Discovering applications of higher order functions through proof planning

Cook, A.; Ireland, Andrew; Michaelson, Greg; Scaife, Norman

doi:10.1007/s00165-004-0054-5

Cited by 3 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The strict topology and execution model of Hume programs provide strong guidance in structuring proofs, and we therefore believe we can implement special purpose LCF-style tactics, or possibly proof plans (as in e.g. [8,26]), to prove preconditions. To illustrate, [12] implements a special purpose (post facto) transformation tactic for Hierarchical Hume in Isabelle/HOL.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Hume box calculus: robust system development through software transformation

Grov

Michaelson

2010

Higher-Order Symb Comput

View full text Add to dashboard Cite

Hume is a contemporary programming language oriented to systems with strong resource bounds, based on autonomous concurrent "boxes" interacting across "wires". Hume's design reflects the explicit separation of coordination and computation aspects of multi-process systems, which greatly eases establishing resource bounds for programs. However, coordination and computation are necessarily tightly coupled in reasoning about Hume programs. Furthermore, in Hume, local changes to coordination or computation, while preserving input/output correctness, can have profound and unforeseen effects on other aspects of programs such as timing of events and scheduling of processes. Thus, traditional program calculi prove inappropriate as they tend to focus exclusively either on the coordination of interacting processes or on computation within individual processes.The Hume box calculus offers a novel approach to manipulating multi-process systems by accounting seamlessly for both coordination and computation in individual rules. Furthermore, the "Hierarchical Hume" extension enables strong locality of the effects of program manipulation, as well as providing a principled encapsulation mechanism.In this paper, we present an overview of the Hume box calculus and its applications in program development. First of all, a base set of rules for introducing, changing, composing, separating and eliminating Hume boxes and wires, possibly within hierarchies, is presented. Next additional strategies are derived and a constructive approach to program development is illustrated through two examples of system elaboration from truth tables. Finally, at a considerably higher level, the use of the Hume box calculus to verify a generic transformation from a single box to an equivalent multi-box program, offering a balanced parallel implementation, is discussed.

show abstract

Section: Discussionmentioning

confidence: 99%

“…7, and its limitations and future directions in Sect. 8. This paper represents a substantive elaboration of [14], where the box calculus was first presented.…”

Section: Overviewmentioning

confidence: 99%

Hume box calculus: robust system development through software transformation

Grov

Michaelson

2010

Higher-Order Symb Comput

View full text Add to dashboard Cite

show abstract

“…Our approach may prove simpler to extend to such languages, providing that state is explicit and variable usage and update are de ned. One exception to the program calculation approaches is Cook's approach, which uses higher-order uni cation to discover and introduce common sequential pa erns [12]. is approach is, however, limited in that it uses a proof assistant as an intermediate representation, and further requires that functions have a structure that mirrors the structure of the pa ern to be discovered.…”

Section: Related Workmentioning

confidence: 99%

In search of a map: using program slicing to discover potential parallelism in recursive functions

Barwell

Hammond

2017

Proceedings of the 6th ACM SIGPLAN International Workshop on Functional High-Performance Computing

View full text Add to dashboard Cite

Recursion schemes, such as the well-known map, can be used as loci of potential parallelism, where, e.g., map is replaced with an equivalent parallel implementation. is paper formalises a novel technique, using program slicing, that automatically and statically identi es computations in recursive functions that can be li ed out of the function and potentially performed in parallel. We de ne a new program slicing algorithm, built a prototype implementation, and demonstrated its use on 12 Haskell examples, including benchmarks from the NoFib suite and functions from the standard Haskell Prelude. In all cases, we obtain the expected results in terms of nding potential parallelism. Moreover, we have tested our prototype against synthetic benchmarks, demonstrating our prototype has quadratic time complexity. For the NoFib benchmark examples we demonstrate that relative parallel speedups can be obtained (up to 32.93× the sequential performance on 56 hyperthreaded cores).

show abstract

“…As part of the PMLS project we have used proof-planning to construct a synthesiser which extracts HOFs from arbitrary recursive functions [6]. Figure 2.…”

Section: Performance Prediction Examplementioning

confidence: 99%

Empirical Parallel Performance Prediction from Semantics-Based Profiling

Scaife

Michaelson²,

Horiguchi

2005

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. The PMLS parallelizing compiler for Standard ML is based upon the automatic instantiation of algorithmic skeletons at sites of higher order function use. PMLS seeks to optimise run-time parallel behaviour by combining skeleton cost models with Structural Operational Semantics rule counts for HOF argument functions. In this paper, the formulation of a general rule count cost model as a set of over-determined linear equations is discussed, and their solution by singular value decomposition, and by a genetic algorithm, are presented.

show abstract

Discovering applications of higher order functions through proof planning

Cited by 3 publications

References 9 publications

Hume box calculus: robust system development through software transformation

Hume box calculus: robust system development through software transformation

In search of a map: using program slicing to discover potential parallelism in recursive functions

Empirical Parallel Performance Prediction from Semantics-Based Profiling

Contact Info

Product

Resources

About