Profilage mémoire d’applications OCaml

Abstract : Automatic memory reclaim is a common feature of programming languages. It offers to the programmers some guarantees about reliability, but, by taking them away from the implementation details of the memory management, it makes it more difficult to understand and control memory allocations. OCaml is a multi-paradigm statically typed programming language developed at INRIA. It uses a garbage collector (GC) in charge of the automatic reclaim of unused memory and OCaml users need tools for analyzing the memory behavior of their programs. As part of this thesis, we studied the memory behavior of applications written in OCaml and more specifically the case of memory leaks: still accessible memory blocks (hence not reclaim by the GC) that are no longer needed for the computation. In this thesis, after describing some known methods for analyzing the memory behavior of applications, we present an instrumentation method of the OCaml toolchain and runtime. This instrumentation then is used by tools that we have designed and implemented, and used to categorize, visualize and locate memory allocations. Results and tools presented in this thesis were obtained as part of a collaboration between ENSTA ParisTech, INRIA and OCamlPro SAS, and have been successfully used in an industrial context.
Document type :
Theses
Complete list of metadatas

Cited literature [84 references]  Display  Hide  Download

https://pastel.archives-ouvertes.fr/tel-01122262
Contributor : Cagdas Bozman <>
Submitted on : Tuesday, March 3, 2015 - 3:12:57 PM
Last modification on : Thursday, July 4, 2019 - 4:00:50 AM

Identifiers

  • HAL Id : tel-01122262, version 1

Collections

Citation

Çağdaş Bozman. Profilage mémoire d’applications OCaml. Informatique [cs]. ENSTA ParisTech, 2014. Français. ⟨tel-01122262⟩

Share

Metrics

Record views

344

Files downloads

920