First off, I missed the very thing I was most looking forward to, in large part due to getting caught up in a great BoF, run by a friend and former colleague. This was my fault, I had fully intended to have dinner with the Microsoft team. My apologies to them and to the other guests. More about this a little later.
I am uploading pictures/photos/movies, including about 15 minutes of Ray Kurzweil’s keynote. Video is not that good (taken from far away in a dark room), the audio isnt that good either, but you should be able to hear most of it.
Mr Kurzweil is very interesting. He makes some of the same cases we have been making about data growth rates, computational demand rates, though he has been doing it far longer than I. I am not sure I like his “paradigm shift” bits. Buzzword enabled arguments.
Short version: Exponential data growth rates across the board, exponential computational power growth rates across the board. The latter may come to an end in current technology, the former will never end, and is in fact accelerating. Moreover, with additional power, new things become possible, new markets are open.
He demoed his portable reader. Very cool (though a part of me wondered if it was “staged” with a pre-recorded segment … would have been a cool mashup regardless).
Basically his idea is that technology is an enabler, computing technology is an infinitely flexible enabler, and its growth allows us to explore new approaches that have not been possible before then. Yes. Absolutely.
Give a scientist or engineer a more powerful tool and let them create: not just more productively, but give them the chance to ask the “what-if” questions.
This is where accelerators come in.
I walked by the MD-Grape bit. In a little SGI case. These units allow researchers to ask “what-if” questions, by supplying so much computational power to molecular dynamics simulations, that the research may be able to directly interact with the simulation, to change it, to measure things, to re-run with varying conditions. Now apply this to misfold diseases, or similar things of medical interest. It may be possible for researchers to find structural agonists/antagonists for misfold based disease. Multiple metabolic chemical pathways may be simulated under a variety of conditions, and again, it may be possible to use the computing power to design solutions or treatments for metabolic pathway diseases.
I saw a DRC unit. Interesting approach.
I had a little talk with some group. NDA, can’t talk about it. Way cool, and you know who you are …
In the BoF meeting, the ISV basically stated that they were starting to look into acceleration technologies. She made exactly all the same points we had been making to the VCs over the years.
HPC is moving down-stream. It has been, and will continue to do so. Every time it takes another large step, price performance gets much better. Doug Eadline has been making some of these points in his value cluster and core-2 cluster bits.
see next post for part 2 …