Some mathematical problems can be broken down into arrays of smaller pieces. An array processor is very efficient at working vector and array problems, versus a single or dual processor working one part of the array in a serial fashion.
One example: Atmospheric modelling and weather prediction. If the whole atmosphere is divided into 1 cubic kilometer chunks, then an array of 25.76 billion processors -- one assigned to each of the 25.76 billion cubic kiometer cubes in the atmosphere could go to work to model the atmosphere. Each processor would be responsible for all the conditions of its own cube, taking in data from each adjacent cube (shared memory or comm ports to and from the processors). The 'conditions' would be temperature, moisture level, clouds, cloud types, amount of sunlight passing through (or none at night), air density, etc, etc, etc, All the interactions would be calculated, so the model would be accurate.
.
2007-06-25 07:01:49
·
answer #1
·
answered by tlbs101 7
·
0⤊
0⤋