T. C. Bell, J. C. Cleary & I. H. Witten, Text Compression, Prentice Hall Adv. Ref. Series, 1990.  Home/Search   Document Not in Database   Summary   Related Articles

....size (total number of pages in the database) and k be the cache size. In typical databases, ff is large and k ff. In this section, we describe our three simple, deterministic prefetching algorithms based on practical data compressors. An elegant discussion of the data compressors appears in [BCW]. We describe our prefetchers in Sections 3.1 3.3 in their generic form, as pure prefetchers that can store their entire data structure in cache. These prefetchers make k suggestions for prefetch ordered by their relative merit. To make these suggestions the algorithms use O(k) time. Sometimes ....

.... traversed. For example, in Figure 1, at node x we can store counts of 1, 3, and 1 at the three transitions (instead of the probabilities) The same comment holds for the PPM and FOM algorithms described below. In our simulations, we use a heuristic for LZ that parallels the Welsh implementation [BCW] of the Lempel Ziv data compressor. While LZ is at a leaf, instead of fetching in k pages at random, it resets its current node to be the root (that is, it goes to the root one step early) However, it updates the transition counts for both the leaf node and the root. 3.2 Algorithm PPM Although ....

[Article contains additional citation context not shown here]

T. C. Bell, J. C. Cleary & I. H. Witten, Text Compression, Prentice Hall Adv. Ref. Series, 1990.

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