As human beings, we take a lot of things for granted; breathing, for example, or our heart beating when it should. I don’t imagine many of us actually think about the neural pathways transmitting messages to our brain and then to the involuntary muscle groups that make sure we keep breathing, blood keeps pumping, etc. We just accept that these basic functions will adapt and carry on regardless of what exterior stimuli we send its way –or in some cases, don’t send its way.
The same could be said for our expectations regarding computers. We expect to plug them in, turn them on, and they’ll just work. Not only that, but when we upgrade our system, we expect to work faster and better and have more bells and whistles than its predecessor, why? Because that’s what we’re lead to believe by the sales guys who talk about the system speed, RAM, size of processor, etc. You really don’t hear them talk much about the on-chip and system software that is the PC’s own central nervous system.
Unfortunately, just like the quadriplegic whose ability to breathe is compromised due to insufficient neurotransmitter function to diaphragm, today’s quad core processors can’t function to capacity because their ability to ‘talk’ to each other is compromised due to insufficient software functionality. I touched on this very issue back in August after reading a Newsweek article suggesting Moore’s Law doesn’t matter without the proper programming in place.
I’m not one to support the naysayer, so I was happy to see that work is being done to right the situation and make Moore’s Law matter again. I’m talking about the European PRO3D Consortium recently launched by Leti. The main thrust of this consortium is not to address technology issues – those are well underway – but to now address the software issues of running a 3D manycore system by taking the holistic approach required to simultaneously develop hardware and software for 3D system architecture.
What I find curious is why, as a byproduct of the previous 2D architecture, this software issue was never addressed. IDMs were cranking out higher functioning cores without making sure they could function to capacity. The transition to 3D integration, however, gives us the opportunity once again to abandon the traditional 2D approach and develop the whole 3D system solution simultaneously. I suspect that as a result, we could experience a huge leap in PC function due not only to the 3D technologies being integrated, but that this time, all the pieces will be in place to see what these babies can really do.
I have to admit, up until now, I’d never given much thought to the notion that there is software running behind the scenes making sure for example, that the cooling system functions, or that the processors communicate. But of course, when you think about it, it makes sense. It’s just one of those things we take for granted.– F.v.T.
