Previous Lecture Slides for Lecture 2 ENCM 501: Principles of - PDF document

slide 2/22 ENCM 501 W14 Slides for Lecture 2 Previous Lecture Slides for Lecture 2 ENCM 501: Principles of Computer Architecture Winter 2014 Term ◮ introduction to ENCM 501 Steve Norman, PhD, PEng ◮ course organization Electrical & Computer Engineering ◮ review of computer organization basics Schulich School of Engineering University of Calgary 14 January, 2014 ENCM 501 W14 Slides for Lecture 2 slide 3/22 ENCM 501 W14 Slides for Lecture 2 slide 4/22 Today’s Lecture ENCM 501 Course Topics ◮ introduction to computer system design goals and performance measurement (textbook, Chapter 1) ◮ brief list of ENCM 501 topics ◮ brief overview of ISA principles (parts of Appendix A) ◮ what does “computer architecture” mean? ◮ memory system design and performance assessment ◮ trends in computer system performance (parts of Appendix B and Chapter 2) ◮ classes of computers ◮ aspects of instruction-level parallelism (parts of ◮ trends in computer technology Appendix C and Chapter 3) ◮ preliminaries for energy and power use ◮ aspects of thread-level parallelism (TLP) (parts of Related material in Hennessy & Patterson (our course Chapter 5) textbook): Sections 1.1–1.5. ◮ introduction to programming with TLP (not covered in textbook) slide 5/22 slide 6/22 ENCM 501 W14 Slides for Lecture 2 ENCM 501 W14 Slides for Lecture 2 What does “Computer Architecture” mean? (1) What does “Computer Architecture” mean? (2) It is surprisingly hard to come up with a simple, short Now let’s move a down one or two levels of abstraction . . . definition of computer architecture . It’s kind of an “umbrella” ◮ Given the ISA, how exactly are instructions handled by term that includes a bunch of related ideas and activities. processors—how deep are pipelines; can instructions be Let’s start at the level of instructions . . . executed out-of-order? ◮ What instructions are available to applications How is the memory system organized to minimize loss of programmers? This is often called instruction set clock cycles in fetching instructions and reading and architecture , or ISA . writing data? This category of concern is sometimes called ◮ What additional instructions are provided to operating microarchitecture or organization . system kernel programmers? (Examples: Instructions to ◮ Given a microarchitecture, what are good ways to query system state when an interrupt occurs, to manage virtual memory hardware, to control I/O devices, and so implement it at the integrated circuit and printed circuit on.) board levels? These are hardware design problems.

slide 7/22 slide 8/22 ENCM 501 W14 Slides for Lecture 2 ENCM 501 W14 Slides for Lecture 2 It’s good to have a broad perspective on Trends in computer system performance (1) architecture The next slide shows a plot of “benchmark” performance Obviously, ISA choice dictates much about microarchitecture, scores for various computers, showing the years various and microarchitecture dictates much about hardware. systems were introduced. But the influences also work in the opposite direction, from “Performance” here means roughly 1 over the time taken to lower to higher levels of abstraction. complete a collection of processor-intensive tasks. (We’ll look Cost of fabrication (a hardware issue) makes some much more carefully at performance measurement in future microarchitectures attractive and others less attractive. lectures.) Physical size of components may also matter. (The text on the plot will be pretty much illegible in the Aspects of microarchitecture matter when a new ISA is classroom, but we can still make a few important points by designed or an existing ISA is extended. Preference for looking at it.) relatively simple, clean microarchitecture might rule out some useful instructions. ENCM 501 W14 Slides for Lecture 2 slide 9/22 ENCM 501 W14 Slides for Lecture 2 slide 10/22 Trends in computer system performance (2) 100,000 Intel Xeon 6 cores, 3.3 GHz (boost to 3.6 GHz) Intel Xeon 4 cores, 3.3 GHz (boost to 3.6 GHz) Intel Core i7 Extreme 4 cores 3.2 GHz (boost to 3.5 GHz) 24,129 Performance ratio, 2010 compared to 1978: about 24,000 to 1. In Intel Core Duo Extreme 2 cores, 3.0 GHz 21,871 19,484 Intel Core 2 Extreme 2 cores, 2.9 GHz 10,000 14,387 AMD Athlon 64, 2.8 GHz 11,865 other words, what took about 7 hours in 1978 took about 1 second AMD Athlon, 2.6 GHz Intel Xeon EE 3.2 GHz 6,043 6,681 7,108 Performance (vs. VAX-11/780) Intel D850EMVR motherboard (3.06 GHz, Pentium 4 processor with Hyper-Threading Technology) 4,195 in 2010. IBM Power4, 1.3 GHz 3,016 Intel VC820 motherboard, 1.0 GHz Pentium III processor 1,779 Professional Workstation XP1000, 667 MHz 21264A 1,267 Digital AlphaServer 8400 6/575, 575 MHz 21264 1000 993 From 1986 to 2003, the average annual performance improvement AlphaServer 4000 5/600, 600 MHz 21164 649 Digital Alphastation 5/500, 500 MHz 481 was 52% per year. Digital Alphastation 5/300, 300 MHz 280 22%/year Digital Alphastation 4/266, 266 MHz 183 IBM POWERstation 100, 150 MHz 100 117 From 2003 to 2010, the average annual performance improvement Digital 3000 AXP/500, 150 MHz 80 HP 9000/750, 66 MHz 51 was 22% per year—the pace of improvement has slowed in recent 52%/year IBM RS6000/540, 30 MHz 24 MIPS M2000, 25 MHz 18 MIPS M/120, 16.7 MHz years. 13 10 Sun-4/260, 16.7 MHz 9 VAX 8700, 22 MHz 5 AX-11/780, 5 MHz There have been comparable improvements in telecommunication 25%/year 1.5, VAX-11/785 bandwidth and data storage capacity. 1 1 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 The result is a pattern that has been seen over and over: Computer Image is Figure 1.1 from Hennessy J. L. and Patterson D. A., Computer applications go from impossible to practically unaffordable to Architecture: A Quantitative Approach, 5nd ed. , c � 2012, Elsevier, Inc. cheap and commonplace over periods of several years. slide 11/22 slide 12/22 ENCM 501 W14 Slides for Lecture 2 ENCM 501 W14 Slides for Lecture 2 Classes of computer (1) Classes of computer (2) In Section 1.2, Hennessy and Patterson divide computer systems into five classes. Knowing what they mean will help in following the textbook! Also in Section 1.2, Hennessy and Patterson make some ◮ personal mobile device (PMD): things like smartphones distinctions between various kinds of parallelism in hardware design. and tablets. ◮ desktop: what most of us would call “desktops”, and You can read that material to get a general idea about the also laptops . This is a somewhat unusual definition, but diverse forms of parallel computation, but we won’t worry makes sense as use cases and requirements are broadly about the details until much later in the course. similar. A good “takeaway”: If somebody tells you in a vague way that ◮ servers an algorithm uses parallel processing, you should ask, What ◮ clusters and warehouse-scale computers : systems large kind of parallel processing? enough to support operations like Google, Amazon, etc. ◮ embedded computers : computers built-in to machines such as appliances, cars, telecom infrastructure.