I remember talking to someone about nanotechnology once, and he said that the term may always be used to describe something coming down the pipe, rather than something that exists today. His point was that as nanotechnologies mature into actual products, they tend to get called something else. Kind of like the way even the oldest, clunkiest PCs in our offices today would be considered supercomputers to the IT industry of yesteryear.
As the annual Supercomputing conference gets under way in Austin this week, the show’s organizers sent out the bi-annual list of the Top 500, machines that offer breakthrough performance in the number of calculations achieved at unheard-of speeds. This year, thanks to a University of Toronto project profiled on both ITWorldCanada.com and ITBusiness.ca, Canada made it into the top 100 at No. 54. Which is great, I guess, if it made any difference to anyone else in the local IT industry.
Oh sure, eventually the processing power of these machines makes its way into mainstream hardware, but by that time it’s old news. Right now, the Top500 list offers a glimpse into the future but without encouraging users to explore the full context of what these machines are doing. In some ways the list reminds me of the back-and-forth battle waged between Intel and AMD. Originally it was based on clock speed (“We’re at 2 GHz!” “Oh yeah? Well we’re 2.5 GHz!”), and now it’s about cores (“We’ve got dual cores!” “Oh yeah? We’ll we’ve got double-dual-dynamite cores!”). Meanwhile, users continue to buy PCs with previous-generation chips that basically offer what they need for most software programs.
And yet if you dig a little deeper, the Top500 list offers some great stories about how IT is being applied to complex problems. Some are being used to find cures for terminal illnesses. Others could teach us more about why the solar system functions the way it does. Unfortunately, if you click on the tabs of the actual list, all you’ll see on winners like the Dawn 5000A is a one liner that says “Application area: Research.”
The Top 500 list is one of the key ways the Supercomputing conference speaks to the outside world. Unless you work for a major research facility, that includes a lot of IT professionals using everyday equipment at everyday firms. What’s the point of advancing the hardware if they don’t get inspired by what these machines can do? We’ve tried to flesh this out in our coverage, but it’s a major area in which the actual teams behind the supercomputers – through blogs, wikis or other publishing mechanisms – could provide us a rich education in truly leading-edge technology.
Forget the Top 500 list of supercomputers. It should be the Top 500 list of supercomputer projects. As with any business investment in IT, the computer is not as important as what it allowed the organization to do. Or put it another way, I’m less interested in whether someone ranks No. 1 on the Top 500 list than what they did on the way there.