Who decided that it was desirable, even necessary, to have the processing power of a small server behind the dashboard of my car?
There was a time when cars were simply a convenient, if energy inefficient, means of getting from A to B. Not anymore. Today we spend so much time in our cars that we turn them into mobile offices and playrooms for our children.
Consequently, it no longer seems out of place to talk of having a 2,000 million instructions per second (MIPS) dual-core microprocessor behind the dashboard powering our in-car multimedia equipment.
And quad-core processors will take that performance to a staggering 4,000MIPS within 12 months.
That sort of processing power will give us truly useful things such as image recognition and eventually collision avoidance technology. But right now the main application seems to be providing whizzy 3D graphics for your sat nav.
Desireable as that might be to some drivers, I cannot help feeling we are still in the phase of technology looking for an application when it comes to applying multi-core processors to anything other than a desktop PC. Even here the true performance gains can be questioned.
Embedded systems developers are not as easily convinced as the consumer electronics product designers. They want to see true performance and efficiency gains from multi-core processors before they adopt them wholesale. But the lack of effective program partitioning algorithms, needed to achieve the performance benefits of multi-core processors, means the embedded market is taking a measured, not to say cautious, approach to multi-cores.
But the microprocessor companies need not worry. They seem to have created sufficient demand for multi-core technology in the consumer market to more than fill the void before the embedded guys jump on the bandwagon.
So I am off to the computer centre to purchase my new sat nav terminal. It must have 3D map graphics and I would also like an Internet browser and digital TV reception too!
Leave a comment