Guaranteed Optimization: Proving Nullspace Properties of Compilers (2002)  (Make Corrections)  (6 citations) Todd L. Veldhuizen, Andrew LumsdaineProceedings of the 2002 Static Analysis Symposium (SAS'02)   Home/Search   Context   Related

View or download: iu.edu/~tveldhui/papers/2002/...null.ps Cached:  PS.gz  PS  PDF This document uses CoBlitz to cache paper downloads. If your firewall is blocking outgoing connections to port 3125, you can use these links to download local copies.   PS.gz  PS  PDF   Image  Update  Help From:  iu.edu/~tveldhui/papers/ (more) Homepages:  T.Veldhuizen

Rate this article: (best)   Comment on this article

@inproceedings{ veldhuizen:sas:2002,
  author = "Todd L. Veldhuizen and Andrew Lumsdaine", 
  title = "Guaranteed Optimization: Proving Nullspace Properties of Compilers",
  url = "citeseer.ist.psu.edu/veldhuizen02guaranteed.html",
  url = "http://osl.iu.edu/~tveldhui/papers/",
  booktitle = "Proceedings of the 2002 Static Analysis Symposium (SAS'02)",
  publisher = "Springer-Verlag",
  year = "2002",
  pages = "263-277",
  volume = "2477",
  conflocation = "Madrid, Spain",
  conftitle = "Static Analysis Symposium",
  series = "Lecture Notes in Computer Science",
  abstract = "Writing performance-critical programs can be frustrating because
   optimizing compilers for imperative languages tend to be
   unpredictable. For a subset of optimizations - those that simplify
   rather than reorder code - it would be useful to prove that a compiler
   reliably performs optimizations. We show that adopting a
   ``superanalysis'' approach to optimization enables such a proof. By
   analogy with linear algebra, we define the nullspace of an optimizer
   as those programs it reduces to the empty program. To span the
   nullspace, we define rewrite rules that de-optimize programs by
   introducing abstraction. For a model compiler we prove that any
   sequence of de-optimizing rewrite rule applications is undone by the
   optimizer. Thus, we are able to give programmers a clear mental model
   of what simplifications the compiler is guaranteed to perform, and
   make progress on the problem of ``abstraction penalty'' in imperative
   languages." }