Re: Interprocedural optimization and code reuse

Compiler optimization to automatically take advantage of
the memory hierarchy (e.g., registers & cache) is a pretty
hot topic these days, the goal being to provide high-level
architecture-independent programming for users.


Here are some recent references:


      Gannon, Jalby, and Gallivan, "Strategies for Cache and Local
      Memory Management by Global Program Transformation", Journal
      of Parallel and Distributed Computing, 5(5):587-616, October
      1988.


      Callahan, Carr, and Kennedy, "Improving Register Allocation
      for Subscripted Variables", in Proceedings of ACM SIGPLAN '90,
      White Plains, NY, June 1990. SIGPLAN Notices 25(6):53-65.


      Wolf and Lam, "A Data Locality Optimizing Algorithm", in
      Proceedings of ACM SIGPLAN '91, Toronto, Canada, June 1991.
      SIGPLAN Notices 26(6):30-44.




In particular, interprocedural optimizations to enable
memory optimizations (via cloning and inlining) are described in:


      Briggs, Cooper, Hall, and Torczon, "Goal-Directed Interprocedural
      Optimization", Dept of Computer Science Technical Report
      TR90-147, Rice Univ, Nov 1990.


      Hall, "Managing Interprocedural Optimization", PhD Thesis,
      Dept of Computer Science Technical Report TR91-157,
      Rice Univ, April 1991.




Hope that this helps get people started :-)


Chau-Wen Tseng
Rice Univ
--

Return to the comp.compilers page.
Search the comp.compilers archives again.