These are free software. They are distributed in the hope that they will be usefull, but without any warranty. You can distribute them and/or modify them under the terms of the GNU general public license as published by the Free Software Foundation.

Software Packages

  • Naplet: Mobile agent system for network-centric applications
  • LPR: A Library of Run-time Loop Parallelization
  • Accelearated News Surfer
  • DSA: Distributed Shared Array: An integration of message passing and multithreading

Naplet: Mobile Agents for Network-Centric Applications

Naplet system is a Java-based secure and flexible mobile agent framework in support of network-centric distributed applications. It features a structured navigation mechanism, a flexible inter-agent communication facility, and secure interfaces to legacy and priviledged Internet services.

  • Overview and Architecture
  • Naplet 0.18 API
  • Download the software: Naplet-0.18 Source Code
  • Naplet-0.12 uses GNU make utility to simply the system deployment. The package also contains installation instructions on systems without make utility.
  • See for more details

LPR: A Library of Run-time Parallelization Algrorithms for Loops with Dynamic Dependences

Loop parallelization focuses on cross-iteration dependence analysis and parallelism exploitation. For loops that have statically insufficiently defined access patterns, compile-time analysis must be complemented by techniques capable of exploiting parallelism at run-time. The objective of run-time parallelization techniques is to exploit enough cross-iteration parallelism and meanwhile reducing the run-time overhead. LPR contains a group of run-time parallelization algorithms for loops that have indirect access patterns on symmetric multiprocessors. They follow the inspector/executor scheme and exploit parallelism at either array reference and iteration level.

  • Reference: Time stamping algorithms for parallelization of loops at run-time
  • Download the software: LPR code

Accelerated News Surfer to tolerate Web latency

With the explosive growth of WWW applications on the Internet, users are experiencing access delays more often than ever. Recent studies showed that pre-fetching could alleviate the WWW latency to a larger extent than caching. Existing pre-fetching methods are mostly based on URL graphs and limited to retrieval of URLs that are frequently accessed before. SmartBrowser deploys a novel pre-fetching technique to retrieve documents whose URLs have never been accessed before. It relies on keywords in document captions to capture user access patterns and on neural networks over the keyword set to predict future requests. It features a self-learning capability and good adaptivity to the change of user surfing interest.

  • Reference: Neural net based prefetching to tolerate Web latency
  • Source code to be released soon