From: ptkwt@...1.aracnet.com (Phil Tomson) Date: 2002-01-25T07:16:48+09:00 Subject: Re: looking for an example problem to demonstrate TaskMaster In article , Rich Kilmer wrote: >How about searching for a file in the filesystem (or a distributed grep). > >-Rich Kind of assumes you've got the same filesystem available to all clients... Phil > >> -----Original Message----- >> From: Jack Dempsey [mailto:dempsejn@georgetown.edu] >> Sent: Thursday, January 24, 2002 1:12 PM >> To: ruby-talk ML >> Subject: RE: looking for an example problem to demonstrate TaskMaster >> >> >> How about some sort of search for a value? As you add more nodes to search >> the time goes up linearly. >> >> > -----Original Message----- >> > From: Phil Tomson [mailto:ptkwt@shell1.aracnet.com] >> > Sent: Thursday, January 24, 2002 12:57 PM >> > To: ruby-talk ML; undisclosed-recipients: >> > Subject: looking for an example problem to demonstrate TaskMaster >> > >> > >> > I'm looking for suggestions here... >> > >> > I was thinking of an application where the user would enter a range of >> > numbers to find primes in. The script would break this range down into >> > smaller ranges and use TaskMaster to send these tasks (finding primes >> > withing the given subranges) to multiple client machines. But >> that's not >> > a real good demo since as the numbers in the ranges get larger >> > the time it >> > takes to test each number in the range goes way up (probably >> > exponentially). So I'm looking for a problem that's a bit more >> linear so >> > that as you add more client machines it's obvious that the problem is >> > solved faster. It should also be something pretty easy to >> understand (and >> > program) since it's intended to be a demo. >> > >> > Phil >> > >> > >> >