Recently, I had the opportunity to present on behalf of Adaptive Computing with our marketing automation provider, Act-On Software at their annual user conference (I Heart Marketing). This was an interesting experience, given that the audience was not limited to technology companies. In fact, they covered the gamut of industries and company types. I was justifiably concerned about stumping for an HPC company at a broad users conference, given most of the users might not know much about HPC, let alone know what it is. The cloud side of our business was a given – it’s such an industry buzz word, that other companies were saying they were “cloud,” when really their delivery model would just use some cloud resources. The really interesting observation came when I spoke of HPC, and to be honest, it sparked some very interesting conversations.
I talked about Adaptive’s involvement in HPC, and some of our customer’s use cases. I always love telling people; “We can tell you who some our customers are, but then we’d have to kill you.” The interesting thing is most people would bring up specific examples of HPC usage or data derived from HPC research, and it wasn’t too frequent I conversed to a deer in the headlights, which I expected more of. I had great conversations about weather modeling, cancer research, satellite image data, manufacturing, intelligence (some not positive), and much more. So, what does this anecdotal evidence suggest about HPC?
As many mainstream pundits have noted, HPC seems to be going mainstream—meaning the research and benefit that HPC provides is becoming more and more well known. I might be willing to wager that the average man-on-the-street knows about at least one Supercomputing use case, more than he might be able to define what exactly “the cloud” is. I felt somewhat validated after my marketing conference.
I’ve seen some great articles published the last little while from The Register, and we even had our own CEO Rob Clyde appear on Bloomberg TV stumping for HPC. The more interesting thing to correlate is not that HPC is becoming more known with the populace, but that HPC in and of itself might be going more IT mainstream. The Register has pointed out that more and more HPC systems are going to an x86 architecture, but why? Perhaps HPC is becoming more commoditized, or perhaps HPC is preparing for migration into more standard IT data centers, not just staying a resource silo. Maybe it’s just that more development and efficiency are being bred into x86.
In any regard, I think we’ll see some interesting changes in HPC going forward, especially in more commercial-centric resources, that bring HPC more into the IT mainstream.