecture Intro II & Trends Steve Ko Computer Sciences and Engineering University at Buffalo.
CPSC 614 Computer Architecture Lec 2 - Introduction EJ Kim Dept. of Computer Science Texas A&M University Adapted from CS 252 Spring 2006 UC Berkeley Copyright.
1 Lecture 15: DRAM Design Today: DRAM basics, DRAM innovations (Section 5.3)
1 Multiprocessors. 2 Idea: create powerful computers by connecting many smaller ones good news: works for timesharing (better than supercomputer) bad.
Memory Hierarchy.1 Review: Major Components of a Computer Processor Control Datapath Memory Devices Input Output.
CS 501: Software Engineering Fall 2000 Lecture 19 Performance of Computer Systems.
1 Interfacing Processors and Peripherals I/O Design affected by many factors (expandability, resilience) Performance: — access latency — throughput — connection.
CS CS 5150 Software Engineering Lecture 25 Performance.
1 Lecture 13: Cache Innovations Today: cache access basics and innovations, DRAM (Sections )
1 CS 501 Spring 2005 CS 501: Software Engineering Lecture 22 Performance of Computer Systems.
Lecture 37: Chapter 7: Multiprocessors Today’s topic –Introduction to multiprocessors –Parallelism in software –Memory organization –Cache coherence 1.
CptS / E E COMPUTER ARCHITECTURE
Introduction to Database Systems 1 The Storage Hierarchy and Magnetic Disks Storage Technology: Topic 1.
F1020/F1031 COMPUTER HARDWARE MEMORY. Read-only Memory (ROM) Basic instructions for booting the computer and loading the operating system are stored in.
* Definition of -RAM (random access memory) :- -RAM is the place in a computer where the operating system, application programs & data in current use.
CSE820 Lec 2 - Introduction Rich Enbody Based on slides by David Patterson.
Memory. Random Access Memory Defined What is memory? operating system and other system software that control the usage of the computer equipment application.
CS136, Advanced Architecture Introduction to Architecture (continued)
Semiconductor Memory 1970 Fairchild Size of a single core –i.e. 1 bit of magnetic core storage Holds 256 bits Non-destructive read Much faster than core.
Topic: Disks – file system devices. Rotational Media Sector Track Cylinder Head Platter Arm Access time = seek time + rotational delay + transfer time.
Eng. Mohammed Timraz Electronics & Communication Engineer University of Palestine Faculty of Engineering and Urban planning Software Engineering Department.
2007 Sept 06SYSC 2001* - Fall SYSC2001-Ch1.ppt1 Computer Architecture & Organization  Instruction set, number of bits used for data representation,
Lecture 03: Fundamentals of Computer Design - Trends and Performance Kai Bu
Computer Architecture Lec 2 - Introduction. 01/19/10Lec 02-intro 2 Review from last lecture Computer Architecture >> instruction sets Computer Architecture.
Hardware Trends. Contents Memory Hard Disks Processors Network Accessories Future.
1 Computer System Organization I/O systemProcessor Compiler Operating System (Windows 98) Application (Netscape) Digital Design Circuit Design Instruction.
1 CS 501 Spring 2006 CS 501: Software Engineering Lecture 22 Performance of Computer Systems.
Egle Cebelyte. Random Access Memory is simply the storage area where all software is loaded and works from; also called working memory storage.
CPE 731 Advanced Computer Architecture Technology Trends Dr. Gheith Abandah Adapted from the slides of Prof. David Patterson, University of California,
CS 5513 Computer Architecture Lecture 2 – More Introduction, Measuring Performance.
Super computers Parallel Processing By Lecturer: Aisha Dawood.
– 1 – CSCE 513 Fall 2015 Lec08 Memory Hierarchy IV Topics Pipelining Review Load-Use Hazard Memory Hierarchy Review Terminology review Basic Equations.
1 CS : Technology Trends Ion Stoica and Ali Ghodsi ( August 31, 2015.
CS 152 L18 Disks, RAID, BW (1)Fall 2004 © UC Regents CS152 – Computer Architecture and Engineering Lecture 18 – ECC, RAID, Bandwidth vs. Latency
Authors – Jeahyuk huh, Doug Burger, and Stephen W.Keckler Presenter – Sushma Myneni Exploring the Design Space of Future CMPs.
Csci 136 Computer Architecture II – IO and Storage Systems Xiuzhen Cheng
Sep. 17, 2002BESIII Review Meeting BESIII DAQ System BESIII Review Meeting IHEP · Beijing · China Sep , 2002.
COMP381 by M. Hamdi 1 Clusters: Networks of WS/PC.
1 CS 501 Spring 2003 CS 501: Software Engineering Lecture 23 Performance of Computer Systems.
1 Adapted from UC Berkeley CS252 S01 Lecture 18: Reducing Cache Hit Time and Main Memory Design Virtucal Cache, pipelined cache, cache summary, main memory.
Computer Organization Yasser F. O. Mohammad 1. 2 Lecture 1: Introduction Today’s topics:  Why computer organization is important  Logistics  Modern.
Spring EE 437 Lillevik 437s06-l22 University of Portland School of Engineering Advanced Computer Architecture Lecture 22 Distributed computer Interconnection.
1  2004 Morgan Kaufmann Publishers Page Tables. 2  2004 Morgan Kaufmann Publishers Page Tables.
CS203 – Advanced Computer Architecture
Ch1. Fundamentals of Computer Design 1. Formulas ECE562 Advanced Computer Architecture Prof. Honggang Wang ECE Department University of Massachusetts Dartmouth.
1 Lecture 20: OOO, Memory Hierarchy Today’s topics:  Out-of-order execution  Cache basics.
Personal Computer (PC)  Computer advertisement specification Intel® Pentium 4 Processor at 3.06GHz with 512K cache 512MB DDR SDRAM 200GB ATA-100 Hard.
William Stallings Computer Organization and Architecture 6th Edition
Hardware Technology Trends and Database Opportunities
Types of RAM (Random Access Memory)
Reducing Hit Time Small and simple caches Way prediction Trace caches
RAM, CPUs, & BUSES Egle Cebelyte.
HY425 – Αρχιτεκτονική Υπολογιστών Διάλεξη 02
Lec09 Memory Hierarchy yet again
CS : Technology Trends August 31, 2015 Ion Stoica and Ali Ghodsi (
Virtual Memory Main memory can act as a cache for the secondary storage (disk) Advantages: illusion of having more physical memory program relocation protection.
CS775: Computer Architecture
CPE 432 Computer Design 1 – Introduction and Technology Trends
Input-output I/O is very much architecture/system dependent
Lecture 20: OOO, Memory Hierarchy
Computer Evolution and Performance
AKT211 – CAO 07 – Computer Memory
DRAM Hwansoo Han.
Lecture 20 Parallel Programming CSE /27/2019.
Presentation transcript:

Bandwidth Rocks (1) Latency Lags Bandwidth (last ~20 years) Performance Milestones Disk: 3600, 5400, 7200, 10000, RPM

Bandwidth Rocks (2) Latency Lags Bandwidth (last ~20 years) Performance Milestones Memory Module: 16bit plain DRAM, 16b Page Mode DRAM, 32b, 64b, 64b SDRAM, 64b DDR SDRAM Disk: 3600, 5400, 7200, 10000, RPM

Bandwidth Rocks (3) Latency Lags Bandwidth (last ~20 years) Performance Milestones Ethernet: 10Mb, 100Mb, 1000Mb, Mbit/sec Memory Module: 16bit plain DRAM, 16b Page Mode DRAM, 32b, 64b, 64b SDRAM, 64b DDR SDRAM Disk: 3600, 5400, 7200, 10000, RPM

Bandwidth Rocks (4) Latency Lags Bandwidth (last ~20 years) Performance Milestones Processor: ‘286, ‘386, ‘486, Pentium, Pentium Pro, Pentium 4 Ethernet: 10Mb, 100Mb, 1000Mb, Mbit/sec Memory Module: 16bit plain DRAM, 16b Page Mode DRAM, 32b, 64b, 64b SDRAM, 64b DDR SDRAM Disk: 3600, 5400, 7200, 10000, RPM

Bandwidth Rocks (5) Rule of Thumb for Latency Lagging BW In the time that bandwidth doubles, latency improves only by factors of 1.2 to 1.4 (and capacity approximately triples in the time that bandwidth doubles)

Bandwidth Rocks (6) Improvement in Time for BW to Double CPUMemoryLANDisk Latency / Year Capacity / Year Bandwidth / Year Capacity Improvement Years for 2X Bandwidth Latency Improvement

Bandwidth Rocks (7) Recent Improvement in Time for BW to Double CPUMemoryLANDisk Latency / Year Capacity / Year Bandwidth / Year Capacity Improvement Years for 2X Bandwidth Latency Improvement (Last 3 Performance Milestones)

Bandwidth Rocks (8) 1.Latency helps BW, but not vice versa Spinning disk faster improves both bandwidth and rotational latency Higher linear density helps BW, not Latency 2.Bandwidth hurts latency Longer queues help BW, hurt Latency 3.Distance limits latency Speed of light and computers on network Size of DRAM block => long bit and word lines => most of DRAM access time 6 Reasons Latency Lags Bandwidth

Bandwidth Rocks (9) 6 Reasons Latency Lags Bandwidth 4.Moore’s Law helps BW more than latency Smaller, faster transistors communicate over relatively longer lines: limits latency Faster transistors, more transistors, more pins all help BW 5.Software overhead hurts latency Long messages amortize overhead; overhead more for short messages 6.Bandwidth is easier to sell E.g., 10 Gbit/s vs. 10  sec latency Ethernet

Bandwidth Rocks (10) 4 Ways to Cope with Latency Lags Bandwidth 1.Caching (Leveraging Capacity) Processor caches, file cache, disk cache 2.Replication (Leveraging Capacity) Read from nearest head in RAID, Distributed caches in SMP, Distributed ISPs 3.Parallelism and Out-of-Order Processing (Leveraging Bandwidth) Processors, Disk controllers, Messages 4.Prediction (Leveraging Bandwidth) Branches, Prefetching: disk, caches

Bandwidth Rocks (11) Conclusion If everything improves at the same rate, then nothing really changes –When rates vary, we see real dislocation that in turn requires real innovation Hardware and software developers should plan that performance advances will emphasize BW over latency