Hawking Is First User of "Big Brain" Supercomputer 93
miller60 writes "Calling your product the 'Big Brain Computer' is a heady claim. It helps if you have Dr. Stephen Hawking say that the product can help unlock the secrets of the universe. SGI says its UV2 can scale to 4,096 cores and 64 terabytes of memory, with a peak I/O rate of four terabytes per second and runs off-the-shelf Linux software. Hawking says the UV2 'will ensure that UK researchers remain at the forefront of fundamental and observational cosmology.'"
hawking's been hacked. (Score:5, Funny)
i'm convinced that someone else is controlling what his computer-chair-interface says. perhaps it's even...bum bum bum....a super advanced AI, tricking us all into giving it access to a supercom...oh no! it's too late!
Re:hawking's been hacked. (Score:5, Informative)
From the article it sounds like it. I find it hard to believe that Hawking wrote the line "soon to be supercharged with Intel’s MIC technology" they have him quoted as saying. Sounds like a PR flack programmed his speak-and-spell.
Re:hawking's been hacked. (Score:5, Funny)
I find it hard to believe that Hawking wrote the line
Me too. Maybe dictated.
Re: (Score:1)
Well then you should see Hawking in Intel's promotional video's --- he is making a lot of money promoting Intel lately. At the end of one video he state that Intel has always powered his wheelchair -- the video is quite gimmicky but it's Hawking so... I have no problem believing he has sold his soul to Intel, They are probably paying for his flight into space.
Re: (Score:2)
Wow...I thought they went out of business a LONG time ago....
Re: (Score:2)
Well, Hawking's first voice computer was a PC interfaced to a DECtalk running something under DOS. It's been updated many times since then (other than
Re: (Score:1)
Re:hawking's been hacked. (Score:5, Interesting)
Regardless of the impression given, there is a need for celebrity scientists. Second rate scientist with a first rate promotional skill is just what the TV needs. Plus the promotion time couldn't be better spent in the lab either.
These people inspire the young to become scientists and raise public awareness for what ever cause made the news that week.
How on earth did you bring Germaine Greer into this? That's wonderful..
Re: (Score:2)
My point is, good TV needs a person that can read really dull text out and give it a bit of stuff. It doesn't matter if they are not in the top 10% of the scientific world.
I'm not sure what level the peop
Re: (Score:2)
Re: (Score:1)
Everyone who has watched Weekend at Bernie's is already well aware of what's going on here...
Too bad SGI was gutted (Score:2, Interesting)
It's really too bad that the company currently known as SGI has only the name in common with the SGI of yore. Truly some pioneering work done there, although they did fail to keep up with the "G" portion of their name in the late 90s. Imagine what the world would look like had they bought out nVidia way back when? Probably, we'd all be running SGI video cards, and Monoprice would sell Craylink cables. Microsoft would be a struggling software company, Linux would still be a pipe dream, and SVR4 (with some BS
Re: (Score:1)
amen. Though (and maybe you know this) - up until the late 90s, nvidia was comprised of little more than SGI engineers at varying stages of disgruntlement.
If you've ever wondered how a hack like huang could be so successful now you know - he just stole technology the legal way - by hiring away from a company whose management were too incompetent to care.
Wrong. Both LInux and MS would be doing fine (Score:5, Interesting)
because
A) Silicon graphics had little influence one way or the other on the progress of Linux (or Windows so the same applies) even when they were a big player.
and
B) Your average home user would not be willing to pay the multi thousand dollar price tag of an SGI system just to have a version of Unix wirth decent graphics at home.
Unfortunately both SGI and to a lesser extent Sun missed the signs that x86 PCs were going to rapidly catch up woth the abilities of their workstations and instead of dropping prices to sane levels continued to carry on business as usual as if it was still 1990. And the end result is what you see.
Re:Wrong. Both LInux and MS would be doing fine (Score:5, Informative)
Unfortunately both SGI and to a lesser extent Sun missed the signs that x86 PCs were going to rapidly catch up woth the abilities of their workstations and instead of dropping prices to sane levels continued to carry on business as usual as if it was still 1990. And the end result is what you see.
As a general statement, if they had kept up their research with their own processors, instead of trying to catch a ride on the Itanic [wikipedia.org] they would have kept ahead on capability for quite a while longer. It wasn't really that they missed the signs, it was that they stopped their own progress to all get on the same bus, not noticing that it was in the slow lane, with its blinkers on and belching smoke. By the time it got up to speed, x86 had already whizzed by.
Re: (Score:2)
SGI hired a Microsoft exec much as Nokia did.
Re: (Score:2)
As a general statement, if they had kept up their research with their own processors, instead of trying to catch a ride on the Itanic they would have kept ahead on capability for quite a while longer.
Itanic might have been a mistake, but there's no way SPARC/MIPS/Alpha could have stayed way ahead of Intel given the huge volumes Intel ended up getting in the late 90s. A large fraction of the cost of CPUs is the cost of the R&D and the fabs. Those got more and more expensive from one generation to the next and the only way Intel was able to keep up was because of the boom in the PC market at that time.
Re: (Score:2)
The PC market only boomed because the unix workstation manufacturers saw the consumer and small business market as beneath them. More fool them.
Re: (Score:2)
Silicon graphics had little influence one way or the other on the progress of Linux
Wrong on three counts. 1) OpenGL has a huge influence on Linux and 2) A number of Irix design elements were incorporated into Linux and 3) SGI engineers were (and still are) huge contributors to Linux, providing much of the memory scaling infrastructure for one thing, and the respectable if now somewhat dated XFS filesytem.
Re: (Score:2)
Wrong.
1) OpenGL had an influence on parts of X windows which framebuffer aisde has nothing to do with the linux kernel.
2 & 3) XFS had been pretty much an irrelevance in the linux ecosystem , hardly anyone used it. As for the memory scaling - I think you're overestimating their contribution to the total.
Re: (Score:1)
yes.
SGI couldn't buy out Nvidia - they had laid off the graphics designers who went to Nvidia...
This was during the time of their MS VP that did the same thing to SGI that has just been done to Nokia.
And he sold much of SGI graphics patents to MS.
Re: (Score:2)
It's really too bad that the company currently known as SGI has only the name in common with the SGI of yore.
Not actually true. The company formerly known as Rackable Systems also has many of the former SGI engineers in common, and the Altix technololgy. What's gone is the original corporate, the flashy headquarters (now inhabited, hermit crab-like, by Google) and the graphics business, which is now owned by we, the people.
Re: (Score:1)
Re: (Score:2)
I've wondered... (Score:3, Interesting)
I assume that most great cosmologists aren't expert computer programmers with specialties in high performance computation, and that most great programmers specializing in high performance computation aren't great cosmologists.
So how do these people get their ridiculously complicated physics stuff crunched by ridiculously complicated machines?
Re: (Score:2)
I assume that most great cosmologists aren't expert computer programmers with specialties in high performance computation, and that most great programmers specializing in high performance computation aren't great cosmologists.
So how do these people get their ridiculously complicated physics stuff crunched by ridiculously complicated machines?
This: the humble computer scientist (no, programmers are not necessarily, and only very infrequently, computer scientists) will act as a liason between the two.
Re: (Score:1)
That is pretty much never the case. Physicists have long been proud makers of their own tools and software is no different.
Re: (Score:2)
That is pretty much never the case. Physicists have long been proud makers of their own tools and software is no different.
You are begging the question. Computer Scientists aren't necessarily making tools, but are the architects of computer models. A computer scientist may not ever write a single line of discrete code, but at a higher level, direct programmers to code modules that fit in the grand design. Also, it matters not what the discipline the scientist comes from, if they are doing computer science, they are computer scientists.
Re:I've wondered... (Score:5, Interesting)
So how do these people get their ridiculously complicated physics stuff crunched by ridiculously complicated machines?
Because they know the equations for their ridiculously complicated physics stuff, and most physicists are expected to be literate with computer programming. I have two PhD friends, both in Physics (meteorology and cosmology) who are now both hardcore coders due to their training.
Re: (Score:2)
PhD 1: Hmmm... I have this ridiculously complicated physics stuff that I need crunched by ridiculously complicated machines
PhD 2: Give it to the grad students and tell them they won't be fed until the ridiculously complicated machines are satisfied.
Grad Student: I can survive 2 weeks without food, but only 2 days without water. If I save my pee and distill it in the lab, I can stretch out my survival to 4 days. 4 days is just enough. Those PhDs won't kill me this time!
Re: (Score:1)
So how do these people get their ridiculously complicated physics stuff crunched by ridiculously complicated machines?
a) Their algorithms aren't complex but they typically have a lot of data.
b) They're not real hardcore programmers so they program in BASIC, Python, etc. They need bigger machines to compensate for the inefficiency of their toolset.
Re: (Score:2)
No, the algorithms too can be quite complex, depending on the field. And nobody with a sane mind uses BASIC. Usually something will be tested using a small test-problem in Matlab or Mathematica (or even Excel for very simple stuff) that have built-in graphing capabilities and mathematical libraries to get a feeling of how everything works. If everything looks good but runs slow, the model will be re-written in Fortran or C that will do the number crunching. Usually the scientist will write a serial version
off the shelf (Score:2)
Re: (Score:2)
Who keeps Linux on a shelf?
I do. The easiest way to install it is from a CD or a DVD. I keep a collection of older distros in case somebody has an old Windows machine with the OS so corrupt that there's no fixing it without a reinstall and they don't have their Windows disks.
Re: (Score:2)
No.
Re:Off the shelf Linux? (Score:4, Informative)
Is the standard linux kernel optimised for 4096 cores...?
Imagine a Beowulf Cluster ...
Re:Off the shelf Linux? (Score:4, Funny)
amidoinitrite?
Re: (Score:1)
last I checked (a long time ago), yes. To clarify: it doesn't need any specific optimizations for whatever number of cores.
Re: (Score:3)
Yes!
http://xkcd.com/619/ [xkcd.com]
Is he a real genius (Score:1)
I have always doubted Hawking's "genius". The guy repeats what other lesser known (even the well known scientists) scientists have already said or claimed.
Is this a matter of just feeling sorry for the guy because of his handicap? I am only going by what I have heard the guy say, I do not know him personally nor have I ever worked with him to determine if he is a real deal. Nor do I really believe what other scientists say about the guy.
I do not dislike the guy, but I just feel he repeats what others have p
Re: (Score:3, Insightful)
Maybe you should do some research into this and present your findings.
Re: (Score:2)
Excess CPU cycles? (Score:3)
Re: (Score:1)
Re: (Score:2)
It means that it is able to consolidate my complete pr0n collection in a single day!
Re: (Score:2)
It means that it is able to consolidate my complete pr0n collection in a single day!
Ah ha! So that's how they lured Stephen Hawking into endorsing this machine.
In case anyone's been living under a rock, here's [huffingtonpost.com] an interesting tidbit.
DR Stephen Hawking.. (Score:1)
Okay, while I won't dispute that he has a PhD I think it a little insulting not to recognise that he's a PROFESSOR!!
He should ask it (Score:1)
Unlock secrets of universe (Score:2)
Every computer can help to unlock the secrets of the universe.
Way to go, England! (Score:2)
Aww, how cute! England wants to come out and play. While I am happy to see that our mates from across the pond are getting themselves a nice little number cruncher - it is still little. NASA has a setup called Pleiades [nasa.gov]:
Total cores: 112,896
Total memory: 191 TB
But here's the real hard-to-fathom point. The sport two 11-dimension hypercube interconnect configurations using Infiniband QDR and DDR networking (mostly DDR). Now DDR is ~4GB/s and QDR is ~8GB/s, but Inifiniband is rarely singily-connected. Usually yo
But... (Score:1)
Re: (Score:2)
how many frames per second does it get in Crysis?
Only 60 FPS, but the screen resolution is 10,000,000 x 10,000,000 projected on the inside of a sphere
SGI? (Score:2)
As in Silicon Graphics Incorporated?
I thought they went out of business.
Ah. They are Rackable [wikipedia.org]
Exceptions Per Second (Score:2)
Just imagine how many Windows BSOD's this "Big Brain" could possibly generate. Thank Satan it's uses a sensible (TM) Operating System.
Hawkings@BigBrain:~$ cat ~/.bash_history (Score:1)