Dujardin, E., Amiel, E., Simon, E.: Fast Algorithms for Compressed Multi-Method Dispatch Table Generation, TOPLAS Journal, vol. 20, no. 1, Jan 1998, p116-165.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....one method argument are used for the dispatching. Multi method dispatching has been identified as a powerful feature in object oriented languages supporting multi methods such as Cecil [5] CLOS [4] and Dylan [8] Several recent results have attempted to deal with d ary dispatching in practice [1, 6, 7, 14, 15]. Ferragina et al. 18] provided the first non trivial data structures, and, quoting this paper, several experimental object oriented languages ultimately success and impact in practice depends, among other things, on whether multi method dispatching can be supported efficiently . Our main ....

Eric Dujardin, Eric Amiel, and Eric Simon. Fast algorithms for compressed multimethod dispatch table generation. ACM Transactions on Programming Languages and Systems, 20(1):116--165, January 1998.

....actions. This is a conservative approach and one can identify situations where inlining in multi dispatch Java would provide correct results. Identifying these opportunities will yield higher overall performance. Other multi dispatch techniques exist, including compressed n dimensional tables [1, 12], look up automata [9, 10] and efficient multiple and predicate dispatch [7] A comprehensive exploration of these techniques using Java is incomplete at this time. Another significant improvement for multi dispatch is to incorporate our code testing tool into the javac compiler. At this time, ....

E. Dujardin, E. Amiel, and E. Simon. Fast algorithms for compressed multimethod dispatch table generation. ACM Transactions on Programming Languages and Systems, 20(1):116--165, January 1998.

....solving the multi method dispatching problem, is believed to be space and time intensive. 1;2 That is a serious concern because user programs spend significant amount of time in dispatching. Hence, considerable research effort has been focused on multi method dispatching within the OO community [A91, D96, CTK94, ADS96, LM98]. Multi methods and multiple dispatching. Computation in OO systems proceeds by invoking methods (procedures with several arguments) on class instances. Determining the correct method definition that gets applied on any such invocation constitutes the so called dispatching problem. The correct ....

....is a combinatorial explosion in the space usage as d increases. Since even the case d = 1 uses a prohibitive amount of space, this solution is clearly useless for d 2. No better theoretical bounds are currently known. Several recent results attempted to deal with d ary dispatching in practice [A91, D96, ADS96, CTK94, LM98]. However, even the case d = 2 is not well understood and designing effective heuristics for table compaction schemes or for using automata to prune the search space is still open. To summarize, the main algorithmic research goal is to design algorithms with interesting, provable tradeoffs ....

E. Amiel, E. Dujardin and E. Simon. Fast algorithms for compressed multi-method dispatch tables generation. Technical Report # 2977, INRIA, Sept. 1996.

....The standard table can be compressed by selector aliasing and class sharing, and the conflict table by class sharing alone. 4 Multi Method Dispatch Techniques This section provides a brief summary of the existing multi method dispatch techniques. 1. CNT: Compressed N Dimensional Tables [AGS94, DAS96] represents the dispatch table as a behavior specific k dimensional table, where k represents the arity of a particular behavior. Each dimension of the table is compressed by grouping identical dimension lines. The resulting table is indexed by type groups in each dimension, and mappings from ....

....Each object file can compute its own partial dispatch tables, and the dispatch tables can be rapidly merged together at link time. SRP is the only table based multi method dispatch technique that currently provides completely incremental dispatch table modification. The algorithms presented in [DAS96] for CNT require a global type ordering, which in turn implies complete environment knowledge and precludes incremental maintenance. No incremental algorithm is currently known for LUA (besides the naive approach of recomputing the entire lookup automata) Since MRD effectively collapses a ....

[Article contains additional citation context not shown here]

Eric Dujardin, Eric Amiel, and Eric Simon. Fast algorithms for compressed multi-method dispatch table generation. In Transactions on Programming Languages and Systems, 1996.

....languages by replacing references to product types with types. 3 Multi Method Dispatch Techniques This section provides a brief summary of the existing multi method dispatch techniques. Refer to the citations for additional information. 1. CNT: Compressed N Dimensional Tables ( AGS94, DAS96] represent the dispatch table as a behavior specific k dimensional table, where k represents the arity of a particular behavior. Each dimension of the table is compressed by grouping identical dimension lines into a single line. The resulting table is indexed by type groups in each dimension, ....

Eric Dujardin, Eric Amiel, and Eric Simon. Fast algorithms for compressed multimethod dispatch table generation. In Transactions on Programming Languages and Systems, 1996.

....time efficient [1] We will show how this single receiver technique can be generalized to multi method languages in Sect. 5. 4 Existing Multi Method Dispatch Techniques This section provides a brief summary of the existing multi method dispatch techniques. 1. CNT: Compressed N Dimensional Tables [9 11] represents the dispatch table as a behavior specific k dimensional table, where k represents the arity of a particular behavior. Each dimension of the table is compressed by grouping identical dimension lines into a single line. The resulting table is indexed by type groups in each dimension, and ....

....data structures in Fig. 8e. Since ffi is a 3 arity behavior, Expr. 3 becomes Expr. 4. M [ I 1 [ I 0 [ B [ ffi ] num(E) num(D) num(B) 4) Substituting the data from Fig. 8e into Expr. 4 yields the method ffi 1 , as shown in Expr. 5. M [ I 1 [ I 0 [ 7 4 ] 3 ] 1 ] M [ I 1 [ I 0 [ 11 ] 3 ] 1 ] M [ I 1 [ 5 3 ] 1 ] 5) M [ I 1 [ 8] 1 ] M [ 15 1 ] M [ 16 ] ffi 1 5.4 Optimizations Single I . For simplicity of presentation, we defined an Index Array per arity position. Actually, we only need one Global Index Array, I , to store all level 0 to level (k Gamma ....

[Article contains additional citation context not shown here]

Dujardin, E., Amiel, E., Simon, E.: Fast Algorithms for Compressed Multi-Method Dispatch Table Generation, TOPLAS Journal, vol. 20, no. 1, Jan 1998, p116-165.

....3 CNT: Multi Method Dispatch Using Compressed N Dim. Tables Before presenting the two new techniques, we will summarize an existing multi method table based technique. The technique was first presented in [AGS94] Efficient algorithms for performing the proposed compression were presented in [DAS96]. Since they involve compressing each dimension of an N dimensional table, we will refer to it as multi method dispatch using compressed N dimensional tables, or CNT for short. 3.1 Data Structures The basic idea of the CNT technique is to form a compressed k dimensional table, D k m , for each ....

....numbers. The value stored at index j of the i th argument array represents which group the type with number j belongs to in the i th dimension of the k dimensional table. If a type does not belong to any group in a particular dimension, the index for that type is a special empty indicator. [DAS96] observes that many entries within an argument array will be empty , and that it is possible to compress two or more argument arrays into a single group argument array as long as, for every index j, only one of the original argument arrays has a non empty value. Since this compression results in ....

[Article contains additional citation context not shown here]

Eric Dujardin, Eric Amiel, and Eric Simon. Fast algorithms for compressed multimethod dispatch table generation. In Transactions on Programming Languages and Systems, 1996.

No context found.

Dujardin, E., Amiel, E., Simon, E.: Fast Algorithms for Compressed Multi-Method Dispatch Table Generation, TOPLAS Journal, vol. 20, no. 1, Jan 1998, p116-165.

No context found.

Dujardin, E., Amiel, E., Simon, E.: Fast Algorithms for Compressed Multi-Method Dispatch Table Generation, TOPLAS'96 Conference Proceedings, 1996.

No context found.

Dujardin, E., Amiel, E., Simon, E.: Fast Algorithms for Compressed Multi-Method Dispatch Table Generation, TOPLAS'96 Conference Proceedings, 1996.

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