• 沒有找到結果。

最後 祝你快樂得高分!   

N/A
N/A
Protected

Academic year: 2022

Share "最後 祝你快樂得高分!   "

Copied!
1
0
0

加載中.... (立即查看全文)

全文

(1)

Syllabus Computer Architecture (計算機結構) Jan.

2010

碩士班、資工四

--- Class Lecture Meeting Time: Every Tuesdays (週二)

SF305

(9:00AM -12:00NOON) Office Hours: Tuesdays 1:00PM-3:00PM (Email me for personal meeting) Instructor: Dr. Joseph Arul (周賜福) Email: arul@csie.fju.edu.tw

Homepage: http://www.csie.fju.edu.tw/~arul ( 周 賜 福 ) (All class materials, homeworks, and solutions will be posted on this web page. The Powerpoint presentation will be available as a pdf format. One can download the materials and print it out.)

Office: Room SF 615 聖言樓 Phone: 2905-3896 Course Objectives:

The purpose of this course is to provide you with a solid foundation in computer systems architecture and parallel computer organizations. This course is generally considered as a foundation to further study and research in computer systems. We will also survey several different approaches to parallel architectures, issues, trade-offs; Cache and pipeline issues. We will not go into significant depth; however the course will provide insights into these issues.

Prerequisite (Background): Digital Logic, Computer organization or equivalent. (If you want to refresh your knowledge on computer organization, you can also look at my notes on the web.) Primarily, I would like to see that you know how a basic CPU works (including instruction fetch, decode, execute cycles, microprogramming), instruction sets and choices (including RISC vs CISC, address modes), memory organization (including virtual memory, memory interleaving), some I/O (including interrupts, polling, daisy chaining, serial and parallel I/O), ALU (including multiplication, division and floating point algorithms), synchronization and mutual exclusion (like semaphores), some understanding of networks and protocols, some understanding of compilers and runtime support, Operating system concepts like process scheduling, virtual memory, protection domains, etc.

Text Book: Computer Architecture (4th Edition) “A Quantitative Approach”

By: John L. Hennessy and David A. Patterson (Chinese version or English version) (Third Edition on 64 bit addresses is also available.)

Other Useful books to refer:

1. K.Hwang, “Advanced Computer Architecture: Parallelism, Scalability, Programmability”.

2. K.Hwang and F.A Briggs, “Computer Architecture and Parallel Processing.”

Grading Policy: The final grade will be composed of the following weights. Make sure that you don’t copy the homework assignments. The Projects that you want to do must be written and given by OCT-13. Projects must be completed and submit a written report before the final exam and make a presentation in the class about your results and findings in the projects.

1

(2)

If you are absent more than 3 classes without permission, you are likely to fail in the course.

Class attendance will be taken either in the first hour or on the last hour of the class.

Activity Total Points

Homeworks (Few) 10 Pts

Quizzes or Project 15+10 Pts

Mid-term Exam 30 Pts.

Final Exam 35 Pts

Grand Total 100 Pts

You must receive at least 70 pts to pass the course (Grad students) 60 Pts for the Under Grad.

Ph.D students must do significant amount of work to pass the course.

Exams: Exams will cover lecture notes, textbook materials and homework materials.

Documented crisis or one week’s prior notice required for consideration of exam or projects submission and presentation make-up. Mid-term and Final exam will be closed book- open notes.

Homework Policy: A few home works will be assigned throughout the term. Home works are due at the start of the class on the due date. (Kindly note very carefully) After two weeks of the due date, homework will not be accepted. If you pass the due date and the time, 25%

will be detected. Copying of the homework will be penalized resulting in failing of the course.

Turn in Policy: Turn in all work to the instructor, preferably before class. Make sure that the work is done on your own without discussing with other classmates. Refer to the Academic Honesty below.

Absence and Participating Policy: Class attendance and effective, constructive participation is important to your performance in the course and makes up a portion of the grade based on the instructor’s discretion. If you cannot make it to the class on the date something is due, please make arrangements to have someone else bring your assignment to me before or at the start of the class. Also note the exam absence policy mentioned earlier. Don’t wait for the last day to cover up all the materials. It will be too hard.

Academic Honesty: Follow the University policy on academic honesty. The instructor’s academic honesty policy is very strict; instances of academic dishonesty will be penalized, ordinarily by at least a failure on the assignment and likely failure in the course (in addition to any University penalties). Unless otherwise stated, all work is to be individual work.

Violations of the individual work policy will be regarded as instances of academic dishonesty.

Beware of copying materials from the Internet (Don’t forget to cite the materials.) Reading assignments: Reading assignments will be informed after each class.

Computer Architecture Course Outline 1. Fundamentals of Computer Design

Defining Computer Architecture

2

(3)

Dependability

Trends in Cost and Power in Integrated Circuits.

2. Instruction Level Parallelism and Its Exploitation Concepts and Challenges

Basic Compiler Techniques for Exposing ILP Reducing branch costs with prediction.

Overcoming Data Hazards with Dynamic scheduling 3. Limits on Instruction Level parallelism

Studies on Limitations of ILP

Multithreading: Using ILP support to Exploit

Performance and Efficiency in advanced Issue Processors.

4. Multiprocessors and Thread Level parallelism Symmetric Shared Memory Multiprocessors

Distributed Shared Memory and Directory Based coherence Synchronization – The Basics

5. Memory Heirarchy Design

Eleven advanced Optimizations of Cache performance Memory Technology and Optimizations.

6. Storage Systems

Advanced topics in Disk Storage.

I/O performance, Reliability Measures, and Benchmarks.

Case studies

7. Dataflow architectures and its trend will also be covered.

Recent architecture developments.

Caution: Read the materials covered in this class every week and don’t postpone.

Project Hints

Projects: Obviously, the projects should be related to Computer systems or Parallel Processing Organizations. Preferably try to use some simulator that can be found in my web page. Try to do the project alone. If you need to do with someone else, you need my permission. Some topics that you can decide to work on are as follows:

1. You can investigate various cache design alternatives. There are several cache simulators available for your use. One such simulator is used by the authors of the textbook and is available for use. See the textbook and available for use. See the textbook on how to obtain the simulator. You can explore the impact of various cache parameters on the performance – somewhat similar to what the textbook talks about.

2. You can examine the various implementations of the Intel Pentium architecture by Intel, Cyrix, AMD and compare the features such as cache organization, pipeline structures,

3

(4)

instruction issues, branch prediction, etc. If possible show some benchmark comparisons and identify the contribution of these various features to the overall performance.

3. Similarly, you can look at modern RISC architecture (with dissimilar instruction sets) like DEC Alpha, HP-RISC, MIPS R-10000, Sun Solaris and Intel Pentium MMX and report on the various architectural features. Again, if benchmark results are available, you should try to report on the contribution of these features to the overall architecture.

4. Special purpose architecture designs such as dataflow architecture, non-blocking or blocking multithreaded architectures and so on.

5. Write a survey of various branch prediction techniques available. If possible define a taxonomy for the features, and if benchmark data is available, outline the performance effects of the various techniques.

最後 祝你快樂得高分!   

GOOD LUCK! Try to get above 90 pts in this course.

4

參考文獻

相關文件

bgez Branch on greater than or equal to zero bltzal Branch on less than zero and link. bgezal Branch on greter than or equal to zero

Look at all the words opposite and complete the following networks. Make two or three other networks to help you to learn the words on the opposite page. Match the adjectives on

a single instruction.. Thus, the operand can be modified before it can be modified before it is used. Useful for fast multipliation and dealing p g with lists, table and other

• 57 MMX instructions are defined to perform the parallel operations on multiple data elements packed into 64-bit data types. • These include add, subtract, multiply, compare ,

• 57 MMX instructions are defined to perform the parallel operations on multiple data elements packed into 64-bit data types.. • These include add, subtract, multiply, compare ,

• Non-vanishing Berry phase results from a non-analyticity in the electronic wave function as function of R.. • Non-vanishing Berry phase results from a non-analyticity in

Monopolies in synchronous distributed systems (Peleg 1998; Peleg

Corollary 13.3. For, if C is simple and lies in D, the function f is analytic at each point interior to and on C; so we apply the Cauchy-Goursat theorem directly. On the other hand,