Big data processing will be the future of innovation. Our greatest discoveries will come not from carefully conceiving a theory and testing it, but by throwing every possibility up against a CPU wall and seeing what sticks.
And as GigaOm points out, we're quickly reaching a point where the CPU in its current form just won't do.
The shift is a move from creating scads of information in a format that can be stored cheaply, to being able to process and analyze that information more cheaply as well (all the while adding new layers of data thanks to a proliferation of devices and networks). The challenge is that under the current computing paradigm, adding more processing is problematic both because it's becoming more difficult to cram more transistors onto a chip, and those chips and their surrounding servers are sucking up an increasing amount of power.
The solution? Rethink its design. GPU computing, while great for supercomputing, is inferior when it comes to raw number crunching [To clarify, GPUs are better for parallel sets of data that utilize the same operation/algorithm, while the standard CPU is better for processing a single, large, complex set of data. Thanks, Nitesh] And packing more and more transistors in different configurations--Intel's 3D chips, for example—will only go so far. HP has been reconsidering every aspect of the processor, and think they found the future of computing in a new circuitry component: the memristor.
HP's answer is its concept of nanostores, chips that tie the memory and the processor together using a completely new kind of circuit called a memristor. The basic premise for HP is that 80 percent of the energy inside a data center is tied to moving data from memory to the processor and then back again. We're already seeing the trend of moving memory closer to the processor (that's what the addition of Flash inside the data center is about) to speed up computing.
And even beyond the preservation of Moore's Law or market relevance, the need for companies pushing CPU design forward is important for the future of science. Take for example, this New Yorker piece on quantum computing and the leading mind behind it, David Deutsch. It's big idea is that if we build a working quantum computer, it could theoretically process more numbers than there are believed particles in the universe.
What's that good for? It could prime factorize absurdly large numbers in a matter of seconds. And it could prove the validity the Many Worlds Interpretation. The theory essentially supposes that there is a different universe for every possible permutation of anything in the universe. One scientist, Peter Shor, developed an algorithm for quantum computers that would potentially support this theory, if we ever had a quantum computer powerful enough to run it on.
The theory also explains how quantum computers might work. Deutsch told me that a quantum computer would be "the first technology that allows useful tasks to be performed in collaboration between parallel universes." The quantum computer's processing power would come from a kind of outsourcing of work, in which calculations literally take place in other universes. Entangled particles would function as paths of communication among different universes, sharing information and gathering the results. So, for example, with the case of Shor's algorithm, Deutsch said, "When we run such an algorithm,, countless instances of us are also running it in other universes."
[Image via Flickr/velvetkevorkian]