From: Vasileiadis Spyros Newsgroups: comp.parallel.pvm Subject: Re: What good is it, anyway!? Date: Thu, 22 Jul 1999 01:39:03 +0300 Organization: GRNET - Greek Research Nerwork Message-Id: <37964C07.846071F6@it.teithe.gr> References: <7n22d0$u14$1@nnrp1.deja.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-7 Content-Transfer-Encoding: 7bit Xref: ukc comp.parallel.pvm:8585 Well i fell the same way my friend so I decided to go one step ahead (if this is possible) I am now trying to figure out a way toy hack linux kernel so that new created threads will be distributed to the clusters without effecting the behavior. So every existing program that simply forks will run in parallel without reprogramming or recompiling!!!!! The programmer will just think that he is working on shared memory multiprocessor system and not on a distributed memory parallel clustering system any syggestions are wellcomed ambros97@my-deja.com wrote: > > I'm a wee bit disappointed about what I've seen so far. > To make a long story short, if you want to take advantage of PVM you > need to rewrite your application. > > How about rewriting the application in such a manner that you can run > one instance on each computer? That will give you the same "parallel" > tasking advantage but without the syncing overhead. > > Seti@home is a good example: You don't need to make this PVM capable, > simply run one instance of the program on each client. If you have to > redesign the software, you're running into a lot of work anyway. > > I was expecting to see something like an interface that tricks a > multithreaded application into believing there are as many processors > available as there are in the cluster. Now that would certainly help! > You write your problem solving app once with multithreading > capabilities and off you go. My renderer XYZ would be lightning fast! > > If anyone would like to comment, please also send to my mail address. > > - Ambrosius > > Sent via Deja.com http://www.deja.com/ > Share what you know. Learn what you don't.