The reports that are popping up all over the web that IBM, along with Georgia Tech, has demonstrated the world’s fastest ever chip are, sadly, not quite as exciting as you might at first think. Not least because this wasn’t a chip at all, but rather a transistor, and even the least technical minded of people will realize there’s something of a difference between the two. Looking beyond that basic misunderstanding of the facts, the news still fails to excite me as much as it seems to have excited everyone else and here’s for why:
Yes, it did run at more than 500GHz. Yes, that is 125 times quicker than the current batch of commercial chips. Yes, you do need a near absolute zero cold room (with a temperature of minus 451 Fahrenheit, or minus 260 degrees Celsius if you prefer) to achieve that speed. And yes, you did read that right. Within a normal environment you’ll be more likely to see a still none too shabby 350GHz which is impressive considering that your average commercial chip at the moment manages less than 4GHz, but sadly not record breaking.
Still, building a transistor made from silicon laced with germanium that runs so fast is good news and bodes well for a super-speedy computing future. It also suggests that there’s plenty of performance still to be squeezed out of silicon, albeit under rather unrealistic conditions and using very expensive base materials. What it doesn’t answer is how all the other computer bottlenecks can be overcome in order to exploit anything like the kind of performance a chip based around such technology could produce. But that will, no doubt, come with time, research and funding.