Moore’s Law suggests that we get a doubling of computing power on chips of a given size every couple of years. Do we really need an ever increasing number of CPU cycles?
Yes of course. In future, tasks that currently take weeks of supercomputer churning will be accomplished in near real-time by relatively common equipment.
Today’s invention is an example. Imagine a drill press which has access to a massive database of engineering components. It also has a supercomputer built-in that can grab or generate a 3-D mesh of the barcoded component currently on its table.
A laser shines down the drill axis and is imaged by a camera so that the system knows where and how big a hole is about to be formed.
In under a second, it calculates the stress distribution of the part, post drilling, under typical loading conditions, and projects that onto the surface as a ‘heat map.’
Too many stress hotspots generated? Move the drill to one side and recalculate. Similar things could be done for fluid flows as part of the process of engine tuning.