Eigil Rosager Poulsen. Representation analysis for efficient implementation of polymorphism. Technical report, Department of Computer Science (DIKU), University of Copenhagen, April 1993. Master Dissertation.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....as interaction with ML module system and efficient propagation of type information through many rounds of compiler transformations and optimizations. Many people have worked on eliminating unnecessary wrapper functions introduced by representation analysis. Both Peyton Jones [17] and Poulsen [18] let the programmer to tag some types with a boxity annotation, and then statically determine when to use boxed representations. Henglein and Jorgensen [10] present a term rewriting method that translates a program with many coercions into one that contains a formally optimal set of coercions. ....

Eigil Poulsen. Representation analysis for efficient implementation of polymorphism. Master's thesis, DIKU, University of Copenhagen, 1993.

....as interaction with ML module system and efficient propagation of type information through many rounds of compiler transformations and optimizations. Many people have worked on eliminating unnecessary wrapper functions introduced by representation analysis. Both Peyton Jones [20] and Poulsen [21] let the programmer to tag some types with a boxity annotation, and then statically determine when to use boxed representations. Henglein and Jorgensen [13] present a term rewriting method that translates a program with many coercions into one that contains a formally optimal set of coercions. ....

Eigil Poulsen. Representation analysis for efficient implementation of polymorphism. Master's thesis, DIKU, University of Copenhagen, 1993.

....as interaction with ML module system and efficient propagation of type information through many rounds of compiler transformations and optimizations. Many people have worked on eliminating unnecessary wrapper functions introduced by representation analysis. Both Peyton Jones [20] and Poulsen [21] let the programmer to tag some types with a boxity annotation, and then statically determine when to use boxed representations. Henglein and Jorgensen [13] present a term rewriting method that translates a program with many coercions into one that contains a formally optimal set of coercions. ....

Eigil Poulsen. Representation analysis for efficient implementation of polymorphism. Master's thesis, DIKU, University of Copenhagen, 1993.

....entire SML language (Caml has a much simpler module system than SML) and how to efficiently propagate type information through many rounds of transformations and optimizations. Many people have worked on eliminating unnecessary wrapper functions introduced by representation analysis. Poulsen [Pou93] proposes a way to tag each type with a boxity annotation, and then statically determine when to use boxed representations and when to use unboxed representations. The major problem of his technique is that it is not easy to extend to the SML module system. Henglein and Jorgensen [HJ94] presents ....

Eigil Poulsen. Representation Analysis for Efficient Implementation of Polymorphism. Master's thesis, DIKU, University of Copenhagen, 1993.

....to its unboxed representation where polymorphic operations that need parts of their arguments to be boxed use explicit boxing operations to ensure this. Other translations may elide some of these boxing and unboxing operations directly; e.g. the type inference based translations of Poulsen [Pou93] 1.2 Coercion calculus Beyond offering yet another translation we seek to formulate and answer the more fundamental questions that underlie the very purpose of boxing analysis and, more generally, similar static analyses: Given two completions for the same program, which of them is better What ....

....1.4. Computing boxing completions 5 1.4 Computing boxing completions Although the rewriting based algorithm gives a way to compute formally optimal completions it is not the most efficient way to do this. We show that an efficient algorithm exist which uses a method similar to that of Poulsen s [Pou93] but which in contrast to Poulsen s algorithm always finds optimal completions. This algorithm makes it easy to see that our work is an improvement Poulsen s. Together with the fact that also Leroy s algorithm [Ler92] does not in general produce optimal completions (see section 1.3) this shows ....

[Article contains additional citation context not shown here]

Eigil Poulsen. Representation analysis for efficient implementation of polymorphism. Master's thesis, DIKU, University of Copenhagen, 1993.

....is uniform across instances. Most implementations of ML impose a similar restriction, requiring that values of unknown type be boxed (stored on the heap and represented by a pointer) Early implementations used a LISP like representation in which all values are boxed [5] later implementations [31, 32, 30, 24, 43] seek to minimize boxing by taking advantage of whatever type information is manifest in the program. Despite these recent improvements, current implementations still resort to pointer representations for unknown types. Furthermore, current implementations make use of tag bits on values to assist ....

....the typing properties of intensionally polymorphic operations. A number of authors have considered problems pertaining to representation analysis in the presence of poylmorphism. The boxing interpretation of polymorphism has been studied by Peyton Jones Launchbury [30] by Leroy [32] by Poulsen [43], and by Henglein J rgensen [24] with the goal of minimizing the overhead of boxing and unboxing at run time. Of a broadly similar nature is the work on soft type systems [3, 11, 23, 48, 53] which seek to improve data representations through global analysis techniques. All of these methods ....

Eigil Rosager Poulsen. Representation analysis for efficient implementation of polymorphism. Technical report, Department of Computer Science (DIKU), University of Copenhagen, April 1993. Master Dissertation.

....However, they only supported a limited set of unboxed types (ints and floats) and restricted these types from instantiating type variables. Later, Leroy suggested a general approach for unboxing values based on the ML type system [28] Leroy s approach has been extended and implemented elsewhere [38, 24, 42], including the SML NJ compiler. It does not support unboxed array components nor flattened, recursive datatypes. Tolmach [46] combined Leroy s approach with tag free garbage collection. However, he used an ad hoc approach to propagate type information to the collector. Other researchers have ....

Eigil Rosager Poulsen. Representation analysis for efficient implementation of polymorphism. Technical report, Department of Computer Science (DIKU), University of Copenhagen, April 1993. Master Dissertation.

No context found.

Eigil Rosager Poulsen. Representation analysis for efficient implementation of polymorphism. Technical report, Department of Computer Science (DIKU), University of Copenhagen, April 1993. Master Dissertation.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC