Abstract
Algorithmic skeletons are higher-order functions which provide tools for parallel programming at a higher abstraction level, hiding the technical details of parallel execution inside the skeleton implementation. However, this encapsulation becomes an obstacle when the actual algorithm is one that involves iterative application of the same skeleton to successively improve or approximate the result. Striving for a general and portable solution, we propose a skeleton iteration framework in which arbitrary skeletons can be embedded with only minor modifications. The framework is flexible and allows for various parallel iteration control and parallel iteration body variants. We have implemented it in the parallel Haskell dialect Eden using dedicated stream communication types for the iteration. Two non-trivial case studies show the practicality of our approach. The performance of our compositional iteration framework is competitive with customised iteration skeletons.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsNotes
- 1.
Skeleton iteration should not be confused with parallel for-loops or parallel map, where a sequential block is executed in parallel by multiple threads, instead of several times. We focus on computations defined by algorithmic skeletons which are by themselves already parallel and will be executed several times in sequence.
- 2.
The original Eden definition specifies that top-level lists are communicated as streams. In this work, we use a modified Trans class which gives programmers more control of streaming through separate stream types.
- 3.
The superscripts in our types are merely annotations to indicate implicit constraints on the list lengths. Fixed sized lists could however be implemented e.g. using the recent Haskell library Vec, see http://hackage.haskell.org/package/Vec
- 4.
Eden supports explicit placement of computations in a multi-node parallel system. We have omitted placement aspects from our code for simplicity throughout.
- 5.
Similar to whileControl, but forwards the initial input directly to the iteration body, thus doing at least one iteration before termination.
References
Berthold, J., Dieterle, M., Loogen, R.: Implementing parallel google map-reduce in Eden. In: Sips, H., Epema, D., Lin, H.-X. (eds.) Euro-Par 2009. LNCS, vol. 5704, pp. 990–1002. Springer, Heidelberg (2009)
Bu, Y., Howe, B., Balazinska, M., Ernst, M.D.: The HaLoop approach to large-scale iterative data analysis. VLDB J. 21(2), 169–190 (2012)
Bird, R.S.: Using circular programs to eliminate multiple traversals of data. Acta Inform. 21, 239–250 (1984)
Cole, M.: Algorithmic Skeletons: Structured Management of Parallel Computation. MIT Press, Cambridge (1989)
Dean, J., Ghemawat, S.: Mapreduce: simplified data processing on large clusters. CACM 51(1), 107–113 (2008)
Dieterle, M., Horstmeyer, T., Loogen, R.: Skeleton composition using remote data. In: Carro, M., Peña, R. (eds.) PADL 2010. LNCS, vol. 5937, pp. 73–87. Springer, Heidelberg (2010)
Ekanayake, J., Li, H., Zhang, B., Gunarathne, Th., Bae, S., Qiu, J., Fox, G.: Twister: a runtime for iterative mapreduce. In: HPDC ’10. ACM (2010)
Ewen, St, Tzoumas, K., Kaufmann, M., Markl, V.: Spinning fast iterative data flows. PVLDB 5(11), 1268–1279 (2012)
Loogen, R., Ortega-Mallén, Y., Peña-Marí, R.: Parallel functional programming in Eden. J. Funct. Program. 15(3), 431–475 (2005)
Loogen, R.: Eden – parallel functional programming with Haskell. In: Zsók, V., Horváth, Z., Plasmeijer, R. (eds.) CEFP. LNCS, vol. 7241, pp. 142–206. Springer, Heidelberg (2012)
Leyton, M., Piquer, J.M.: Skandium: multi-core programming with algorithmic skeletons. In: PDP. IEEE Computer Society (2010)
MacKay, D.: Information Theory, Inference, and Learning Algorithms. Cambridge University Press, Cambridge (2003). See chapter 20, p. 284ff
Peña, R., Rubio, F.: Parallel functional programming at two levels of abstraction. In: PPDP’01, pp. 187–198. ACM (2001)
Zhang, Y., Gao, Q., Gao, L., Wang, C.: iMapReduce: a distributed computing framework for iterative computation. JOGC 10, 47–68 (2012)
Acknowledgements
The authors thank the anonymous referees for their helpful comments on a previous version of this paper. Jost Berthold was partially supported by DSF under contract number 10-092299 (Hiperfit).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Dieterle, M., Horstmeyer, T., Berthold, J., Loogen, R. (2013). Iterating Skeletons. In: Hinze, R. (eds) Implementation and Application of Functional Languages. IFL 2012. Lecture Notes in Computer Science(), vol 8241. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-41582-1_2
Download citation
DOI: https://doi.org/10.1007/978-3-642-41582-1_2
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-41581-4
Online ISBN: 978-3-642-41582-1
eBook Packages: Computer ScienceComputer Science (R0)