Multithreaded Parallelism: Languages and Compilers
Languages and compilers for multithreaded parallelism: pH, the lambda calculus, type systems, monadic I/O, Bluespec and term rewriting.
Writing programs that run many threads at once is hard when the programmer must manage every detail by hand. This course studies languages and compilers that expose parallelism implicitly, through functional programming in pH, a parallel dialect of Haskell. It covers the lambda calculus and its variants, type systems, higher-order functions and non-strictness, then adds side effects and nondeterminism, turns to the Bluespec hardware design language, and ends with term rewriting and operational semantics.
What You Will Achieve
- Write implicitly parallel programs in a functional language
- Reason about programs with the lambda calculus and its variants
- Explain Hindley-Milner typing, polymorphism and overloading
- Handle side effects with M-structures and monadic I/O
Program Overview
Programming in pH and the Lambda Calculus
Implicit parallelism and its formal basis.
- Expressing parallel computation
- The lambda calculus with constants and let bindings
- Hindley-Milner types, polymorphism and overloading
- Algebraic types, arrays and I-structures
I/O, Side-effects and Nondeterminism
Adding state without losing parallelism.
- M-structures
- The lambda-S calculus with side effects
- Monads and monadic I/O
Bluespec and Term Rewriting
A hardware design language and operational semantics.
- Bluespec for hardware design, simulation and synthesis
- Modules and type classes
- Term rewriting systems
Course Completion
Job Outlook
These ideas inform work on programming languages, compilers, parallel systems and hardware design languages.
Good for: Computer science students interested in functional languages, type systems and compiling parallel code.
Keep in mind: A fall 2002 course with lecture notes, four problem sets, midterm topics and project suggestions; the assignments use pH and Hugs, and some files came from MIT’s Athena computing environment.
Bottom line: A deep, theory-rich course on implicit parallelism that still teaches durable foundations.
Frequently Asked Questions
Which book does it use?
Implicit Parallel Programming in pH by Rishiyur S. Nikhil and Arvind, with readings assigned chapter by chapter.
Which languages do the assignments use?
pH, described as parallel Haskell, and Hugs. The supporting files for the problem sets are posted with them.
What kind of project do students do?
Yes. Students completed a project, and a list of suggested project topics is posted with the course materials.



There are no reviews yet.