From: "aallen" Subject: Re: Honours Topic Newsgroups: comp.parallel.pvm References: <36C24E38.697011FE@redgum.bendigo.latrobe.edu.au> Message-Id: <01be5880$c7407120$880f37a6@genisis> Date: Mon, 15 Feb 1999 01:13:23 GMT Xref: ukc comp.parallel.pvm:8031 In software development, there is certainly a couple: - thread design and implementations in O/S (could use NT, Unix & Linux as examples), Posix vs Non-Posix complaint designs, issues with each approach, benefits, etc. - Also consider the issues surrounding applying parallelism to problems. One of the key issues on why "supercomputing" has been limited is the inability to apply parallelism on a larger scale with more problems. Why is this true? What constrains a problem and the ability to apply parallelism to it? What makes a problem amenable to this approach? What does the future hold for this approach? - In networking you could certainly analyze the latency issue within a Supercomputing environment and compare that against other types of latency including software, etc. How much time is spent where, when - where should the average joe spend his/her time to improve program design/system design/etc. Nathan Martyn wrote in article <36C24E38.697011FE@redgum.bendigo.latrobe.edu.au>... > Hello, > > Could anyone suggest a "niche" in the parellel computing domain that > would be suitable for an honours > topic. My interests lye somewhere is software development and networks. > Any ideas would be > appreciated. > > Thank you. > > -- > > Regards > > NM > > reply : j6873276@ironbark.bendigo.latrobe.edu.au > >