PowerMac a "Supercomputer" according to PBS

[bowens]

I watched an episode of NOVA on PBS about Hurricane Katrina and they showed a clip of a guy doing some models back in 2004. Here is the exact quote: "The fictional hurricane Pam was created by Ivor Van Heerden in his supercomputer..." He was using a PowerMac. I just thought this was kind of funny. I know you all know that you have been using supercomputers, but it's good to here someone else on the outside commend Apple for it's "superness."

Anyway, I thought you guys might like that. Here's a link to the video if you want to watch it. http://www.pbs.org/wgbh/nova/orleans/program.html It's pretty interesting. Although I liked the one about the DARPA race for unmanned vehicles better. http://www.pbs.org/wgbh/nova/darpa/program.html It's amazing that they had 5 robots complete this 130 something mile course. It's incredible. The farthest any went the year before was 8 miles.

 

[kevin1182]

Not so sure about that

Hi,

Although you may be right, I have a different theory on that. I am a meteorology graduate student who has run plenty of numerical simulations of hurricanes, and I believe that this individual probably *ran* the model on a "supercomputer" and for the interview was showing output from the model on his Mac. Most models are run on computing clusters to allow for a greater number of computations to be made in less time. For instance, model runs that I have done have been performed on up to 64 nodes at the same time; running the same thing on one node would obviously take a lot longer.

So, while he was looking at output on his Mac (which I do on my Macbook as well) I wouldn't be surprised if he ran the model on his "supercomputer" which may be what the folks at PBS were referring to when they used that term.

But I do think the Mac is a super computer, don't get me wrong

Kevin

 

[bowens]

That's true. I didn't think about that. It's funny though, since I got my mac last year, I have noticed them everywhere from commercials to tv shows to movies. I rarely ever see pcs on tv except for pc commercials.

 

[mkrishnan]

Kevin, is any technology like XGrid that pulls processing power from distributed desktop class CPUs in use by your group?

(At Michigan also... my lab group when I was in engineering did work on massively parallel machines... typically though the user activity was either at a Sun workstation or some Windows PC....)

 

[kevin1182]

Hi,

I do not believe that xgrid is being used....I am by no means an expert in the system at all. If you were curious you could check the cluster out at:

http://www.pams.ncsu.edu/pco/hpc/linux.html#specs

From what I know, it is just a cluster of processors running redhat linux. We access the cluster remotely by using ssh. After running the model we usually move files off of the cluster to local machines/storage for data analysis; the cluster is just used for its horsepower and the ability to run large computing jobs quickly.

Kevin

 

[mkrishnan]

Thanks! My cousin went to NCSU!

I was just wondering... when I was in engineering at Michigan (I'm in psychology at Florida now), it was kind of a hot topic as to whether all the lab computers that are in partial / non-use could effectively be harnessed together to provide additional flexible high performance computing environments. XGrid is the Apple technology, but there have certainly been other methods for years.... I'm kind of surprised they've never caught on at Universities. Think about how many perfectly good Pentium 4s are sitting around in labs not being used? The folding people have certainly put an elaborate form of this to good use.

 

[bowens]

mkrishnan, how long have you been at Florida? I graduate from there in May with a bachelor's degree in Business Management. I still live about half an hour away where I grew up.

 

[mkrishnan]

mkrishnan, how long have you been at Florida? I graduate from there in May with a bachelor's degree in Business Management. I still live about half an hour away where I grew up.

I got here for grad school in the summer of 2004. I'll probably be here through May or June 2008, and then go away for an internship year, and then come back to collect my Ph.D., with luck, in April or May 2009.