Lecture Slides


   Introduction to course (1 lecture)

Getting started: A simple one-pass compiler (2 lectures)
   Stack Machines: SaM
   Required reading: Parsing
   Recursive-descent parsing and code generation

Back-end optimizations
   Intermediate Representations (1 lecture)
   Control-flow graphs

   Scalar optimizations


   Solving fixpoint equations (1 lecture)

   Introduction to the Galois framework (1 lecture)
   Distributed Graph-Word2Vec (1 lecture)

   Live variables and copy propagation (2 lectures)
   Dataflow analysis frameworks
   Notes on dataflow

   Partial redundancy elimination

   Lazy code motion (slides from Cooper and Torczon, Rice University)
        Papers: Lazy Code Motion by Jens Knoop et al (PLDI 1992)
                     Drechsler and Stadl's version of Lazy Code Motion (SIGPLAN , May 1993)
   Strength Reduction
         Paper: Operator strength reduction by Cooper et al (TOPLAS September 2001)

   Dominators,control dependence,SSA (1 lecture)
         Papers:  The Program Dependence Graph and its Use in Optimization by Ferrante et al
                       Efficiently computing the SSA form and control dependence by Ferrante et al (TOPLAS 1991)
                       Algorithms for Computing the SSA Form by Bilardi et al (JACM, May 2003)

   Register allocation (1 lecture)

   Points-to analysis (1 lecture)
         Papers: The ant and the grasshopper: fast and accurate pointer analysis for millions of lines of code
                        by Hardekopf and Lin (PLDI 2007)   (efficient Andersen-style analysis)
                       Flow-sensitive pointer analysis for millions of lines of code by Hardekopf and Lin (CGO 2011)
                         (uses flow-insensitive analysis as a pre-processing step)
Loop transformations (4 lectures)

  Cache models for locality analysis
  Auto-tuning for cache performance: ATLAS
        Papers: A comparison of empirical and model-driven optimization
                               Yotov et al, PLDI 2003
  Dependence analysis of affine loops
         Papers:  The Omega test: a fast and practical integer programming algorithm for dependence analysis
                                 Bill Pugh, CACM August 1992
                        Access normalization: loop restructuring for NUMA computers
                                 Wei Li and Keshav Pingali, ACM Transactions on Computer Systems (TOCS), 1993.
                        A data locality optimizing algorithm
                                Michael Wolf and Monica Lam, PLDI 1991.
  Vectorization: slides from David Padua's tutorial
  Unimodular loop transformations
  Fractal Symbolic Analysis
          Paper: Fractal Symbolic Analysis

Programming models for parallelism

The operator formulation of algorithms
Papers:  The TAO of parallelism in algorithms
                                  Pingali et al. PLDI 2011


Student presentations