NASA Installs Linux Supercomputer

I can almost.. by aaron_ds · 2003-11-19 20:14 · Score: 2, Funny

Imagine a beo...oh never mind.

Uh-oh... by Hanzie · 2003-11-19 20:18 · Score: 5, Funny

The article explicitly says they're using NUMA archeticture.

Obviously, it's running SCO's intellectual property. SGI doesn't really own NUMA, they only wrote it. Deep down, it's really a derivative of vi.

--
********* sig: If you don't like the law, get filthy stinking rich, and buy a better one.

Re:Uh-oh... by Bernie · 2003-11-19 22:17 · Score: 1

Cool; vi is a derived work from one of the professors at my institution in the UK.

Does this mean Queen Mary owns NASA?
Re:Uh-oh... by 4of12 · 2003-11-20 01:44 · Score: 1

SGI doesn't really own NUMA

Of course not.

Anyone that's read Cliver Cussler knows the National Underwater and Marine Agency (NUMA), headed by Admiral Sandecker, is part of the U. S. government.

While Dirk Pitt has done most of the work, NUMA has saved the world from catastrophe on numerous occassions.

--
"Provided by the management for your protection."

Altix by Preach+the+Good+Word · 2003-11-19 20:20 · Score: 5, Informative

SGI's Altix handles up to 64 processors on a Linux kernel using the patches they release as opensource. As SGI hacks away at their bigmem and numa patches, they'll be able to handle more and more processors. The plan is to eventually graft enough IRIX technology to support just as many processors on Altix as they do with MIPS processors in Origin with IRIX.

Even if you aren't a fan of Itanium2, Linux, or NUMA, these patches are bringing some nifty high-end tech to the free software arena.

Re:Altix by rf0 · 2003-11-19 20:40 · Score: 2, Insightful

The thing is this Altix is 256 CPU machine in a single system image. They are looking to take this to 512 in a single image. That is some serious scaling

Rus

--
Cheap UK and US VPS
Re:Altix by Anonymous Coward · 2003-11-19 21:10 · Score: 5, Interesting

The only IRIX technology they're grafting is the SCSI layer. With the 2.6 kernel, there really isn't that much they need to do. They're testing their 512 processor systems on linux 2.6 as we speak, btw.
Re:Altix by MrMr · 2003-11-19 21:58 · Score: 1

They claim to do quite a bit more on their oss page
Re:Altix by EyeSavant · 2003-11-20 01:31 · Score: 1

Are you sure there is not that much to do?
I don't think there is a problem with the scaling.. or they are fixing it anyway.
The problem would be the NUMA Does the 2.6 kernal support NUMA (Non-Uniform Memory Access)?
For NUMA to work efficiently you need to get the data and the processors close together. Most (if not all) previous the time Linux is used it does not matter if you shuttle the processes from processor to processor. In dual etc processor systems the memory acess time is the same. At the moment SGI have build a module ontop of linux to do this. They cannot do it in the kernal as they need to use a standard linux distribution. (you cannot get support from people like SAP (who make payroll software etc) if you are running a custom version of linux. Not important for you or me, but rather important for big business.
Re:Altix by GooberToo · 2003-11-20 01:39 · Score: 4, Informative

Tons and tons of work has gone into 2.6 to make it fully NUMA aware and scalable. Last I heard, there were still some minor memory allocation node biases left in the scheduler, however, those are actively being worked on. Furthermore, they were considered corner cases and not general NUMA processing problem domains.

IIRC, one or more the developers has a 128 (or larger, I forget) CPU NUMA system that 2.6 periodically gets tested against. Many smaller NUMA systems are commonly used by several others. It seems that many, but not all of the NUMA optimizations, also help SMP systems as well. As such, the developers have not been shy about embracing it. When the O(1) scheduler was writen, it was a very short period of time before they started adding HT and NUMA optimizations to it.
Re:Altix by EyeSavant · 2003-11-20 01:42 · Score: 3, Informative

The technology is fantastic. It will be even more fantastic when it works :-(. At the moment you have to be nuts to buy one of these, as they are complete bleeding edge technology. In Amsterdam they "upgraded" from a 1000 processor mips machine to a 416 processor Altix/Itanium2 machine. On paper an itanium processors should be 5 times faster than the old mpis processors. At the moment we are lucky to see 2 or 3 speedup. And that is AFTER you have tuned the damn things for itaniums.
It seems to be quite normal for code to run slower on the new machine than the old... So you have a tough porting job to do...
The problem with the itaniums is that all the hard optimisation stuff has been moved from the hardware to the compiler. Plus the compiler is buggy as hell at the moment... I am currently trying to track down why my code fails on the new machine when I turn the optimization on.. My money is on a compiler bug to go with the many other compiler bugs we have seen before.
I am sure they will be a nice machine eventually, but if you buy one now you are nuts.
Re:Altix by fitten · 2003-11-20 03:29 · Score: 1

On paper an itanium processors should be 5 times faster than the old mpis processors. At the moment we are lucky to see 2 or 3 speedup. And that is AFTER you have tuned the damn things for itaniums.
It seems to be quite normal for code to run slower on the new machine than the old... So you have a tough porting job to do...

Is this speedup per processor or overall on your application? In any case, any time you move to a new system, you may have to do some tuning of your software for the new machine. For example, the Sun E10K+ has memory latencies very similar across the system. You can write code that works fine there, but when you move to a NUMA machine it may run very poorly because of memory access patterns and memory locality to the processor. You'll need to tune your software to make it run better.

The problem with the itaniums is that all the hard optimisation stuff has been moved from the hardware to the compiler. Plus the compiler is buggy as hell at the moment...

So... since it's "easy" to be done in hardware and "hard" to be done in software, hardware engineers are better at software optimization than software guys? ;) Buggy compilers are always a pain, but ignoring that, there's no reason why
the compilers can't get better over time to produce better code. I know that doesn't help you right now though :)
Re:Altix by SuperQ · 2003-11-20 03:32 · Score: 1

yep.. it's great.. we have a 48proc altix at work, and we have another one on order. (1.3ghz I2)

Rumor has it.. SGI has 128proc's in beta in their labs. That would match the capacity of the origin 3800's that our Alitx replaced. (we had 2 96proc 3800's)
Re:Altix by SuperQ · 2003-11-20 03:35 · Score: 1

yea.. my co-worker who does our Origin/Altix admin stuff is always bitching about chasing Intel compiler bugs.. even on the 8.0 compilers.
Re:Altix by drinkypoo · 2003-11-20 03:37 · Score: 1

I don't suppose there's any chance that the Linux XFS team is planning to bring us guaranteed rate I/O? I'm sure that would require some modifications to other parts of the kernel...

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Re:Altix by oudzeeman · 2003-11-20 03:53 · Score: 1

The Altix isn't a 256 CPU single image system, at the current time anyway.
Altix nodes come with up to 64 Itanium2 processors. Each node has a single OS image. A 256 CPU Altix supercomputer would be made up of at least 4 nodes(you could use 8 32 CPU nodes,etc). Each of these nodes has its own OS. There is shared memory accross nodes however, using SGI NUMALink between nodes.
I looked into the Altix when I thought we might want a 64 processor Itanium system, but our main code isn't network bound, so there was no way to justify the extra cost of the Altix.
Re:Altix by EyeSavant · 2003-11-20 04:12 · Score: 2, Interesting

Is this speedup per processor or overall on your application? In any case, any time you move to a new system, you may have to do some tuning of your software for the new machine. For example, the Sun E10K+ has memory latencies very similar across the system. You can write code that works fine there, but when you move to a NUMA machine it may run very poorly because of memory access patterns and memory locality to the processor. You'll need to tune your software to make it run better.
The speedup I quoted is per processor speedup, but AFTER they had done some tuning. After you just port but before the optimization it seems quite normal to get slower code rather than faster... but from what you said that is fairly common. I am not sure it should be so expected going from one NUMA MIPS machine to a NUMA Itanium2 machine. Apart from the fact that the Itaniums are not very well balanced..
I do think they will get it right eventually, but it is still bleeding edge stuff unfortunately.
Note this is what the computer people at the computer centre told me not my personal code. The current state of my personal code is that it runs at 1/4 of the per processor speed on the new computer, (i.e it is running about about 1/10 to 1/20 of the speed you would expect) and gives the wrong answer... Clearly those things could be related... but I am destinctly underwhelmed at the moment.
So... since it's "easy" to be done in hardware and "hard" to be done in software, hardware engineers are better at software optimization than software guys? ;)
No that is not quite what I meant. The best way has to be some combination of the two IMNSHO. In the Itanium they started with a blank piece of paper, and asked the hardware guys what they wanted. Like normal engineers they made the tough stuff (the optimisation) someone elses problem. All the itaniums do is do exactly what the compiler tells them to do. So the compiler has to work very very hard to get optimsed code, and you always have more information at runtime than you do at compile time.
What we are seeing from this is exactly what you would expect, something with a VERY good peak performance that performs well on some benchmarks, but royaly sucks for real world applications..
Buggy compilers are always a pain, but ignoring that, there's no reason why the compilers can't get better over time to produce better code. I know that doesn't help you right now though :)
I am sure they will get better... Intel are patching their compiler about every month, and have a huge number of people working on it. Whether the whole EPIC idea will work on the other hand I don't know.... They are going to struggle very hard to get anything like the peak numbers out of it for normal applications..
Re:Altix by Wesley+Felter · 2003-11-20 05:17 · Score: 1

Yeah, it's sort of funny to see posts to linux-kernel from SGI people complaining about things that break on 512-way machines.
Re:Altix by bdrago · 2003-11-20 05:47 · Score: 2, Informative

This is incorrect.
The Altix at NASA is a true 512 processor single system image - it is not a cluster of smaller nodes.
Disclaimer: I work for SGI.
Re:Altix by GooberToo · 2003-11-22 03:14 · Score: 1

That's an interesting idea. Just the same, does that fall under the realtime XFS extensions? Does SGI currently offer guaranteed I/O rates natively? How do you assign I/O? On a per process basis? Part of an API that each process must use to request a minimal I/O level? How does the layer know how much I/O it can guarantee? Seems like the HA and associated driver, VM, and VFS caches would all make for a huge difference in what it could guarantee. How would an implementation cope with all of these variables and still deterministically guarantee bandwidth?

Re:Ocean? NASA? by mwilli · 2003-11-19 20:20 · Score: 1

hmmm, that's what I was thinking too. NASA(national auronatuical and space administration), I guess aeuronautical works with the ocean too, but well, it just doesn't seem right.

--
My sig beat up your sig.

All research welcome by fr0m · 2003-11-19 20:21 · Score: 2, Offtopic

It is fairly obvious that changes in our environment has increased rapidly during the past years.
Last winter has been one of the coldest in a few hundred years in Sweden. I was there (south part) during christmas and the warmest temperature we had was -24. The same goes for the summer here in Europe. So damn hot. Here in Paris we've had thousands of deaths due to the heat.

Something strange is happening. All research about our "new" environment is welcome. Ocean or otherwise. What are your thoughts?

Re:All research welcome by lebha · 2003-11-19 23:58 · Score: 1

But the meteorogists say that the superwarm summers and supercold winters fit in "normal variation". Whom to believe in when your gut reaction combined with the flood of information about the greenhouse effect and global warming make you feel like "this is it, climate change in action" and the scientists say otherwise?
Re:All research welcome by Zloopy · 2003-11-19 23:58 · Score: 1, Informative

I think it's mostly BS. The environmentalist wackos use every occurance of what should be called "weather" as a proof of (man made) global warming.

Last winter was indeed very cold, and The Swedish Environmental Protection Agency (Naturvardsverket) ran lots of ads in newspapers and on billboards about global warming, and how warm the weather is and that you must cut down on driving or else all snow will melt.

When it was the coldest winter in at least a generation, people laughed at the ads, but environmentalists used the same strategy as always. The unusual weather was caused by the global warming.

In the 1970's the enviros where sure that an ice age was coming up. In the 17th century it was unusually cold in the nordic region, and that was called "the little ice age". In the 10th century it was a lot warmer than it is today.

Just picking a few unusual events and saying "It is fairly obvious that changes in our environment has increased rapidly during the past years." doesn't make it true.
Re:All research welcome by adamfranco · 2003-11-20 03:40 · Score: 1

I think it's mostly BS.

While this may be true, since we only have one earth/climate to live in why not err on the side of caution? Research, developement, and implimentation of less/non-polluting energy sources is a huge market waiting to be developed. Yes, for now clean energy costs more, but those costs are going to creating gennerally good jobs. Its not just a loss.

Also, think about the "worst-case scenarios":

If global warming is BS and we move our economies to cleaner energy, then at worst we spend a bit more and have less smog, less acid rain, and cleaner rivers.

If global warming isn't BS and we do nothing, then we are left with artificially bad weather, rising seas, droughts, and such along with the large amount of general pollution created.

Yes we save some money (though most of the savings are captured by large industries, not households), but wouldn't you rather we screwed up in the clean direction?

--
"When ideology and theology couple, their offspring are not always bad but they are always blind." -- Bill Moyers
Re:All research welcome by Just+Some+Guy · 2003-11-20 04:26 · Score: 1

Something strange is happening.
Say it with me:
Weather is not climate.
Weather is not climate.
Weather is not climate.
It is impossible to correlate one specific weather event with a change in climate. While it may be related, it could just be a statistical anomoly.
Last winter has been one of the coldest in a few hundred years in Sweden.
What was the cause of that winter? Was it part of a trend, or was it a random fluke? If it was part of a trend, could the current weather be related to a similar natural climate change? If it was just an anomoly, why couldn't the recent winter also be one?
The bottom line is that global warming may be real, but you can't use one particular event as proof.

--
Dewey, what part of this looks like authorities should be involved?
Re:All research welcome by SQLz · 2003-11-20 05:06 · Score: 1

Yeah because a couple hundred years means something to earth.
Re:All research welcome by Zloopy · 2003-11-20 06:11 · Score: 1

Yes, for now clean energy costs more, but those costs are going to creating gennerally good jobs. Its not just a loss.
Of course it's a total loss. Where do you think the money used to develope your things would have ended up otherwise? My guess is that it would be spent on purchasing more energy (lower prices raises demand) and also on other goods and services.
The jobs not created in the clean energy business will end up in those areas instead. Not only that, people will also get more goods that they want, compared to unnecessary energy research no one asked for (or at least wanted to pay for).
Buying expensive stuff which is just as good as cheaper goods for the sake of creating extra jobs, is just like digging holes in the ground and filling them up again, and saying that more jobs are created that way.
I also agree with the Anonymous Coward regarding the precautionary principle.
Re:All research welcome by Dick+Faze · 2003-11-20 07:49 · Score: 1

If global warming is BS and we move our economies to cleaner energy, then at worst we spend a bit more and have less smog, less acid rain, and cleaner rivers.
Actually, we spend a TON more for marginal, arguable improvements (see Catalytic Convertor, creating less of a more toxic chemical than the one it filters). The costs are incredible, but you and I can argue about it, since we're not the ones who will end up hungry, unemployed, or homeless because of it.
If global warming isn't BS and we do nothing, then we are left with artificially bad weather, rising seas, droughts, and such along with the large amount of general pollution created.
And if there is a Santa Claus then we'll really regret all that bad stuff we did some day. No point in speculating about this crap. Its not true.
Yes we save some money (though most of the savings are captured by large industries
Who will have saved millions of dollars to use to grow their businesses and hire more employees, maybe even be able to afford another month's pension fund payment.
Re:All research welcome by nametaken · 2003-11-20 15:41 · Score: 1

My thoughts are, "Excuse me, but screw Paris. Oh, life is shit... just look at our weather? Let's spend that money on keeping our space shuttles intact on re-entry." Besides, now that the French government doesn't have to spend anything on pumping that POS concorde, they can figure out why it's so damn hot there.

Re:Ocean? NASA? by javiercero · 2003-11-19 20:25 · Score: 4, Informative

Fluid dynamics and environmental studies are also part of NASA's research mission.

NASA was going to roll this out earlier... by AvantLegion · 2003-11-19 20:26 · Score: 5, Funny

... but their copy of Mandrake 9.2 broke their supercomputer's LG CD-ROM drive.

Re:NASA was going to roll this out earlier... by chadpnet · 2003-11-19 21:21 · Score: 1

*jabs knife in a little deeper and twists* :)

Patches ? by noselasd · 2003-11-19 20:26 · Score: 1, Funny

Whewre can I get the patches from SGI to make 2.4 run on such insane hardware ?

Re:Patches ? by BJH · 2003-11-19 20:29 · Score: 5, Informative

Don't know if this is what SGI is using, but the status of NUMA in the kernel and associated patches for it is shown here.

Re:Ocean? NASA? by Preach+the+Good+Word · 2003-11-19 20:29 · Score: 1

How is space trivial?! Besides, they have done a pretty good job of exploring the nearest objects.

Re:Ocean? NASA? by Stubtify · 2003-11-19 20:31 · Score: 3, Funny

Didn't you see the movie "The Abyss"?

or the tv show "Seaquest DSV"?

the ocean seems to be a gateway to the stars...

In other news..... by micaiah · 2003-11-19 20:35 · Score: 5, Funny

SCO sues Nasa for using Linux.

Darl McBride stated yesturday, "Since Nasa is using Linux we now own the entire universe and are claiming our rightful ownership."

Re:In other news..... by screenrc · 2003-11-19 20:57 · Score: 1

Strange, I thought I have more rights on
Linux than SCO. At least, I have the right
to distribute the Linux operating system to
others. SCO has violated the GPL; thus, they
don't even have distribution rights.

(Oh, assuming that is they recognize the
validity of the GPL, for if not, that will
be one additional reason they cannot legally
distribute Linux; although, almost everyone
else on Earth is free to distribute!)
Re:In other news..... by McGarnacle · 2003-11-19 21:01 · Score: 1

Darl McBride stated yesturday, "Since Nasa is using Linux we now own the entire universe and are claiming our rightful ownership."
That's just silly

--
I disagree with what you say, but will defend to the death your right to tell such LIES!
Re:In other news..... by klafhat · 2003-11-19 21:28 · Score: 3, Funny

SCO sues Nasa for using Linux. Darl McBride ...

In yet other news, Darl McBride is now so afraid how NASA is going to respond, that he have hired a couple of bodyguards.

--
Tell me more, tell me more
Re:In other news..... by dipipanone · 2003-11-20 00:35 · Score: 1

Strange, I thought I have more rights on Linux than SCO.

So send 'em an invoice for $699 in the next post. Give 'em a taste of their own medicine.

Re:Ocean? NASA? by SpaceCadetTrav · 2003-11-19 20:35 · Score: 1

It turns out that the study of other planets also requires some investigation into our own.

--
Life in Orange County

Not just that... by freakmn · 2003-11-19 20:39 · Score: 1

As if NASA needs any more funding problems, $699 goes a long way for monkey-food in space.

Not just $699, but $699 per processor IIRC. With 512 processors, that comes out to around $357,888 plus tax.

--
warning: This post is likely to contain gobs of dripping sarcasm. Consume at your own risk.

Re:Not just that... by peragrin · 2003-11-20 00:48 · Score: 1

NASA is a goverment agency and as such is tax-exempt. For Intelligent reasons the goverment doesn't tax itself. one of the few sensible policies it has.

--
i thought once I was found, but it was only a dream.

they also have. by rf0 · 2003-11-19 20:42 · Score: 4, Funny

A 512 CPU and 1024 CPU IRIX system. The 512 one is referred to as the small one :)

Rus

--
Cheap UK and US VPS

Not quite "Supercomputing" by quigonn · 2003-11-19 20:43 · Score: 4, Interesting

The SGI Altix 3000 is not quite a supercomputer. Our local university got the very first model for production use of the Altix 3000's successor, the Altix 3700, in last April or so, and it made it in the TOP 500 supercomputer list in last June, but it fell out of the current list. And the 3700 is even faster than the 3700, so what's so special about it?

--
A monkey is doing the real work for me.

Re:Not quite "Supercomputing" by Repran · 2003-11-19 21:05 · Score: 1

What's special about it is that it supports more processors - namely 5120 - on a single image under Linux than any other supercomputer before.

--
-- Contradictions only exist in thought - not in reality.
Re:Not quite "Supercomputing" by Repran · 2003-11-19 21:07 · Score: 1

Sorry - that should have been 512 not 5120...

--
-- Contradictions only exist in thought - not in reality.
Re:Not quite "Supercomputing" by quigonn · 2003-11-19 21:13 · Score: 1

This doesn't really impress me since it's a NUMA architecture (with all the NUMA-specific drawbacks). But of course, a cat /proc/cpuinfo | lpr would be extremely long. ;-)

--
A monkey is doing the real work for me.
Re:Not quite "Supercomputing" by Realistic_Dragon · 2003-11-19 21:40 · Score: 2, Funny

And the 3700 is even faster than the 3700, so what's so special about it?

If it's faster than itself, then with some handy infinite recursion you can prove it's infinatly fast!

Sounds pretty special to me.

--
Beep beep.
Re:Not quite "Supercomputing" by martingunnarsson · 2003-11-19 21:41 · Score: 2, Funny

"And the 3700 is even faster than the 3700, so what's so special about it?"

It's obviously faster than itself, that sounds pretty special to me!

--
Martin
Re:Not quite "Supercomputing" by Error27 · 2003-11-19 21:42 · Score: 5, Informative

The thing that is special about the NASA computer is that it is a single image system and not really some cluster type thing. Mostly people say that Linux 2.4 scales well up to 8 processors, but this system has 256 processors.

SGI is working on scaling the kernel to even more processors. For example, Erik Jacobson from SGI recently noticed that 'cat /proc/interrupts' doesn't work if you have 512 CPUs in your system. Frankly when I saw that I thought it was a joke, but I guess it must be real if they already have paying customers.
Re:Not quite "Supercomputing" by quigonn · 2003-11-19 22:02 · Score: 1

My university's Altix 3700 doesn't impress me, either. A supercomputer that impresses me? Big Mac.

--
A monkey is doing the real work for me.
Re:Not quite "Supercomputing" by FrostedWheat · 2003-11-19 22:18 · Score: 1

"And the 3700 is even faster than the 3700, so what's so special about it?"

It's obviously faster than itself, that sounds pretty special to me!

Those pesky overclockers again! What will they do next.. overclock there toaster??
Re:Not quite "Supercomputing" by grey1 · 2003-11-20 00:23 · Score: 1

this comment looks perfectly reasonable ("just because your campus' super computer is not in the top 500 shouldn't relegate it to atari 2600 status") so what's it doing at -1?

--
"we demand rigidly defined areas of doubt and uncertainty!"
Re:Not quite "Supercomputing" by GooberToo · 2003-11-20 01:44 · Score: 1

Ya, I followed that thread too. I thought it was fairly interesting myself. Seems the kernel runs out of kernel memory to handle the kernel-like printf statement. The fix is to convert it to an iterative approach whereby, many smaller chucks of memory simple get reused. Appears the infrastructure was already in place from another part of the kernel. So, seems it will be a fairly quick fix for them.

I knew it wasn't a joke (or certainly hoped not) and immediately thought, WOW! Serious scalability!

how does this compare? by kjba · 2003-11-19 20:46 · Score: 3, Insightful

"NASA plans to study the ocean's future with the help of the world's first supercomputer of its kind to run on the Linux operating system."

Nice to know that it is the fastest Linux supercomputer, but how does this compare to the other top-ranked supercomputers in the world?

Re:how does this compare? by DeathPenguin · 2003-11-19 21:22 · Score: 1

Like Lightning, a Linux-based Opteron cluster ranked at #6 in the newly released 2003 list?

By "of its kind" the article meant that is used for oceanography. By "fastest" they mean that it excelled in ECCO.
Re:how does this compare? by EyeSavant · 2003-11-20 01:21 · Score: 1

That is not that hard.. There is an 416 processor SGI Altix at Sara in amsterdam, that is currently 41 in the top 500. (www.top500.org) The problem with these machines is that they run benchmarks really well, but if you try doing anything real with them you do not get anything like the proper performance... That is more a problem of the Itanium2 than anything else though. And the Intel compiler for itaniums is buggy as hell. This is bleeding edge technology :-(.
Re:how does this compare? by jelle · 2003-11-20 03:32 · Score: 1

Maybe it wasn't originally delivered as a single system image, but they could still install this on it instead of beowulf/pvm/mpi/mosix and then it becomes a single system image cluster.

--
--- Hindsight is 20/20, but walking backwards is not the answer.
Re:how does this compare? by ryanwright · 2003-11-20 07:10 · Score: 1

Nice to know that it is the fastest Linux supercomputer

It's not. They lied. This is the fastest Linux supercomputer:

http://www.top500.org/list/2003/11/
Pacific Northwest National Laboratory
Integrity rx2600 Itanium2 1.5 GHz, Quadrics Interconnect
1,936 processors
8633 rmax / 11616 rpeak

--
-Ryan, with the unoriginal sig
Re:how does this compare? by jelle · 2003-11-20 11:22 · Score: 1

Agreed, it won't have special hardware for NUMA, but it still is a single system image. The only way you and your apps will notice the difference is by the speed. The OS can make memory appear shared between processes, independent from whether or not the memory is shared across a special NUMA bus, or ethernet. The difference is only visible from the latency and bandwidth the app gets.

--
--- Hindsight is 20/20, but walking backwards is not the answer.
Re:how does this compare? by jelle · 2003-11-22 09:10 · Score: 1

"The only way to do it without modifying the cache controller to do distributed cache coherency is to turn off the data cache completely."

Not true, because pages can be marked read-only, or even marked to fault on a read, at which point the OS can use mutexes/semaphores across any type of network to ensure data coherency. And IIRC, both openssi and openmosic either already can do that, or will be very soon (follow the links, or search for migShm).

Again, latency will suck compared to shared memory on SMP or true NUMA, but it will be shared memory with no other difference than the latency and bandwidth.

--
--- Hindsight is 20/20, but walking backwards is not the answer.

Study the oceans FUTURE? by Anonymous Coward · 2003-11-19 20:52 · Score: 1, Interesting

Jeez, if NASA has a method of studying the FUTURE of anything, you'd think somebody would have the vision to study the future of the stock market or even the sports books in Vegas.

Re:Study the oceans FUTURE? by Gumber · 2003-11-20 09:15 · Score: 1

Info about the future of stock markets or sports books in Vegas may be more immediately monetizable than info about the future of the oceans.

But...

The future of the oceans is a lot more financially valuable than either of them. Trends in the Oceans could have major economic impacts that send stockmarkets spiraling to their deaths.

this does not even make the top500 so by ... by Anonymous Coward · 2003-11-19 20:53 · Score: 1, Interesting

this does not even make the top500 so by definition it is NOT A SUPERCOMPUTER.

top500.org does show that for 3.2 million of macs and under 2 million of infiniband and hardware racks you can get to the THIRD position in top500,org using macintohses and the mac os x.

Re:this does not even make the top500 so by ... by Anonymous Coward · 2003-11-19 23:03 · Score: 1, Informative

Eh, are you kidding? Most standard Altix are listed on TOP500, and this is not a standard Altix so it should be on among the top 100. And you cannot compare a simple cluster (the Big Mac) with a real supercomputer (the Altix) with a single shared memory system.

Bad Influence by Anonymous Coward · 2003-11-19 20:56 · Score: 5, Funny

Isn't using a supercomputer that has hundreds of very hot processors to simulate climatic change going to directly cause a change in the climet be ejecting large quantities of hot air?

Re:Bad Influence by Querty · 2003-11-19 22:35 · Score: 1

Isn't using a supercomputer that has hundreds of very hot processors to simulate climatic change going to directly cause a change in the climet be ejecting large quantities of hot air?
That would be a manifestation of the Heisenberg Uncertainty Principle, thereby making this the first commercial supercomputer to exhibit Quantum effects. I can read the headline now: "Linux runs Quantum supercomputers!"

Re:Ocean? NASA? by IvyMike · 2003-11-19 21:03 · Score: 4, Informative

To quote one of the articles:

"Using the vantage point of space, NASA gains an understanding of our home planet that we could never achieve were we bound to the Earth's surface," notes Dr. Ghassem R. Asrar, associate administrator of NASA's Earth Science Enterprise. NASA's remarkable 45-year history and vast scientific and engineering resources have helped the agency launch numerous research missions to understand and protect planet Earth.

Sounds pretty reasonable to me.

3000 Procs @ $699 each. by kidgenius · 2003-11-19 21:05 · Score: 2, Funny

3000 Processors @ $699 ea ~ $2.1 million
3000 Processirs @ $1399 ~ $4.2 million

So that's where SCO was planning on getting its money for Linux. It all makes sense now

Re:3000 Procs @ $699 each. by kidgenius · 2003-11-19 21:08 · Score: 1

My bad...i didn't RTFA...only 512 processors so take all numbers and divide by 6

Re:i live across the 101 from ames by Mr.+Hankey · 2003-11-19 21:07 · Score: 1

Nah, that's not the NAS. That's just the power grid next to the 80x120' wind tunnel. Their 1024 processor box takes a lot more juice than this anyway.

Slightly on topic, it's nice that Linux is making headway there as well. There are only so many companies building 512 CPU boxes though, and Linux is SGI's best bet for the future.

--
GPL: Free as in will

Re:Ocean? NASA? by quigonn · 2003-11-19 21:10 · Score: 2, Interesting

Well, they pay lots of people for doing other stuff than space research. Just think of Larry Wall, who was working on Perl during his time at the NASA JPL.

--
A monkey is doing the real work for me.

Nice to see SGI still making sales by Space+cowboy · 2003-11-19 21:14 · Score: 4, Insightful

... I have something of a soft-spot for SGI, and it's nice to see them still making high-profile sales - it'll do their government profile no end of good :-)

512 processors running a single image is pretty cool :-))

Simon

--
Physicists get Hadrons!

Re:Nice to see SGI still making sales by jbrocklin · 2003-11-20 00:55 · Score: 1

I agree - I too have a soft-spot for SGI, and am glad to see that they are still getting customers in a market that fewer and fewer people really understand a need for. Linux & SGI aside, large systems like this are just plain cool!

Almost but not quite... by Powercntrl · 2003-11-19 21:37 · Score: 4, Informative

Don't hit the post button if your joke requires a life support system such as:

"oh, never mind"
Printed backspace symbols^H^H^H^H^H^H^Hcharacters
...pause (Think about it longer, you'll find a new way to make an old joke funny!)
A comment relating to the moderation system or karma
Rehashing all your old Slashdot memes are belong to Natalie Portman's hot grits in Soviet Russia goatse.cx posts YOU!
Using any form of Slashdot cliche as an attempt at humor
Ending your post with @^T#G@#YHB^#@$NO CARRIER

--

---
DRM is like antifreeze, to the MPAA/RIAA it's sweet, to the consumers it's poison.

Re:Almost but not quite... by achurch · 2003-11-20 15:44 · Score: 1

Don't hit the post button if your joke requires a life support system such as:
"oh, never mind"
Erm, but that is the joke . . .

Re:Ocean? NASA? by basingwerk · 2003-11-19 21:48 · Score: 1

I can think of a couple of reasons why NASA might be interested in this research area. First, the ocean and the rest of our planet is orbiting the sun in, well, space. Second, few organisations have the infrastructure to engage large-scale science programs like this one. If the 'Gulf Stream Conveyor' theory is true and the next ice age is due to start in 20 years time due to global warming, this research could be the rather more important than, say, finding out if germs exist on mars.

--
I stole this .sig

Space toilet by xixax · 2003-11-19 22:21 · Score: 1

This whole exercise probably costs far less than the shuttle's zero-G toilet.

Xix.

--
"Everything is adjustable, provided you have the right tools"

Re:Space toilet by salesgeek · 2003-11-20 00:24 · Score: 1

This whole exercise probably costs far less than the shuttle's zero-G toilet

Whew. Thank God it wasn't as expensive as the coffee maker.

--
-- $G

How is this "News" for nerds? by ODD97 · 2003-11-19 22:34 · Score: 3, Informative

If I'm not mistaken, NASA *invented* the bewulf cluster. And it ran Linux then, too.
Clicky

--
The emperor is naked.

Re:How is this "News" for nerds? by ODD97 · 2003-11-19 22:36 · Score: 1

Pardon.

Invented the _term_ "Beowulf".

*cough*

--
The emperor is naked.
Re:How is this "News" for nerds? by ODD97 · 2003-11-19 22:52 · Score: 1

One more try.

Developed the first actual computer cluster, and coined the term "Beowulf".

Why aren't there 5 or 6 preview screens? Or personal editors?

--
The emperor is naked.
Re:How is this "News" for nerds? by Styx · 2003-11-20 00:07 · Score: 4, Informative

It isn't a beowulf cluster.

It uses a single system image for all processors, as opposed to a beowulf, which has separate system images for all cluster nodes.

--
/Styx

Heads are going to roll by dcordeiro · 2003-11-19 22:43 · Score: 2, Funny

Linus is going to ask the responsible for the /proc/interrupts code how on hell didn't he tested that at home:
" Why didn't you test /proc/interrupts with 512 CPU in ?"

DUH!!

This new supercomputer has paid off already... by Anonymous Coward · 2003-11-19 22:47 · Score: 5, Funny

... as it has completed it's simulation of the ocean in order to predict it's future:

cold and wet.

Re:Ocean? NASA? by cpghost · 2003-11-19 22:51 · Score: 1

Wasn't it NOAA, who is responsible for studying oceans?

--
cpghost at Cordula's Web.

Re:Correction by waitigetit · 2003-11-19 23:37 · Score: 1

NASA plans to study the ocean's future...

I think I can guess the ocean's future if they use FreeBSD. It must be dying.

--
I could care less, but not without a lobotomy

Japanese govement ordered 4 x 64 CPU Altix 3700. by zzztkf · 2003-11-20 00:19 · Score: 2, Informative

According to SGI Japan, Ministory of Education and Science
has orderd 4 Altix 3700 computer to make up 4 node
super computer, November 18th.

Each node, which altix3700 is equipped with 64 cpu. Total
main memory has reached 1.9TB.

It's also said that hardwares will be installed and in
operation in the early half of 2004.

/proc/cpuinfo anyone? by prestwich · 2003-11-20 00:29 · Score: 2, Interesting

So anyone got a copy of /proc/cpuinfo from this mother?

(Oh and do you do something special to 'top' so it doesn't give you 512 lines of CPU state?)

Re:/proc/cpuinfo anyone? by drinkypoo · 2003-11-20 03:43 · Score: 1

They don't need to change top, NASA has all those nifty video walls that We the people like paying for so much. The admin will just open a >512x80 xterm...

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Re:/proc/cpuinfo anyone? by Anonymous Coward · 2003-11-20 05:11 · Score: 1, Informative

You do a "t" from within top (Toggle display of summary information) to turn off the cpu summary.

more than cool - very good engineering by grey1 · 2003-11-20 00:38 · Score: 1

Oh, I agree, and I'd go further - a single image over 512 procs is more than cool, it's very good engineering.

As with many things, the compute problems that get thrown at "supercomputers" or big clusters or whatever, will vary enormously. Some will require lots of CPU but have little need for a large network connection. Others will work much better with this sort of highly-connected system - low-latency, high bandwidth, single system image. There are some parts of problem space that best fit machines like the Altix (or the IRIX equivalent - Origin350/Origin3000).

--
"we demand rigidly defined areas of doubt and uncertainty!"

Imagine... by GooberToo · 2003-11-20 01:47 · Score: 1

...a cluster of these!!! ;)

Single System Image is Nice by 4of12 · 2003-11-20 02:01 · Score: 3, Insightful

The thing that is special about the NASA computer is that it is a single image system

I did parallel code development on Sun SMP boxes. Starting up jobs, seeing what was going on, killing zombies, debugging was all easier on one system than through different boxes you'd have to ssh over to see.

Even though I was using MPI and getting ready for a distributed memory architecture for the really big runs, the development was easier on the SMP box that showed a single system image.

I haven't used things like OpenMOSIX, and Don Becker, early pioneer of Linux ethernet drivers (not many other folks can claim a complete decade of experience with Linux networking), founded a company called Scyld that sells Linux clusters with single system image.

Sometimes it's convenient to see the whole box as if it were one, even though efficient programming dictates that you become aware of the different costs of data access (network, main memory, cache, disk). Practically speaking, developing and running parallel jobs is a higher user productivity proposition on a single system image.

--
"Provided by the management for your protection."

Re:Single System Image is Nice by jelle · 2003-11-20 03:29 · Score: 1

Here is a free single system image clustering system for Linux. IIRC it was HP that helped that project get a jumpstart.

--
--- Hindsight is 20/20, but walking backwards is not the answer.

uh-oh...lok out below! by whovian · 2003-11-20 02:17 · Score: 1

NASA engineers worked with SGI to expand the Altix systems' capabilities. It has achieved a high speed of 2.45 trillion operations per second.

Unfortunately, NASA will (mis)manage this cluster to running at 0.245 trillion ops per second.

--
To-do List: Receive telemarketing call during a tornado warning. Check.

Re:Ocean? NASA? by BuilderBob · 2003-11-20 02:36 · Score: 2, Informative

The research is being done at the AMES labs in Moffett field, California. Home of the climate model for Mars and Solar system modelling and numerical modelling in general. As well as some X-projects (as in X15 and X33).

NASA doesn't just send things into space anymore.

256 Procesors, that's not parallel computing!!! by GNUALMAFUERTE · 2003-11-20 02:37 · Score: 2, Funny

I am at work just at email reading and reloading Slashdot, the radio at high volume, Watching an Iron Maiden '92 live video, while i'm on the phone with yet another client that is mad 'cause we deactivated http upload for php, Now, *That* is _real_ Multitasking!!!!

--
WTF am I doing replying to an AC at 5 A.M on a Friday night?

Re:256 Procesors, that's not parallel computing!!! by GNUALMAFUERTE · 2003-11-21 03:04 · Score: 1

God, i don't know if this is funny or not ... i mean, there are things that we shoudn't play with, like war, my sister, and massive file system corruption : ) ...

Yes, we are a kind of Windoze NT ... we can do 2 things at the same time, but one of them (or both) will breake.

There is a research that says that there a person with a normal inteligence can manage to have 7 ideas / concepts / manage 7 situations at the same time. So, the standard deviation shows us that if that's true, most people should be beetween 7 - 2 and 7 + 2; but once again, reality says that that's not true, and most people are far aways from 7 ....

--
WTF am I doing replying to an AC at 5 A.M on a Friday night?

Sun and Report Writers?? by WindBourne · 2003-11-20 02:50 · Score: 1

I wonder how this will impact Sun. Sun has been making a big deal about Linux not scaling and that Solaris is. Yet solaris does not run on a SSI system of this size.
Linux keeps winning all the rounds.

--
I prefer the "u" in honour as it seems to be missing these days.

Look out NASA.. by Nonillion · 2003-11-20 03:02 · Score: 1

Because SCO is probally going to sue you ;)

--
"I bow to no man" - Riddick

Prodigy by lordmage · 2003-11-20 03:22 · Score: 1

IBM's new Prodigy Linux Comercial

--
I can program myself out of a Hello World Contest!!

No carrier? by TWX · 2003-11-20 04:06 · Score: 1

But where will those naval aviators land?

I used to have a 7MB taglines database, which was quite an accomplishment for someone who only had 4.77MHz of processor power and a 30MB hard disk. A large number of them were NO CARRIER jokes.

Ah, the days of Blue Wave Offline Mail Reader, FrontDoor, and Fidonet...

--
Do not look into laser with remaining eye.

Re:Ocean? NASA? by Jack+Auf · 2003-11-20 04:11 · Score: 1

Who modded this as a Troll? How exactly did you think he was trolling?

I was wondering the exact same thing - what are NASA doing spending money and resources studying the ocean when they can't seem to keep space research on track.

--
"They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety" - BF

You'll want "hinv" by green+pizza · 2003-11-20 04:52 · Score: 1

I haven't used an Altix, but from the docs it seems sgi's patches provide a lot of the tools and goodies from IRIX. On a montster Origin, "top" gives 4 nicely formatted status lines above the process details. For specific details on each processor and other aspects of the machine, there's "osview". There's also a handy script called "hinv" that prints out a hardware inventory of the machine. "hinv -m" even prints out the part numbers and laser-burnt serial numbers for the various boards in a system.
I believe the Altix software environment provides all three of these utlities. Each of which gives pages and pages and pages of nicely formatted, human-readable details with the right flags.

and now SCO... by josepha48 · 2003-11-20 05:24 · Score: 1

... sues NASA and all SCO executives miraculously disapear off the face of the earth and are last seen hurdling towards the sun in a firey death ball..... (LOL)

--

Only 'flamers' flame!
Does slashdot hate my posts?

Re:Is NASA reading tea leaves? by sfjoe · 2003-11-20 09:03 · Score: 1

Shoot, that's good enough for me. Anyone want to by my copy of yesterday's OS?

Thank, but I already have a copy of Windows 2000.

--
It's simple: I demand prosecution for torture.

Re:Aeronautics and Space = Water? by sfjoe · 2003-11-20 09:06 · Score: 1

What is the National Aeronautics and Space Administration doing spending huge sums studying water?

This actually is a good point. Why isn't NOAA doing this instead?

--
It's simple: I demand prosecution for torture.

Slashdot Mirror

NASA Installs Linux Supercomputer

111 of 189 comments (clear)