91.406 / 91.534 - Compiler Construction I

Fall 2002

Lecture: Wednesday and Friday, 2:30-3:45 in Olsen 415

Class web page (this page):: http://www.cs.uml.edu/~dimock/courses/compilers/Fall2002/index.html
Class discussion board:: http://teaching.cs.uml.edu/cgi-bin/ikonboard/ikonboard.cgi
Class Discussion Boards > Prof. Dimock's Courses > 91.406/91.534-Fall2002
(When you first get to the board, you need to register: In the line "ť Welcome Guest Log In :: Register" in the upper left of the first screen, clock on Register. Or further down, see "Not Registered? Register Now!" and click on Register Now.)

Term Project

The term project is to implement parts of a compiler for a subset of the Java language. The compiler will produce code for the MIPS-II instruction set architecture. The term project is based on the Minijava project, with some differences documented here.

Syllabus and schedule

Assignments

Homework 5. (Ungraded: please review before exam). Get it here in ps format, or here in pdf format.

Project 8. Simple machine code generation: Kanga to MIPS. Assignment date: 4-Dec. Due date: 12-Dec.
Read assignment in ps format here, or in pdf format here. Project files are here.
Documentation for SPIM may be found online: James Larus's Assemblers, Linkers, and the SPIM simulator (pdf). This is Appendix A of Hennessy & Patterson's Computer Organization and Design: The Hardware/Software Interface, published by Morgan Kaufman. It includes documentation on the MIPS R2000, its assembly code, and the SPIM simulator.
A briefer version of SPIM documentation is available in pdf or ps format, and a slightly revised version is available in html format.

Homework 4. Assigned 27-Nov. Due 6-Dec in class before 2:35PM. Get it here in ps format, or here in pdf format.

Project 7. Register allocation and translation from SPiglet to Kanga. Assignment date: 22-Nov. Due date: 3-Dec.
Read assignment in ps format here, or in pdf format here. Project files are here.

Homework 3. Assigned 20-Nov. Due 27-Nov in class before 2:35PM. Get it here in ps format, or here in pdf format.

Project 6. Liveness analysis. Assigned 13-Nov. Due Thursday, 21-Nov before midnight.
Read assignment in ps format here, or in pdf format here. Project files are here.
Project 5. Fill in missing cases in spggen.c, a simplifier for Piglet. Assigned 1-Nov. Due Tuesday, 12-Nov before midnight.
Read assignment in ps format here, or in pdf format here. Project files are here. The actual description of the SPiglet subset of the Piglet language and the simplification process is here, or in pdf format here. The document from Project 3 on the input and output formats of the ast program may be useful in understanding piglet.cg: See below.
Project 4. Fill in missing cases in pggen.c, a translator from Minijava to Piglet. Assigned 23-Oct. Due Thursday, 31-Oct before midnight.
Read the project description in ps format here, or in pdf format here. Project files are here. The document "Minijava Code Patterns in Piglet" by Nigel Horspool is the best reference on the Piglet language. Get a copy here, or in pdf format here. Other references on Piglet are the original description from the Minijava project, here. The Minijava project's grammar for Piglet may be useful in addition to piglet.cg. The document from Project 3 on the input and output formats of the ast program may be useful in understanding mj.cg and piglet.cg: See below.
Project 3. Fill in missing cases in type checker for Minijava. Assigned 9-Oct. Due Monday, 21-Oct before midnight.
Read the project description in ps format here, or in pdf format here. You will also need a copy of the typing rules for Minijava: here, or in pdf format here; and a copy of instructions for understanding the input and outputs of the ast program: here, or in pdf format here.
Homework 2. Assigned 25-Sep. Due 2-Oct in class before 2:35PM. Get it here in ps format, or here in pdf format.
Project 2. Modify parser for Minijava. Assigned 25-Sep. Due Monday, 7-Oct before midnight. (Note change from 3-Oct).
Read the project description in ps format here, or in pdf format here. Rules are the same as for the lexical analyzer art of the project (below).
Homework 1. Assigned 13-Sep. Due 20-Sep in class before 2:35PM. Get it here in ps format, or here in pdf format.

Project 1. Lexical analyzer for Minijava. Assigned 13-Sep. Due 24-Sep before midnight.
Read the project description, rules, and background in ps format here, or in pdf format here. Template for lexer and support code are available here. Testing on programs in the JavaPrograms directory should yield the results in the Project1/results directory.

Turning In Assignments

Written homework is generally due a week after being assigned. Written homework may be handed in class before 2:35PM on the due date. If you can't get to class for some reason, homework can be handed in by pushing it under the door of the instructor's office (Olsen 215). Homework is due in the office before the instructor leaves for class. Late homework may be penalized. Homework more than a few days late may not be accepted at all.
Project portions are normally due in the evening. Project portions are submitted using the "submit" command on saturn.cs.uml.edu or venus.cs.uml.edu.
All code for the term project should compile without errors, warnings, or suggestions when compiled with the gcc compiler using the -pedantic and -Wall flags.
You may develop code using any C compiler that you like, but check that the code using gcc -pedantic -Wall before turning it in.
You will work in pairs on the programming assignments. When you code in pairs, you should work together: both people are involved in creating all the code as per extreme programming pairs.
Written homework is to be done individually, without reference to any other student's work, or answer sheet, from this year or from any other year. Developing in pairs is to be done without reference to any other pair's work. References to the literature should be credited. If you need help or have an idea to share with other students, please use the bulletin board. Cheating may result in a failing course grade.
Late assignments: 20% off any time after the deadline to 24 hours after the deadline. An additional 20% off for each subsequent 24 hour period. Assignments will not be accepted after answers have been handed out. In the case of project portions, there is little time before the answer to the last portion is made available for use in the next portion.

Grading

Grading breakout:

Resources

Nonstandard software is found here. Instructions for installing the software on your own linux box are found here: there may be bugs in the procedure, you have been warned (but please report any that you find).

Books

The course textbook is Compilers: Principles, Techniques, and Tools, by A. V. Aho, R. Sethi, and J. D. Ullman. The book is available from the UMass Lowell Bookstore.

Class

The lectures will go beyond the contents of the textbook to cover some tools and techniques that have come into common use since the textbook was last revised.

Manuals

The Flex lexical analyzer generator is documented here.

here

The Gperf perfect hash generator is documented here. (You will not have to use gperf, but it is used for some automatically generated code that is given to you for the term project.)
The Ast abstract syntax tree generator from CoCoLab is documented here (.ps). (You will not have to use ast, but you will be making heavy use of code generated by ast.)
We will be using some standard data structures (atoms, tables, bitmaps, ...). To avoid the coding overhead of writing these, we will use imlementations by David Hanson, described in C Interfaces and Implementations. The quick reference guide (pdf) is somewhat terse, since the author would presumably like to sell you the book. However, you should be able to understand the interfaces from the quick reference guide.
SPIM is a simulator written by James Larus for the MIPS-II instruction set architecture.
Documentation for SPIM may be found online: James Larus's Assemblers, Linkers, and the SPIM simulator (pdf). This is Appendix A of Hennessy & Patterson's Computer Organization and Design: The Hardware/Software Interface, published by Morgan Kaufman. It includes documentation on the MIPS R2000, its assembly code, and the SPIM simulator. There is a slightly revised version of the SPIM documentation in html here.

General questions

Use the class discussion board.

Handouts

11-Sep-02: Regular grammar, Regular Expression, and C code for Java floating-point number .pdf.gz.
18-Sep-02: Some slides on grammars and LL(1) recognition .pdf.gz or .ps.gz
??-Sep-02: Some slides on LR(1) recognition .pdf.gz or .ps.gz
06-Nov-02: Some notes on register allocation, not slides yet .pdf.gz or .ps.gz
06-Nov-02: Hosking's notes on graph-coloring register allocation .pdf.gz or .ps.gz
06-Nov-02: "Linear Scan Register Allocation" by Massimiliano Poletto and Vivek Sarkar. TOPLAS 21(5) Sep. 1999, pp. 895-913 .pdf.gz or .ps.gz
15-Nov-02: "The Design and Application of a Retargetable Peephole Optimizer" by Jack W. Davidson and Christopher W. Fraser. TOPLAS 2(2) Apr. 1980, pp. 191-202 .pdf.gz or .ps.gz
15-Nov-02: "Code Selection through Object Code Optimization" by Jack W. Davidson and Christopher W. Fraser. TOPLAS 6(4) Oct. 1984, pp 505-526. .pdf.gz or .ps.gz
06-Dec-02: Slides on instruction scheduling .ps.

Office hours

here