SC|04 Report

Haven't posted for a few days since they've cut the attendees off from the conference center and consequently free wireless access, presumably for the bandwidth challenge. Over the past three days, we've been talking to a number of vendors to see what needs to be done to build a HPC center at UTD. I've also been talking to various labs to see what people have been doing in as far as combining HPC and biology (for my work at UT-Southwestern); a lot of work is being done on molecular dynamics problems as well as whole genome work, but nothing seems to be immediately useful for the lab (which is currently involved in SNP association studies as part of the Dallas Heart Study). On the upside, a potential collaboration may be in the works.

There also seems to be a great deal of work being done on grid network and services. This concept of distributed computing seems to provide an attractive solution for providing HPC services although I'm not entirely clear on how a consistent business model could be built around it. Government funded systems I can see, but how a company could make money by directly paying for grid services is something I can't.

Day three (Wednesday) was fairly interesting. I attended a talk regarding the making of Shrek 2 and the influence of HPC on the animation industry. It's amazing the amount of work that goes into producing a single shot - not just in terms of raw computing power, but also the amount of thought that goes into mixing real and cartoon physics. For example, the number of layers and simulation needed to animate rain falling and scattering off objects.

Moreover, day three had one hell of a party. First, the Apple folks brought us over to the Omni William Penn Hotel for a "brainwashing" session, i.e., presentations trying to convince us (and fairly convincingly) that Apple can be and is a player in the HPC market. Then came the SGI party. We (Aidan and I) traded in our casino "money" for extra tickets and proceeded to be a little inebriated. This was clearly not an academic conference, with all that free food and booze.

Day four was just a "clean up" session. Visited a few remaining booths, picked up some extra goodies, and then attended a last party for the technical conference attendees. Also, got to see and play around with Blue Gene, which is currently rated as the fastest supercomputer in the world; much fun there.

Finally, Wanda Sykes had a good comment on Leno Tuesday night: "You had to cheer for Bush [in the last election]; otherwise, it would be like booing at the Special Olympics."

Anyways, sitting in Pittsburgh International fairly tipsy and I'm about to board my plane to head back to Dallas. More later.

SC|04 Day One

5:30 pm EST
Sitting at 30,000 feet on my way to Pittsburgh writing this. Tonight is the first night of the SC04 conference, and seven of us from UTD are headed there as part of a reconnaissance force. The school has a lot of research money from TI (as part of that $300 million deal), and there's an effort to spend part of that money on starting a high performance computing (HPC) center at UTD. Our jobs is to take a look at what's out there and to pick the brains of those in the business - in other words, get an idea what's needed to get this project off the ground.

11:00 pm EST
Just got off the floor and the opening ceremony. So far, it appears to be one big trade show, though most of the interesting presentations don't start for a few days. The expected big players are here (IBM, Hitachi, NEC, SGI, Cray, various government labs (Fermi, LANL, etc.), and universities), and the products seem really cool. I'll probably spend more time on the floor over the next few days getting more details (when I'm not touring Pittsburgh). Interestingly, Dell and Microsoft were also on the floor. MS I can see since they've announced their HPC initiative, though I'm skeptical whether they can get it right for a long time. Dell I really can't understand; given the amount of grief the 6650s at work have given me, I really can't see how they could even consider themselves a viable contender in this field. Then again, maybe they'll do something right and surprise me (but I really doubt it). In any case, security at the convention center is getting a bit antsy about the five college kids hanging around late at night so I'll post more tomorrow. (Oh, and it's really nice for the university to pay for us to stay at the Westin until Friday...even if the hotel did screw up and only gave us one room for seven people.)

Election night!

Finally, it's election night, and I'm watching the best in election coverage -- The Daily Show: Indecision 2004 AKA "Prelude to the Recount". It's a bad sign regarding the state of our media that one of the better sources we've got is a show on a network called "Comedy Central". Granted, it's far from being accurate news, but given the press situation from the last two or three years, I've pretty much given up on accuracy in the media. On the upside, at least Jon Stewart doesn't try to present himself as a serious journalist while acting as nothing more than a pundit. For those of you who've seen Stewart's appearance on CNN's Crossfire, I think you'll understand what I mean.

Some highlights:

• Al Sharpton came out swinging away at the Republicans, and in the words of my roommate, "Man, he is pissed!"
• Bob Wiltfong discussed e-voting, specifically how we've combined the obscurity of the 17th century voting system with the lack of accountability of modern electronics. Somehow, it ended with the comparison of determining results and Tron, which I don't entirely understand.
• There were a few more short pieces, which while amusing, didn't really stand out.
• Finally, there was Stephen Cobert's call against national healing; after all, hasn't all this divisiveness made our nation more politically active?

Now at 12:10 AM CST on 3 November 2004, it looks like Bush will take this election, which I find really disheartening. Of course, this is based on exit polls (you'd think we'd have learned from the last Presidential election), so it's entirely possible that I'll wake up tomorrow and find out that the polls are wrong.