Distributed applications using Java threads

There is much to be gained from an in-depth study of making existing multi-threaded programs more portable. A good example of where this is useful is in porting multi-threaded applications that have been written to use the Solaris thread API. At MVC much of the available hardware is Silicon Graphics based rather than Solaris based, and given that the Solaris threading API is one of the most prevalent, there is a strong motivation to gain information on porting existing Solaris based code to more portable threading models such as the POSIX thread API under C, and also to Java threads. This is to allow the code to be used on not only the Silicon Graphics machines, but also on other systems. The emphasis of the project has been on the production of case studies to document the steps necessary to port numerical applications to other threads models, and also the advantages and disadvantages of doing so; including performance analysis of the ported code. Much of the work has been dedicated to one study in particular, which involves the solving of Fredholm Integral Equations of the second kind, and follows on from an earlier case study at Sun Microsystems where the code on which the study is based was developed. 1