Computer Science 252. Graduate Computer Architecture. (4 units)

New! Stuffit Archive of All Lectures in PowerPoint

Click here to download the 5 MByte bin hexed archive file. Expands to over 8 MBytes(!).

Catalog Description

Three hours of lecture and one hour of discussion per week. Prerequisites: CS 152. Graduate survey of contemporary computer organizations covering: early systems, CPU design, instruction sets, control, processors, busses, ALU, memory, pipelined computers, multiprocessors, and case studies. Term paper or project required.

Expanded Description

This course focuses on the techniques of quantitative analysis and evaluation of modern computing systems, such as the selection of appropriate benchmarks to reveal and compare the performance of alternative design choices in system design. The emphasis is on the major component subsystems of high performance computers: pipelining, instruction level parallelism, memory hierarchies, input/output, and network-oriented interconnections. A new focus will be the architectural design issues of highly portable, power-limited computing systems. Students will undertake a major computing system analysis and design project of their own choosing.

Course Grading

Homeworks (two person teams): 30%
Exams (two in-class midterms): 30%
Project (two person teams): 30%
Class Participation: 10%

Instructors, Fall 1995

Lecturer: Randy H. Katz, Professor

637 Soda Hall, 642-8778, randy@cs.Berkeley.edu
Office Hours MWF, 2 - 3 PM

Teaching Assistant: Giao Thanh Nguyen ("Yao")
Here is his picture.

Office: 330 Soda, gnguyen@cs.Berkeley.edu
Office Hours: T 3- 4 PM
Giao's CS 252 home page

Location

Lecture: MWF 1 - 2 PM, 310 Soda Hall
Discussion: T 2 - 3 PM, B1 Northgate; Th 3 - 4 PM, 343 LeConte

Textbook

J. L. Hennessy and D. A. Patterson, Computer Architecture: A Quantitative Approach, 2nd Edition (NEW!), Morgan Kaufmann Publishing Co., Menlo Park, CA.

Enrolled Student Snapshots

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Course Projects

An Evaluation of Correlation Based Branch Prediction on the Alpha Architecture, Alok Agrawal and Michael Chu
Power and Performance Tradeoffs in Microprocessor Cache Design, Jennie Chen and Bruce McGaughy
Optimizing the QR Eigensolver for the IBM Power2 Architecture, Tzu-yi Chen and Andrey Zege
Benchmarks for Graphics/Video Applications , Stephen Chenney and Alok Mittal
Latency Hiding in Uniprocessors using Multithreading , Brent Chun and Franklin Cho
Branching on Superscalar Machines: Speculative Execution of Multiple Branches, Richard Fromm and Bassam Tabbara
Accelerating the RISC Processor Using Programmable Logic, Sriram Rajamani and Pramod Viswanath
Reducing Power Consumption for the Next Generation of PDAs: It's the Network Interface!, Mark Stemm, Paul Gauthier, Daishi Harada
Harware/Software Architectures for TCP/IP Acceleration of UNIX Workstations, Roy Sutton and Sameer Jalnapurkar
A Comparative Analysis of Branch Prediction Schemes, Zhendong Su and Min Zhou
Slow Fourier Transforms on Fast Microprocessors, Taku Tokuyasu and Bernt Pfromm
Compressed Reduced Instruction Set COmputing (CRISCO), Patrick Warner and Geroncio Galicia
Instruction Level Power Analysis of the ARM60, Anna Reznik

Tentative Course Lecture Plan

Week 1 (28 August - 1 September)

Lecture 1: Course Introduction and Overview; See some images of the first day's crush of students by clicking here!
Lecture 2: Technology Trends;
Lecture 3: Amdahl's Law, Rules of Thumbs, Performance and CPI Formula;

Week 2 (6 September - 8 September)*

Lecture 4: Benchmarks and Performance Metrics;
Lecture 5: Cost, Price, and Price for Performance;

Week 3 (11 September - 15 September)

Lecture 6: Instruction Set Architecture and the 80x86;
Lecture 7: Introduction to Pipelining, Structural Hazards, Forwarding;
Lecture 8: Pipelining, Control Hazards, Static Branch Prediction, Interrupts, MIPS R4000;

Week 4 (18 September - 22 September)

Lecture 9: Instruction Level Parallelism and Advanced Pipelining (Scoreboards);
Lecture 10: Case Study: CDC 6600 Scoreboard;
Lecture 11: Case Study: Tomasulo Algorithm;

Week 5 (25 September - 29 September)

Lecture 12: Dynamic Branch Prediction, Superscalar, VLIW, Software Pipelining;
Lecture 13: Trace Schedule, Conditional Execution, Speculation, Limits of ILP;
Lecture 14: Midterm Review and Catch-Up

Week 6 (2 October - 6 October)

FIRST MIDTERM EXAMINATION
Lecture 15: Memory Hierarchy, Motivation, Definitions, Four Questions about Memory Hierarchy;
Lecture 16: Memory Hierarchy Misses, 3 Cs and 7 Ways to Reduce Misses;

Week 7 (9 October - 13 October)

Lecture 17: Memory Hierarchy: 5 Ways to Reduce Miss Penalty (including Second Level Cache);
Lecture 18: Memory Hierarchy: Main Memory and Enhancing its Performance;
Lecture 19: DRAM-Specific Memory Organizations, Virtual Memory, Alpha 21064 Memory Hierarchy and Performance, Fallacies;

Week 8 (16 October - 20 October)

Lecture 20: What Every Computer Architect Should Know About Computer Arithmetic (Guest Lecturer Israel Koren)
Lecture 21: I/O, Storage Devices, Metrics, and Productivity;
Lecture 22: I/O, A Little Queuing Theory, and I/O Interfaces;

Week 9 (23 October - 27 October)

Lecture 23: I/O, I/O Busses;
Lecture 24: I/O, RAID;
Lecture 25: I/O, ABCs of File Systems, I/O Benchmarks;

Week 10 (30 October - 3 November)

Lecture 26: I/O, UNIX File System Performance on Workstations and Mainframes;
Lecture 27: I/O, Tertiary Storage Systems;
Lecture 28: IEEE Floating Point Standard (Guest Lecturer Israel Koren)

Week 11 (6 November - 10 November)

Lecture 29: Networks, Intro, Implementation and Performance Issues;
Lecture 30: Networks, Architectural Issues including Topologies, Practical Issues and Protocols;
Lecture 31: Multiprocessors, Flynn Categories, Small vs. Large Scales, Cache Coherency;

Week 12 (13 November - 17 November)

Lecture 32: Multiprocessors, Snoopy Caches and Directory Schemes;
Lecture 33: Midterm Review and Catch-up
SECOND MIDTERM EXAM

Week 13 (20 November - 22 November)*

Lecture 34: Multiprocessors, Performance Snoopy Cache vs. Directories;
Lecture 35: Project Status Reports

Week 14 (27 November - 1 December)

Lecture 36: Multiprocessors, Synchronization and Consistency Models;
Lecture 37: Portable Computers, Technology Underpinings;
Lecture 38: Portable Computers, Low Power Processor Issues;

Week 15 (4 December - 8 December)

Lecture 39: Portable Computers, Processor Case Studies (AT&T Hobbit, Acorn RISC Machine, Low Power Power PC);
Lecture 40: Course Summary, Future Predictions
Lecture 41: Project Poster Session

Handouts

Handout 0: Background Questionaire
Handout 1: Homework Assignment #1, Due 8 September 95
Handout 2: Project Suggestions from Previous Semesters
Handout 3: Project Suggestions Fall 1995
Handout 4: Homework Assignment #2, Due 22 September
Handout 5: Project Survey #1, Due 15 September
Handout 6: Result of Project Survey, as of 18 September
Handout 7: Project Survey #2, Due 29 September
Handout 8: Result of Project Survey #2, as of 2 October
Handout 9: Homework Assignment #3, Due 13 October
Handout 10: Project Checkpoint #1, Due 13 October
Handout 11: Homework Assignment #4, Due 30 October
Handout 12: Project Checkpoint #2, Due 17 November
Handout 13: Final Project Report Specification