Slashdot Mirror

1) This was fully deployed in only 15 weeks.
(Link)

2) This number was using only 16 of the 20 systems, so a full benchmark should be larger too.
(link)

3) The storage attached holds 44 LoC's (link)

1) This was fully deployed in only 15 weeks.
(Link)

2) This number was using only 16 of the 20 systems, so a full benchmark should be larger too.
(link)

3) The storage attached holds 44 LoC's (link)

I would agree with you for the most part. However, I think she genuinely harbors some sort of grudge against IBM outside of the IBM vs. SCO theatrics. Take an article about Itanium from August - "Itanium-Armed Start-up To Menace IBM's Precious Mainframe Monopoly". The article itself is not particularly venomous towards IBM, but it certainly isn't an unbiased report on future competition for IBM. It just seems kind of absurd to go out of your way to prepare an article about a start-up company 'menacing' IBM's 'monopoly' when IBM already has plenty of more noteworthy competition.

But, as you said in another post, one has to wonder if she's doing this purely for the troll-value of it all.

When I worked on a rather large IT company in the old days[a year ago] with a lot of clients I found that Radmin was the best choice. The speed and refresh rate of it is great. Much more than in any standard VNC programs I have tried. Plus the option of using a radmin computer as a gateway to another radmin computer is great when trying to help cusomers over a vast WAN. Plus it has shell and file transfer possibilities. And I found it useful for checking how my EVE Online pilot was doing from work.

Only problems is that is is Windows Only.

For adminin linux or unix computers I just use SSH and an X tunnel. Heck. I even ran Maya between continents using that method. Both systems running Irix.

To admin the Windows servers at work I use a combination of an SSH tunnel and RDP.

Mips: MIPS IV Instruction Set Section A.2.1

Alpha: Cannot find an authoritative resource for proof, but the way I understand it is that NT was IMPOSSIBLE to run on a big-endian-only CPU, hence the #1 reason it never made it beyond rumor stage on Sparc.

Solaris and Endianness.

I remember back in '96-'97 timeframe hearing from a number of Sun vendors about experiments with NT on Ultrasparc, but could never get a demo (and we had Sun workstation vendors falling all over us to give us hardware at the time).

They say necessity is the mother of invention, well surly coolness is the father. 3D displays are fantastic for some things, obviously not that useful for all though. I don't know if that's an actual photo on the article maybe I need to RTFA but if it is it's a lot more clear than most 3D displays I've seen. I'd be interesting to see how it stacks up in usability for medical and engineering professionals vs something like SGI's Reality Center products.

I wonder how this compares to the one NASA is building, which is being collaborated with Intel and SGI. Since you can't base performance simply on the number of processors, it should be interesting.

Of course, we geeks know why the parent post is so funny: it wasn't a UNIX system in Jurassic Park, it was an IRIX system! (Running the fsn 3d filesystem naviagator.)

MIPS is dead in workstations? What about these
http://www.sgi.com/products/workstations/tezro/
http://www.sgi.com/products/workstations/fuel/

MIPS is dead in workstations? What about these
http://www.sgi.com/products/workstations/tezro/
http://www.sgi.com/products/workstations/fuel/

You want to change resolutions or monitor profile? Reboot.

No, restart X. Or run a second X server. Or hit Control-Alt-Numpad+ to go to the next resolution setting. Or use an X server that implements the RandR extension. (I think most drivers do in Xfree86 4.3).

You want to change font DPI? Reboot.

No, restart X. Or run a second X server. I admit that I don't know of an extension that can change the font DPI on the fly. You may be able to do some magic by restarting the font server to get different fonts in all new applications, but I haven't tried it.

Your system has scheduled fsck to run, either out of routine or out of improper shutdown? You need to reboot after it's done.

One word: ext3. Or another word: XFS. A third word (although some people consider it an obscenity, myself included): ReiserFS. There's even the IBM option.I've been running ext3 on my laptop over three years and many unclean shutdowns, and I never have to wait for an fsck, or reboot afterwards.

About the only thing that requires a reboot these days is a kernel or hardware upgrade. You go ahead and keep on rebooting the system, if that's all you know how to do. Just don't confuse your lack of troubleshooting skills with a deficiency on Linux's part.

Mix the principle from this Random Lava Lamp Generator, a high speed multi-mega pixel digital and film with a suffiviently high resolution, pipe resulting image as raw key-data, and you're done...

or if you prefer "better randomness" (sic) use HotBits cesium decay generated random numbers and pipe...maybe not gigabytes of data ( (about 30 bytes per second, to be precise, sucks...but then...

Mix both obtained key with the obscure, non repetitive algorithm of your choice (a simple XOR will be enough) and you can start having pretty impressive figures...

Generating gigabytes of data is quite easy, an I agree with you that making sure they are truly random is much more difficult, but it's not really that impossible 8)

The real problem is making sure both people get the SAME timepad, the problem here being to have both people REALLY in synch when startng the capture to generate the pad...

which is another problem entirely...

There also is/was:

• KAI C++
• Compaq's Alpha compiler
• SGI pro64

KAI C++ is basically dead (Intel bought them a couple years back, which sucks, as KAI was about the best C++ compiler I have seen), and I'm not totally sure about Compaq's Alpha compiler, but it looks like the SGI Pro64 compiler is still alive (and GPL'ed now... interesting).

Xsan from Apple is pretty much StorNext for OS X, only the licenses are a lot cheaper.

Worth mentioning is SGI CXFS, which runs on most popular platforms, although I belive some like OS X are still in beta. Licensing is expensive ($2500+ per node), the setup is complicated and you only get support if you're running SGI storage systems.

Also from IBM is SANergy. This runs alongside your deployed filesharing protocols like NFS & SMB, and "fuses" the actual data transfer over your SAN. Licensing is expensive, and from what I've heard development has been discontinued.

There's also Melio but that's only for Windows and looks like it's going to be version 1.0 forever.

http://www.sgi.com/company_info/newsroom/press_rel eases/2004/june/fleet_numerical.html

CRTs also have other advantages, of higher viewing angle than LCDs and higher color gamut and larger dynamic range. I still prefer using a LaCie Electron Blue CRT to do my Photoshop work. Most video cards RAMDACs are limited to 8-bit (exceptions I can think of are SGI Onyx IR and Tezro systems and other specialized gfx cards), CRTs could easily do 10-bit, 12-bit color per pixel. If you're in the print and publishing industry you'll still want to use a CRT unless the 21" CRT hulk is crowding your workspace, or a radiologist who needs 10-bit or greater grayscale to discern which legions in the x-ray have a tumor. I don't want my radiologist looking at no 8-bit (256 shade) video card and monitor for my X-rays. I haven't seen any specs on what the OLEDs can do. All that stated is that its comparable, which is vague and ambiguous to me.

SIGGRAPH 2004, LOS ANGELES (August 10, 2004)--The latest version of the OpenGL® specification, incorporating support for the OpenGL Shading Language application programming interfaces (API), was announced today by Silicon Graphics and the OpenGL Architecture Review Board (ARB) at the SIGGRAPH 2004 industry tradeshow. One of the most important and enduring standards in the computer industry, OpenGL® 2.0 presents a revolution in graphics by providing high-level access to the programmable features of modern graphics processors and is an important step in creating photo-realistic, real-time 3D graphics.

"Open Inventor support many of the thoings I want, but still, it's not part of the specification. That means that on the enterprise customers machines that run FreeBSD, Solaris and Windows there is no truly portable way to run things."

No portable way to run things? Why not? Inventor is open-source. Just download, compile and run/use it. Problem solved.

Yes, Opterons are sweeter when it comes to routing. Here's excerpt from commentary on Linux netdev:

We can certainly do better than that on Opterons. Robert
reports a 1.3 Mpps rate on a dual opteron 1.6Ghz. Our numbers on Xeons
are less than 1Mpps.

I suppose the inventors of computer graphics www.es.com and OpenGL had nothing to do with it.

Doom1 was an excellent 2.5D game when the big iron was already doing full-scene antialiased 3D.

Hey don't forget about these, even though it seems most people have! :(

A better one is here where Lava lamps are used to generate true random bits.

Too bad the website for it appears to be off line. SGI used to be cool, too...

don't forget the SGI Altix range - scaling to 256 processors in a single image

and there was the recent PR about the contract with Nasa for a 10,240 processor server (though that's not 10,240 in a single image...)

don't forget the SGI Altix range - scaling to 256 processors in a single image

and there was the recent PR about the contract with Nasa for a 10,240 processor server (though that's not 10,240 in a single image...)

You could look to SGI. Their Altix range is up to 1024 Itanium 2 processors in a single supercomputer, and they are putting 20 512 * processor nodes together in a cluster of linux supercomputers for NASA while also working on doubling up the maximum single machine cpu count to 2048.

As always, the answer is "it depends."

SGI Triples Backup and Restore Record, Scalable Technology Protects Large Data Environments

This is for tape storage backup, and 10.1TB backup to tape isn't too bad. I know for connecting to data storage SGI uses 2Gb fibre channel. But, other than that, I'm kinda clueless about data storage centers, in general.

Because it's a 20-nodes cluster, each node being a supercomputer with 512 CPUs.

The article was written, unfortunately, by a rather clueless journalist. Here's a link to the proper information:

http://www.sgi.com/newsroom/press_releases/2004/ju ly/supercomputing_ctr.html

http://www.sgi.com/newsroom/press_releases/2004/ju ly/supercomputing_ctr.html

Rus

The MIPS/IRIX systems have VPro graphics, yes. But those are not from NVIDIA. VPro for MIPS/IRIX is the last chipset to be developed inhouse.

The confusion comes from the fact that Sgi marketing thought it would be a good idea to give both the PC and Irix graphics cards the same brand: VPro.

They currently don't have anything newer for their workstations, but their newest Onyx (visualization system) computers use a couple of Ati cards for their graphics. It's called the UltimateVision.

>>Of course Nvidia's card is going to do better. Doom3 has a specialized codepath for nvidia hardware, while the ATI card does not.
>>If a codepath were written for the X800 series of cards, I'm sure the scores would be closer to each other.

The ATI card doesn't have a special codepath. NVidia exposes their advanced functionality through the NV_vertex_program2 and GL_NV_fragment_program2 extensions. ATI doesn't have any special extensions, they use the GL_ARB_vertex_shader and GL_ARB_fragment_shader extensions(which NVidia also supports, at a lesser speed).

OpenGL extension registry

IA64's dead? Better get on the phone to SGI and let them know!

Maybe you only have a 2D GUI.

It's a UNIX system! I know this!

Yes, I know it's still a 2D display. It's supposed to be funny.

Indeed. Someone correct me if I'm wrong, but the IBM zSeries with Linux can easiliy run into the millions. And don't forget these guys.

Remember, IRIX was the first OS to scale a single Unix OS image across 512 CPUs. And now they've eclipsed that, with Linux.

Just to clear things up:
SGI's first 1024p single-image supercomputer was an Origin 3800 running a customized "XXL" IRIX 6.5 kernel. This was in August 2002, almost 2 years ago.
http://www.sgi.com/newsroom/press_releases/2002/au gust/nasa.html

The point here is that if performance continues to grow like it is today, they will be selling these machines for $1,000 at Walmart in just 14 years. It will be about the same size as the computer you own now.

The problem with 1024CPU is much more then just the operating system. It is a mess of communication hardware needed to wire everything together. It is about special power feeds and air conditioning, and sometimes floor loading requirements.

Take a quick look at the end of this PDF. It talks about heat output and the need for 3 phase 240V power coming into this computer. It is not unusual to hire both an electricial and a cooling expert when you talk about installing one of these babies. Not for the Home user, and never will be, however, idential compute power comming in just 14 years, so get ready...

The link to the press release as of July 14.

CC.

The ATI Rage M6 chipset is underpowered, and is unlikely to be supported by Quartz Extreme, much less Core Image. Apparently, it's a Rage 128 derivative. found in the 550 and 667 MHz Powerbook G4s. Quartz Extreme relies on hardware support for GL_EXT_texture_rectangle, which Apple notes is only supported by "Radeon, Radeon Mobility, Radeon 7500 Mobility, Radeon 8500, Radeon 9000, Radeon 9200, Radeon 9600, Radeon 9700, Radeon 9800,GeForce 2MX, GeForce 4MX, GeForce 3, GeForce 4Ti, GeForce FX." Presumably, you can verify this for yourself by running a GLInfo program.

Apparently, it's less flexible than the proposed GL_ARB_texture_non_power_of_two.

Core Image is much more sophisticated, and levergaes the vertex and pixel fragment capabilities of an current generation OpenGL card.

Apparently, they offer DMF 2.8 for their Linux-based Altix line of servers, which aren't so horrible in and of themselves.

Apparently, they offer DMF 2.8 for their Linux-based Altix line of servers, which aren't so horrible in and of themselves.

I've read most of the nVidia discussions, and I've never heard of that, so lose the tone. Also, earlier today I stumbled across this in the Mesa FAQ; apparently in 1999 SGI opened up a sample OpenGL implementation. (Not that I'm sure this is the same thing you're talking about.)

This is a really nice ability, and I'm glad someone has done it.

SGI had done a dual user configuration.

I think this has already been reported on /. around November 2003, about the time sgi introduced this thing. They also sell the proper graphics hardware to attach 10 of those to a single box. No need to use a wimpy PCI based Parhelia.

--Troy

In a typical workload, there may be a lot of small files open in a fast succession. Each time the file open's there are locking exchanges. It is interesting to see that they ( like XFS from SGI ) are positioning this for HPC applications where the files tend to be large and tend to stay open for a relatively long time. This indicates that they're not content with the performance of their locking mechanisms.

I know this!" Oh, other Jurassic Park references?

(and before anyone replies, i know that the 3d file manager for irix actually does exist...)

I meh at you VNC-using neophytes. Real men xmove.

There are also some nice pictures of Huffman's origami here. The pictures also show Huffman himself doing the folding.

If you could rebuild everything from the ground up (and had tons of money to throw at it), you'd most likely want to build a system based on a very expensive vendor solution.

Assuming that you can't do that, your best bet would be to go with some sort of parallel filesystem, the likes of Lustre, GFS, Ibrix, GPFS or CxFS. The architectures of these vary, but the basic principle they share is performance scalability based on increasing the number of data paths to the disk. So if you have, say, 100 nodes on a high-speed network, you take 10 of them and attach them to your SAN. The parallel filesystem spans the entire SAN and so requests from the nodes can reach any bit on the SAN from any of the ten paths. If you need more performance, you add more paths: controllers, HBAs and storage nodes. I know GPFS scales linearly in performance based on the number of paths to the data, and I believe the others scale well also.

I haven't hit 50 TB on disk (I have on tape, but your post suggests that tape wouldn't give you the performance you need), but I have set up several 4-8 TB GPFS filesystems that could easily grow to 50 TB if I had the spindles.

Good luck finding a solution; symlink-based solutions on a convnentional *NIX filesystem are a nightmare; I sympathize.

I really miss the custom attributes that were such a unique feature of BeOS - I don't believe any other OS has implemented such a scheme.

Some have. XFS on IRIX and Linux can too, but on Linux support for those attributes suck.

Eh? I think you mean Ariana Richards. Also, the 3D file navigator used in Jurassic Park actually exists.