June 12, 2004

• OK, let me see if i understand this correctly. there are these (old) supercomputers out there and the public has access to them once a week?
• Maybe they are old. But they are still very powerful. It's a Godsend for those who don't have normal access to supercomputers and at least want to try to learn how to use them.
• One question, is there any distributed software like SETI@home where you can actually test your own simulations? Please, someone answer. Don't make me do a Curious George out of this.
• dumb question: what can a supercomputer do that a regular computer can't? for the average public, i mean.
• They could generate a lot of heat. No, seriously. A "super" computer's hat trick was the ability to perform certain vector calculations simultaneously, in parallel. (A conventional aka von Neumann computer performs all calculations sequentially: first one, then another, then another, etc.) Certain classes of problems requiring massive amounts of computation (e.g., weather simulation and forecasting, molecular modeling) could be solved semi-tractably by supercomputers. Nowadays, these problems are usually cooperatively attacked by groups of small computers. The vectorizing approach still has its strengths and adherents, however.
• but once again, what's in it for John Q. Public? why would the average person out there need access to supercomputers? to play?
• I'm not so sure wadaya mean by average Joe. You mean someone who is not a Phd. in computer science or something similar who probably already has access to supercomputers in universities or research centers? I'm that kind of regular Joe. Have no access whatsoever to that kind of hardware. Yet I'm a computer science fan who likes to learn new things and do (probably useless) research on his own and would like to have a least a chance to don some kind of experiments with vector programming and the like. Then again. There are many Phd.s and graduate students who have no access to supercomputers who would be very glad to have this kind of opportunities. Or, by your standarts. An average Joe is someone who doesn't use computers for anything else other than regular stuff? Dont' get me wrong. I'm not starting a grudge with you Sidedish. I just wanna make clear that computer science is not rocket science and in the bottom we are all regular Joes. Some more regular than the others. ;)
• why would the average person out there need access to supercomputers? pr0n! the 'average person' out there would never need a supercomputer. But there might be some 2nd or 3rd world countries that might need access to a powerful computer, X-Prize contestants civilian space travel or companies' commercial interests. on preview, what Zemat says
• ah! third-world countries. OK. yeah, i'm just trying to understand why there SHOULD be public access to supercomputers. i just figured only scientists/computer folks would know how to run them, what to do with them, etc. i remember when university of illinois got its supercomputer (yes, i'm old) and it was quite a huge deal. maybe that's why i'm puzzled by this. it's gone from being a huge deal in a major university, to folks collecting the crays and offering public access on the internet.
• why shouldn't there be public access to supercomputers? I can never see me using a supercomputer for any valid* reason, but why limit it to just acedemia, governments and super-mega-corps? Maybe some nut out there can validate his(her) FTL theory, prove dark matter, or create a longer lasting light bulb. *I'd pay 5 bucks for a certificate that read:" I have an account on a Cray Y-MP Supercomputer system and all I did was veiw pr0n," just because I could.
• One of the ideas I liked the most from Permutation City is that of public access distributed computing. Where anybody with access to the Net could buy processing time from a vast global array of supercomputers to do whatever they wanted with it. That way, a loner geek/programmer managed to create a new form of life inside a virtual environment. I'm not saying anything more about the novel. It's more complex and wild that you could ever imagine, but completely OT. Anyway. One project I started working on but never lifted from paper was implementing some sort of client for a distributed system like SETI@home that could be used to run scripts that could take avantage of a distributed environment. Now I just expect someone else can come with something like it, heh.