DF00100 Advanced Compiler Construction

(PhD course, up to 9hp, early autumn 2026)

Goals
Give Ph.D. students or practitioners knowledge about advanced compiler technology, including program analysis, intermediate representations, compiler optimizations, code generation, compiler frameworks, run-time systems, and compilation techniques for parallel and embedded systems.

Prerequisites
Basic course in compiler construction, corresponding to the undergraduate courses Compilers and Interpreters or Compiler Construction.
Basic course in data structures and algorithms.
Basic knowledge in processor architecture.
For the labs: Programming in C++/Linux.

Contents and schedule

Lectures, lessons, lab introduction: 2 intensive weeks (mostly in september 2026) in Linköping.
The course will be given on-site, but some moments (lectures, lessons) may also be transmitted on zoom for external participants. The student presentations and exam will be on-site in Linköping.

The preliminary schedule is given below. Slides are currently still as of 2023 and may be updated shortly before/after each lecture.

Date	Time	Location	Topic (with links to slides (PS/PDF) where available)	Lecturer
Week 37 - Focus on Program Analysis and High-Level Optimizations
Tuesday 09/09/2025	13:15-15:00	Knuth / Zoom	Opening remarks (PDF, updated for 2023). Introduction: Multi-level intermediate representations. Principles of code generation, stack organisation and call sequences, local common subexpression elimination, DAGs, lowering. Short introduction to SSA form. Survey of some compiler frameworks. (PDF, updated for 2023)	Christoph Kessler
09/9/2025	16:00-17:00	Knuth / Zoom	Control flow analysis (PDF, changed to 1x1 format)	Christoph Kessler
10/9/2025	09:15-10:00	Zoom	Lab introduction: LLVM Compiler Framework. Lab part 1, Lab part 2	TBA
Wednesday 10/9/2025	10:15-12:00	Knuth / Zoom	Control flow analysis (cont.) / Data flow analysis (PDF)	Christoph Kessler
Wednesday 10/9/2025	13:15-17:00	Knuth / Zoom	Dependence analysis, loop transformations, loop parallelization (PDF, upd.)	Christoph Kessler
Thursday 11/9/2025	11:00-12:00	Knuth / Zoom	Lesson 1: Program representation, control and data flow analysis	Christoph Kessler
11/9/2025	13:15-17:00	Knuth / Zoom	Abstract interpretation. (PDF)	Welf Löwe
Friday 12/9/2025	09:15-12:00	Knuth / Zoom	Interprocedural Analysis. Points-to Analysis. (PDF)	Welf Löwe
Week 38 - Focus on SSA, Low-level Optimizations and Code Generation
Monday 15/9/2025	09:15-10:15	John von Neumann	SaS seminar (not part of the course, but possibly of interest for some attendants): Automated Graph Generation for the Assurance of Software-intensive Cyber-Physical Systems	Daniel Varro
Monday 15/9/2023	13:15-17:00	Knuth / Zoom	Instruction selection (PDF) Register allocation (PDF)	Christoph Kessler
Tuesday 16/9/2025	09:15-12:00	Knuth / Zoom	Instruction scheduling (PDF). Side note: Space-optimal scheduling for trees (PDF, for self-study)	Christoph Kessler
16/9/2025	13:15-15:00	Knuth / Zoom	Lesson 2: Dependence analysis and loop transformations	Christoph Kessler
16/9/2025	15:15-17:00	Knuth / Zoom	Software pipelining (PDF)	Christoph Kessler
Wednesday 17/9/2025	09:15-11:00	Knuth / Zoom	Software pipelining (cont.) Auto-tuning (PDF) Code generation for embedded processors, DSP processors, clustered VLIW architectures. (PDF, only partly covered). Not lectured in 2023, for reference only: Integrated code generation (PDF) Compilation for actors and stream computing: task fusion, scheduling, mapping, DVFS (MP4 recorded lecture 25min, watch off-line) Data transfer fusion optimization (PDF);	Christoph Kessler
17/9/2025	11:15-12:00	Knuth / Zoom	Lesson 3: Instruction scheduling, software pipelining	Christoph Kessler
17/9/2025	13:15-15:00	Knuth / Zoom	Guest lecture: EMCA C Compiler, an LLVM based custom compiler at Ericsson	Mattias Eriksson
17/9/2025	15:15-17:00	Knuth / Zoom	Guest lecture: SkePU precompiler: Experiences with using the LLVM-clang compiler framework for source-to-source translation (PDF)	August Ernstsson
Thursday 18/9/2025	09:15-11:00	Knuth / Zoom	Guest lecture: The Open-source Modelica Compiler (PDF)	Martin Sjölund
18/9/2025	11:15-12:00	Knuth / Zoom	Assignment of papers for student presentations.	Christoph Kessler
18/9/2025	13:15-17:00	von Neumann / Zoom	SSA, construction and destruction, Memory SSA. SSA-based optimizations. (PDF, upd.)	Welf Löwe
Friday 19/9/2025	09:15-12:00	Knuth / Zoom	More SSA based optimizations; Chi functions in lazy memory SSA based analysis (PDF)	Welf Löwe
Week TBA:
TBA	09:00-17:00	Donald Knuth	Student presentations
Week TBA:
TBA	09:00-13:00	John von Neumann	Written exam (TEN1)	Christoph Kessler

Organization
Lectures, lessons, compiler framework programming labs, and student presentations of compiler research papers.

Lectures are, where possible, given in block format in 2 intensive weeks, usually 09:15-12:00 and 13:15-17:00 every day.

Lessons are problem solving sessions, discussing exercises to complement the lectures. The exercise sets are available here.

Lab series (3hp) using the LLVM compiler system.
Lab introduction (2021)
Lab instructions, part 1 (2021)
Lab instructions, part 2 (2021).
The lab assistant will supervise labs and correct lab reports.

The presentation part of the course consists of student presentations of about 30 minutes each (25 minutes presentation plus 5 minutes questions) on a recent paper in compiler technology, with a written summary of 1-2 pages. It also includes opposition of one other presentation.
See this link for detailed instructions and the list of selected/assigned student papers.
For admission to the presentation it is required that at least 50% of the lectures and lessons have been attended.

Written/oral exam: Thursday 27/4/2023, 09:00-13:00, room: von Neumann.

No aids are allowed, except for a dictionary from English to your native language.
For admission to the exam it is required that at least 50% of the lectures and lessons have been attended.

Old exams:
Part 1 (PDF), Part 2 (PDF), Solution proposal part 2 (PDF)

Literature
You should be able to follow the course mainly based on the slide material only. The following books and papers are recommended as additional reading wherever necessary:

Books:

No textbook covers the entire course contents completely. But parts of either of the following books and articles can be useful as accompanying text for a major part of the course (these are also available in the local library):

• Alfred Aho, Monica Lam, Ravi Sethi, Jeffrey Ullman: Compilers: Principles, Techniques, and Tools. Second edition, Addison-Wesley, 2006.
  The course book of the undergraduate compiler course. If you already have it, use it.
• Steven Muchnick: Advanced Compiler Design and Implementation. Morgan Kaufmann, 1997. (Quite expensive, but recommended reading. The LiU TekNat library has 2 copies.)
  Errata for early printings
• alternatively: Keith Cooper, Linda Torczon: Engineering a Compiler. Morgan Kaufmann, 2003.
• Y.N. Srikant, P. Shankar (ed.): The compiler design handbook: optimizations and machine code generation, CRC Press, 2003.
  (Available in the LiU TekNat library, also the first edition..
  More complete and more up to date than Muchnick's book, but a collection of independent chapters rather than a monograph, and expensive.)
  In particular, chapters 5 (Optimizations for memory hierarchy), 7 (Energy-aware compiler optimizations), 11 (SSA form), 16 (ADLs for retargetable compiling), 17 (Instruction selection), 18 (DSP VLIW compiler), 19 (Instruction scheduling), 20 (Software pipelining). The first edition contains a chapter on dataflow analysis.

• C. Kessler: Compiling for VLIW DSPs. Book chapter in the Handbook of Signal Processing Systems, Third edition, 2019. Preprint, 38 pages, handed out to registered students.
• Matt Godbolt: Optimizations in C++ Compilers. Comm. ACM 63(2), Feb. 2020.
• David Bacon, Susan Graham, Oliver Sharp: Compiler Transformations for High-Performance Computing. ACM Computing Surveys, December 1994, Volume 26 Issue 4.
• David A. Padua and Michael J. Wolfe:
  Advanced compiler optimizations for supercomputers.
  Comm. ACM 29(12), Dec. 1986.
• Vicki Allan et al.: Software Pipelining. ACM Computing Surveys 27(3): 367-432, Sept. 1995.
  (Chapter 18 of Srikant's book contains a short version of this article.)
• Christoph Kessler, Sebastian Litzinger, Jörg Keller: Static Scheduling of Moldable Streaming Tasks with Task Fusion for Parallel Systems with DVFS. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems (TCAD) 39(11), pages 4166-4178, Nov. 2020. DOI: 10.1109/TCAD.2020.3013054
  
  • Video presentation from EMSOFT'20 (25min, .mp4)
• Christoph Kessler: Global Optimization of Operand Transfer Fusion in Heterogeneous Computing. Proc. 22nd International Workshop on Software and Compilers for Embedded Systems (SCOPES-2019), St. Goar, Germany, May 2019. ACM. DOI: 10.1145/3323439.3323981

• Randy Allen, Ken Kennedy: Optimizing compilers for modern architectures. Morgan Kaufmann, 2002. (Covers e.g. dependence analysis, loop optimizations, cache optimizations, parallelization, interprocedural analysis and optimization.)
• Michael Scott: Programming Language Pragmatics. Morgan Kaufmann, 2000. (Excellent summary of concepts in programming languages and their implementation in compilers. Chapter 13 gives a summary of code optimization.)
• Joseph A. Fisher, Paolo Faraboschi, Cliff Young: Embedded Computing - A VLIW Approach to Architecture, Compilers, and Tools. Morgan Kaufmann, 2005.
  Available in the LiU TekNat library.
• Peter Marwedel, Gert Goossens (Eds.): Code Generation for Embedded Processors. Kluwer International Series in Engineering and Computer Science, 1995.
• Rainer Leupers: Retargetable Code Generation for Digital Signal Processors. Kluwer, 1997. (Available in Kvartersbibliotek B.)
• Rainer Leupers: Code Optimization Techniques for Embedded Processors. Kluwer, 2000. (Available in Kvartersbibliotek B.)
• Hennessy, Patterson: Computer Architecture, a Quantitative Approach, Third Edition. Morgan Kaufmann, 2003. (Chapter 4 gives an overview of compiler optimizations for ILP.)

• slides on DFS
• slides on trees and tree traversals

• Short introduction to SSA (slide set)

Examiner
Christoph Kessler
Welf Löwe (guest examiner)

Examination

• TEN1: Written or oral exam, 4.5hp (date see schedule).
  No aids are allowed.
  Admission to the exam requires formal registration and presence in at least 50% of all lectures and lessons.
  Passing the exam is mandatory for getting any credit points on the course.
  
• LAB2: Compiler framework lab with written report 3hp
• PRE1: Paper presentation, opposition and written summary 1.5hp (Christoph) Deadline (summary): 6 April 2023
  For admission to the presentation it is required that at least 50% of the lectures and lessons have been attended.

Credit
Up to 6hp for the entire course, provided that the written exam is passed.
Admission to the examination moments (exam and presentation) requires formal registration on the course and presence in at least 50% of all lectures and lessons.

Note: Those who pass the course but are not IDA graduate students receive a paper certificate about their result. For potential acceptance of the certificate towards other graduate education programs please contact your program coordinator before joining the course.

Comments
This has been a CUGS Advanced Graduate Course 2002-2014.
The course is usually offered every other year.
It is not open to undergraduate students.