Slashdot Mirror

Amazingly, it was also true. That was an SGI file browser running on Irix. See http://www.sgi.com/fun/freeware/3d_navigator.html

And vice versa. Which makes it hard to compare the two.

And the future is multiprocessors...

According to whom? Sun? There's a lot of high end work that's massively parallel, and clusters on commodity hardware are becoming huge. That said, if you want to pay for a 256 CPU Linux box, SGI will sell you one. Including hot swappable CPU support.

I wouldn't be surprised if page locking wasn't available in Linux too.

Give everyone credit for "GLX".

Personally, I only give GLX's credit for SGI and the contributors. The only estabilished definition for GLX, of course, being the glue layer between OpenGL and X11, which is pretty widely used these days, coming right out of box with XFree86 and all.

Lesson number whatever: If you come up with a cool term, at least bother to google for it =)

Yes, there are codes that would be faster on 1000 Itanium2 vs 5000 Opterons, but you would never runs these on cluster, because they would be faster still shared memory system.

This machine basically is a shared memory system: OK, it isn't, but the performance is the same. Remember your Athlon or Pentium 3? It probably had a memory bandwidth of around 1GB/sec.

Guess what the bandwidth between any two nodes of Thunder is? You guessed it, slightly under 1GB/sec. QsNet II is not _quite_ as fast as SGI's NUMALink 4 (which they use to make their shared memory Altix systems but it's in the same ballpark.

Don't be silly, Apple is a very small fish in a very big pond compared to real supercomputer companies like SGI.

What the big companies do is let you login to their machines free of charge, over the internet, and take their machines for a spin. See e.g. http://www.testdrive.compaq.com

On top of that, what the companies with a *serious interest* in high performance computing (like SGI and Cray) do is let you email your program to them: they will spend days or weeks (up to you) tuning it, telling you how it works on their hardware, and even telling you that you're better off with the competition, if that ends up being the case. What these companies understand very well, that Apple doesn't, is that people who spend $10M+ on computing equipment, typically with public money, need to show due diligence in their choice of hardware. This means THOROUGH evaluation: benchmarks, benchmarks, benchmarks - and not _standard_ benchmarks, but the benchmarks that matter: the user's own software.

Take a look here to see a real-life snapshot of this kind of process.

Any company that tried to "fudge the numbers" would be caught out, and that looks VERY bad. So the companies instead do all they can to help with the evaluation process, and hope that they get chosen. If not, there's always the next sale. A big supercomputer is sold somewhere every day or two, after all.

Virginia Tech's "X" doesn't come into this category: for example, no actual scientific work has been performed on the machine so far apart from benchmarks/system development, etc. Dr Varadarajan is probably going to get a pretty nasty grilling in a couple of years, when the University asks "so, what did we get for our money?" Some P.R., and not a whole lot more it would seem. I wish them luck...

Apple is not really a supercomputer company: and really, who cares? Apple has been, is, and will probably remain a great computer company for a good long while yet. But don't expect Apple to help you out a great deal if you pop them an email going "say, how does the G5 run on the ASCI Purple benchmarks?". That's just not what Apple are about. (example: they are happy to sell you a box with a 64-bit CPU, but they could care less that they don't have a 64-bit OS to go with it..)

XFS supports defragmentation
http://oss.sgi.com/projects/xfs/manpages/xfs_fsr.h tml

Well, I can't recommend a book, since I learned C++ first, but I have a warning...there is no definitive API. I imagine that would be a stumbling block for Java-to-C++ converts.
However, might I recommend this for the STL?
I think that in searching for a book, you should probably look for one that highlights the differences between the languages. That helped me when I was learning Java, and it's probably the quickest way you can pick it up. Have a look at Amazon and just search under books for "java c++". The first few entries are aimed at this. I can't vouch for these books, but hey, that's what I found.

"...more detail to security will have to be maintained to ensure the safety of the systems you are running."

Oh you mean like SELinux, UML Linux, File System ACL's, and Chroot jails? Oh I feel much more comfortable about Linux's security future than Windos.

They still sell workstations. Top of the line in desktops is a 2-CPU 800MHz RISC processor. Does anybody care?

They sell Itanium-based servers. Nobody buys them. (Well, actually they've sold a total of 13,000 Inanium CPUs. Not systems, CPUs. This is somewhere under 1% of the server market.) They sell MIPS servers running Irix. Nobody buys them either. SGI isn't even listed in the top five server vendors any more.

Almost all their buildings in Mountain View have been taken over by others.

a ReiserFS partition
c'mon. everybody knows that xfs is way better than reiserFS

No, you're right. I was replying to a poster whom I'm now certain is some sort of anti-OSS troll (and is thus now on my foes list, grrr). He was trying to be whitty and criticize a "Linux user" for using KDE and the MS-isms therein. I made a simple parody of his post but at least two people seem to have missed my point. Maybe because the original post was soon modded as troll, maybe because my post was so short and I didn't qualify it with an explaination. I simply tried to come up with some "features" of Windows that were developed outside of MS (which isn't hard, really).

An anonymous post got the general gyst though.

• Multitasking - Multics, Unix, VMS. This feature goes way back to time-sharing mainframes. It's probably difficult nowadays to ascertain who really did it first.
• GUI - Xerox PARC, Apple, MIT (X), Acorn RiscOS.
• Mouse - Stanford, Xerox PARC, Apple.
• Audio hardware - erm, not really sure. Apple macs and Amigas had built-in sound hardware long before the x86 PC did.
• Accelerated graphics hardware (2D/3D) - Silicon Graphics, Sun, HP, Apollo, Evans and Sutherland.
• 3D API's - IrisGL/OpenGL, PEX.
• LANs and the Internet - Xerox PARC (LAN concept), 3Com (ethernet), Apple (Appleshare?), Unix (TCP/IP), Novell (IPX/SPX).

Show some respect for your elders, or at least know them ;)

U3D is actually just the Shockwave3D file format that Intel are trying to ram through a standards body somewhere.

VRML was based on SGI's Open Inventor file format. You could change the header line in a VRML file then get it to open in the fancy SGI apps. I believe VRML was a subset of the IV format. Still haven't seen link-a-tron in VRML though.

I think VRML failed because the browsers were too slow, too beta, or too buggy when it came to scripting. It's still very useful for ad hoc data visualizations though.

...which is good but relies on SGI's FAM which itself relies on a heap of other stuff. Definitely not for the faint of heart.

Sorry but the best filesystem for high I/O by far IMHO is XFS on SGI Irix. When you run I/O via the distributed filesystem on a san where 40 machines mount to the same filesystem at once.

Perhaps you need to take another look at at std::vector

Both Xsan and CXFS are cross-platform: you can attach heterogenous (Windows, Linux, Irix, Solaris, Mac OS X, possibly others) systems to the one filesystem, and have it all work. The interesting part is that CXFS needs an SGI Irix box at the centre to deal with the metadata updates (as I understand it). Xsan also needs a metadata server, but it's unclear whether it needs to be an OS X box, or if it'll work with other operating systems at its core. If the former, it's understandable. If the latter, it'll be a good chance to make it into the enterprise in a big way.

Either way, it looks like Apple is making some serious, steady steps towards the enterprise market. They're very much the underdogs; people looking at this sort of thing like to see a track record before buying; but still... interesting times, indeed.

Actually, they're still using the cool logo, just not on the front page of their site. See here, for example.

That was an actual program. 3D File System Navigator.

I wish that people that pretend to be computer experts would do the teeniest bit of research.

How about this gem: First introduced in 1995, Microsoft's DirectX application programming interface (API) was designed to make life easier for developers by providing a standard platform for Windows-based PCs. Before the arrival of DirectX, developers had to program their software titles to take advantage of features found in individual hardware components. With the wealth of devices on the market, this could become a tedious, time-consuming process.

I'm glad he cleared that up for us. Because this little known company called SGI didn't develop OpenGL back in 1992. In fact, were it not for MS, we would still be in the computer graphics dark ages.

I'm not trying to troll here. I am just pissed that people pretend to be experts when they don't have a clues what they are talking about.

I wish that people that pretend to be computer experts would do the teeniest bit of research.

How about this gem: First introduced in 1995, Microsoft's DirectX application programming interface (API) was designed to make life easier for developers by providing a standard platform for Windows-based PCs. Before the arrival of DirectX, developers had to program their software titles to take advantage of features found in individual hardware components. With the wealth of devices on the market, this could become a tedious, time-consuming process.

I'm glad he cleared that up for us. Because this little known company called SGI didn't develop OpenGL back in 1992. In fact, were it not for MS, we would still be in the computer graphics dark ages.

I'm not trying to troll here. I am just pissed that people pretend to be experts when they don't have a clues what they are talking about.

What people seen to be missing is that Linux doesn't only do multi-processor architecture via clustering a' la' Beowulf.

In the traditional clustering architecture you have a lot of standalone machines that operate each in their own memory space an each use their own operating system image.
This is great when a problem can be broken into a lot of nearly independent pieces.
When you have a lot of interconnected pieces as might happen in matrix problems - finite elements or big matrix inversions it is best to have a single block of memory accessed by many processors simultaneously. This cuts down on the penalty for inter processor communication.

Linux has had NUMA from since the mid 2.4 days.

The SGI Altix uses this technology with 256 processors operating with a single Linus OS image on a single block of memory!!!

I can just imagine some ad involving flying penguins zipping past a jet or something..

For a certain class of computational problems,
a cluster will not cut it. However, a big
shared memory machine does not have to be
a CRAY, it can be a linux shared memory
machine, as the Altix product range
from SGI.

The SGI Shared-memory linux clusters are genuine HPCs.

Note: SGI uses Itanics in the Altix, possibly because Intel gave them access to everything they needed to build the memory interfaces they needed for this chips. I'm wondering when we'll see an AMD/64 Altix, and if not, why not.

All those points sound good, which is why UNIX(tm) salemen still use them. Too bad they're lies.

1. It takes fewer Unix boxes to accomplish the same job as the ever multiplying rabbits^W x86 machines.

Wrong. In 2000 when SGI offered me an O2 for $25,000, it would've taken 5 of them to do the same job as one $3,000 Dell Pentium2.

Reliability, uptime, maintainability, etc. are all worth paying extra for.

Those are all areas where an x86 PC trounces Sun or SGI systems. With Sun, the annual maintenance fee can often exceed the lifetime ownership cost of a PC variant.

I suppose that in the 4 years that've gone by since a UNIX dealer has last tried to sell me anything, they might've gone more into a desperation mode and started reducing prices to get competitive with Intel/AMD systems again. But they just won't be able to catch up. Because x86 is used for home/business as well as the server room, it's got an unmatchable economy of scale subsidizing R&D and manufacturing costs.

Long term, Unix machines still win the day.

True, but only because Unix!=UNIX(tm). Dell P5s with Linux are Unix, and they win.

You missed a few:

http://www.hp.com/products1/unix/java/

http://h18012.www1.hp.com/java/alpha/

http://www.sgi.com/software/java/

http://www-106.ibm.com/developerworks/java/jdk/ind ex.html

http://www.apple.com/macosx/features/java/

So it looks like we have JVMs for, at least, Linux, Solaris, Windows, OS X, Irix, AIX, HP-UX, Tru64, OpenVMS, OS/2, and z/OS.

What was the list of platforms for C# and .NET again?

I've seen 64+ processor Linux boxes. Their made by SGI, and they scale up to 256 processors currently and can use up to 4Tb of memory. One of my friends at my LUG works for SGI and did a presentation on them. Take a look at SGI's site

Look at number 8 on this list:

TPC Top Ten non-clustered

Notice Sun is nowhere to be found...NDA result for their 72 way E15K is around 480,000 tpmCs. The 32 way Itanium running SuSE Linux eats it alive @ 609,000 tpmCs.

Also have a look at:

SGI Altix

Up to 256 CPUs in a single system image (Sun can't do that even with their E25K not yet shipping) and the Altix will be/can go to 512 CPUs...I beleive NASA has the first one of this size?

What's wrong with the old UNIX systems? Solaris still boasts of some functionality that Linux will probably take a few months :) to program and test. Think 128-cpu scalability, hot-swap CPU...

For large systems, with lots of concurrent processes and users and large amounts of SAN and other I/O, Linux is totally unfit compared to the real thing.

Tell that to SGI and their 256 cpu, single system-image servers

Another poster said:
The only thing that keeps Solaris alive is that companies buy a system with Solaris on it for the hardware.

That is 180 degrees out of whack. People don't buy Sun for the hardware, Sun has consistently trailed the performance and reliability curves at all levels of the marketplace. People buy Sun hardware to run Solaris to run the applications that Solaris supports. Linux is rapidly eating away at that advantage with old apps being ported daily and new ones popping up on Linux first. Sun can't compete on hardware alone and they know it, which is why their response to linux has been so schizophrenic - they can see the end coming, but they would rather fight and postpone than accept the envitable.

Indeed and I love the fact that they say: "As seen in "Jurassic Park"!" http://www.sgi.com/fun/freeware/3d_navigator.html

"This is UNIX" not Linux. It was IRIX on SGI Indys. The 3D filesystem is still available for download.
FSN

Yes, you mean the SGI FSN 3D Navigator. Perfectly real.

You mean the SGI FSN 3D Navigator?
http://www.sgi.com/fun/freeware/3d_navigator.html

A good replacement is the open-source fsv:
http://fsv.sourceforge.net/

Actually, FSN is a real program and can be found here.

nah that was SFN

this is true. mod up. here is the project's webpage.

Please...

This may be an old news, but the details of that machine is here. That's some stuff to drool over. Some excerpts:

... provide a combination of 4TB of online storage and more than 20TB of nearline storage as a global storage repository ...

... create and manage up to 100TB of data ...

And now this machine is up for a rent. Here's the company website.

Have a look at companies like SGI, Nvidia, Sun, Microsoft, Sun and IBM, and the OpenGL logo. All of these companies logos are either abstract shapes or just an rearrangement on the companies name.

Funny you should mention that. I went to SGI's website on your recommendation and on the front page is an ad that clearly features a picture of a penguin to emphasize their point of production quality linux

Have a look at companies like SGI, Nvidia, Sun, Microsoft, Sun and IBM, and the OpenGL logo. All of these companies logos are either abstract shapes or just an rearrangement on the companies name.

Funny you should mention that. I went to SGI's website on your recommendation and on the front page is an ad that clearly features a picture of a penguin to emphasize their point of production quality linux

SGI's IRIX has this tool with little versions of each desktop in a scrollable resizeable pane. You can hover your mouse over the little windows and icons in the desktops and tooltip-like name of the application shows up. You can drag little icons fro one desktop to the next. It's like you own little computer . . . it's so cute. This tool has been a standard part of IRIX for at least 8 years (throughout most of IRIX-5.* and all of IRIX-6.*). The window manager is 4Dwm. I've never in my life seen a desktop pager for Windows. They must hide it or make you pay for it.

Hmm I think I'd like to do this, email a moviemail to a server, or send what is normally a lousy video stream from a FOMA phone to a server and do it there. Seen this sort of algorithm around in Siggraph once upon a time, wonder if there is anything linuxy that can be bent into shape to use the cycles on my new VPS..

(To me this looks like what was documented in Graphica Obscura - projective warping of multiple photos - by SGI researcher Paul Haeberli. Actually his site has lots of info (I haven't seen code though) for doing wild things with color, depth of field, resolution, and so on using neat algorithms, style, and mucho fast computers. But now we can approach more closely the power he had 10 years ago. I love how slashdot forces me to look for sites I loved and lost. This seems to be a repetitive cycle for me with a period of a year. Help Mr. Gerlernter!

Here is more info from the original article.
It is a quickie so if someone wants to take a shot at translating the whole thing it might be good.

This is jointly announced by NEC and Nara Advanced Science and Technology University Graduate Program, which were working together as an example of biz-academic collaborations the government has been trying to foster.

It's based on two technologies, "mosaicing" and "ultra high resolution imaging". Mosaicing is defined as making an image of a flat surface or virtually flat distant scene with a wider angle than the camera is normally capable of capturing, by changing the position and angle of the camera, and later composing the resulting images into a single one. (So this is just a definition of a mosaic)

Ultra-res imaging technology is defined as oversampling by turning the object through slightly different angles and composing the resulting images into a single one. (So this is like Magellan's oversampling).

It says they were aiming at using consumer video cameras and camera-equipped phones to make a low-cost, low-annoyance way to do imaging, with a goal of say 15 megapixels, or like what you would get with an A4 page scanned at 400dpi.

The development was done without any special sensors or whatnot, and claim they are able to get similar quality to what a scanner would get by just using a consumer video camera to scan an A4 sheet of paper with this technique.

Then there's some marketing speak and it is presented as research results, no discussion of exactly what the system is or if it will be provided to the public.

A journalled filesystem does not journal the data (well, some do, read on). A journalled filesystem journals metadata changes (permission bits, file renames, etc). It is entirely possible that the data in those files gets corrupted. In particular, we use the excellent XFS file system from SGI. Because XFS uses such aggressive caching data can be corrupted during a failure.

Now, some journalled filesystems (e.g. an option in ext3) journal file data as well. This doesn't help but can actually cause *worse* problems as seemingly correct but stale data can be left behind.

The very fact that arch makes no use of the fsync system call on the server side (since it has no dedicated server) means that data loss is much more likely than with Subversion which ensures that data is written first to log files and then replayed into the respository DB.

As for repository size? Is that really an issue that even matters? Unless arch users write gigabytes of source code I don't see the problem here. How much does a couple of SCSI disks and a RAID controller really cost that guarding source code isn't worth it?

What do the arch guys use for a server that this is such an issue? Can you even buy less than 9GB SCSI disks anymore? Is spending $2000 for a decent server really out of line?

SGI has a pretty good STL reference here. It's quite comprehensive if you're looking for a reference for the standard containers and algorithms.

Most people I know (including me) use SGI's page.

I think the exciting thing about this is that we will finally have the possibility to multipipe our graphics systems. A gamer can have a very good system with one card installed, but a graphics workstation might have 4 installed. Not going to 4 monitors as others have mentioned, but all splitting the load for one display for 4 times the performance.

As it is I have to have 4 full machines and a wicked fast network to do the same.

fire

> Not necessarily easy, but it would sure help in cases like this.
Not necessarily complicated:
Monitor the "/home" directory for changes with FAM a "FAMMoved"-event denotes the renaming of a home directory. Make approriate changes in necessary files (e.g. passwd/shadow). A "FAMDeleted"-event could trigger a dialogue, which confirms the removal of the user from the db, a "FAMCreated"-event the dialogue for adding a new user.

The problem is, for every such solution there are surely hundred of other places to break the system, especially as root. So such work would probably in vain.

I would be more impressed if Verisign restarted the Pathfinder instead of Sitefinder.

Its from SGI.

I believe it was used for the onscreen displays - where the girl says 'this is a unix system, we have these at school' and it's a 3D maze type user interface etc.
That's Silicon Graphic's FSN (fusion) file manager from the late 1980s. http://www.sgi.com/fun/freeware/3d_navigator.html. There's also a modern clone of it somewhere.

Okay, I'll bite. Read how the aglorithms in the STL are implemented. Stroustrup's book has excellent examples that are more digestable than cracking open the STL source if you don't feel like approaching that.