Forgot your password?
typodupeerror
Linux Software

Quad G4 Boards 61

Posted by CmdrTaco
from the yum-yum-gimme-some dept.
su-geek writes " Synergy Microsystems offers QUAD G4 boards as well as QUAD G3 boards which all run SMP linux. Not only that but you can add more than one board to the dual PCI backplane." I didn't find any evidence of running Linux, and they don't appear to be motherboards: they look like addons, but still, they look nifty. I want a couple CPUs dedicated to running OpenGL screensavers in my root window or something.
This discussion has been archived. No new comments can be posted.

Quad G4 Boards

Comments Filter:
  • Quad G4.
    If you've got Quad G4's @ 500 mhz and some reasonable cache on them, that makes for one really freakin fast CPU setup.

    No more waiting for the Gimp.

    Of course all of this is moot if Quake isn't tweaked to take advantage of it...

    z
  • by NevDull (170554) on Saturday April 22, 2000 @07:13AM (#1116955) Homepage Journal
    For those of you who want to know more about how to go about putting together a working system with Linux and VMEbus components, see www.vmelinux.org [vmelinux.org].

    From their site:
    This project's primary offering is the Kernel Level Linux Device Driver that interfaces between the Unix Shell environment and the VMEbus. The driver is compatible with the Tundra Universe PCI-VME bridge integrated circuit.
  • I enjoy experimenting with running Linux on many different platforms. So far, I'm runnning Linux on x86, SPARC, and MIPS. I haven't tried Linux/PowerPC yet because the machines from Apple and IBM are exorbitantly priced.

    Are there any options for those of us who are interested in PowerPC *without* running MacOS or AIX?
  • by fbrehm (136312) on Saturday April 22, 2000 @07:14AM (#1116957)
    Those things don't look like motherboards because they aren't. The VME and CompactPCI bus are industrial grade busses. The Synergy Micro boards are "Single Board Computers" that plug into a bus and control the boards that plug into other slots of the bus. These other boards can be anything from digital I/O for controlling machines, to video interfaces for machine vision applications. These systems go into applications for which your desktop PC (or Mac) would quickly fail. For example, my company is bidding on a project with a requirement that the boards in a CompactPCI (or is it VME, I forgot) withstand accelerations of 20G AND be able to survive being cleaned with a fire hose!
  • It's about time this happens. With constant yield problems due to an unreliable fab process this has been delayed way too long. AppleInsider.com [appleinsider.com] has an article about this from October. In any case one has to wonder what Apple's product strategy with these things will be. Are they trying to make a push into the high end server market?
  • by Anonymous Coward
    It's a sad fact, but many many games today are being cruelly mistreated. Many CPU's are available to them, but thanks to the actions of many programmers, these games have to go without the many benifits that come from SMP.

    What can you do? Support PMesa! With PMesa your games will render faster, as they will take adantage of all the CPU power you have! No need to rag on game developers. Here's an example of a game that isn't SMP ready:

    Q2 itself may not be tweaked for SMP but if a PMesa enabled graphics driver is rendering, the rendering process is distributed! Sure, all the Q2 data itself still stays on the first CPU, but at least the REAL hardcore CPU hog, the rendering, is going out to those other CPU's.

    (: This strange post brought to you by too much reading of the Utah-GLX development lists and all the SMP posts on /. recently!
  • The G4 is not considered a supercomputer. I have no idea where you get that from; the current High Performance Computing export restriction thresholds are considerably higher the sustained MTOPS that the G4 can acheive. I recommend checking out the Cox Report and "High-Performance Computing, National Security Applications, and Export Control Policy at the Close of the 20 th Century" (Goodman, Wolcott, Homer) for more information. Not that this matters much, considering that every nuclear weapon currently in service was designed on a computer with less power than the average desktop Pentium 2, Athlon, or G4...
  • i can't imagine how fast that is, those have to rip, maybe i'll get quad g4's with a gig of ram and oc3 lines for my server, now if i only had over 10 grand...

  • Can these boards be bought from anywhere in the UK ? It is not feasible to buy them from America as the import taxes would be massive. I know, I have been caught before :(

    I tried accessing the supplied link from Slashdot but cannot get through.

    Can I get these boards for other chips such as Athlon or Spitfire (when it comes out)./p

  • by Anonymous Coward
    Look for an older Mac from a used hardware resller or an auction house (i.e. eBay). If you decide to pick up a Mac make sure it's a PPC box with a PCI bus.. The 6100's, and a couple other models, use nubus rather than PCI which will force you to use MkLinux.

    There are other PowerPC machines either available or under development. Check out this link [applefritter.com] for a few of the POP efforts out there...

    Personally, I've been looking at the quad 604e/G3/G4 PCI cards from Total Impact [totalimpact.com] for a while now. They're expensive! However, they aren't single board computers. They sit in a PCI slot in your Mac/PC and act as additional processors. Total Impact announced a LinuxPPC based server box [totalimpact.com] (they don't have much info on their site about it) using up to 13 internal G4 processors and scalable using an additional backplane...
  • These systems go into applications for which your desktop PC (or Mac) would quickly fail.

    ... but these devices have their limits too. Hardened against EMP, yes, but they are still unable to withstand Slashdotting... check it out.

  • heheheh, I got a feeling that's not your room.
  • Well... actually, I just cleaned it yesterday, so it's pretty clean, actually. Now, I *could* use a firehose for my car..

  • by Anonymous Coward
    There is IMHO a much better system for VME accesses under Linux at ftp:://vlab1.iram.es/pub/, this driver allows you to write board specific device drivers which can be loaded as kernel modules and handle VME interrupts in the kernel, queue DMA operations, etc. Several research institutes are already using it, some for fairly complex data acquisition tasks.

    Note that there are other manufacturers who have announced quad G4 VME or CompactPCI boards: CETIA, Force Computers, and perhaps others.

    VME and CompactPCI boards are more expensive but hey, you pay for the high reliability and hot plugging capability given by the indirect connectors, the fact that you can put many more boards on a single backplane (9 for CompactPCI, 21 for VME), the capability to operate in severe environments (whether temperature, vibration, acceleration, orientation, dust, etc), in some cases the warranty (5 years for Motorola VME boards), and many other details which are not required in desktop systems whose main design goal is planned obsolescence.
  • I know where he got it: Apple ads.
  • I think I read a story about something like that on /. a while back-i don't remember precicely, since it was a while ago and i haven't slept in 31 hours. I think it was something about passing one gigaflops or somethink...
  • Sorry: couldn't resist. ;) I do however think that getting decent computers that you can run outside in less than decent conditions is going to be a Big Thing. Like where I live, relative humidity rarely dips below 90%, salt spray is ubiquitous, and electronic componentss in non-air-conditioned areas last about, oh, a year if you use'm every day to dry them out. I'd want a nitrogen-packed, o-ring sealed computer even if I didn't take it outside. Good luck with the car.
  • I want a couple CPUs dedicated to running OpenGL screensavers in my root window or something.

    Why don't you just get a decent video card, rob? GLGears runs great in my root window on a Celeron 500 with an ancient millineum II 4MB card.

    moderate down for dissin' rob
    --

  • You betcha! New Macs are painfully high end, but there is a huge used and refurb market. Try these links: http://www.macresq.com/product/desktops.html http://www.lworld.com/prod02.htm
  • In any case one has to wonder what Apple's product strategy with these things will be?

    This is a Synergy Microsystems product, not an Apple product, so I doubt that Apple has any product strategy with these things. Apple has some dual proc machines in the works for use with OS X however. They might try to push into the high-end server market, or at least the high-end graphics market with those.
  • Q2 itself may not be tweaked for SMP but if a PMesa enabled graphics driver is rendering, the rendering process is distributed! Sure, all the Q2 data itself still stays on the first CPU, but at least the REAL hardcore CPU hog, the rendering, is going out to those other CPU's.

    Only if you do software rendering. Otherwise 90% of the rendering is done on the graphics card. You might say geometry calculations need to be moved there, but that will only help minimally, and that is being moved to the graphics card also.

    SMP will not be a commonplace feature. Not everyone will use it, so the game companies will not take much advantage of it. Without game support, SMP is useless, and SMP based PMesa drivers are just as useless.


  • They list a collection of so-called "firsts" for them and the pertinent list is below. Also, additional or upgrade CPUs being accessed via PCI or another bus is far from uncommon. 1st Gen PowerMacs can get G3's through the PDS. Others get it in the NuBUS slot!! Amiga's take it like a man in their ZORRO slots (Zorro The Gay Blade?)to get fast PowerPC CPUs. Don't see why these can't work.

    Tyler

    Multiprocessor CPU Boards
    Quad and dual PowerPC-VME,
    dual CompactPCI and
    dual conduction-cooled
    P0-PCI
    Fast, secondary data bus
    Linux
    Linux running on VME &
    CompactPCI PowerPC
  • For the record, all x100 series PPC's are NuBus (ie, 6100, 7100, 8100) NuBus seemed like a good idea at the time, but in retrospect as one of those bad decisions that apple feels it needs to make to balance those good decisions... MkLinux will run on NuBus machines btw.

    The important thing to note here, though, is that the 601e was really just a toe-wetter by AIM and when you get right down to it, it's performance vs. the MC68040 is nothing to email home about. You can get a comparable pre-PPC Mac with a 68040 for a lot less money and only a little less power.... Mind you, if Linux is your goal you'll be up the creek. The solution there is to move to Open BSD, which claims to support "all" the MC680x0 family of processors. That's a wee bit of a lie, though, since the MC680LC0 is not only unsupported but fully un-functional. You'll only find that chip in "higher end" LC macs (ie the LC630).

    If you want the good on older Macs, you should check out lowendmac.com Don't trust me to get the URL right, that's what google is there for.
  • by Anonymous Coward
    It's working perfectly. We compared it with a dual pIII-600 (classic) at work for a an application we wrote and the quad g3 did better. I have a K7-700 on my desk and I want to bench the application on it as well. The application could be considered a general-purpose application with a lot of FPU processing going on for Fourrier transforms. It hasn't been optimized for any platform yet. The system ended up costing in the very low 5 digits. I don't remember if it was in U$ or CND$ though. I wish I had a good benchmark program because I would bench the quad g3 and some other fun machines as well. By the way the board is incredibly small (VME).
  • Actually, Carmack did a hand coded assembler version of the Q1 or Q2 renderer for Dual pentium machines. IIRC the best he could get was between 3% and 10% improvement over 1 CPU. It was due to the braindead way the Pentium CPUs did cache invalidation or something.

    The other comment about most rendering being done on the video card is well taken. It doesn't make much sense to start writing SMP software renderers on the dawn of every PC having a T&L card. Games _should_ be written with SMP in mind, but you'd effectively only want one CPU doing gfx. Even on massive SGI's you tend to only want one thread working with a given rendering context, for context switching reasons.

    Additional CPUs should be used for better responsiveness. The Sega Saturn machine used dual primary CPUs; this was in addition to its graphics and audio processors (it had like 8 processors total). It was said that in some fighting games, you'd assign the AI of one character to one CPU, and the other cpu would do everything else.

    I'm not sure what the best way to utilize SMP in a gaming environment would be, but I can think of some scenarios. A dual processor configuration could be used sort of like an application-specific read ahead cache. Let one CPU be running the game, and when it realizes that it needs to spool a new world tile off of disk soon, it sends a message to the other CPU telling it to do it and to get everything set up. If you had enough ram, and the game was written properly, it'd be like "double buffering" but with game levels. At the end of a level you'd just move the "current level" pointer :) It makes even more sense for something like Asheron's call or a game that has a contiguous world environment. Obviously the whole world can't fit into ram at once, but you could use a smart algorithm to decide when the the other CPU should page sections of world in and out of main memory, for seamless access by the other CPU.

    Of course, most game developers are buying engines from other companies, and just cranking out different textures and FMV sequences. I wouldn't expect a whole lot of game innovation any time soon, save for from a couple of well known companies :)
  • I have an IBM PC clone. I believe it is nigh-invulnerable and could withstand a fire hose.

  • Four G3's outperform dual PIII 600's? I guess thats cool, but a dual PIII 600 is probably less than 3 grand. It sounds like G4's would be better suited for what you are doing since you mentioned that your program is FPU intensive. If you can get libraries optimized for G4's (AltiVec) it will be like having four 1000 mHz pentiums. Nothing that sits on a desktop will be able to touch it.

    -b
  • We should all know by now that anything not related to linux on slashdot is offtopic...
  • The Total Impact boards are SBCs; they just happen to plug into a host machine. Thus if you put 3 of those cards in a Power Mac, you do not have a 13-processor Mac; you have a 13-processor cluster made of 4 separate machines. (Anyone remember the Radius Rocket?)

    Some apps love clusters, but most apps don't get any benefit.
  • I dunno about these days, but last time I checked to support a VME backplane, you're looking at like 100-200 grand ... these are for high-demand industrial applications. Of course I want one anyway :)

    This stuff is usually MilSpec, i.e. can deal with extremely harsh environments and keep functioning. They used to blow up one (or two) of these for testing nukes and other stupid stuff like that.
  • Bad marketing practices (eg, convincing consumers that Mhz==horsepower) have created a topheavy PC architecture in terms of processors being so much faster than RAM its getting ridiculous fast (at least that's what I hear).

    With chipsets surpassing the Ghz, RAM is becoming pathetically slow. Intel thinks RDRAM is the answer, but its ridiculously expensive, and I hear conflicting views as to whether it's even worth the sky high price. Is *DRAM getting so old that it might fall out of the picture? Maybe some mobo mfgr will offer a mobo that accept up to 128 megs of off die L2? or something equally silly sounding today, but neccesary tommorrow? What I'm really the most scared of is that AMD will miss the fast *DRAM boat entirely, and Intel will clutch the high end market for another ten years, pulling in ridiculous prices because there's no viable alternative...

    And anyway, WTF are the x86 manufacturers doing in the god damn high end server market??? Isn't that reserved for risc machines? I mean, if I was going to spend 10 big ones on a high end server I'd buy something that was designed to rip up the job, not play f'ing Doom. Seriously. Shouldn't anyone with delusions of building a server put their money into a machine designed to do work not play games? And dont remind me that x86's are "universal computers" too, you know what I mean.

    If you need to run windows so you can manage the sucker, tack on an extra $300 and get a damn k6-2.

    Ok rant's almost over...

    What really pisses me off is that it seems like the whole IT world is losing IQ through a hole that could admit a sperm whale. x86's are for Personal Computers. RISC's are for serving data, compile farms, and real work in general. An x86 might be a viable stand in if you're strapped for cash in the short term, but if you're talking about spending huge sums of money, its just a waste to buy a 600mhz quad Xeon with a zillion bells and whistles.

    Isn't it?

    Am I wrong? or has the whole world gone mad?

    "A witty saying proves nothing." -Voltaire

  • I know what I would do!!!

    DISTRIBUTED.NET!!!!

    Just imagine (4) G4 chips cracking away at RC5-64!!

    We are talking like 15+Mkeys/sec here!!! Now that would rock some serious ass!!!
  • How in the *bleep* does the above post rate being moderated as a troll?!?!?
  • Ah - the famous 'moderator smoking crack syndrome' strikes again.
  • by RISCy Business (27981) on Saturday April 22, 2000 @01:16PM (#1116988) Homepage
    Okay, first off, what is it?

    This isn't a motherboard, it's not a PC, it's not some piece of trash laptop component. This is a very very powerful VME board. Industrial computing only - no, you can't afford the chassis to hold it. It's usually in the $20k+ range for a basic VME chassis able to handle this type of board, assuming I'm looking at my current pricing sheet. (I don't think it is, but it can't be that much cheaper.)

    Now, it says it'll run VxWorks. Linux isn't on there. And I wouldn't waste my time - Linux won't work on it. Unless somebody's been REAL busy with the PPC tree, FORGET IT. PERIOD. END OF ARGUMENT HERE. Don't bother flaming me, I'll be more than happy to just delete it. LINUX DOES NOT RUN CORRECTLY ON SMP POWERPC AND HAS NOT FOR MANY MANY MANY MANY MANY MANY VERSIONS. For a VERY long time, it wouldn't even run on two processors. And when it did, there was *exponential* performance degredation on multiple processor systems. And now you think that MAGICALLY you're going to use *FOUR* PowerPC 750's (NOT G3s) with Linux? If it wasn't Slashdot, I wouldn't believe the incompetence and idiocy of claiming to run Linux.

    And as if THAT weren't enough reasons, did we mention that LINUX DOES NOT SUPPORT VME? Last I checked, Linux doesn't even support CompactPCI! And somehow you're going to just run Linux on this? Give me a goddamn break. LynxOS won't even run on this, unless they've added VME support. That's why it says "VxWorks" and not "Linux" and not "Windows." The fanatics need to get their heads out of their rectums and realize that every time they make some blanket statement crap like this, they only make an ass of themselves.

    This board doesn't even have a controller - what, you're going to magically plug in a SCSI disk to the chipset directly? The board boots and runs off of NVRAM Flash. This isn't some cracked-up PC. This is INDUSTRIAL equipment. How do I know these things? Probably because I have a CompactPCI (similar to VME somewhat) system built out for engine and onboard computer diagnostics and tuning. How much did it cost? I lease for a reason. What's it run? Not Linux, that's for damn sure. It runs LynxOS for a reason - not only because Linux isn't fit for the job (shut your mouths, zealots. I don't see any powerband analysis software for Linux any time soon.) but because LynxOS is *designed* for things like this.

    Guess I probably pissed a lot of people off. Good. Maybe those of you who were offended by my comments will shut up, get your heads out of the clouds, and come back to earth sometime this year. Linux is not the do-everything OS, and it never will be at this rate. With supporters that claim it can do anything without any proof to back it up, and claiming it incorrectly half the time, I don't see any profit for ANYONE from it anytime soon. RedHat's stock isn't trading at 26 for no reason at all.

    With supporters like this, I think I'll support FreeBSD instead. At least they have some respectability left.

    =RISCy Business
  • Hmmm, Quake III runs dual-proc. I seem to recall John Carmack's .plan claiming that there was up to a 50-70% speed up - mmmmmm :-)


    This would make sense. I remember when the Voodoo2 came out, benchmarks on lower spec PCs showed no speed up over the old Voodoo cards. The reason was, that the slower pentiums couldn't kick out data fast enough to keep up with the graphics cards.


    I also remember hearing that back in the days of Wolfenstein/Doom *half* of the processor time was spent in user input routines. I don't know if this is an exagerated figure, but both these facts would support the fact that splitting graphics rendering/world update into seperate threads will speed the game up.


    OTOH - you're probably right - at least in the short term it is unlikely to many games players will be owning SMP machines, so this will not have too big an impact.

  • uuh..linux can support upto 16 cpu's ..check the kernel archives..someone put a dmesg up there for 16 cpus.
    compactPCI/VME are usually real time...true..but linux IS supported on there - go to the page and you'll see linux support in bold letters.
    SCSI will work nicely on a VME board...and linux can survive on the flash.
    linux VME patches are also available.

  • oh my God.
    the following post is horribly offtopic, but here goes anyway:

    thank you. your post was like a breath of fresh air.
    especially on a topic that's slightly Apple-related, since such topics bring the flamebait masters out in droves.. i generally read anything on slashdot connected to Apple even tangenitally not expecting to find a single comment attatched that is the least bit coherent, accurate or informed.
    Your post was all three of these, and even better didn't mention apple! Apple really doesn't have a whole lot to do with this article and not a lot of posts in this thread seem to be aware of that.

    Again, thank you. it's gotten to be very rare that i feel like a better/more informed person after reading a slashdot post, but your post did it to me. Ignore that second reply. :P
  • Last I checked, Linux doesn't even
    support CompactPCI!

    Excuse me? I have with my very own hands brought Linux up on a Ziatech compact PCI board, thankyouverymuch! The only difference between standard PCI and cPCI is the layout of the bus: PCI uses an edge-card connection, while cPCI uses a 96 pin connector with controlled impedance. From the software perspective, there is no difference! Furthurmore, if you look over at Ziatech's home page [ziatech.com] you will see they list Linux quite prominantly.


    Now, as for your comments about LinuxPPC, I cannot comment, as I have no experience there. But you are quite incorrect about cPCI. /. readers, take note and weigh the rest apropriately.

  • Some guys from Canada tried to buy a G4 from my school computer store (University at Buffalo) but they were denied because the G4 is considered a super computer and could not be exported. Perhaps the people at the store were mistaken but it is still interesting.
  • The update to OS 9 (9.0.4) was a long time coming, mostly because of hardware issues, according to various articles at MacOS Rumors [macosrumors.com]. One of the issues was support for new "Mystic" machines, multiprocessor G4 configurations from Apple available possibly as early as this July at MacWorld, according to this article [appleinsider.com] at AppleInsider [appleinsider.com].

    This looks awesome, but I'm worried about system overhead with multiprocessing. If the BSD kernel can offer true SMP support in MacOS X, then I'm all for it, but a multiprocessor G4 seems like a waste under OS 9.0.4.

  • Sorry to see that another moderator had to waste a point undoing some other moderator's moment of mental abberation, glad to see that that same other moderator stepped up to fix the problem and restore seebs karma.
  • I beg to differ on the point "LINUX DOES NOT SUPPORT VME." I'm currently in the process of putting together a data acquisition system for a particle physics experiment. For our test run this summer, we will be running Linux on a 200 MHz Pentium-based VME single-board computer. I can tell you from close firsthand experience that (with an appropriate driver for the VME chipset) LINUX SUPPORTS VME. The VMEbus address space appears as a device file on which you can use all of the usual I/O system calls, including read(), write(), and mmap(). There are two driver implementations out there; my experience is with this one [uni-bonn.de].

    Why use Linux instead of the more conventional vxWorks? As usual, the reason is reliability. vxWorks provides no memory protection between processes. Programming errors quickly lead to a need to reboot, requiring several minutes of experimental downtime. When your accelerator costs tens of thousands of dollars per hour to operate, any downtime becomes inexcusable. In addition, the ready availability of excellent development and debugging tool makes it possible to put together a project like this with limited manpower.

    It is possible to pay a wide variety of prices for a VME crate. A 6U crate with a small power supply should only cost you a few thousand dollars.

  • Are they trying to make a push into the high end server market?

    Well, they do make out quite nicely with WebObjects. Why not make a WebObjects box, ala Oracle's raw iron?

    - Scott
    ------
    Scott Stevenson
  • I'll also point you to this page on the manufacturer's web site [synergymicro.com] which specifically says that they support Linux on their boards. The page has a pretty cool "industrial penguin" logo...

    Admittedly, they do not claim that SMP or real-time features will be ready until Q3 2000, but that's only a few months away.

  • You ain't kiddin. We are looking at quad G4 boards for a project we are starting up now. A fully populated quad G4 VME board with 256MB pushes $60k. Ouch.
  • by mindstrm (20013)
    Put foot back in mouth. Be quiet.
    As many have pointed out, both cPCI and VME are supported in linux. The manufacturer of the board in question even says it works.

    Wow. Glad to know there are lots of 'experts' out there to educate everyone.
  • Well, if you want to compare NuBus to ISA, sure, I'd take a NuBus any day. The big difficulty was with Apple's absolutely *terrible* sense of timing on their hardware development path. They make the move from 68k to PPc and then, like, 18 months later move to PCI. From '94 to '96 it seemed that every piece of hardware I bough became obsolete before I got it out of the static bag. I think that qualifies as a "crappy decision"

    Don't get me wrong: I'm not an "Apple Basher". I've got 4 macs at home (including one of the afformentioned LC630's) and I admin a network of 200 macs ranging from Quadra 610s to G4/450's. I like the hardware, I like the software... I just think that the "hardware revolution" of the mid-nineties was not as smooth as it could have been and it caused a lot of unneccessary (sp?) grief.
  • maybe, the test box they had running did run simple text, but after 2 typing sentences It ran out of memory and segfaulted ;-)
    ---CONFLICT!!---
  • The G4 chip was supposed to be multi coreson one chip, but they're not.
    IBM has a dual core chip working, the Power3 or something like that. The 'G' chips are Apple's names for the Motorola versions of the IBM Power line, so the G5 when it comes out should be dual or maybe quad core.
    ---CONFLICT!!---
  • I think it can be something of an over simplification to say that x86 should only be used for one thing and RISC for another. Since games require lots of floating point computations, using an x86 based machines can provide a cost effective alternative. It may be possible to buy several SMP x86 machines instead of 1 RISC (depending on specification of course) these machines could then be put to work in a cluster. They also offer increased redundancy. If the RISC machines falls over and dies a lot more is lost than if one of the x86's dies.That is just my opinion though. Yours may differ.

  • You could try Darwin (a BSD variant), MkLinux, or if you have a non-G3 PCI PowerPC there's also BeOS - they haven't yet bothered to get it working on G3/G4 cpu's yet, I guess they're not as 'caring' as the linuxppc people.
  • Not necessarily up the creek if you've got a 680x0, do a search for MacMinix...
  • GLGears runs great in my root window on a Celeron 500 with an ancient millineum II 4MB card.

    Indeed. In fact, on decent hardware (read: non-Intel), you don't even need that. I used to run the slinky/staircase xlockmore mode whose name I can't remember in my root window. And that was on a HyperSparc 150. It was fine in wireframe, but a bit too slow to be useful when fully shaded. That was more due to memory starvation than anything else, though (I was running a dual headed X server, which chewed up a lot of RAM).

  • Actually hate to burst your bubble, but beowulf is one of the 'lower end' clustering technologies...but it's the cheapest/easiest/best supported...
  • I agree with the whole 'Linux isn't the do all OS', but I disagree on it not getting close. Think of it. Linux has made more progress in it's exsistance than Microsoft has, and they had the head start. Part of what makes Linux great is those zealots. Those zealots are what make it a do everything OS(or at least close). When they need something, they program it, then they share it. Which is better than I can say for some high and mighty 'my-OS-is-more-expensive/craptacular/unsupported than yours, therefore my kung foo is superior to yours' sysadmins\suitgeeks. Well that was an unintented flame-o-rama.
  • No, the zealots usually aren't coders. They are users who don't right code, don't understand the concept of Free Software (or the less philosophical open source), and don't care.

    They decided that Linux was cool, and that they are better people for using it. They realized that /. gives Karma for championing Linux at all costs, and post accordingly. Also, many of them are young (high school), and therefore have not been exposed to industrial equipment.

    Like the sig, most slashdot readers think that the entire Internet could run on beige x86s, these people approach computing from a very narrow view.

    They assume that because Linux is more stable than Win98, it is the most stable system on the planet. Also, because they don't crash it, they assume that it doesn't crash. They don't appreciate the value of real time OSes, and other things in the field. While people are working and hacking away at a real time Linux, the zealots don't care and champion Linux for real time work anyways. They compare a Beowulf cluster to an actual supercomputer (while it is a cheap way to get LOTS of processing, it isn't a supercomputer... it doesn't do what supercomputers do...), assume that Linux can replace S/390s, etc., etc.

    Linux is getting hurt by the zealots, because they turn off potential users that are interested and could become interested in free software by criticizing all that disagree.

    Further, just to alienate some people, I'm going to disagree with Linux being where it is because the zealots hacked code. Linux is where it is by an even MORE amazing set of circumstances than those that led Microsoft to it's position. Linux is here because RMS and company were finishing up the GNU tools around the time that Linus's kernel started to function. The combination formed an OS, and as the kernel got polished, the tools became useful. A few other projects worked, and Linux gained in popularity because of the GNU System and other free software.

    Linux did NOT get anywhere by obnoxious teenagers insulting others for running another OS.

    Alex
  • You have WAY too much time on your hands! :)
  • What was I saying about those admins?? Listen, not all of us can afford your supercompters and industrial prts and what not. And while some zealots aren't coders (the more obnious ones aren't), some are. I'd take 15000 annoying Linux zealot before I'd take one sysadmin thinking he is king of the universe because he has an industrial grade board. Big whoop. The rest of us will be busying advancing humanity. And some of the other posters say it is Linux compatible. I don't know how true it is, but I'll trust that drivers could be written for it if it isn't...
  • And one more note about Linux and Karma... I'll have you know I found not one lick of karma for that last post and I don't care. The more people realize that karma is an artificail standard created by the administration so that more people will side with them on issues, the better.

For God's sake, stop researching for a while and begin to think!

Working...