article · 2003

Alias analysis for intermediate code

Sanjiv K. Gupta, Naveen Sharma

Cited by 0Abstract and details only

Abstract

Most existing alias analysis techniques are formulated in terms of high-level language constructs and are unable to cope with pointer arithmetic. For machines that do not have ’base + offset ’ addressing mode, pointer arithmetic is necessary to compute a pointer to the desired address. Most state of the art compilers such as GCC lack the mechanism to determine aliasing between such computed pointers. Few other existing alias analysis techniques described for executable code can handle pointer arithmetic but require large memory when applied to intermediate languages such as RTL. In this paper, we describe a method of disambiguating the computed pointers within a procedure at the intermediate code level. The method is similar to the techniques described for executable code but requires significantly less amount of memory. We have experimented our method with the GCC RTL and it reduces the code size of array manipulating benchmarks by approximately 4-7 % for the machines that do not have ’base + offset ’ addressing mode. 1

Cite this paper

APA 7

Gupta, S. K., & Sharma, N. (2003). Alias analysis for intermediate code.

MLA 9

Gupta, Sanjiv K., and Naveen Sharma. “Alias analysis for intermediate code.” 2003.

Chicago

Gupta, Sanjiv K., and Naveen Sharma. 2003. “Alias analysis for intermediate code.”

Harvard

Gupta, S.K. and Sharma, N. (2003) ‘Alias analysis for intermediate code’.

BibTeX

@article{gupta2003alias,
  title = {{Alias analysis for intermediate code}},
  author = {Gupta, Sanjiv K. and Sharma, Naveen},
  year = {2003}
}

Made from the index record. Check names, capitalisation and page numbers against the paper. Need another style? AI for Research has over 10,000.