Another action-packed day at the 2010 GPU Tech Conference in San Jose… toward the end of the day, we sat down with Andy Keane, NVIDIA’s GM of the Tesla GPU business unit and talked about a wide variety of issues. It’s been a busy week for Andy, given that his product is the focus of attention all this week.

In our few minutes together, we talked about how the system vendor announcements from IBM, Cray, and T-Platform put the NVIDIA GPU in a different light for many potential buyers. Having their products as standard catalog offerings from major system vendors marks a milestone for GPU computing in general and Fermi specifically.

Having a Fermi blade offering for IBM’s BladeCenter opens up a vast new market for these products, given the large numbers of BladeCenter chassis in IBM’s customer base. To me, this move by IBM, along with moves in the same direction from Dell and HP, paves the way toward wider enterprise adoption of GPU computing. For you hardware buffs, I plan to get some video of these new products and post it in a video blog tomorrow.

We also discussed the idea of using GPUs as a shared enterprise resource – like NAC (“Network Attached Compute”). It’s a model that Andy thinks has some validity – but, then again, he could be humoring me. Speaking of humor, we got him on tape calling a someone an idiot – a moment funny enough that I had to leave it in. (He was kidding.)

On the downside, I forgot to get him to say, “GPU-riffic!” It occurs to me that I’d be much further along in life if I could remember something, or even just stick to a simple plan, for longer than five minutes.