Positive supercompilation for a higher order call-by-value language

Jönsson, Peter; Nordlander, Johan

doi:10.1145/1594834.1480916

Cited by 7 publications

(5 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From the very beginning the development of supercompilation has been conducted mainly in the context of the programming language Refal [43,44,45,54] based on syntax and semantics similar to that of our presentation language L . A number of model supercompilers for subsets of functional languages based on Lisp data were implemented with the aim of formalizing some aspects of the supercompilation algorithms [24,27,42,50]. The most advanced supercompiler for Refal is SCP4 [43,44,46].…”

Section: On Supercompilationmentioning

confidence: 99%

Finite Countermodel Based Verification for Program Transformation (A Case Study)

Lisitsa

Nemytykh

2015

Electron. Proc. Theor. Comput. Sci.

View full text Add to dashboard Cite

Both automatic program verification and program transformation are based on program analysis. In the past decade a number of approaches using various automatic general-purpose program transformation techniques (partial deduction, specialization, supercompilation) for verification of unreachability properties of computing systems were introduced and demonstrated [10,19,30,36]. On the other hand, the semantics based unfold-fold program transformation methods pose themselves diverse kinds of reachability tasks and try to solve them, aiming at improving the semantics tree of the program being transformed. That means some general-purpose verification methods may be used for strengthening program transformation techniques. This paper considers the question how finite countermodels for safety verification method [34] might be used in Turchin's supercompilation method. We extract a number of supercompilation sub-algorithms trying to solve reachability problems and demonstrate use of an external model finder for solving some of the problems.

show abstract

Section: On Supercompilationmentioning

confidence: 99%

Finite Countermodel Based Verification for Program Transformation (A Case Study)

Lisitsa

Nemytykh

2015

Electron. Proc. Theor. Comput. Sci.

View full text Add to dashboard Cite

show abstract

“…They report average performance boosts on a set of programs of 40%. Supercompilation has also been investigated in a call-by-value setting and can give good speedups there too [6,8].…”

Section: Related Workmentioning

confidence: 99%

“…The focus in this paper has been on a practical algorithm that can supercompile programs without risking code explosion. We believe that our previous proofs of correctness of a closely related algorithm [7] should go through without much adjustment for the algorithm presented.…”

Section: Future Workmentioning

confidence: 99%

Taming code explosion in supercompilation

Jönsson

Nordlander

2011

Proceedings of the 20th ACM SIGPLAN Workshop on Partial Evaluation and Program Manipulation

Self Cite

View full text Add to dashboard Cite

Supercompilation algorithms can perform great optimizations but sometimes suffer from the problem of code explosion. This results in huge binaries which might hurt the performance on a modern processor. We present a supercompilation algorithm that is fast enough to speculatively supercompile expressions and discard the result if it turned out bad. This allows us to supercompile large parts of the imaginary and spectral parts of nofib in a matter of seconds while keeping the binary size increase below 5%.

show abstract

“…We prove that our supercompiler both terminates and does not alter whether a program terminates or not. The complete proofs are available from a companion technical report (Jonsson and Nordlander 2008), but they do not reveal anything unexpected.…”

Section: Correctnessmentioning

confidence: 99%

“…We show that our positive supercompiler does remove intermediate structures, and can improve the performance by an order of magnitude for certain benchmarks. We have left out the full details of the instrumentation of the runtime system and the transformed result of each benchmark for space reasons, but they are available in a separate report (Jonsson 2008). All measurements were performed on an idle machine running in an xterm.…”

Section: Benchmarksmentioning

confidence: 99%