Ueda, K., Optimization of Concurrent Logic Language Implementations. Report of the contract research, Institute for New Generation Computer Technology, 1994 (in Japanese).  Home/Search   Document Not in Database   Summary   Related Articles

....to the program size n and the complexity of the data structures used (in terms of the size of the grammar to generate them) The size of the data structures to be generated 3 does not matter. The practical cost of separate analysis is worse only by O(log n) times. The readers are referred to [10, 9] for mathematical details and basic theorems. As mentioned before, KL1 supports data structures called vectors (one dimensional arrays) Vectors can take any terms as elements. Several vector operations are provided, but under the static mode system, the most fundamental operation turns out to be ....

....through the new array NewV. All the other elements of V are inherited to NewV and thus remain accessible. In contrast, an access operation that simply returns a new reference to the Ith element of V, such as Prolog s arg, will impose far stronger mode constraints on V and its elements; see [9] for the reasons. The moral of the above result is that, in the array processing in Moded Flat GHC, array elements should be removed once accessed and the resulting blank should be filled with another value. For instance, when performing some operation on an array element and storing the result ....

[Article contains additional citation context not shown here]

Ueda, K., Optimization of Concurrent Logic Language Implementations. Report of the contract research, Institute for New Generation Computer Technology, 1994 (in Japanese).

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