From: drouri Newsgroups: comp.parallel Subject: an answer for teds@intex.com, Date: 20 Dec 1998 06:45:28 GMT Organization: CITU - Universite d'Orleans - FRANCE Approved: bigrigg@cs.cmu.edu Message-Id: <75i6e8$hja$1@encore.ece.cmu.edu> Originator: bigrigg@ece.cmu.edu Hi, a solution of your problem in a good execution time can be the REDUCTION ALGORITHME: first, each machine does an iteration an to calculate the average: 50 machines get a data from one of here neighbours and add it to their own data second, 25 machines (of the past 50) gets the (25) sums calculated below by the other 25 machines and add it to the sum they calculate third, 12 machines (of the 25) gets the 13 sums calculated below and add it tto the sum they calculate (one of the 13 will get 2 datas) fourth, 6 machines (of the 12) gets the 6 sums calculated below and add it to the sum they calculate fifth. 3 machines (of the 6) gets the 3 sums calculated below and add it to the sum they calculate sixth, one machine (whitch will own the result) gets the 2 sums calculated below and add it t the sum she calculates seventh and last, this machine calculate the average. Good bye, -- Drouri Said E-mail: drouri@lifo.univ-orleans.fr -- Articles to bigrigg+parallel@cs.cmu.edu (Admin: bigrigg@cs.cmu.edu) Archive: http://www.hensa.ac.uk/parallel/internet/usenet/comp.parallel