It’s interesting to see the future of chip technology – those little things that run our world now – beginning what appears to be something of a divergence.
In a recent Newsweek article entitled “Moore’s Law Doesn’t Matter” Daniel Lyons reminds us that computer chip performance has doubled every 12 to 24 months per Moore’s Law for about 50 years. He points out expert claims that we’re bumping up against the laws of physics. The problem isn’t so much capacity as speed. As PC chips reached 3 GHz speeds, they’ve found you can’t make them any faster, or they overheat and melt.
What to do?
Well, for some time nanotechnology adherents have claimed that they represent the future. Nanotech involves computing at the atomic level, and among its many benefits are miniscule scale and the fact that no power source is required. There are plenty of other benefits to chips that are the tiniest fraction of a human hair (technically, nano represents one-billionth of a meter). In essence, nanotechnology solves the heat/size problems of traditional chip technology, providing a path to the future.
Not so fast say the folks at Intel, Microsoft, et. al. These days we’re seeing dual-core and even quad-core chips that represent parallel paths of processing. Within a decade says Justin Rattner, CTO of Intel, we’re likely to see 100 cores on a chip.
The problem, he says, is we don’t know how to program them very well. Says Craig Mundie, Microsoft’s chief research and strategy officer, in the Newsweek article, “For 50 years we’ve done things one way, and now we’re changing to a different model.” It seems neither operating systems nor programming languages nor development tools are really set up for parallel processing. Programmers have spent their lives programming for a single processor, that is in serial, not parallel. There’s a big learning curve ahead. Say’s Microsoft’s Mundie: “It’s the biggest challenge we’ve ever faced.”
That might be saying a lot, considering the source. And while technically the applications for Nano vs. Parallel are different today, my bet is they will continue to converge, and the technologies will compete for dominance.
It will be an interesting race to see which tech-horse we will be riding into the middle of the 21st century.