A Type-Based Approach to Parallelization
| dc.contributor.author | Dana N. XU | en_US |
| dc.contributor.author | Siau Cheng KHOO | en_US |
| dc.contributor.author | Wei Ngan CHIN | en_US |
| dc.contributor.author | Zhenjiang HU | en_US |
| dc.date.accessioned | 2004-10-21T14:28:52Z | en_US |
| dc.date.accessioned | 2017-01-23T06:59:54Z | |
| dc.date.available | 2004-10-21T14:28:52Z | en_US |
| dc.date.available | 2017-01-23T06:59:54Z | |
| dc.date.issued | 2003-10-01T00:00:00Z | en_US |
| dc.description.abstract | Parallel functional programming plays an important role in parallel programming [16]. Type system has signi.cant impact on program analysis [23]. In this paper, we show how to automatically and correctly synthesize parallel programs from sequential functional program based on the concept of a type system. Our type system captures the parallelizability of a program, in a modular fashion, by exploring the ring structures of the programs operators. It handles programs de.ned by self-recursive functions with accumulating parameters, as well as a limited form of non-linear mutual-recursive functions. In contrast to the Damas-Milner type system (the typical type system) that is constructed from the evaluation rules of the underlying language, our type system is constructed from a set of meta-rules that are used to transform sequential programs into a special normal form suitable for parallelization. The idea of this paper has been implemented and used to generate parallel code of a form, called mutumorphism, a general parallel computation model. Transforming into such a form is an important step towards constructing e.cient data parallel programs. | en_US |
| dc.format.extent | 519493 bytes | en_US |
| dc.format.mimetype | application/pdf | en_US |
| dc.identifier.uri | https://dl.comp.nus.edu.sg/xmlui/handle/1900.100/1438 | en_US |
| dc.language.iso | en | en_US |
| dc.relation.ispartofseries | TR30/03 | en_US |
| dc.title | A Type-Based Approach to Parallelization | en_US |
| dc.type | Technical Report | en_US |