Slashdot Mirror

By contrast, NCSA's surprise entry in November 2003's list, Tungsten, achieved 9.82 Tflops for $12M asset cost.

When I looked here, I found this: ``Tungsten entered production mode in Novermber 2003 and has a peak performance of 15.36 teraflops (15.36 trillion calculations per second).''

To me, that looks faster than System X, not slower.

Let's see: NCSA stands for ``National Center for Supercomputing Applications''. ``NCSA is a key partner in the National Science Foundation's TeraGrid project, a $100-million effort to offer researchers remote access ...''

Looks as if the NCSA has a huge budget. I'd guess that ``gold-plated everything'' and ``leave no dollars unspent'' are basic specs for everythig they buy.

What can we learn about Virginia Tech? How about this:

System X was conceived in February 2003 by a team of Virginia Tech faculty and administrators and represents what can happen when the academic and IT organizations collaborate.

Working closely with vendor partners, the Terascale Core Team went from drawing board to reality in little more than 90 days! Building renovations, custom racks, and a lot of volunteer labor had to be organized and managed in a very tight timeline.

Yes, System X and the Apple hardware is pretty neat, but don't use the price/performance ratio of these two systems as a metric for the relative worth of Linux and OSX clusters.

It's unfair and meaningless to compare volunteer labor and academic pricing and scrounging on a limited budget to bureaucratic design, bureaucratic procurement and an unlimited budget.

Actually, the reason hurricanes lose energy over land is the same reason they gain energy at sea, they are gigantic heat engines. Hurricanes form when an area of low pressure and wind shear is over waters above 80 F. The low pressure causes the hot, humid air from the ocean to converge together, which pushes it high into the atmosphere and causes thunderstorms. The horrific winds of hurricanes are caused by the coriolis force acting upon the extreme convergance of air. Trees and small things less than 500 ft would mainly affect the wind gust speed by compressing and deflecting the air around (although I'm not sure what a wind farm would do). Mountains however would rip apart a hurricane. Hurricanes lose energy whenever they don't have a nice fuel supply, which is why hurricanes, even though they form in the Pacific, never hit California with it's 70 F waters. More info over at the NOAA.

Then is Open BSD going to stop using GCC ? I mean, GCC is GPL so it is using GPL software to create their system, right ?

LLVM is a BSD-licensed C/C++ compiler (*) and is generating code which outperforms GCC's on some benchmarks, so at some point, OpenBSD may want to consider using LLVM.

(*) currently using GCC's GPL'd C/C++ parser, until someone wants to write a new one :)

On the other hand, as another poster wrote, when a big-time college sports program is successful, all of that gets swept aside. I had some exceptional professors in college. None, however, could be expected to pack in 70,000+ paying customers to attend their lectures each week- and certainly not to draw a TV audience in the millions. The sponsorships, the increased alumni support, the increased visibility of your school on national television- and even more directly, the huge sums of money that come with success. If Urban Meyer leads the Gators back to BCS-bowl glory just once, he makes 14 million dollars instantly for the University of Florida, paying off their investment in him. Also, while it may sometimes seem that way, there's no reason a school's academic and athletic achievement need to be negatively correlated. Look at Cal this year- did they put academics on the back burner over in Berkeley so they could have a great football season? Duke and Stanford have terrible academic reputations because they put together good basketball teams each year, right? Heck, my alma mater currently has the #1 men's basketball program in the nation, and I'd argue that all of the media attention they've been getting for it is a major positive for the school. (I will not discuss the Illini football program at this time.)

I will admit that Americans take our sports too seriously, and unfortunately we don't limit ourselves to taking highly paid "professionals" (the word belongs in quotes until athletes prove themselves worthy of the term again), but make way too much of games played by college students and even children. You hate to crush dreams, but it needs to be impressed on kids that all the "professional" athletes in the U.S. put together wouldn't fill the seats at an NBA arena. The question is, though, how do you change a child's heroes? I played football in high school, and had fun, but I never considered college ball, much less pro- I wouldn't have made the pros unless there was a lockout, and then the replacement players walked out, and their replacements walked out and so did their replacements. And probably not even then. I mean, I wanted to be a scientist pretty much since about the fourth grade. That's what I was excited about, and I made choices appropriate to making that happen. How we can get everyone else excited by math and science- not just creating mindset of it leading to a decent, sensible job, but making it genuinely exciting in the way that sports are, I wish I knew.

in a Central IL city that houses a major university with a top-ranked Engineering school and borders a city named similarly to an alcoholic beverage

Is there a good reason why you don't refer to UIUC by name, or are you just trying to be cute?

...which is precisely the proper way to group this sort of study as far as I'm concerned (as opposed to offering a MS with a specialization in "cyber security").

Here's the not-yet-extant Univ of Illinois - Urbana Champaign certificate program in Computer Security description

They'll probably supplement it with Network Security stuff, Intro to Crypto, ...

It's also damn handy when your old school IT manager refuses to allow installation of any other browser besides IE.

If that IT manager was really "old school", they would insist on using nothing but verion 1.0 of NCSA's Mosaic browser.

I have a BS CS degree from the University of Illinois (UIUC). That qualifies me to wash dishes using several layers of abstraction :-). The most important things are:

1. Experience, as everyone says, but what they really mean is your list of skills and ability to apply them.
2. Your first job out of college sets the direction for your career. Which job you take is more important than what your degree is. Where you work is more important than where you went to school.
3. Communications skills can either make or break you. They don't matter if your other skills are just right and you are the only one like you around. But your skills are never just right and there are usually lots of people like you around. What does "communication skills" mean?
   • The ability to speak and write in language that is pleasing to your target audience and appropriate for the setting. In other words, speak geek or PHB-ese as needed.
   • The ability to use the technology (email, phones, presentation software, etc.) effectively
   • The ability to sell. Everybody sells. This means being able to think on your feet and come up with effective arguments for your position, whether that's "buy our stuff!" or pitching your plan to the boss. Slashdot is the perfect training ground, except you're probably sitting.
   • The ability to speak and make a presentation to a group
4. Always have an escape route. When someone quits or gets fired, send them a note or give them a call. You never know where they'll be when you get canned. When you get a job, casually tell all your old contacts.
5. If you get canned, tell everyone immediately. Resist the urge to hide it. It will get easier as you refuse to be embarassed by not having a gig.
6. Never, ever, (ever!) lie on your resume. At best it will get you a job for which you aren't ready, and at worst it can ruin your career and land you in jail.

Not completly correct most plants convert solar energy at an ratio of about 1-3% however some like sugar cane perform at 8%.

http://www.uwsp.edu/geo/faculty/ritter/geog101/mod ules/ecosystems_biomes/biogeography_eco_energy.htm l

http://www.life.uiuc.edu/govindjee/whatisit.htm

http://employees.csbsju.edu/SSAUPE/biol116/Ecology /energy-flow.htm

However photosynthesis converts light at 100%, all sorts of biochemical reactions make the overal much lower.

Check out the first PLATO patent, issued in 1968 for a Versatile Display Teaching System. BTW, PLATO was conceived in 1959.

Plasma Panels, TUTOR and Teaching - Oh my

http://en.wikipedia.org/wiki/PLATO_system

http://web.library.uiuc.edu/ahx/uaccard/adminhist/ showsg.asp?rg=7&sg=13

The easiest way is to download something like IESPYAD which puts a whole bunch of domains into the restricted sites zone in IE. Just open the data file and start browsing. You can download it here:

https://netfiles.uiuc.edu/ehowes/www/resource.htm# IESPYAD

Another alternative is one of the many HOSTS files out there. Unfortunately, many of those also contain sites that serve ads, so you'll have to filter them yourself. Here are a few:

http://www.mvps.org/winhelp2002/hosts.htm
http://www.dozleng.com/hpguru/

And I know there's Camlp4 and MetaOCaml, but they're really too kludgy for my taste, though still better than the insanity that is C++ template metaprogramming. (Ick!)

[Stuff about lackluster optimization]

[Stuff about COCaml interface]

Speaking of changing the intensity of the colors individually... https://netfiles.uiuc.edu/bkpeters/www/LEDBed/inde x.html

Ummm... My Win machine is running SP4. Oh, you mean XP SP2. Not on my machines, man... The highest I'll go on my personal machines is 2k. Wow, thanks for that little tidbit. Now we all know that xp is t3h sux0rz. also, don't forget lynx, links, and Mosaic

I probably should not respond to a guy who does not even bother to use google, but here goes:

Look at this page. This is not complex stuff. I remember this from high school chemistry. The diagram on the bottom shows water. Plug in your temperature and pressure, and look to see where you are at (solid, liquid, gas). Below approximately 0.01 atmospheres, water can go directly from liquid to a gas. In fact, at such low pressures, you cannot have liquid water at all!

The first browser was called Mosaic, not Mozilla.

Mosaic was a separate, freely-available browser from which the company known as "Mosaic Communications" stole the name and the original developers. They very quickly changed their name to "Netscape" but kept the "Mozilla" name internally. The domain 'mcom.com' is still owned by Netscape/AOL/Time-Warner/etc.

The real Mosaic codebase itself apparently found its way into Internet Expolorer, if you believe IE's "help/about" screen.

See: http://archive.ncsa.uiuc.edu/SDG/Software/Mosaic/N CSAMosaicHome.html

No, Netscape was based on Mozilla.

OK, so you haven't been around that long. Netscape 6 and later was based on Mozilla. Mozilla itself started when Netscape open sourced Netscape 4. So Netscape is Mozilla's daddy.

If you go further back, they're all of course descendants of NCSA Mosaic.

Now how is this really different from IBM's project?

A skeptic might think that IBM simply want to have a foot in the door of these big anarchic distributed projects.

Despite the stunning power available to this kind of distributed computing, it is less useful than it appears. In my research area (computational biology), the effort of parallelizing an algorithm and collating the results is seldom worth the dividend in speedup. Supercomputers generally run idle at most universities, for this very reason.

Folding@home was a nice success story, and there are further applications of those models, e.g. simulations of prion aggregation (mad cow disease, Alzheimer's, etc). But (IMO) this is the exception, rather than the rule. Anyone who thinks that parallelization is a quick & easy panacea to difficult computational problems in general is living in a dream world (and I say that as a proud owner of several Macs with parallelized RISC CPUs *and* go-faster stripes).

I've lost count of the number of times I've heard these cheap parallelization ideas floated (another example is building cheap clusters out of console hardware which I reckon I first heard in 1996!). And every other month someone offers me supercomputer time... the problem is in redesigning the algorithm to work in parallel. Certain algorithms, such as MCMC, are better suited to this treatment than others.

Of course, then you have to persuade a bunch of other scientists that Your Algorithm is the most deserving, which is a political issue (but hey, if it saves those CPUs from being used for the eminently futile task of looking for bug-eyed aliens, maybe it's a good thing...)

Wearing socks + sandals is definately not going to score you any sex.

As seen here https://netfiles.uiuc.edu/bkpeters/www/LEDBed/imag epages/image10.html

He's wearing sandals with socks... he's completely guaranteed to be sex-free!

You can tell by the fact that he wears socks with sandals that he's not trying very hard to hide his geekosity with fashion.

And if you look at the pics at this link you can read something written in his girlfriends yearbook, and if you look *REALLY* carefully you can see her name is Michelle, AND you can find out the email address of her friend.

*sits back and watches the geek frenzy*

i dunno i think i see some rolls rolls there.

Hey, you guys should back off. At least he's getting some

what a dork?
https://netfiles.uiuc.edu/bkpeters/www/Prom/imagep ages/image4.html

Just goes to show what you can do with the money saved from not having to buy the LM317s.

Step 2: Tell women about your groovy pot stash and invite them back to frob the big knob.

That's one of the questions that I had at the back of my mind when I got my first PC, a Radio Shack TRS-80 Model I (with the 16K expansion!) in 1977. I could program the thing in BASIC, and learned some other rudimentary stuff, but really I didn't understand it. It seemed magical.

The question stayed with me through high school, until finally in college I learned about transistors, NAND gates, latches, full adders, microcode, machine code, assembler, compilers, UNIX, and how it really worked.

But it still seems magical.

On second thought I shouldn't say no success. There have been successes in computing special, less physical cases, for example in treating the stars as frictionless dustballs not possessing magnetic fields. But these features are very important in determining the rotation structure of stellar fluids, and are probably essential in modeling the physically correct binary merger. The general problem remains to be solved, and the goal is to figure out what physical processes produce gravitational waves, so that we know what to be looking for experimentally.

Here are some visualizations of previous merger simulations:
1
2

On second thought I shouldn't say no success. There have been successes in computing special, less physical cases, for example in treating the stars as frictionless dustballs not possessing magnetic fields. But these features are very important in determining the rotation structure of stellar fluids, and are probably essential in modeling the physically correct binary merger. The general problem remains to be solved, and the goal is to figure out what physical processes produce gravitational waves, so that we know what to be looking for experimentally.

Here are some visualizations of previous merger simulations:
1
2

Andreas(R):

While you haven't provided enough details to comment in length, I do have some experience with what you're planning. A couple of years back I started a programming system (XPS) which was rather audacious in scope. After two years of working on it, I realized that I too needed a "back end" compilation system. I looked at various alternatives like GCC (too complex), research compilers (low quality), open source virtual machines like Mono (immature at the time). I was quite surprised when I looked at UIUC's LLVM compiler toolkit. I thought it would be just another half-baked compiler system from a University that never got finished when the Ph.D student left. Instead, I found a well designed, working, *toolkit* for compiler construction. While LLVM still lacks some features, its core is very solid and easily extensible. I've been working with it for a year now and its been quite a pleasure. Check it out at http://llvm.cs.uiuc.edu/

You might want to look at LLVM. One of the ideas behind the project is to utilize information from the entire lifetime of the program to perform optimizations.

2. Doesn't pay for itself? No new reactors? Then why is France building more plants? And why is over 78% of French electricity demand produced by nuclear power?

Nuclear power works. Get over it.

I spent many nights in junior high "hacking" in the PLATO labs at the University of Illinois (UIUC). One of the grad students there at the time, the unspoken Hacker King, was one Rob Kolstad. We wrote (ok, so the other guys wrote and I pretended to write) software for student instruction, and were rewarded with computer time.

Anyway, back on topic: we used that time mostly to play a game called "moria" ("MOR-ee-uh" or "mor-EYE-uh"). It was a multiplayer, 3D action game drawn in bitmap graphics and text. Wireframe walls and corridors. You formed teams, managed your resources, fought battles to gain experience, and the rest.

Ah, nostalgia.

LLVM does exctly that. Though it calls itsel a "compiler infrastructure" or a "ompilation framework" it defines a low-level language, which can be used as intermidiate language in compilers, yet has an external (on-disk) bytecode representation that can be interpreted, JIT-ed, or compiled on a target system. The language is an SSA based representation, so the final compilation can perform various target specific optimizations.

LLVM does exctly that. Though it calls itsel a "compiler infrastructure" or a "ompilation framework" it defines a low-level language, which can be used as intermidiate language in compilers, yet has an external (on-disk) bytecode representation that can be interpreted, JIT-ed, or compiled on a target system. The language is an SSA based representation, so the final compilation can perform various target specific optimizations.

Here is an example of what they can look like.

If you totally can't stand NFS, you *might* want to give OpenAFS a try. It's probably total overkill, but is is really neat, and once you get an AFS cell up and running it is very good. It just takes a lot of work when you first start out of the gate. I did it for these guys, and we wrote a bunch of documentation, and about a year later, things are really good. AFS rocks.

If you totally can't stand NFS, you *might* want to give OpenAFS a try. It's probably total overkill, but is is really neat, and once you get an AFS cell up and running it is very good. It just takes a lot of work when you first start out of the gate. I did it for these guys, and we wrote a bunch of documentation, and about a year later, things are really good. AFS rocks.

If you totally can't stand NFS, you *might* want to give OpenAFS a try. It's probably total overkill, but is is really neat, and once you get an AFS cell up and running it is very good. It just takes a lot of work when you first start out of the gate. I did it for these guys, and we wrote a bunch of documentation, and about a year later, things are really good. AFS rocks.

If you totally can't stand NFS, you *might* want to give OpenAFS a try. It's probably total overkill, but is is really neat, and once you get an AFS cell up and running it is very good. It just takes a lot of work when you first start out of the gate. I did it for these guys, and we wrote a bunch of documentation, and about a year later, things are really good. AFS rocks.

Sure, the motor to spin the disc for movies uses some juice, but not nearly as much as processing 3D games and sound.

Fair enough. I was going to disagree, but when looking for sources (you helpfully didn't provide any), I found a nice paper (Google HTML version) on laptop power usage. The Pentium M they tested used 1 to 3W in idle, and 3 to 14W underload. The optical drive used 2.8W when just spinning and 5.31W when reading. When running 3DMark, the system used about 30W, during playback of an audio CD it used about 19W, the largest single component draw coming from the optical drive. Sadly they did not test the power consumption during DVD playback.
Of course, those numbers will be different on a PSP. It doesn't use a CPU nearly as powerful as a Pentium M 1300, but at the same time the optical drive will be smaller making for lighter media. I thought the optical drive would make more of a dent, but now I guess movie streaming and rendering and games will use about the same amount of power, assuming the game doesn't continuously access the CD, as well. I don't think you're correct in assuming that gaming uses considerable more power, though. Keep in mind that in addition to constantly spinning the optical medium, the PSP also has to decide the video which is a non-trivial task for a handheld mobile CPU.

Ok, googling ...

It was a concert on 02/23/82 at the University Of Illinois Auditorium - Champaign, IL. It only seated 1,936 people at the time. It was packed, as I recall. The balcony above our heads was unstable, and the whole place smelled of bong water.

I've been to lots of concerts, and I know they're loud. I liked my music loud even more then than I do now. Could you have missed the point more?

I would counter a few of your underlying assumptions with the following references - note that electron tunneling pathways affect protein folding dynamics and that quantum interference plays a critical role in photosynthesis. See also Zeilinger's biomolecule matter-wave interference experiments.

Of course, future computing architectures can incorporate these 'spooky' features.

I would counter a few of your underlying assumptions with the following references - note that electron tunneling pathways affect protein folding dynamics and that quantum interference plays a critical role in photosynthesis. See also Zeilinger's biomolecule matter-wave interference experiments.

Of course, future computing architectures can incorporate these 'spooky' features.

I usually follow that with an installation of Enough is Enough, SpywareBlaster, the combination of which pretty much neuters IE (but provides an easy way to add the sites that only work with IE to Trusted Sites from a menu), and Firefox, making it the default browser.

Why not download it, run it for a day and let us know what you think. :)

Let us not forget CERN's early work with the www client and wwwd server. In particular, the work of Tim Berners-Lee. That link includes some web history.

Let us not also forget NCSA Mosaic, which became a "killer app" in the early/mid 1990s, before being spun off as SpyGlass.

My memory is faulty, but I believe more than half of the NCSA team left the project and formed NetScape. Can anyone correct this?

The web as we know it also owes a debt to previous research in hypertext systems dating back decades, as well as existing document-markup systems.

To those who keep Mozilla alive today:
I salute you, but do take too much pride in yourselves:
Never forget that you stand on the shoulders of giants.

...but it's still worth remembering that Netscape changed the world not once (by making the first really good browser)...

What was wrong with Mosaic?

Here's the video used by te Ig Nobel researchers:

http://viscog.beckman.uiuc.edu/grafs/demos/15.html