On a Technique for Transparently Empowering Classical Compiler Optimizations on Multithreaded Code

• Autores: Pramod G. Joisha, Robert S. Schreiber, Prithviraj Banerjee, Hans-J Boehm, Dhruva R. Chakrabarti
• Localización: ACM transactions on programming languages and systems, ISSN 0164-0925, Vol. 34, Nº 2, 2012
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • A large body of data-flow analyses exists for analyzing and optimizing sequential code. Unfortunately, much of it cannot be directly applied on parallel code, for reasons of correctness. This article presents a technique to automatically, aggressively, yet safely apply sequentially-sound data-flow transformations, without change, on shared-memory programs. The technique is founded on the notion of program references being “siloed” on certain control-flow paths. Intuitively, siloed references are free of interference from other threads within the confines of such paths. Data-flow transformations can, in general, be unblocked on siloed references.

    The solution has been implemented in a widely used compiler. Results on benchmarks from SPLASH-2 show that performance improvements of up to 41% are possible, with an average improvement of 6% across all the tested programs over all thread counts.