Slashdot Mirror
Big Mac achieves around 14 TFlops with 128 Nodes
mzs writes "The Virginia Tech G5 cluster has achieved around 80% of its peak performance in preliminary Linpack testing with 128 nodes according to Jack Dongarra at the Top 500. "They're getting about 80 percent of the theoretical peak," Dongarra said. "If it holds, and it's unclear if it will, it has the potential to be the world's second most powerful machine." Typically getting 60% of peak in the Top 500 lists is quite good. If the Big Mac cluster achieves 60% of peak it would displace the 2,300 2.4 GHz Xeon cluster at LLNL for the number three spot on the current list."
Anyone remember Happy Meal Ethernet and Big Mac Ethernet?
Imagine a beo .... son of a bitch. They beat me to it.
- Sometimes you're the pidgeon, sometimes you're the statue.
Title is wrong - they get 80% efficiency on 128 nodes. The 14 TFlops number is if that efficiency is held through the full size of the machine (2000+ processors).
My apologies; however, I have to ask:
But will it run Windoze?
Thanks in advance,
W00t
So, looking at this, I am wondering if the federal constraints on computer exports are still in place? This Apple supercluster shows that just about anyone now could afford to build a supercomputer giving smaller countries access to compute cycles never before dreamed of for relatively few $$'s
Visit Jonesblog and say hello.
When will Slashdot add or change the G4 icon to G5?
A programmer is a machine for converting coffee into code.
Imagine a McDonalds cluster of these.
[ducks and runs]
There are weenies that will say "Psstt.. you know that #2 computer in the Big 500? It only has one button on the mouse!"
Trolling is a art,
A few non-Mac related news stories once in a while would be nice. Sheesh. Back to our regularly scheduled Macsturbatory hourly updates on VA Tech.
If the Big Mac cluster achieves 60% of peak
They ALREADY achieved 80% of peak, is the submitter listening?
What is it using for disk space? MCR @ LLNL is using the Lustre file system with DataDirect Networks storage.
How many libraries of congress is this per second?
that the 80% holds as more and more nodes are added. Wouldn't each additional node have to be 100% efficient for this to happen... (ok so let's whip out the math induction and prove this is not possible)
----
In Soviet Russia, the overlords welcome you!
It would be cheaper to just give the scientists each a Trans Am so then at least the general masses would be aware of the scientists' penis envy.
This is tremendous advertising for apple, but what about clusters of Power4's and 5's? why wouldnt they out-perform this cluster? at 14tflops... would 2200 macs be fairly equal to the earth simulator?
And lastly with IBM seeing their G5's at 3Ghz in 8 months or so.... do you think we will also see a 50% increase in power? i.e. 21tflops for 1100 machines? i know this doesnt translate directly.... but holy mackeral.
--Idiots, Every single one of YOU, A flaming mass of conglomerated morons, hey wait a second, isnt that how RAID works?
Big Mac also achieved around 14 KTons with 128 kids.
An Indian-American Hindu committed to non-violent thought/speech/action alarmed by the global explosion of radical Islam
yeah blah blah blah #3 on the list but
does it run the Doom 3?
...If the Big Mac cluster achieves 60% of peak ...
Big Mac cluster? What's that do, display in real time the flow of special sauce molecules over the surface of a flame broiled patty?
I lost my concept of community when my community lost all concept of me.
You want a /., I'll help you. I'm reloading as fast as possible on your site.
Just to elaborate, the 14 TFlops is if Big Mac achieves 80% of peak with the full 1100 nodes. 14 TFlops with just 256 PPC 970 cpus is completely implausible. I cannot believe I wrote that, sorry about it my mistake.
that litany doesn't make any sense
itunes for windows is out and /. wont accept my story! ugh! it a much bigger story
here is the link
a href=http://www.macrumors.com/events/musicevent200 3.html> story
geez im just gonna hve to wait until they accpet someone else's story
The war with islam is a war on the beast
The war on terror is a war for peace
IBM has 49 of the top 100 from the last list. Who cares whose on top if you make almost half of the 100 fastest computers in the world?
Yes, this is a troll post. I am pissed the submission after the Napster 2 one wasn't the itunes4Windows announcement.
Going on means going far
Going far means returning
then learn how to post a link
Southend.org
I do not fear computers. I fear the lack of them. Isaac Asimov (1920 - 1992)
2003-10-16 18:22:14 iTunes for windows launched (articles,apple) (rejected)
AAAGH!
The war with islam is a war on the beast
The war on terror is a war for peace
"The choice of InfiniBand as the interconnect for this cluster filled a critical missing piece in creating a world class cluster, based solely upon industry standard computing components," says Hassan Aref, dean of Virginia Tech's College of Engineering. "Unlike its predecessors, InfiniBand is an industry standard and it provides the fastest high performance interconnect available today."
www.supercomputingonline.com/article.php?sid=4316
Time to replace the G4 icon with a G5 pic dont you think??
Like this for example > http://www.apple.com/g5processor/
Those who laugh at you for you having a Mac.. are the people who constantly call you to fix their PC.
Break into the labs and use this baby to run some RC5 cracks! My team will be unstoppable!
Karma Whoring for Fun and Profit.
Great feat - IBM!
;)
Why is everyone so obsesses with Apple when it comes to G5? What does Apple have to do with the G5?! This is the kiddie version of IBM's success CPU named Power4, if I'm not totally incorrect.
Apple, to me, is a group of cosmetologist hangarounds.
Would you like fries with that?
-- Dr. Eldarion --
The submitter stated 128 nodes. This is wrong. The article states 128 processors which would be 64 nodes.
-You may license this sig for only $6.99.
Okay, I RTFA'ed, and it looks good. Hope they can keep the 80%. I wonder if I can pester U of Wisc at Madison to fund a $10 million Apple supercomputer.
By the way, I just downloaded iTunes for Windows from Apple's site. Apple is really basking in the sunlight here....
Big Mac achieves around 14 TFlops...
All this time the answer to my supercomputing needs was at McDonalds. What a burger!
Or do they have a note from their CS professor that their dog ate their last cluster...
What exactly do you mean by "Don't touch this button?"
It should be noted that the final result doesn't need to be anywhere near 80% efficiency to assure it second place on the list. The third place machine is only around 7 TFlops, meaning Big Mac could operate at as low as 40 to 45 percent efficiency and still take second.
-You may license this sig for only $6.99.
I wonder if they are using GCC3.3 or IBM XL C. Apparently the performance from code generated by the IBM compiler far exceeds that generated by the latest PPC GCC.
Click!!
Also moderate down the other reply post by tekiegreg.
Want more info on crappy posts, read my journal entry (click the link in the current sig).
-- Fighting mediocrity one bad post at a time.
Of course, it might get out of hand....everyone knoes it takes two hands to handle the whopper!
Imagine Running iTunes on this thing... ::hint hint::
Too Many Hacks
The open-source community should abandon its piecemeal approach to securing Linux-- and soon.
By Hal Flynn Oct 15 2003 03:23PM PT
When I was a kid growing up in Arizona, I used to spend most of my time playing with the other kids that lived up the street. We would do the things that young American boys do: beat each other up, call each other names, and perform really dangerous tricks with our bikes. This was life before the discovery of girls.
One summer, one of the kids found the site of a freshly demolished building, which had been recently bulldozed. There were several heaps of rubble and huge mounds of dirt; perfect for jumping. When the friend came back and told us all about the site, we quickly jumped on our bikes, ready to go; all except one kid, who had a strange bulge in his bike tire. Knowing I could probably fix the problem, or at least ask my Dad to help me fix it, I jumped on his bike and quickly rode it home.
I dug around in Dad's toolbox and went to work on taking the bike tire apart with a screwdriver. After hearing the noise, my Dad came outside to see what I was doing. Upon discovering that I was trying to fix the tire, Dad said to me, "Make sure you fix it right the first time son. Otherwise, you'll look like a donkey."
Those words come back to me when I look at the direction Linux security is taking.
Recently, a post to the Bugtraq mailing list by security researcher Zaraza reminded the community of a problem in inetd, the Internet Super-Server. Inetd times out after receiving a large amount of connections within one minute, and refuses all connections for ten minutes afterwards.
The Linux community should participate in the Trusted Computing Group... or start its own.
This isn't a new problem. Daniel J. Bernstein (DJB), the University of Illinois at Chicago professor known for his venomous tirades and clever coding solutions, stepped up to solve this issue years ago by creating a special software package called ucspi-tcp, consisting of tcpserver and tcpclient. True to DJB form, he distributes the package through a Web page that details exactly how crappy inetd is, and why ucspi-tcp is better.
True to the UNIX way, another program that performs a specific task very well was designed and released, rather than depending on the application itself or lower-level system internals. DJB just introduced another program with an entirely new set of limitations, when programmers should move to designing applications that are network-aware. In essence, a hack was resolved with another hack.
The modular kernel in UNIX systems is another example of a hacked solution. When implemented, it had the best of intentions. It was designed to conserve kernel memory by using modular pieces of code that can be inserted as needed, and removed, while the system continues to run. In theory, this is a great idea.
However, modular code has one big drawback. It is possible for the administrative user to load modules on the fly in kernels configured to support it. This sounds a lot like a feature, until one considers a person without legitimate administrative access: If a person of questionable integrity can load a module, the system can be made to lie at its lowest level.
Trustworthy Linux
A few different groups of people recognized this problem. For example, those in the Solaris community decided that the best way to handle this is to use Role-Based Access Control to strip the administrative user of the ability to load kernel modules. This functionality could also be extended to other parts of the system as well.
In the Linux community, the general consensus was that this issue is best dealt with by creating a static kernel without loadable modules. Remove the functionality, and you remove the potential abuse.
But this was just another new hurdle on the obstacle course, and it was only a matter of time before someone cleared it: in this case a hacker named Silvio Cesare, who proved with an alarming degree
Most indubitoubly or however the hell it's speeled.
2. Funny.
"Chain" is the word you are looking for.
Anyone know the theoretical peak of the 2300-node xeon cluster at the Lawrence Livermore lab? I've read elsewhere that it was (to be) a 1920-node cluster with a theoretical peak of 9.2 teraflops.
This article reports that the 2300-noder operates at 7.6 teraflops, but i was wondering what percentage of its theoretical peak that is.
Look out honey, 'cause I'm using technology; Ain't got time to make no apology
ha ha ha!
Interestingly, a fun number to compare against is SETI@Home's array power, which is approximately 15 teraflops. [See the SETI@Home FAQ]
Although they don't run Linpack, and therefore can't be considered on Top500 the same way, it's still cool to know that SETI would still place second on the supercomputing list. Back in 2001, they were averaging a very large number of teraflops as well, (>10TF) the figure is on the internet somewhere. In 2001, that was greater than the top three supercomputing sites combined.
It would be interesting to see the power of the Seti array using today's processors.. which are arugably far faster than 2001's, despite the short amount of time.
Still, SETI outperforms this mac cluster, although it's obvious that SETI's distribution model is clearly not usable for the same problems that need to be solved.
Tired of legitimate data sources? Try UNCYCLOPEDIA
The story about the Big Mac is great, but the link to the "desktop supercomputer" chip is what really captured my attention
, 00.html
http://www.wired.com/news/technology/0,1282,60791
I'm on a chair.
This puts NORAD's WOPR to shame...
From article: Dongarra said the cost is so low he questioned whether the college got a special discount.
At $5.2 mil for 1100 machines, I think they paid full market price; that's over $4,500 per machine, and currently Apple is selling dual 2 Ghz G5's for ~$3000. And that's with lots of extras that they wouldn't want in a cluster (ATI 9600, CDRW, etc), which hopefully they convinced Apple they didn't need... (else they've got a whole lot of Mac keyboards sitting around!)
I wonder how much of the cost was the actual machines, and how much was infrastructure and networking stuff (I can just see 1,100 Macs all powered off one extension cord and a bunch of surge protectors).
And what OS are they using? I haven't read about that anywhere but I would venture to guess it's some commodity os (linux) they don't need to mention. And yes at 80% efficiency they have to use altivec optimalization .. if it's gcc hopefully we'll see better altivec support sooner rather than later.
Of those to whom much is given, much is required.
Especially since the article is specifically about G5's and not G4's. What gives?
It's a Dune refernce pal. Paraphrasing the "Litany Against Fear"
"I must not fear. Fear is the mind killer"
Moderators, here's a handy list of crappy posts to get you started!
post my story .. itunes for windows is out!
www.itunes.com
iver been submitting it for thelast hr and it gotten rejected! ugh!
The war with islam is a war on the beast
The war on terror is a war for peace
What will the effect of the 'magical' software they have to detect memory errors be?
Most expensive super computer EVAR.
Repeat after me: IT RUNS MACOSX
PDF presentation link.
the article suggests that maybe va tech got a big discount from apple because of the huge cost difference from other sooperputers so I went to the apple online store.
I pulled the modem out of the default dual 2ghz g5 mac and upped the ram to 4gigs, as I vaguely remember the big mac node config being from some other article somewhere. Cost 5,320. Tried to up the quatity to 1100 in my cart but the web form would only allow 3 digits, so I did 110. Got 563,200. That looks like a volume discout because my calculator spits out 585200 for simple multiplication. So if we go to 1100 the apple online store may automatically give a steaper discount still.
And remember: these are joe blow prices. they probably paid educational prices. AND my figures use apple's expensive ram, not third party prices.
So it looks like they did get a discount, but not anything special. Looks like the kind of deal any school would get if they pieced one together online today.
The only special treatment they probably got was the shipping priority.
"Let him go, Ralph. He knows what he's doing." --Otto Mann (simpsons)
Opteron cluster of 2800 CPUS/1400 nodes.
I'm picking nits here but...
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
not going to say it...
ARGH!
Imagine a beowulf cluster of these!
MacDonald's sells its 86th billion Big Mac (thats 86GBflops - 86 GigaBurgers flopped on a grill) Did you have a Big Mac today?
I did the store thing twice to get my numbers. Second time I for got to replace the superdrive with the combo. That accounts for the automatic discount. Unit price is 5,120 on that config * 110 is 563200 which is spit out in the cart.
Sorry folks. I'll sheepishly retreat into the corner now
"Let him go, Ralph. He knows what he's doing." --Otto Mann (simpsons)
Well that changes my perception. at first I just considered it a big blue supercomputer but ... the 2nd fastest supercomputer on the planet runs OSX!
Of those to whom much is given, much is required.
In fact, he kinda demoed it first on a Mac.
Hey - You're right! It looked a little odd - so I zoomed it up a bunch and sure enough: 12 stripes! I wonder which colony they dropped?
(Reminds me of the time I was driving home from work and some guy had painted a US flag on the side windows of his Chevy Blazer. As a former teacher, it saddened me that this "patriot" figured we only had 36 states - six rows of six stars. Also, in his world, there were 15 colonies/stripes.)
WOPR Supercomputer
it's a giant heatsink... but it the best looking heatsink I've ever seen.
"Things are more moderner than before- bigger, and yet smaller- it's computers-- San Dimas High School football RULES!"
Okay folks with mod points, this is the third post to this thread with minus_273 pleading to have his iTunes story accepted. here is the first and here is the second. Is this what Slashdot has come to, people begging to have their stories (about something that is hardly news to anyone who has been keeping up) accepted just so they can get some karma and add a number to their "accepted" stories?
Bah... probably still tied to an ATI Rage video card, though...
Here's how I figure that:
A G5 proc can do 4 64-bit FP ops per clock cycle (2 FPUs with each capable of doing a multiply/add op), so that's 8 GFlops/s per 2GHz proc, or 16 GFlop/s per dual-proc node. For 128 nodes, that's a little over 2.048 TFlop/s peak. Dongerra said they were getting 80% of peak, which would be 1.638 TFlop/s.
The thing that I *still* haven't heard explained in all the reporting on the VA Tech G5 cluster is how they're scaling the interconnect out to 1100 nodes. The biggest InfiniBand switch you can get right now is 128 ports, so somehow they're going to have to slave a whole bunch of IB switches together. There's a lot of ways they could do that, but the cheapest way (with minimal bandwidth between switches) could really hurt them on big parallel codes like the Parallel Linpack Benchmark. It's even worse for codes like parallel FFTs that are bound by bisection bandwidth (where all the nodes are talking to all the other nodes at full bandwidth simultaneously). It's not clear to me if VA Tech is planning on putting in an IB network that preserves bisection bandwidth or not.
"My life's work has been to prompt others... and be forgotten." --Cyrano de Bergerac
Ok you frothing Macaholics. *Someone* explain to me *how* such a largely clustered machine was put together with 80% efficiency *WITHOUT* ECC. This issue has been posted several times on this story, without any satisfactory answer.
If there are too many nodes the error rate goes up super-linearly (since interconnect errors have to be factored in.) There's no getting around this. If they are just running *without* ECC and hoping the results just come out correctly, then this is worthless -- its amounts to "fluke calculations".
Were did the number for the 'theoretical peak'
com from?
It looks suspiciously similar to 1,100 * ~14
gigaflops (which is the measured performance
of 1 G5 Powermac).
There's no science in that!
You'll see McDonalds suing Apple over the name 'Big Mac'. Argument: Consumer are too stupid to distinguish between a computer and shit put between two buns.
We need to beef up security around the Whopper.....
how long does it take to copy a 17 M file from one folder to another?
= 9J =
Had no idea hamburgers could run that fast.
Must be the "special" bun in the middle.
Okay folks with mod points, this is the third post to this thread with minus_273 pleading to have his iTunes story accepted. here is the first and here is the second. Is this what Slashdot has come to, people begging to have their stories (about something that is hardly news to anyone who has been keeping up) accepted just so they can get some karma and add a number to their "accepted" stories?
= 9J =
I want to know about how many volkswagen bugs it is, can this thing fit in my house, or do I need to rent a warehouse?
Also, I don't trust it until it comes with a 'Performance Rating' from Cyrix, How many Pentiums is this beast?
"Sometimes, I think Trent just needs a cup of hot chocolate and a blankie." -Tori Amos on Nine Inch Nails
...as the cheaper Dell supercomputer...
Exactly, Dells are cheap. Would you really want to cut corners on a supercomputer? A friend of mine wrote "Dell" on his Logitech mouse, and it broke the next day.
My double whopper is going to beat that!
Cat got your tongue? (something important seems to be missing from your comment ... like the body or the subject!)
LOL!!@! MACS CAN'T PLAY GAMES LOL!!@!!!!!!
You are very clever. Now STFU and actually read something about Macs that isn't 8 years old.
Department of Homeland Security: Removing the rights real patriots fought and died for since 2001
So let me get this straight, 2200 2ghz G5s barely outperforms 2300 2.4ghz P4 xeons...
Probably sould have used AMD Opterons, at least they would have saved some money....
= 9J =
Wow!
Imagine emulating a Beowulf cluster of Pentiums on this!!!!!!!!!
Yeah, it cost twice as much, and only outperforms it by a factor of 4.
who got the better deal?
My sig is blank, I typed this by hand.
Do these boneheads proof their articles for accuracy? To start with, 128 processors -ne 128 nodes.
Does anyone know a site that lists the floating point values of common processors like the Pentium 4 and Athlon series? I would like to calculate the power of therotical clusters but I can't find the Gigaflop/Teraflop information anywhere. Anyone?
We don't know yet what its performance will be.... ... not 4x ok...
the xeons = 7.6Gflops
At peak, the g5s = 17.6
According to the article, optimistic (80%) = 14.1 Gflops
Realistic (60%) = 10.6 Gflops.... thats about 40% more for double the price... i guess thats acceptable seeing how pretty the g5s are...
Uh, huh. So, if those statements were true, the G5 would be 12 times faster than a Xeon. Quite a feat of engineering that would be. But I don't think so. The folks at VT are extrapolating their 64 node performance to a 2300 node cluster.
This tells us two things:
Basically, you can't believe anything anymore anybody says about the supposedly stellar performance of PPC: the hype has simply drowned out all reality. Personally, I found the G5 acceptable but nothing extraordinary. Run your own benchmarks and see how well it works for you.
you're mixin' apples (ahem) and oranges. the UT cluster is NOT the xeons @ 7.6 Gflops.
DANGER DANGER
I saw 'War Games', the 1984 high tech, block buster, starring Matthew Broderick. He hacks into the advanced government computer with his 1200 baud modem to play computer games. The computer was called WOPR (pronounced whopper). And unfortunately he played the game called 'THERMONUCLEAR WAR'. My GOD WHAT HAVE WE DONE!
Sure it's all probably by the same person, but it's funny as hell regardless.
Kasushige Goto
Being a bio guy, maybe I am niave, but isn't the fact that one get set up a top 50 super computer in about two weeks something of an acheivement? Yes, it is only 80%, but that still makes it on the list. Wow. _Iowa
"He who laughs last, didn't get the joke."-Cap
You can throw all the computer horsepower you want at a simulation, but you need some way to validate the results. Of course if you have a perfect understanding of sub-atomic physics and can develop a large scale software project with zero bugs then you should be good to go, but back in reality we need a way to validate.
You didn't get it, 1100 2ghz G5 trounce by a wide margin 2400 2.4 GHz P4 Xeons...
The Athlon can do 2 double-precision floating point ops per cycle. It has 3 fpus but can only issue 2 floating point ops per cycle.
Also in many cases with scientific code FMA instructions are pretty common. They aren't going to be running everyday programs on this thing. If you code with the architecture in mind many algorithms can easily be written to take advantage of that instruction.
A favorite story of mine concerning the difference between having a calculator and knowing how to use one: As the scientists in the Manhattan project were waiting in the trenches/bunkers for the first trinity test, one of them started ripping little pieces of paper and just before the bomb went off, tossed them in the air. After the bomb went off, he paced off the distance they had been blown, did a bunch of calculations and gave the first estimate of the yield of the bomb. Accurate to within 20-40% I recall.
There is no trap so deadly as the trap you set for yourself
-Raymond Chandler, The Long Goodbye
Enrico Fermi on Back of the Envelop Calculations
To put money into a conference booth, usually means that a company is making some serious consideration for that market.
So here's the trivia question:
When did Apple Computer start being an exhibitor at the International Conference for High Performance Computing and Communications?
Apple's been aware of the Top500 list for a several years.
Supercomputers used to be 32-bit with the Top500 list accepting 32-bit results. At the time Apple started attended the SC Conference, I asked Jack D. if he would accept the 32-bit benchmark results for the Top500.
At the time Jack said no -- adding back 32-bit would be confusing to the community.
It took some effort and many voices to convince Apple to pay attention to the scientific market. It's nice to see them headed towards the Top500.
--Hopscotch
HINT: 9991
This really is a win for Altivec, which is after all the spawn of AIM, but one would have to credit Apple for not giving up on it after Moto faltered, as well as the near flawless G5 design. IBM should be especially pleased, perhaps enough to incorporate Altivec into the Power6.
I like the WOPR, fuck the Big Mac!
Error 666 - SCO source has been found in your Linux kernel. Please remove it.
Formerly kdsolutions
AIST is getting a similar scale Linux cluster using Opteron processors, a bit larger (are all the G5 dual proc boxes? If so things are still in the same ball park). about 1100 dual opteron systems, and a few other Intel boxes tossed in. One good article is here.
I have submitted this as story, but it evidently never news like the G5 cluster... Linux and AMD no longer our favorite 'underdogs' anymore?
XML is like violence. If it doesn't solve the problem, use more.
Great, someone knows how to cram fifteen thousand computers into a specially-designed warehouse and make them go fast. Meanwhile, in another part of the world known as `reallity,' people realize there are more important things to worry about than who has the larger penis. More computers != research. There's a lot of people who could be more productive in their use of the electricity, man hours, hardware and system resources that these computers are using. I submit to you that if they had given all of these Apple computers to universities and research facilities all over the world, man's greater goals would have been much closer to being achieved than they would have been with all of these machines crammed into a warehouse.
Costs millions more than the Virginia Tech project and at its highest can only do 11.7 TFlops - 1/3 less performance, but then it's just our tax dollars at work.
iTunes has problems. If you read the forums, people are reporting that iTunes REARRANGES your existing MP3 files into it's own directory scheme. Is that rude or what? Granted there is an option to turn it off, but what a bad default setting if you don't catch it!
.99 a song is still too much money. But maybe I could live with that.
For me,
But it also won't download or let you create MP3 files from music you download from them. I've got a car player with a 40GB hard drive that only supports MP3 and WMA files. I'm sure there's a way to make MP3s out of the AAC files, but I have a particular dislike for artifical hurdles like that. Computers are supposed to make things easy. The majority of portable music players do MP3, why support Windows if you're not going to support the players that Windows people use? Yeah I know, "copy protection", "DRM", mumble mumble... Well I don't give a shit about copy protection, I want to listen to MY music WHEN and WHERE I want to listen to it.
But the proverbial straw that broke the camel's back is that iTunes only supports Win2k and WinXP. I run Win98SE (same installtion for 2.5 years on a heavily-used gaming/Internet machine.) I'm sure it's easier for them to only support 2k/XP, but there's no technical reason it couldn't run on Win98SE. I'm not changing OS's to run iTunes, that's too much to ask.
Experienced nuke engineers have been on the open market for the last 12 years, 59 counting spies that sold out the Manhattan project to the commies in the first place. There is a lot of knowledge out there if the price is right.
09f911029d74e35bd84156c5635688c0
can it run Doom 3?
It would be only fair to note why VirtualPC for the G5 isn't just a simple port or recompile: IBM removed the little-endian mode in PowerPC 970, which was what VirtualPC extensively uses. Implementing a mode where VirtualPC emulates a little-endian CPU on a big-endian CPU, while technically possible, is an enormous task, and it even might not be feasible (as in "resulting emulation runs too slow").
Headlining news stories like this are the cause of overblown expectations and inevitable dissapointment when the actual stats come out. Of *course* 5% of procs will run at close to 80% of theoretical maximum; here Wired is doing a simple linear extrapolation of this to all 1,100 procs which is completely unrealistic. Big Mac will be lucky to reach 60% of tmax, and more likely will achieve closer to 50%, being relatively untried and untuned.
This is undoubtedly an incredibly cost-effective machine, but it will be lucky to get in the top 10 at all. Number 2 is a pipe dream.
Anonymous, is that you? I thought I told you to clean up your room and stop playing on that damned computer. It is past your bedtime and I don't want to have to come back here and tell you a second time -- I'm warning you...One more troll out of you and I'm taking away that damned Mac SE! -- Oh, what? You didn't want your little friends to know that you use that old thing? Don't make me tell them about the Justin Timberlake posters on your wall.
Okay, I love you too, dear. Sweet dreams.
Mother Coward. Making the world a safer place for Mac Zealots since 2003...
If I did my calculations correctly, the VT cluster is already in 82nd place, with a score of 0.809 TF (beating out an IBM 252 sized cluster of Itanium 2 1 GHz chips) at 0.798 TF. The quote was for 128 processors, or 64 nodes at 80% effeciency--17.4 TF/1100*64*.8=0.809 TF I've seen some suggestions that using stock GBit Ethernet would lower the rate 30%, so a "stock" system would come in at .567 TF, or 146th place, slightly behind a SP Power3 375 MHz 16 way/ 520 setup.
...that VPC (and SoftWindows before it) ran just fine on earlier PPC processors that didn't have the little-endian mode. It did help, but it's hardly a deal-breaker.
-fred
Sign #11 of Slashdot overdose: You see the phrase 'moderate Republican' and you wonder if that would be a +1 or a -1.
Oddly enough, there are dozens of countries which treat their people worse than IRAQ did. And, in fact, the average person in IRAQ is and will (the US government admits) worse off now than they were under Sadaam.
I'm all for just wars, if you can find me one. But one where the US government lied to its populace about the justifications for the war, rushed into it without any international support, and alienated basically every other country on earth except for Britan (the citizens of which are now lynching their leadership because of it) isn't it.
Oddly enough, I was okay on the war in Afghanistan, because, well, the leadership there were pretty much a bunch of bastards who tried to kill lots of us. Of course, we destroyed the country, then promised to help rebuild it, and then completely failed to even BEGIN to keep our word, so people are starving to death daily there and other countries are cleaning up after our mess.
But being the world's only superpower is evidently all about being able to kill lots of people, destory their homes, ruin their livelihood, distrupt their way of life, and then not feel sorry for them, because they were dumb enough not to be born Americans.
God bless America.
-fred
Sign #11 of Slashdot overdose: You see the phrase 'moderate Republican' and you wonder if that would be a +1 or a -1.
...yet...
-fred
Sign #11 of Slashdot overdose: You see the phrase 'moderate Republican' and you wonder if that would be a +1 or a -1.
It was not and is not illegal to export computers.
Computers above a certain processing power cannot be exported to certain countries; countries you also can't sell arms to.
The same went for encryption.. it's not that you were not allowed to export strong crypto; you were not allowed to export strong crypto without a PERMIT, which was often attainable.
"Dongarra said the cost is so low he questioned whether the college got a special discount. Lockhart couldn't be reached for an answer"
If you do the math, 1,100 X $2,999 (street price of a DP G5) It looks like it SHOULD have cost only. $3,298,900.00. Now they need racks, so let's figure $800 for a decent rack and shelves to hold 12 CPUs. (That's how many they have stacked in each rack) That takes you to about 92 racks, or $73,000. You are now at $3,372,233.33. Even adding in a building, they are still WELL under the 5.4mil.
Good for them...
My
Actually, not. The $5M cost for the VT cluster includes infrastructure like racks, networking, and facility cooling updates. The $2.5M figure quoted by UT was for boxes only. I don't have the VT figures in front of me but I would guess that the G5 only costs for the VT cluster are likely close to the $3-3.5m mark. The performance being better than 4X and the box costs being close makes VT's cluster a much better deal.