EECE 5183/6083: Compiler Theory/Compiler Laboratory

I look forward to working with you this spring semester. If we all work together, we can make this a fruitful and productive semester. I am hopeful that this will be a great experience for all of us. In this class you will be required to attend and participate in classroom discussions and write a compiler. There are no other homework or exams scheduled. I have outlined a target set of dates for the completion, however other than the final deadlines for the compiler (outlined below), these are merely suggested completion dates that you are advised to target. Good luck. I tend to keep my Tues/Thurs schedule open for student meetings and I work research on MWF so it is much harder to get me to respond on MWF. That said, please feel free to approach me on M/W/F. Just be advised that when i get deep into my research I can be somewhat less than friendly. Please do not take this personal, its just how i work. You can also learn more about my weekly schedule from my google calendar (follow the My Schedule; link in the navigation panel of my dept webpage).

I may occasionally post an announcement on the class canvas webpages. Otherwise, everything besides lectures that you will need for this class is available from my university webpages.

Planning

This course requires strong programming skills. If you are not a strong programmer that is able to build, debug, and manipulate complex data structures, you should strongly consider looking for an alternate course. I will not be helping you with the programming of your compiler solution (beyond discussing the general techniques for implementing a compiler).

• Syllabus
• Schedule
• Possible Topics for Student Lectures

Textbook Readings

1. Chapters 1 & 2: (quick read/skim)
2. Chapter 3: Good overview of lexical analysis
3. Chapter 4.3.3: Eliminating Left Recursion
4. Chapter 4.4: Recursive Decent Parsing
5. Chapter 2.7: Symbol Table Mgmt
6. Chapter 2.8: (and specifically 2.8.4) Code Generation

Class Notes

• Introduction
• Overview of Compilers
• Lexical Analysis
• Building Your Compiler: Lexical Analysis
  • Building a Lexical Analyzer (Chapter 3, open lectures)
• Parsing Intro
• Parsing: LL
• Building Your Compiler: Parsing
  • The following grammars can be input input into the corresponding LL parser generator tool. You might find it useful to study the output from these tools to get additional examples on code that implements a recursive decent parser.
  • Simple set_of_stmts grammar for lola; the output of this is somewhat easier to read than that produced byu antlr4
  • Simple set_of_stmts grammar for antlr4
• Building Your Compiler: Symbol Tables
• Building Your Compiler: Type Checking
• LLVM Overview
  • Here is a nice overview of LLVM.
  • LLVM. The Documentation link will take you to all kinds of documentation. Do not get overwhelmed and try to read everything. The LLVM website also contains an LLVM Tutorial
  • A complete guild to LLVM for programming language creators (another nice tutorial)
  • While the LLVM project does not have general tooling to build frontend analysis, the Clang frontend seems to have gained a some amount of traction for building frontend support multiple languages. Here are some interesting links that might help you learn more about clang.
    • Clang -- An Introduction
    • Understanding the Clang AST.
    • Clang Documentation
  • Projects built with LLVM
• LLVM Assembly
  • The raw LLVM assembly language and the llc assembler (the llvm tool chain also provides an interpreter for its assembly language called lli that you might want to use).
  • A very good gentle introduction
  • A Quick Primer on LLVM Assembly
  • Here is a compiler output explorer. Be careful here, the webpage records your session/configuration and will hold it between invocations. First: The right panel starts with gcc, use the pull-down on it to change to one of the clang selections. Second: In the next row down, selection (Add new->LLVM IR output). Now you should see three columns of code. Don't kill the assembly output column (if you do you will also lose the LLVM output and have to reset the page. Fortunately, if you screw everything up, there is a reset at the very top (More->Reset code and UI layout).
  • Here is another LLVM assembly/IR generator (this is a python package that you have to install/use).
• LLVM IR
  • This section of the Tutorial might be helpful if you want to go with an in-memory translation.
  • Getting started with the LLVM C API
  • A Complete Guide to LLVM for Programming Language Creators (this actually talks about many parts of the LLVM infrastructure including the Assembly Language and the Optimization subsystem)
• Building Your Compiler: A Restricted C for code-generation
• Building Your Compiler: Code Generation

Student Assessment and Grading

• Students must participate in organizing and delivering lectures on an approved topic (generally in the area of compiler tools or compiler issues); failure to do so will result in a grade reduction of one half of a letter grade.
• Grades will be established by project milestones completed.
  • Scanner/Parser: D
  • Semantic Analysis: C
  • Coded Generation: B
  • Runtime System: A
• The project compiler must pass all test programs generating errors for the incorrect programs and correctly executing code for the others. Meaningful errors must be emitted for the incorrect programs and reasonable parser re-synchronization must be achieved for the erroneous programs.
• Projects are normally due by the last day of finals week. However, a one semester extension is granted without penalty. This extension only goes to the last day of classes of the summer semester. You can incrementally turn in projects for grades beginning at the end of this quarter and improve it thereafter (as per the deadlines/penalties above).

Project Description

• Project Description
• Project Input Language Specification
• Test programs (extract by "tar -xzf testPgms.tgz")