Wave Propagation and Deep Propagation for Pointer Analysis

Pereira, Fernando Magno Quintão; Berlin, Daniel

doi:10.1109/cgo.2009.9

Cited by 55 publications

(61 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The source files of each benchmark are compiled into bit-code files using clang and then linked together using llvm-link, with mem2reg being applied to promote memory into registers. We use FI, i.e., Andersen's analysis (using the constraint resolution techniques from [25,27]) as pre-analysis for building indirect def-use chains [14,30,31].…”

Section: Methodsmentioning

confidence: 99%

Region-Based Selective Flow-Sensitive Pointer Analysis

Sui

Xue

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. We introduce a new region-based SELective Flow-Sensitive (Selfs) approach to inter-procedural pointer analysis for C that operates on the regions partitioned from a program. Flow-sensitivity is maintained between the regions but not inside, making traditional flow-insensitive and flow-sensitive as well as recent sparse flow-sensitive analyses all special instances of our Selfs framework. By separating region partitioning as an independent concern from the rest of the pointer analysis, Selfs facilitates the development of flow-sensitive variations with desired efficiency and precision tradeoffs by reusing existing pointer resolution algorithms. We also introduce a new unification-based approach for region partitioning to demonstrate the generality and flexibility of our Selfs framework, as evaluated using SPEC2000/2006 benchmarks in LLVM.

show abstract

Section: Methodsmentioning

confidence: 99%

Region-Based Selective Flow-Sensitive Pointer Analysis

Sui

Xue

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Wave Propagation [24] is used to perform constraint resolution. Global variables are handled context-insensitively in the standard manner as in [20,22]: they are separately tracked without participating in side-effect summarization.…”

Section: Methodsmentioning

confidence: 99%

Query-directed adaptive heap cloning for optimizing compilers

Sui

Xue

2013

Proceedings of the 2013 IEEE/ACM International Symposium on Code Generation and Optimization (CGO)

View full text Add to dashboard Cite

Andersen's pointer analysis becomes more precise when applied with full heap cloning but unscalable for large, heapintensive programs. In contrast, k-callsite-sensitive heap cloning can be faster but less precise for some programs.In this paper, we make one step forward by enhancing Andersen's analysis with QUery-Directed Adaptive (QUDA) heap cloning for optimizing compilers. The novelty of our analysis, called QUDA, lies in performing k-callsite-sensitive heap cloning iteratively, starting with k = 0 (without heap cloning), so that an abstract heap object is cloned at iteration k = i + 1 only if some mayalias queries that are not answered positively at iteration k = i may now be answered more precisely. QUDA, which is implemented in Open64, has the same precision as the state-of-the-art, FULCRA, a version of QUDA with exhaustive heap cloning, but is significantly more scalable. For 10 SPEC2000 C benchmarks and 5 C applications (totalling 840 KLOC) evaluated, QUDA takes only 4+ minutes but exhaustive heap cloning takes 42+ minutes to complete. QUDA takes only 75.1% of the time that Open64 takes on average to compile these 15 programs under "-O2".

show abstract

“…Many approaches based on cycle elimination have been proposed to improve the performance of subset-based analyses [3][4][5][6][7]. Unification-based analyses have near linear time complexity and easily scale to million lines.…”

Section: Introductionmentioning

confidence: 99%