BusinessWeek is running a piece today on the supercomputing for the masses story. They smartly identify that it’s software — not hardware — that is the real catalyst needed to ignite that market
But what may dash the dreams of Intel and other hardware makers is a lack of inexpensive, off-the-shelf software to bring supercomputing to the masses. For now, these sophisticated machines require equally sophisticated and, in many cases, custom-developed programs tended to by highly paid engineers. That’s why Microsoft is building a brain trust and handing out funds to schools doing work in the field. The software giant is underwriting grants to universities to study how supercomputer-style programming can be applied to personal machines.
Even if the software problem is “solved,” apparently not everyone agrees there is a need for supercomputing performance everywhere
“Is this whole infatuation with performance something that has moved beyond what the vast majority of users really care about?” asks Intel’s Rattner. “Are there really a set of applications that require 10, 100, 1,000 times the performance we have today? And if we have it at an attractive price point, will it drive high volumes? It’s still to be determined,” he says. “There are still people who question whether the volume markets are there for all this performance.”
Intel obviously has its hands on way more data and smarts than I do, but I disagree with this assessment. The problem is that it assumes people will want to plug HPC systems into their current environments and continue to solve today’s problems. I think this is wrong in the same way as, say, arguing against mass electrification at the turn of the last century on the basis that gas does just as good a job would have been. Electricity wasn’t only used to replace gas lights in people’s homes, its mass availability led to the feasibility of all sorts of new and improved devices, from doorbells to toasters.
As the usefulness, and usability, of HPC increases, the increased amount of computation available to people has the potential to enable new categories of computation — even somewhere as boring as the back office — for its users. Note that this is only a potential. We have no guarantee the mass availability of HPC will lead to the kinds of revolutions in our culture that electricity has. But it is a potential that is worth investing in.