skip to main content

Pinhão: An Auto-tunning System for Compiler Optimizations Guided by Hot Functions

Siraichi, Marcos Yukio ; Tonetti, Caio Henrique Segawa ; Silva, Anderson Faustino da

J.UCS (Annual print and CD-ROM archive ed.), 2019-01, Vol.25 (1), p.42-72 [Periódico revisado por pares]

Verlag der Technischen Universität Graz

Sem texto completo

Citações Citado por
  • Título:
    Pinhão: An Auto-tunning System for Compiler Optimizations Guided by Hot Functions
  • Autor: Siraichi, Marcos Yukio ; Tonetti, Caio Henrique Segawa ; Silva, Anderson Faustino da
  • Assuntos: auto-tunning system ; compiler ; machi ; optimization
  • É parte de: J.UCS (Annual print and CD-ROM archive ed.), 2019-01, Vol.25 (1), p.42-72
  • Notas: JUCS - Journal of Universal Computer Science Volume Nr. 25 Issue Nr. 1
  • Descrição: The literature presents several auto-tunning systems for compiler optimizations, which employ a variety of techniques; however, most systems do not explore the premise that a large amount of program runtime is spent by hot functions which are the portions at which compiler optimizations will provide the greatest benefit. In this paper, we propose Pinhão, an auto-tunning system for compiler optimizations that uses hot functions to guide the process of exploring which compiler optimizations should be enabled during target code generation. Pinhão employs a hybrid technique - a machine learning technique, as well as an iterative compilation technique - to find an effective compiler optimization sequence that fits the characteristics of the unseen program. We implemented Pinhão as a LLVM tool, and the experimental results indicate that Pinhão finds effective sequences evaluating a few points in the search space. Furthermore, Pinh~ao outperforms the well-engineered compiler optimization levels, as well as other techniques.
  • Editor: Verlag der Technischen Universität Graz
  • Idioma: Inglês

Buscando em bases de dados remotas. Favor aguardar.