Slashdot Mirror
FourHead: One PC, Four Users
LoganGD writes "A reseach group from UFPR university in Brazil, C3SL has managed to make one Linux box run four terminals at the same time. That means four mice, keyboards, displays and users with just one CPU. The way they managed to do that can be found at the FourHead project webpage. The fact that one computer science laboratory can suport up to 60 users whit only 15 PCs is really attractive for low-resource groups and countries."
How is this different from a low-end mainframe?
I wonder what, if any, heat issues would come from running 4 video cards, and the cpu at relatively full power, in a seemingly normal sized case. I would imagine it would get pretty toasty(or very loud with all the fans running, think of the poor power supply!)
I would be interested to see how they handle it!
Imagine a 4-user Beowulf Cluster of these...
Where's the economy in this? The "box" (cpu/ps/hdd) is relatively cheap compared to the display. I guess every dollar/peso/dinar counts.
Wasn't this what Unix (and/or its predecessor, Multics) was designed for?
~ Aero
..looks like a mainframe but smaller and running a good OS. If we're gonna go back to terminals then i'de stick to the monster tower in another room and have numerous terminals. But that wouldn't be /. worthy
Dear aunt, let's set so double the killer delete select all
This has been done a long time ago...i user/
http://cambuca.ldhs.cetuc.puc-rio.br/mult
But if one uses XGGI, its easy to get eight or more users on a single PC.
- A.C.
I would be interested to see how they handle it!
Since they are in Brazil, the easiest way would be to move to a colder country.
Just use a Matrox parhellia (or maybe a bunch of them to power 3/4 screens per cad :) )
IMO the TCO of a 4 head system is still greater then an LTSP deployment, if one 4 head machine goes down you loose 4 workstations, costing time and money, you can use antiquated technology with LTSP, have functional workstations for probably no more then 75 to 150 bucks and beable to loose a machine or two without a derastic impact on the system.
For terminal services server to work, you'd need one computer for the TS server, and four computers each with their own operating system for this to work. So to get the same result with Terminal Services, you'd need atleast 4 (one of the computers in use is also the terminal services server) and ideally 5 computers.
Ever since I discovered the combination of dual head displays and USB keyboards/mice, I've wanted to do this one a windows machine. For gaming it would be absolutely superb(especially with a multiproc machine). No network latency, no configuration hassles, just "hey, I've got a spare keyboard, mouse, and monitor, want to play UT2k3?"
:/
'Course, being windows, I didn't find a way, but this is really cool too. Too bad I have a very proprietary Cyberblade XP ai1 in this machine.
It's been a long time.
Hi there,
1975 called. They want their computer headlines back.
Best regards,
Chairboy
Also, you don't need 4 cards, You could do it with two. NVidia's twinview allows you to run two seperate X-servers off of one card (provided of course that it has two outputs).
Imagine a short time from now when a computer is so fast it can easily handle the needs of a workgroup. It would make much more sense resource-wise to share one system (with one huge pool of resources) than waste it all on the duplications found in multiple machines. I propose the term for this new form of computing to be: compupooling. :-)
Google doesn't do a particularly good translation of Portuguese computer jargon - it refers to Nvidia "plates", a "plate mother", PS/2 "doors", "slide bars" (I assume these are expansion slots) and the list goes on and on.
Karma: Segmentation fault (tried to dereference a null post)
Initially it is necessary to catch the code source of kernel 2.4.25. Makes it sound like an adventure.
-- The morphemes of your disquisition are ascertainable, but they have eschewed an ambit of transpicuous exposition.
Well this is great! Now I don't have to be bothered by the actual downloading for the porn and I can just view it instead of waiting for that cumbersome downloading times.
Person #1 Finds the good porn
Person #2 Exchanges information from Person #1 about where the good porn is and initates the download process.
Person #3 (Me) Views the porn.
Person #4 "Cleans" and maintains the workspace.
Yes! I listen to NYC Speedcore and do math at 3AM. I suggest you try it too.
Not only is this a really basic hack, but people have been using many different methods of sharing computing power almost as long as there have been computers. They could have made this a lot less expensive by just getting some X-terminals from Tektronix or one of the other companies off of Ebay. This is not even difficult on Windows if you buy Terminal Server.
Slow news day I guess.
\/\/oobie
Heh, there was once a time where I had three mice (only 1 cursor though) - I had a normal PS2 mouse plugged in, an old COM port (I think.. never been too sure what port it was) mouse and a wireless USB mouse all plugged in at the same time, and they could all control the cursor.
Hehe, it was good for playing tricks on my parents when they were sitting at the desk with the PS2 mouse and I'm sitting a few feet behind them with the handy wireless USB mouse.
*evillaugh*
It's not like the graphics cards are going to be pushed to the limit - I mean, the CPU wouldn't cope with them all playing ut2004 at the same time. :-) So they shouldn't be generating all that much heat, couple of case fans should cope with it no problems.
As for the noise, it'll still be quieter than 4 separate boxes.
I wonder how it takes care of media access. I mean, if someone put a cd in the drive, surely it becomes accessible to all 4 users? What if there is sensible data on that disc?
It would seem relatively easy (USB sound devices, or cheap PCI sound cards) to add sound to each workstation.
samrolken
I don't think anyone could argue that these create massive amounts of heat. What heat they do create can easily be exhausted by a case fan.
This is definitely a setup for an environment where people are literally running on a shoestring budget. This is a really nice ability, and I'm glad someone has done it.
Man, this is old news. We did this few years ago with five or more people on one machine. All we needed was a really small computer class, some free chairs, one PC, one keyboard, one mouse and one display.
Fastest (or strongest) got the best seats and the one with specs got the keyboard.
Talking about multi-tasking...
hell, just use some old s3's or something.
barely any heat...
world was created 5 seconds before this post as it is.
Never mind that: imagine the resource crunch using only one processor! It looks as though they have four X servers running off of that one poor little CPU. Now, from the translation:
The computers most recent supply to a processing power many times bigger that the demand of many users. Having this in sight, this project looks for to use to advantage better the power of these machines, allowing that the computacinal power of them is shared by some people.
Too much processing power? Well, it certainly could be true if the users were only using the computer for browsing the internet, checking email, word processing, etc., but if all four were doing graphics-intensive (UT2004) or CPU-intensive (manipulating large sets of files) tasks, the slowdown would be considerable.
Speaking of games, these brings a whole new meaning to the term "hotseat".
Without a proper flamewar, Anonymous was undecided on what shell to run.
Considering how we're all supposed to be moving to web based interfaces rather than rich clients, wouldn't the old dumb terminal method suit us better.
Perhaps a method were PCs, ACT like dumb terminals. They are some kind of 'Resource' or extra limb of one 'Computer', running unix?
Might do a hell of a lot to improve security and administration in the long run.
May the Maths Be with you!
did you read the article, or even the summary? there is a keyboard, mouse, and monitor for each user. /. readers are getting more and more lazy
Marge, get me your address book, 4 beers, and my conversation hat.
Comment removed based on user account deletion
I would love to do this, but with only 2 users instead of 4. Anyone know of a good guide that shows how the configuration of a second desktop would be done? I can see this as being an Xfree86 configuration nightmare.
adventure-today.com
So I guess four heads are better than one eh?
(ba-dum-ching!)
(ducks impending flame doom)
Only one of the 4 videocards can use AGP, so the other 3 must be (probably fanless) PCI cards, those don't generate too much heat.
Just think about 3D intensive games, where both top notch CPUs and videocards are running on full load. It produces much more heat but it still can be handled. (Though the newer videocards are getting loud.)
I was thinking, when I was using my KVM Switch, I noticed that you can reverse it's usage, instead of having 4 machines and 1 monitor/mouse. You can have 1 machine KVM'ed out to 4 displays, keyboard, and mice. Once done you could possible just use a terminal with vnc and use port forwarding to get x to work. Not sure it would work but just a thought... --[H]itman_Forhire
-- [H]itman_forhire
No, I won't pardon your ignorance, because you didn't RTFA.
If you want to know how it's a practical solution, maybe you could read the article. Or you don't even need to do that--just look at the picture of four monitors around one computer.
That means FOUR mice, keyboards, DISPLAYS and users with just one CPU.
Read the article and look at the pictures. They each get their own monitor, keyboard, and mouse.
They're not sharing one monitor, they each get their own monitor, keyboard, and mouse. The main server has 4 video cards each going to a monitor. This really isn't that neat though, as it's not a very new idea. So, you're main idea that this is a waste of time is still true.
Hurricane Ivan: A 17th century prison collapsed. All of the inmates escaped.
I'm going to make a quick guess and say that you didn't read the article. The system has four video cards and four monitors. Each workstation gets one monitor, one keyboard, one mouse.
-
Having 4 people sitting around 1 monitor (even a 19inch) all haveing their quarter screen....I can't see where this becomes productive.
You obviously didn't RTFA. It's one computer with 4 monitoes, mice ad keyboards.Haha!
/me slaps the schmucks who mod'd this interesting...
Er, wait. Is this a joke, or did you just not RTFA? Go read the article- there's a picture. It has nothing to do with splitting one monitor into quarters.
Working toward a usable PDA environment in the spirit of Newton OS: Dynapad
the point is that you can have four terminals while using a single computer as the host...
1 cpu and mobo, 4 graphics cards, 4 sets of kb/mice.
there's a commercial solution to do this in windows as well iirc, but just for 2 screens..
world was created 5 seconds before this post as it is.
When it all boils down, it ends up being a bit cheaper to just use terminals
Where can one buy inexpensive X terminals nowadays?
Ranging from:
'xterminals are cheaper' -- anyone care to back that up?
to:
'Windows terminal services can do this' -- don't know where to start on that one, suffice it to say: it can't.
to:
'This is just serial terminals' -- it isn't. RTFA.
I'm sure I missed a few...
"I assumed blithely that there were no elves out there in the darkness"
I have 12 terminals, Shift F1-F12!
Thank you everyone.
And I did in fact miss the whole point of the article, my bad.
Yo Grark
Canadian Bred with American Buttering
Its just damn plain cool.
True, but for the clients, you can pretty much use any crappy computer out there, since it's just a terminal, an input/output device doing no other processing of its own.
Microsoft has TS clients for even Win 3.1.
There's 10 types of people in this world, those who understand binary and those who don't.
Your ignorance could have easily been cured if you read the article and saw the pictures that show 4 separate monitors for each user. :P
I think you mean, "pictures that show 4 seperate monitors, one for each user", you way would have 16 monitors
I have one linux box. In it I have:
2400+ AMD Althon.
800+ Gigs of RAM.
Mpeg2-type TV capture card.
Bttv based TV capture card
2 video cards, a Geforce 4 420MX PCI card, Geforce FX 5900 agp card.
2 harddrives,]
1 CDROM burner, 1 DVD combo drive.
Built it my self.
1 intake fan, dual fan 350watt Powersupply (adjust fan speed automaticly to temp.)
I play ut2004 on it, I rip CD's on it. I encode mutliple MPeg2 streams on it. etc etc etc.
Heat isn't a problem. Everything runs comfortably, even when having both disk drives churning away and cpu pegged at 100% for extended times it isn't a problem.
The single fan in the front is a 120mm fan. Can get annoying at times, but it 's much quiter then most PC boxes that I've seen.
Combining intelligent but plain jane type case design with some tiny little mods it's easy to keep cases cool. As long as it isn't the big nasty messes that most people/companies clobber together.
Ok, so by the time you spend all of the time and effort to do all of this, you could deploy a fully self contained thin client in what, 10 minutes?
The only way I see this as a good idea for any low budget organization is if they get donated lots and lots of monitors, keyboards, mice and computers with graphics cards for this project.
800 Gigs of RAM...? Riiiiiiight...
Not to mention and X-Terminal is not the same as a Standard/Off-The-Shelf PC CRT. It's a special device with it's own video card and enough processing to drive the video card.
This solution uses standard computer peripherals for the "heads"( I/O ) and a standard video card to drive the display. The single GNU/LinuxPC is used to drive/power the 4 "heads" computing resources.
The difference here is also that fact that the "heads" must be local to the computing chasis. With X-Terminals, the "heads" can be anywhere on the network. So for close quarters computing, like a lab, this solution is a money saver. For remote computing, X-Terminals or Terminal Serving( LTSP or WinTerms ) are the preferred solution.
IMHO.
LoB
"Anyone who stands out in the middle of a road looks like roadkill to me." --Linus
They actually have some other nifty features too. Some of the Wyse ones I worked on previously had telnet support so when (not if) the Citrix server went down (oh yeah, they usually support Citrix too) our users could telnet straight from the thin client.
SBC stands for Stupid Bell Company
AT&T stands for All Telephones Tapped
This sounds like a multiplayer gamer configuration. Unlike most shared-CPU systems, everybody has a 3D video card, although they have to be PCI boards. With everybody on the same CPU, latency is a non-issue. Fast FPS games should synchronize perfectly. That tightly synchronized feel will make for much better head to head gameplay.
In brasil, X-Terminals are hard to find and more expensive than monitors and keyboards.
So what? It's still nothing like what these guys are doing!
In countries where computers/power are limited, every single box, no matter how slow it is will be used as a proper machine, not a bloody terminal.
You seem to be missing the point of what these guys are doing.
just the other day I was thinking how I could do something like this... all you need is each tty with its own $DISPLAY variable for X
Observations that "this has been done before" are really missing an important point, that it's being done in a new way. When there are hundreds of software solutions for everything, all for free, then there will cease to be a market for overpriced proprietary solutions. Not only that, but instead of thinking "where can I buy ___," the first thought to come to mind will be "where can I get this in Open Source."
Well, you can, but it's third party, very hardware specific and leaves you stuck with M$ XP. The Linux system demonstrated is a clear winner for schools, libraries, banks, casinos and other places where economical use of hardware is desired. While the Linux system might be difficult to maintain, it can be done. The Windoze solution leaves you dependent on the vendor. The people at Jetway have done an outstanding job but such is the world of proprietary software.
Anyone know of a better system? If you say citrix terminal services, I say, Linux Terminal Project and, cool as those things are, get thee back to 1975.
Friends don't help friends install M$ junk.
Seriously...
multiple comments that mention this was being done for years are all true - for Linux. Serial terminals, remote X access, VNC, network boot, diskless workstations, all that stuff depended on ancient (though still great) UNIX multi-user multi-session framework - multiple users work on the same machine at the same time.
And now Microsoft woke up.
After NEXT, GECOS and a couple others, PC has a GUI! Windows 1.0! Years after Amiga with real multitasking introduces Task Switching and later ('95) first Multitasking. Then the puny '98 "multi-user" (Amiga had that some 5 years earlier, UNIX machines way before that). And now, in 2004 we hear that after users of XP are tired of the pseudo-multisession of Switch User, SP2 is to include REAL MULTISESSIONING! Yeah, right! Two users can work on the same computer at the same time! Yay!
Noticed the catch? The keyword is "two". Yeah. Two sessions ought to be enough...
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
You mean that a multitasking Unix-like OS can handle 4 users AT THE SAME TIME???!!!! It's good to be alive in this modern age!!
...and also getting Linux/XOrg/XFree86 to assign keyboard and mouse devices to the correct users. wscons on Net/OpenBSD are supposed to be good at handling multiple keyboards, but I've never had the patience to get something like that to work.
It seems to me that the only real hurdle there is getting 4 graphics cards to share one PCI bus and memory
that's nothing compared to soviet russia....
I don't think there's any reason it couldn't, if UT2004 had been designed for this. The thing is, I bet a lot of the CPU's effort when playing that is just tracking where everything is. You need to do that whether one person is playing or four. The actual rendering is done by the graphics card; so you've got four times the hardware there. So if you wanted to design a game that one share one game instance across four people on the same machine, I think it would work. (After all, Playstation 2s and XBoxes do this...they just use a split a single display.) I'd think the biggest worry would be bandwidth to the video cards. Most motherboards come with 8X AGP slots now...but only one of them.
BTW, whatever happened to those lightweight computers that were often called "X Terminals"?
If you are thinking of Terminal Services like in Windows... no, you are wrong.
www.ltsp.org is where you get Linux Terminal Services to work. One computer running as server with each 4 clients setup with NO hard drive, or anything. The network card bootprom initiates activity such as sending information about DHCP, where to get packages, etc. Once all the boot up is done, you essentially have a linux workstation running on a computer that doesn't use a hard drive or OS of its own. Its essentially running off of the server.
Great for stuff like computers in a library/internet cafe/k-12 computer lab though. Can't be much slower than my high school's labs of early PowerMacs, can it? And I graduated in 2003 from a school in a relatively prosperous area. This is perfect for a place like Brazil. If you used a relatively powerful computer with enough PCI slots, I am sure you could get more than four users on it as well. Although, a thin-client system might be better, still.
/usr/games/fortune
Also, you don't need 4 cards, You could do it with two. NVidia's twinview allows you to run two seperate X-servers off of one card (provided of course that it has two outputs).
You could do it that way, but regular pci cards are less expensive and requires no fancy software.
Friends don't help friends install M$ junk.
would you mind summarizing the article for me?
Friends don't help friends install M$ junk.
the point is that you can have four terminals while using a single computer as the host...
To ask the question yet again, so what? The only thing different with this is that you have bus connections instead of network connections, allowing dumber terminals than normal. But really, so what?
X terminals never caught on in the PC world, but there's no reason someone couldn't start manufacturing $100 i486-based X terminals. Absent that, just go grab any "obsolete" PC and slap Linux/BSD on it and you have an instant X terminal.
While an X terminal isn't going to let you play DoomQuakeVII at the highest possible FPS, but people who pinch pennies hard enough to need this kind of a solution aren't going to be playing games on their hardware anyway.
Don't blame me, I didn't vote for either of them!
4 plates of NVidia video (TNT2, GForce2 or GForce4), being 1 AGP and 3 PCI or 4 PCI /. appearance.
Pretty funny. Instead of the whole shebang you'd need only 1 piece of a lousy NIC. Ever heard of remote x sessions ? This "solution" is just a waste of video card hardware in the server (which wouldn't need any of them), and at the same time limits the number of "clients" to 4, which is a shame in itself.
Well, it's always fun when newbies discover useful linux capabilties, but I don't think these discoveries merit a
And, as always, Google-ing sometimes can prove to be very useful. E.g. try "remote x access".
Well, I could just be more tired than usual.
I am putting myself to the fullest possible use, which is all I can think that any conscious entity can ever hope to do.
So, what did everybody here think the :0,:1, etc in the DISPLAY environment variable was for?
Might as well head on over to fark i guess.
In 1977 (yes, 27 years ago) I shared a PDP-11 running Unix at Bell Labs with dozens of other users. We all had serial terminals. The PDP-11 had a clock speed of something like 4 MHz (can't remember exactly) and less than a megabyte of memory. Response was better than you'd get today from a 2 GHz, 500MB machine running Windoze, but of course, we just had text on our screens (no GUI, no mice). The only difference with today's story is that the 4 users have GUIs and mice instead of serial terminals. And to support less than 25% of the number of users, the computer has to have 500 times the clock speed and probably 500 times as much memory. If you call that progress, pass me a bucket to throw up into.
Terminal Services requires no special hardware. Install, click "next" a few times, and you can have unlimited people using the same PC at the same time. This has been done for years. We've been doing it at my business for years, in fact.
Oh. My. G0D.
"Who are in control, they are not in control of anything - they don't even control themselves!" - Glen Beck
Most of the time, your PC sits idle while it waits
for you. Every now and then, you sit idle while you
wait for the PC.
Now get a 4x faster PC, and share it 4 ways.
Very seldom will all 4 users need the CPU at once.
So, nearly all of the time, you'll get better
performance. When you need the CPU, most likely
the other 3 users are reading, thinking, chatting,
drinking, picking their nose, or whatever. The
fast hardware is all yours.
At least they're getting more use out of their cpu than most computer users are. Desktop machines out there typically spend most of their time off or running the idle loop.
Why didn't they just use TWO video cards (1 AGP + 1 PCI) that each have two video ouputs (1 DVI and 1 VGA). You could save a little money that way.
This is still a pretty cool project. I doubt something like this would be possible to develop on Windows.
The second reply in the thread gets "redundant," and an apology from the parent poster gets modded "offtopic". Responsible mods don't do this, but responsible mods don't give fools karma in the first place.
Real simple implementation...
/dev/'s dont change on unplug/plug). Configure all 4 monitors as needed, and bam.
1 Recent machine (1 GHz or higher) with USB.
1 Keyboard and 1 Mouse PS/2 variants
3 Keyboards and 3 Mice USB variants (or the PS/2 to USB connecters, use a powered hub. you will need it)
4 Graphics ports (ports, not cards. Some gfx cards have dual monitor support) You probably will use 1 AGP and 3 PCI cards. ATI 128's are good.
4 Monitors, preferrably 1024x768
Now label the Keyboards/Mice/Monitors to which one each is. Read documentation on how to use Linux USB support to "lock" the input devices by UID of device (so your
You just did it.
Please don't post shit
yes, but what about the dedicated hardware to -attach- the keyboards and monitors? The tricky part is that an xterminal provides quite a bit more than keyboard and monitor. GFX card, some networking stuff, basic ROM... If you'd replace the ROM with your custom one, you'd get a neat little computer :) Without all the mainframe horsepower though - obviously.
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
I wonder... If you ran a bouwolf cluster off of 4 PC units running a typical 486 and the heat ratio was in the Dumfeld Range, if the heat sink was titanium, and the dumfelf value was appropriate, could the Sheld Range reach the Afeled Value? I don't think so In MY opinion, the Rafeld would meet the feld. And that alone would conflict with the Whatmatter. You have to KNOW these things. It is as they say required.
"Who are in control, they are not in control of anything - they don't even control themselves!" - Glen Beck
We're buying X terminals for $250 from Synertron Tech the American marketing arm of http://www.lex.com.tw. They have 128 Mb RAM, 3 10/100 Ethernet ports, an 800 Mhz fanless CPU and boot via PXE. Look up the 'Light' system and the rackmounted version of the same thing which costs about $275. You can boot your host with a micro hard drive, too.
Your wireless keyboards and mice are nice, but now that I have four users on my machine, all the wires from the hubs are getting tangly. So will you please build a single hub that accomodates up to 4 wireless mice and up to 4 wireless keyboards, all through a single USB port? That'd be sweet. Thanks!
Actually......
In it's current configuration it would not really work..
But....
Use a 4 processor motherboard, with PCI-X cards (4 of em). And a CRAPLOAD of memory and you really could have a 4 player UT2004 game with latency being almost nothing...
The only wildcard I suppose is whether the combined bandwidth of 4 ATI X600's or Geforce (latest) cards would saturate the PCI-X bus.
Any comments?
Is this even possible with bleeding edge hardware?
Note that HP sells a 4-way linux system built on exactly the same stuff.
It's targeted at schools and universities
The advantage is you HP support and hardware and software all tested to work well in this environment. Was developed in South Africa.
bwahahahaa... which dipshit moderator modded this interesting?? mods are getting shittier and shittier here
You are correct, but, is this hardware commercialy available? And, it would be cheaper than a video card?
Heh, reminds me of my high school days. We used to have some Pentium 150s (brand new back then!) back in high school that ran Linux and X for our Computer Architecture class.
I had a friend in that class and we'd surreptitiously log into each other's machine remotely and start running as many copies of the Swarms screen saver. You could actually run the screen saver on the root window, so your desktop could have the swarms running around as a wallpaper. We used to be able to get a fair number of em running before the other person would notice their compile times taking really long and start an arms race to see who could grind the other's computer to a halt the fastest.
Boy, this is so old. Sun had this ability back in 1989 or 1990, with a third party card that had a video display and keyboard connection. You could add admany cards as their was SBus slots. So in a sparc 1 you could make it 4 headed, (1 using the system video & keyboard, with 3 using the third party video card.) on the Sparc 1000 you could do 4 people per cpu card installed, so up to 16 people. The real problem is the distance that you can run the video and keyboard cables. Not including using X-term and the new X-Term type of thing sun sells now.
You people sicken me. MOD PARENT UP.
This scheme has only 1/4 as many computers to:
* maintain
* upgrade
* de-worm
* fail in the first place
If you dish out a computer each (4x), then you quadruple the chance of independant hardware failure anway; but the number of linked failures:
* worms/ viruses
* lightening strikes
* surges
will just mean 4 times as many computers to fix.
the only benefit to 4 computers for general used is with independant hardware failure where the computers are underutilised so somebody can just use one of the other computers IF they didn't need that particular computer.
But I guess this is offset when multiple failures do occur and staff aren't available to work on all 4 computers at the same time.
Watch TCO tumble with this setup; lets see if the MSFT consultants can quickly churn out a new TCO report to cover this.
Sam
blog.sam.liddicott.com
For those without large IT budgets, you can provide basic computer access to several users using fewer components (just 1 case, 1 CPU, 1 motherboard, 1 HDD, RAM, and 4 cheap graphics cards and heads) and less electrical power. Think developing countries, small businesses, schools, charities.
It could also allow your parents/spouse/children to use the computer/Internet at the same time without having to fight over the computer or buy a whole new second PC.
On the down side it's still difficult to set up, 3D acceleration depends on which cards you use (I couldn't get it working), and not all combinations of graphics cards work. Also with PCIe or whatever it's called, I'm not certain that the Ruby kernel will continue to work.
For those linux geeks interested in experimenting with this, start by checking out the Ruby 2.6 kernel modifications from the Linuxconsole project on sourceforge. Then join the mailing list to get help and advice setting up X etc. There's also a reasonably up-to-date HOWTO in the Linux Doc. Project.
This is not a sig
Good grief; its more scalable than dedicated X terminals which can only serve ONE display. ONE. (1).
"you could run a whole classroom off 1 reasonably specified Linux server"
not forgetting; of course, the classroom full of dedicated X terminals?
To prove my point, you could use this 4 head setup as a 4-head "dedicated" X-terminal; now tell me which is more "scalable"?
Sam
blog.sam.liddicott.com
Cheaper - possibly, though not necessarily. Though attach 30 Xterms to one box and with enough RAM and CPU horsepower it all works. With gfx cards you're limited to number of PCI slots. Commercially available - here's the problem - not anymore. Out of production for some 10 or 20 years already. Only second-hand now. Simply all the "multiple users on one machine" solutions were replaced with The Only OS (which can't do security or multiuser stuff correctly yet, so forget about multisessioning - and why build hardware if The Software doesn't support it) - the article is about a rather ugly cludge abusing abilities of hardware that wasn't meant to do this in the first place, thanks to robustness of a system which could but wasn't asked to do this for quite a while now.
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
I would think that PCI bandwidth might be a bigger concern. Especially if combined with Fibre Channel or Serial ATA storage. I suppose PCI-X will help eliminate bandwidth issuses in the future.
"The world only exists in your eyes. You can make it as big or as small as you want." - F Scott Fitzgerald
Way back when Unix was a phylosophy in someone's mind, and computers were verry slow; the idea was to have one system have multi-user accountable processes and every aspect of the system was to isolate users from eachother via permissions and file/stream ownership. The system resources were extremely limited, and so networks were designed in a somewhat "star" pattern to where one was known as a "dedicated server", that which more beleagured and inexpensive systems would attach and login and interact with their content in that server by making requests et al.
Now, it is great to live in a day when their need not be any thought of a "dedicated server", yet known as a "peer"; the definition of "peer" contrasting that which is hosting both client and server processes active within the same hardware control unit.
Verry nice, but unreliable with consideration to system stability in varioud demanding applications, unless it is noted for the energy cost reductions necessary for avid word-processing useage and noneother...
Ok, so why couldn't you do away with with the hard drive and boot these as thin clients?
The CPU, ram, drive and motherboard are still the big costs in a computer. Two $40 nvidia dual head cards instead of the extra 3x cpu's, RAM, motherboards, NICs, powersupplies, etc. would save quite a bit of cash. Four terminals for the price of one+$40 isn't bad.
I didn't see an explanation why nvidia cards were needed. If they aren't actually needed, use the onboard video of many motherboards and several really cheap PCI cards.
This could also save electricity in large LTSP installations.
If tyranny and oppression come to this land, it will be in the guise of fighting a foreign enemy. - James Madison
Userful also sells a system like this with features geared to libraries and internet-cafes.
They have up to 10 users on 1 box.
http://www.userful.com/
They say it costs 50% less than 10 traditional PCs
http://www.userful.com/products/onsite
We used to do something vaguely similar at my university. We had relatively few high-end Sun workstations (at the time, SparcStation 10s). There was 1 SS10 for four users in some of the labs - three of the 'heads' were inexpensive Linux-based X terminals. This was around 1993/1994 or so (whenever the SS10 was reasonably new). The xterms were built by a local company.
An x terminal isn't a bad way to re-use an ageing machine - an old 486 with a Tseng Labs ET4000 would be fine for a guest terminal, hooked up to the 'real' machine.
Oolite: Elite-like game. For Mac, Linux and Windows
Could it be that windows doesn't support multiple monitors? No.
Or that it can't support 4 different users on them (as opposed to 1 user getting a big display)? Probably not, at the very least, it would be hackable.
Maybe it can't support multiple keyboards, or mice? Again, the most it would need is some hacking.
Or maybe, just maybe, if you posted a webpage, telling someone how to use a single windows license for 4 users, M$ legal would go apeshit on you, and stomp you into a tiny, tiny greasestain?
BINGO!!
Wouldnt it be better to use a Quattro card or something that supports more displays on one card? Seems to be a waste of to use one for each.
I have a Sun SparcStation 10 sitting behind me. It was manufactured in 1993. SparcStations of that time used something called a "Framebuffer", which was a sort of integrated video card and serial port for the monitor, keyboard, and optional mouse. I believe they support something like 3-4 framebuffers, so that means 3-4 monitor, keyboard, and mouse combinations with one single SparcStation. The SparcStation 2, 10, 20, and several others all supported this kind of configuration. You connect the keyboard and mouse into the monitor rather than the system itself.
I still use this SparcStation today. It has a HUGE 128MB of RAM (all eight 200pin SIMM slots filled), a 50Mhz processor, a 2GB hard drive, 19" monitor, and a 10baseT build in Ethernet port. I even have the speaker and microphone parts.
We used to do this on mini's and mainframes. Its come around many times since the PC was born.
In these times of easy fast networking and cheap hardware, it really has no benefit. Consider that four or more people share the same box; you're effectively getting a quarter of the performance, hence four really cheap and crappy older machines with twenty bucks of network cable will do better.
But heres the big rub: the problem with timesharing all along is the app or action of one user severely impacting the others. If you save a few hundred bucks on hardware, thats more than offset the first time someone downs the box and kills off four users work. Think it wont happen? Think the OS is indestructable? Think again...while single user stuff is beaten to death, multiuser stuff is unusual enough that all sorts of cool bugs show up.
The small rub: timesharing creates variable response times. What has been proven to drive users the nuttiest? You got it, variable response times. If they get used to something happening instantly, having it happen in 1, 3, 6 and 10 seconds on a random basis will have them asking "is there something wrong with the system?". People will even get used to consistent, long response times. They dont ever get used to variable.
Interesting recyle of an old piece of technology, but of very little particular use.
We run 74 machines off a single dual 1.13 coppermine server running (blech) SuSE using LTSP3. It provides websurfing, Open Office, squirrelmail and a connection to a queen bee application on an AIX machine. It's solid.
The last round of deployment was a group of HP Vectras we picked up for US$5.00 apiece; 15" monitors were $30.00. Throw in a sawbuck for a little more ram and we were off and running.
Like a lot of incredible tools, LTSP has little Wow!-factor, especially amongst a group that would rather brag about playing 4 FPSes off one box. But it's real stuff that helps real people in countries that can only dream of having the hardware the drooling slashboys throw in the trash.
It worked just fine. Funny how everything old is new again.
Best Buy can have you arrested
My machine has been running for a few weeks and is not noticeably hot, however they are not the latest state of the art graphics cards, especially the PCI one which I found difficult to obtain. I would like to get a twinhead PCI to have 4 monitors, but they are rarer than hen's teeth in the UK.
BTW the graphics setup is OK in Linux or FreeBSD, once the Nvidia drivers are loaded, and with SuSE you must forget all about SAX and edit XF86Config by hand, as per the comprehensive info from Nvidia. The same file, with possibly minor tweaks can then be ported across to the FreeBSD partition, and the FreeBSD version of the Nvidia driver loaded, for those who want to dual-boot it with two decent OSs.
BTW I did this just to get lots of working space on my screens, 3 cheap LCDs at 1280*1024, great for programming and debug, and whatever else you need to do at the same time such as reading manuals, but takes less desktop than the two 1600*1200 CRTs I had contemplated. The Athlon is nothing special, 1600+ I think (not at that machine right now) but it has 2 HDDs, / on one, swap on the other, the usual speedup trick, not that it needs much swapping with 768MB RAM. Definitely not a top performer or games machine, but it works great, and most of the cost was the monitors.
This concept would be great for schools, it may be how Linux or BSD, or both, can make serious inroads into a big market, currently filled by the Criminal Monopoly. I think 6 screens might just about be viable with a top of the range Athlon, and some extra fans in the case, but I will not be building one of these for myself, it would be very OTT.
It is going back to the mainframe in a sense, but in areas where maintenance of the software is a major burden, this reduces that by a factor of 4, approximately. I would love to see this concept trialled in schools in the UK, it could free up money to be spent on other important things. If using a fairly basic PC, and cheap CRT monitors, it ought to be possible to get the cost to a very competitive level indeed, with no extortionate licence fees to you know who.
I t would also be good in offices where they do mainly word processing, hardly a heavy load, one PC for 4 secretaries.....
My 3-head monster has spare disk partitions, but I have not dared to try loading my redundant copy of Windoze 2000. It is rock-solid in Linux, FreeBSD is getting there (I am using 5.2.1 which is not a "stable" version, and have lots of apps loaded, so there is still a bit of tweaking needed, although the GUI etc are solid.)
Just had a thought, did Matrox not do a 4-head AGP card some time ago? Maybe that would do the job also, and leave slots free. I hope we are going to hear more of this sort of thing, on one front we have those who need, or want, performance goimng for multiple CPUs, all the way up to Beowulf clusters, and on the other hand the economy version has several users on one PC. Now whose OS is the scaleable one? Here is the proof!
Think what you could do with a lab full of these running as diskless nodes. That would be hot.
You are trying to think of a way to isolate the "core" keyboard and mouse, traditional to a 1-person computer, from additional keyboard and mice. First of all, XFree86 and virtually all X Servers that grasp the video hardware for the visual output can be given alternate device nodes of where to receive keyboard and mouse input. Yet the problem you speak of remains: distinguishing from the "core" input devices, which the kernel initially receives input.
;-). The first motherboards I noticed to not feature either a PS/2 or DIN style keyboard and mouse was the SGI 160/320 and its related products of the Cobalt series IIRC; their keystroke input was merely USB-oriented in that Pentium III architecture.
In all likeness, the earliest AT and ATX keyboards always featured a PS/2 mouse and a DIN-connector keyboard. The verry design of the x86 PC does not recognize multiple keyboards because it is a verry *singular* or rather solitary architecture in all its intentions: 1 captain/user/king/master/commander/deity...I prefer master
But still the thought of conflict remains. In another post, I discussed seperated Client and Server instances combining into Peer technologies. One of the primary feets of Unix-like environments is Permitions and Ownership. To have multi-users upon a single system would need to establish the environment's restrictions; which is near impossible because the verry nature of USB devices is Plug'N'Play which is impossible to limit to any specific setting and is constrantly changing on the USB's identification chain. We are stuck until we have small IO controllers attach to the USB chain therewith isolates additional input devices from directly interfacing to the USB chain and thus conflicting with the intended "core" input devices which these monolithic/solitary Personal Computers were limited.
XFree86 does just fine in this, if you can isolate each keyboard and mouse and bind it by an X Server unto a specific graphics adaptor. Blame the USB standard, and also blame the Firewire standard; if these allowed us to use jumpers to limit a USB device to any specific identifcation and not a random "as-available-as-necessary" Identification Number over the USB chain then you can distiguish devices for a particular use. But it is futile, because USB and Firewire are not in-line with even the Unix phylosophy to distinguish a User's accountability and Permitions and Ownership. It's guesswork.
In my experience; the Voodoo2 graphics adaptors are an excellent and inexpensive resource for use in multi-user X workstations on a single terminal. For background, Voodoo2 graphics adaptors all lack VGA BIOS and thus you just fill them all into the available PCI expansion slots of a motherboard, configure an X Server to bind them to a specific USB keyboard and mouse (XFree86 works well on this using their Glide wrap-around server), and then make your pleadings that nobody unplug their USB keyboard or USB mouse because that will pretty-much end their X session's input responsiveness.
I hate USB in these regards. It's just a pain in the ass shitmeister fudgepack of a technology that is more usefull to allow snooping on the USB chain's data then being extensible.
I figure that buses would be a bottle-neck. PCI is far less than ideal for 3D gaming, and with four cards, the problems get much worse still.
I love C++
With everybody on the same CPU, latency is a non-issue. Fast FPS games should synchronize perfectly. That tightly synchronized feel will make for much better head to head gameplay.
Limit the gaming to X11 and nothing more; not a problem.
When you attempt multiple users and accelerated openGL, then you have a file permitions conflict that will seizure the system.
Mature 2D Linux games such as Quake 1/2, Doom 1/2, Myth 2, Heretic 2, Heroes of Might and Magic 3, and others will all shine nicely. They work even better when there is abundant RAM, perhaps no less than 256MB per user. Do not use KDE or GNOME, they are bloated for these sessions; limit the users to a window manager and not a desktop environment and it'll be smooth.
I is good to know that a research group is capable of reading the XFree Local Multi-User Howto from the Linux Documentation Project.
This approach doesn't sound new, but should be very useful.
Since most new PC's come with keyboard, monitor and mouse, IT departments have boxes of old keyboards and mice around. We also dispose of perfectly good moniters.
Therefore, the hardware cost is mainly the drives, chips, boards, etc in the case. In wealthy nations, people consider this cheep, but it is still a significant expense to an organization. (For simplicity, I'll ignore the indirect environmental costs of manufacture and disposal.)
My employer has about 200 office employees, many of whom work in adjoining cubicles. Most of them use office apps, telnet sessions and quite a few specialized applications. For each new computer deployed, the PC support staff has to spend considerable time loading and customizing software. The support staff's time is a major factor when comparing TCO.
The implemetation might be expensive, but if this is planned correctly, we'd in theory convert to Linux with free software, one quarter of the PC's and one quarter of the support staff time.
A even better case would be a growing startup company without a big Windows investment. If the only bought one PC for every four users, there would be a big savings in hardware and support.
...where x equals the price of an Apple Mac, and where X is divided by Y, where Y equals 4, 4 people?
Jonathanjk.com
I want to be able to use two mice on the same computer on the same screen ... can this be done?
Are X terminals cheaper? It all depends on how you count the cost.
I've worked in a thin client/server environment and in a networked Windows PC environment, and I think that the first made a whole lot more sense than the second.
Back in 1990, we used a bunch of diskless Sun workstations connected to Convex C1s and C2s, running ConvexOS (a cousin of BSD) and X Windows. I loved this environment. The whole thing worked smoothly without major headaches, and gave me an amazing amount of flexibility in configuring my environment. In other words, the tools helped me get my job done, and otherwise stayed out of my way.
Now I work in a place that has a bunch of silly PCs running Windows 2K, and the environment is chaotic, consumes gobs of administrative resources (and of my time, trying to cope with it), is constantly down either because of viruses transmitted across the company network, or because of "emergency" pushes of anti-virus patches. The sysadmins are constantly trying to police all the PCs to make sure nothing unauthorized is running on them (or trying to make the authorized stuff run). I, for one, would love to go back to 1990 (with upgraded processors and GPUs, naturally 8^).
As some of the responders pointed out, it may indeed be possible to buy cheap X terminals "these days". In fact, any cheap PC will do as a client, as long as it will run Linux/X Windows--you will need to decide how much GPU, CPU and memory you need. But since the server can provide a lot of the number-crunching power, disk storage and interface with the WAN, that might be less than you would need for a standalone PC.
However, by far the greatest advantage I see for the thin client/server architecture is the lower administration cost. In effect, the sysadmin only has to maintain _one_ machine, the host. This means:
Great men are almost always bad men--Lord Acton's Corollary
This could reduce TCO for a lot of applications such as schools and most offices that don't really use the computer to extremes. Save a lot of doe by having 1 CPU for 4 users. This could be very nice.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
Shame he can only address the first 36 of them!
I run a Radeon 9700 Pro, 2 Geforce2 MX200s, and 2 TNT2 16MB cards in my single Antec midtower case with a P4c 2.4Ghz. All are air cooled with stock fans. I have no heat problems.
(brute force multi-monitor setup)
Look, an X-box is invovled here...p g
http://www.c3sl.ufpr.br/fourhead/mesa1.j
Just here to let you know, this 4-user setup defined in this forum's topic is using four nVidia TNT2 graphics accelerators! These are better than GeForce graphics accelerators for driver-related reasons. Way back in the maturation of XFree86, with version 3.3.6, it is possible then and a throughout the 3.x XFree86 branch to configure XFree86Config to "Load """ Utah-GLX's nvidia driver to attain hardware-accelerated openGL. This is a completly different driver approach than DRI's openGL SAL. Utah-GLX provides X Server modules rather than its various competitors providing a /usr/lib/libGL.so.* and any non-standard patch cludged into the X Server. DRI project's openGL acceleration architecture at the moment may also allow mutliple local X Servers, albeit that of the various non-XFree86 such as the capable technology at DirectFB project (which allows accelerated openGL without a X Server; directly using the DRI without an X Server).
Backtracking to Utah-GLX's driver (project page here, this will allow many complex openGL-phile programs to run at the same time given its architecture. I, however, doubt that older XFree86 3.3.6 will scale to this feat; I simply don't know. Yet, the Utah-GLX driver system has been ported to XFree86-4.x; it is a openGL GLX driver package in the form of dynamically loaded X Server modules/extensions and can be manipulated into and without the X Server without having to restart the X Server. It's somwhat parallel to the DRI driver, to provide an alternative, but it is not being maintaned anymore; Utah-GLX is dead and someone needs to commandeer!
I am using three Athlon Thunderbid 700MHz computers with a total 9 nVidia TNT2 adaptors total (three per computer), S-Video composite output to NTSC televisions, and quad-bonded 100BaseTX ZNYX LAN adaptors for verry low-latency threaded shared openGL rendering; I use as Chromium 3D videowalls, by using XFree86 4.3 and Utah-GLX's nVidia openGLX driver.
And yes, Quake3 looks hot!
At Sun we have upwards of 200 people using a single computer with sunrays displaying the graphics and providing i/o. I'm sure we have accounts that beat that number. http://wwws.sun.com/sunray/index.html has more details for the curious. It runs on Solaris today....
My faith on Slashdot is slowly recovering..
Colorgraphic makes a 4 head AGP card (2 radeon 9000 chips, each with dual head). they do offer a 8 head one (4 chips) but is pci only.
Has computing gone forwards or backwards when it takes thousands of times the compute power to support fewer users, doing dumber things. We used to run whole research departments developing mathematical modelling, computational physics programmes on a single DEC VAX 11/750 with 8 MB of main memory and like 80 MB of hard disk space. It was so underutilised that astrophysics would rent out time on the darn thing to geophysics and chemistry.
This was on 4.2 BSD, the mother of all open source operating systems. And we had access to supercomputers at Argonne, NCAR, LANL, LBL and Cornell over the ARPAnet. in the freaking early 1980's.
AND we produced beautifully typeset scholarly papers and theses, full of equations using TeX. Try doing that with Office. Hnf.
Personally, I used to use maple to do the algebraic manipulations, and export to either fortran (to run a numerical simulation to get the results that formed my thesis) or to TeX (in order to publish it). Sure as hell can't do that with the stupid Office (open or MS) programmes you need 15 64MB computers to support only 60 users on in this model. Even if you insist on running a pointy-clicky GUI, with X10 we used to run dozens of graphics terminals off of one VAX
This article just proves that the net progress of computing is actually backwards because the computers certainly are getting bigger/faster/better more slowly than the intelligence and creativity of the users -- now they all need a GUI just to edit text and compile programs. To the point that it's a miracle when you can have more than one person using a computer at a time now. Sheesh!
lol...never mind that you said gigs but 800+? With your smarts I'm surprised you don't know the exact amount of ram you have. Must be a slow day.
-----
One is born into aristocracy, but mediocrity can only be achieved through hard work.
Time sharing systems were common, way back when computers were 10,000 times slower than today. At least now, it will work one hell of a lot better than it did back then.
Oh well, what the hell...
four users, one pc? :|
:D
in my room it's 4 pcs, one user
i'm lucky i guess. IM A FAT PIGGLY WIGGLY NORTH AMERICAN!
*SQUEEELSSS!*
-judging another only defines yourself
THE MONITOR!!!
I have a multi-user setup at home (I'm using it right now) which is described here... it has 3 heads (video/keyboard/sound/mouse for each). I've found it to be an economical approach that involves far less administrative work than any other multi-user configuration -- separate systems with NFS file shares, X terminals, you name it.
There are several significant challenges involved in setting up a multi-head, multi-user system:
The 'normal' kernel USB setup merges multiple keyboards and mice into a single queue. This makes sense when you're using an external keyboard and mouse with a laptop, but it a pain in MHMU. Patches to the USB system are required.
The 'normal' console/virtual terminal system is not MHMU-aware. This has to be patched with the Backstreet-Ruby patches to work properly.
The 'normal' X server can handle a multi-head configuration, but expects all the heads to be used by one user. The prefbusid patch or similar must be applied to fix this.
Apps that use sound are a nightmare-- it's painful trying to get apps to send their sound to the right speakers. With three of us on the system, I regularly get blasted by sound from Flash websites that my kids are visiting. There is no standard for defining *where* sound should go -- yes, there are soundservers such as esd and artsd, but they are not used by all apps (and try finding current documentations for esd!).
If we made a concerted effort, the MHMU patches could be merged into the main source trees for X.org and the Linux kernel, and the distros could be set up to offer MHMU configurations at installation time.
However, the number of people using MHMU configurations is currently too small to really make a push for the integration of these features into the main source trees and distros.
So if you're using this type of system, or would find it beneficial, then let's work together to make this a mainstream option!
(I think that MHMU would be useful for: tellers in banks; kiosks such as library catalogs; computer labs and learning commons in colleges and universities; and family computing setups).
I've studied 2 years Mechanical Engineering in UFPR.
I'm not saying they doesn't exist, but i never seen a Linux Desktop there. 6 months before i've stopped to study there they changed from NT to Windows 2000. Computers with 500mhz running Windows 2000, that computers could be much more productive to students running linux. Of course in Mechanical engineering you need to run programs such as Solid Works, but many computers would be much more usefull if running Linux.
The computers department of UFPR may look at Universities like UFGRS(also here from brazil) witch changed to Linux desktops(in UFRGS there's something about 4000 computers running linux).
In one of the computers rooms for 1 year we have a paper written: "don't execute xxxxx.exe because it's a virus", security was a joke, 10% of the computers are always off because they are with virus, whem the computer was woth Windows NT people goes to the back of the rom and hit "esc" to login and start to do "wrong things"
etc etc etc etc etc etc
Em português, in case you speaka da lingo or want the scripts etc to actually work.
Got time? Spend some of it coding or testing
It could be worse. (-:
I think Chris "Jar'anthe" Le Sueur is referring to JCL, which while powerful was also an unending nightmare of arcana (yes Rhonda, all that incantation actually does is copy a file).
Got time? Spend some of it coding or testing
They'd handle it with water cooling.
Got time? Spend some of it coding or testing
So now, in addition to dealing with the kids using all the bandwidth to download mp3's, you now also have to deal with the nerds who are using all of the CPU time calculating prime numbers.
The only way this will work is if you can "kill -9" other user's processes. Make sure it comes with boxing gloves.
Sorry, I realized that in the previous post I used the abbreviation MHMU without defining it ... MHMU is "Multi-Head/Multi-User" (as opposed to multi-head single-user with or without Xinerama; or multi-user using remote access).
At my university we have some 30-40 X terminals running on one Solaris server (a quad processor server with 4G RAM, aptly named "beast"). The thin clients are Sun Rays. I don't know how much they cost, but I know the X terminal principle is very flexible compared to the 4-terminals-on-one-PC option: you can use a single server to serve terminals that can be spread out over a wide area, instead of next to the PC in question. If there would just be a cheap X terminal solution available, I think it would be a perfect solution for public libraries and internet cafes.
My very first home computer was a Radio Shack TRS 16B+ in the mid 1980's.
It had a Z80 CPU to handle I/O housekeeping chores and an MC68000 main CPU running XENIX (a flavor of UNIX).
It supported four users at the same time - each at their own terminal - with no additional goodies needed at the 16B+.
Mine did have a full load of memory, a larger hard drive, and a few Hayes modems so the other users could be remote, but the modems hooked right into the existing multiple ports on the machine.
For several years it was a minor mail and news server on the web (named tijil).
In what major way it this "new" thing astonishingly different from what I had 20 years ago on my desktop at home?
Take care,
Tomas
you built it yourself?
Wow! youre great. Can I have your autograph?
From a quick look at the site, it didn't really talk about the performance hit that the single processor takes. Would it be 4 times slower, assuming equal usage? You'd probably be able to find 4 machines cheaper, that each meet your requirements, than 1 machines that is 4 times your requirements.
well im running a 1000 megasquirtle doobydoo with a megakilo double widget interface and a pseudo-terrafilter. its got 2 million hexafurtles with a 10 kiloplex wobblywoo and a rechargable virtual combo-backplate monitor. i also fitted an optical finglyfangle with predictive threshold monitoring. its also got a touch sensitive keyboard and a temperature sensitive thermometer.
Built it my self.
------
beware he who would deny you access to information, for in his mind he dreams himself your master
> However running multiple instances of
> X on a single computer is pretty new.
That is exactly what "vncserver" does.
It's a monitor, keyboard, mouse, and integrated monitor stand that hooks back into a computer, via a network. You can even hook up more than 4! Amazing, that. Or is this supposed to be an oooh-aaah for changing a KVM switcher into a multiplexer?
Yes, we understand these tags always apply: fud, dupe, typo, slashdotted, topic name
This reminds me of the system in Key Largo, FL that serves 220 concurrent users from a pair of Compaqs running Red Hat. Notice also that the date on this article is April, 2002.
If you're starting at the low-end, you sure can
get a 4x faster system for less than 4x the cost.
If you start at the high-end, getting a 4x faster
system may be 50x the cost or simply unavailable.
The systems in the article don't look like fancy
gaming systems or enterprise servers. They look
like cheap white boxes. Getting a 4x faster box
could mean going from $200 to $700. The limits of
technology are not being approached.
Can you literally run something on a shoe-string budget? What, exactly, can you get in exchange for a shoestring? Or did you mean a budget literally equal to the value of a shoestring?
Drives me nuts every time someone says 'literally' to modify a phrase that it is literally impossible to construe in any way but figuratively.
~Sub
-1 Troll
-1 Flamebait
+1 Linguistic Merit
+1 Crankiness
Yup.. people need GUIs nowadays. On the other hand, you don't have to be a grad student to use them.
(Figuratively and literally)
So?
I'm pretty sure that instead of buying enough of the right types of extension cables for the keyboards and mouses for this all to work and not have the four users right on top of each other, it may well be cheaper to just buy 'em each a cheapass Dell.
If you're sharing a PC four ways, you're not exactly looking for high performance, and when you figure up the cost of buying monitors, keyboards, mice, and extension cables for everyone, I get to wondering just how much the rest of the computer costs. I mean, Dell's selling their low-end at the moment under $500. Sure, it's low-end, but that's what you're looking for if you're sharing a machine four ways...
Of course, my real reason for thinking this is a bad idea is it's bad enough when someone's hard disk dies and the user's machine is down 'til FedEx shows up the next day. It'd suck even harder to have four users who won't leave you alone...
-JDF
Sure, it's easily hackable on Windows. First, download the latest kernel source.....er, wait a minute...
Keyboard not found.
Press F1 to continue.
There are enough bootleg copies of Visual Studio floating around, and microsoft exposes enough of the API to make it thinkable (hardware vendors need as much for their drivers). Easy? Ok, so maybe I was laying it on a bit thick there, but there are enough windows enthusiasts, if you will forgive me perverting that word, that it might have been done by now. Someone even sold a 2 seat version for win95 (believe it used a custom ISA card for extra kb/mice), but that's not really practical unless you can segregate user processes, I would think.
I dare someone to do it with XP home edition, or 2k pro... the EFF won't make it to you in time to stop the carnage.
We have gone in this direction because computers used to be vastly more expensive and had to be shared. As the price/performance of computing equipment went through the roof and costs continued to fall, it became more cost-effective to give really high levels of computing power to individuals. That most computers spend most of their time waiting indicates the tremendous capacity of what we have, and shows how ideas like this can improve the conditions of those who have not reached the standard of living to where single individuals can afford massive computing power.
Nobody needs a Mercedes Benz, but some people can afford them. When everyone can afford a Mercedes Benz-class automobile, then the quality of automobiles will rise to match. This is the situation of computers now.
The lessons of history teach us - if they teach us anything - that nobody learns the lessons that history teaches us.
Well, this would be comparable today to giving a bunch of users accounts and allowing them to use telnet or ssh to get to a shell account. In a setting like that, I bet your average computer today would support hundreds if not thousands of simultaneous users. With X terminals, I bet you could have a bunch of gui interfaces too.
This is a lot different than that - this is about taking an interface that has been designed assuming there is one user in front of it, and hacking it to support multiple users. Multiple keyboards, mice, and displays on one box, don't normally work this way.
Cool? yes... Practical? probably for someone... but in a world where people hack web servers onto commodore 64s and overclock their Sega Genesis, rank this toward the top of the list.
The lessons of history teach us - if they teach us anything - that nobody learns the lessons that history teaches us.
Commercial software, doesn't count. They licensed this from microsoft, microsoft made it more expensive than just buying extra machines, then they changed the API enough to kill it. I've seen it a few places, but something other than its usefulness (or lack thereof) was causing it to languish in the marketplace. Remember, I'm not a conspiracy theorist if the villain is M$ (haha).
You're missing (ignoring? i hope not.} that old P1's are getting old. Maintenance is a real cost, particularly labour, but also the parts when you've got downtime due to digging around for spare obsolete components.
Think five years ahead. ATX PSUs will still be cheap, AT PSUs will be a PITA. "Keep a stockpile"? Yeah, right; storage space costs a premium in schools. They'll be lucky to have a small closet or even a shelf to devote to future spares. Much cheaper to buy available components from dealers.
The cost of RAM for P1's is already unreal. I've been steadily replacing my free P1's here with cheap Athlon boxes just because of costs like that. The amount you have to futz around with P1's has begun to exceed the value they provide; this isn't a hobby -- we're trying to get work done.
Another issue is getting a budget approved by a non-geek board (the usual kind). This four head setup will get rubber stamped no problem. A "free old P1's" pitch is going to sound like a whacko adventure and be shown the door. Who has time for that? The computer lab needs terminals.
I'm not saying using existing P1's is bad at all. I'm pointing out that there's definitely a place for this 4x arrangement.
The limit is how many PCI cards you can put in and how many connections you can provide via hubs. Stick in some matrox dual head cards and it wouldn't be that hard to get eight - then it all depends on what software those eight users run as to whether it would be usable or not.
Video cards are cheap, and good monitors are often found attached to old systems that can't take a reasonable video card and are thus not good Xterminals.
This was being done for multiuser MSDOS back in the late 1980's. VM/386 and Concurrent DOS supported up to 64 users with special video hardware from Maxspeed. Multiuser Windows NT and SCO Unix were also available on the Maxspeed hardware. Applica is an example of a current product.
You said that they were running four terminals at one time. Why is that so amazing (aside from the fact that they are on one machine). You can do that with LTSP (ltsp.org) and it doesn't take as much configuration and can handle up to 32 terminals per server. Having four monitors, keyboards, and mice attached isn't very useful, especially with the added hw configuration, when you could just do it (faster) with a 200 mhz slim PC and a bootable NIC. If you spend this much money (or time) on doing something like this, wouldn't you like to be able to at least put the terminals in separate rooms? Bottom line: don't bother with the config, seize the carp (carpe dium... ;-) and get LTSP.
Multiple graphics consoles and keyboards attached *directly* to a single mini-computer running a UNIX-type operating system running X-windows *HAS* been done over a decade ago. In this case, they were separate login consoles and, yes, each had their own X-window session, and yes, it was not an xterminal, they were connected directly via video and kbd cables to the main system. I am trying to remember the system in question I think it was a Stardent Titan mini-computer used for graphics visualization.
You'll just create new-age music if two folks start trying to play mp3's at the same time.
"Learning is not compulsory... neither is survival."
--Dr.W.Edwards Deming
Can't this be done more easily and cheaply with dumb X terminals? My wife's old PII 366 laptop happily serves up recipies in the kitchen by loggin in to my main machine via remote XDMCP login. I'm sure an even older machine with just enough hdd space and memory to run X would also suffice. Still, these guys are doing something interesting, even if it is the hard/expensive way ;)
RebateFX.com - Spread rebates for Forex traders
To apply patch, it executes the following command
The best weapon of a dictatorship is secrecy, but the best weapon of a democracy should be the weapon of openness.
Yes, a lot pf people are saying that it has been done before, that a terminal server is cheaper, etc.
I could think of one practival use, apart of cs lab:
What this thing misses still is sound - a usb stereo headset so you can do ip telefony (4 sound cards wouldn't make sense).
Such a setup would be quiet usefull in countries like Brasil. Put a CPU, 4 LCD screens, 4 keyboard, 4 mouses, 4 headsets and a connection to the internet (phone, dsl, wireless, radio, etc), an you can rapidly deploy a internet/phone access in a remote area, an connect to the world.
And administration wise it is easier than manage 5 computers and a router.
The CPU is the single point of failure. But if everything is standardised, ship the same working unit to the remote area as a replacement.
I run it myself. My girlfriend and I have a small appartment, and I'm stuck to the computer almost all the time, and she needs to use it too. We simply didn't have space (or money...) for a second computer, so I'm using a local multiuser setup.
I think it was rather hard to set up, and I had a lot of trouble along the way. I still have some problems, but it may be due to faulty hardware (a Tangtop Generic USBPS2, anybody know those?) For most setups, you need to patch the X server as well as the kernel. The kernel patch is straightforward, but I had some trouble with the X patch.
It has improved a lot in Debian, because now Debian Sarge and Sid ships with the patched X server. Have a look at the isolatedevice option if you run these releases. That's all you need...
I have also run 2.4 with backstreet Ruby up to a few days ago. Now, there's 2.6.7 with the real Ruby patches. It works great!
I'll recommend this setup! For many, I think it is better than a thin client solution.
Employee of Inrupt, Project Release Manager and Community Manager for Solid
While I am not the young man from whom you demanded an explanation, I can offer you som insight on why I do not use Xinerama for my dual-head setup.
I use a standard (non-Xinerama) dual-head config and a dual-head aware window manager (Openbox 2.x). This setup has but one serious flaw, and that is the inability to move windows between the screens. What it offers over xinerama is that it does not require xinerama-aware applications. Popup notifications never show up in between the monitors, for example. The big thing I dislike with Xinerama is that windows can be split over the two screens. I have never wanted to do this. I much prefer the ability to slide parts of windows out of view.
Another big nuisance with Xinerama is that I have yet to find a Xinerama-aware window manager that I like.
My vision of the perfect multihead desktop (which coincides with my vision of a reasonable X architecture) is pretty much X with something like xmove (which I have yet to get working on XFree86/Linux). The window manager could then be told to throw a window to the other screen when dragged to the edge, just as with virtual desktops. The bonus of using an X proxy, or preferably separating the display and server parts of the X server, is that you can move windows between computers, keep gui apps running on servers when the workstation is turned off, restart your X server (e.g. to upgrade drivers) without losing your apps, and so forth.
...ceterum censeo Carthaginem esse delendam.
While the Linux system might be difficult to maintain, it can be done.
.dotfiles, and it Just Works.
... re-formatting every PC in the office, re-installing c:\winnt32\, removing IE, 'fixing the e-mail system', installing new Virus patches, etc. will now be free time you can apply to actually customizing your -one computer office system- to the task of your business, easily, and maybe learn a few Open Source skills on the side, as well.
...) focused on computerizing his/their business process, paired to a single cheap PC that 4 people can be using at the same time, is good Starter Business Manna ... put a smart business app on a Single PC which is dead-easy to maintain (hint: it just runs), make sure Single PC is cheap, and it works, and you've got a 3rd world economic power-station.
... whereas having everything 'local', on a single system with sufficient power (RAM) to cater to the needs of busy people, is entirely feasible, and cuts out a -lot- of the reasons for most modern PC maintenance ...
Every user gets their own session, their own homedir, their own
Once this is set up, and X is running, its done. No maintenance. Multi-head is actually "X" and Unix at its greatest.
Put OpenOffice.org on there and you've got what you need to run the average office, for whatever business.
All that time you're wasting now, in the "difficult to maintain" department, you know
An Apache/PHP/MySQL-based business app being developed by an enthusiastic software hacker (it could be the whole group, with their own PHP scripts
Think of the rows and rows and rows of low- to middle-class businesses you can see in any average big-city, and now add a small, affordable computing system that -4 people can play with at once- in each box in the row.
Most people buy multiple-PC's for the office just so that they can have access to a 'broader computing system' that they then try to tie together with all kinds of other bone-dead 'systems' like 'shared Access databases' and whatnot, over some network (which requires even more hardware and maintenance)
Its gotta get easier to multi-head PC's. It just does. The single-user hardware paradigm needs to be shifted. Microsoft -need- new licenses from new 'complete system' hardware to propel their buggy OS, whereas Linux only -needs- you to use it.
; -- the corruption of government starts with its secrets. a truly free people keep no secrets. --
The performance:cost ratio might well be higher using cheaper individual stations.
if there is a linux driver for this
Year years now video cards have had multi head/monitor displays and matrox has lead the way, even unvieling a tri head display? Could a similar set up be used or something else be devised as to allow one of these matrox cards to be used on 3 different X sessions allowing 3 users to use not just the same CPU but the same GPU. I'd assume that special drivers would need to be written? Now that would be awsome and i think a big money spinner for linux and matrox if they could do it.
Giving IE users a taste of their own medicine since 2005 - http://pods.-is-a-geek.net/
I never tried to plug more than 1 keyb and mouse into a 'doze box but I'm pretty sure it's not a good idea.
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
If you have a URL for these services, please post it here. I know several school administrators that would be interested in such services. I have three questions:
Openbox 2 (as mentioned in my earlier post ;)
...ceterum censeo Carthaginem esse delendam.
I guess I'm talking about relative efforts for each platform. In absolute terms, Windows still blows.
Windoze on it's own is very difficult to maintain if it is connected to a network and exposed to the internet. The only solution to the problem is re-image every day as is done at public libraries. That sucks for the user, in all the ways you noted. Still, on the Windoze side going with the "twin" software will reduce your current effort.
The Linux side difficulty I'm talking about is minor compared, but large for someone like me who's been spoiled by tools like apt-get update and upgrade. If you don't go with Debian stable and want to stay current, you will get changes that blow out your mods that will be difficult to automate away. If you have the money or spare PCs, going with packaged Linux is far easier than customizing.
I'm also spoiled by cheap hardware. I only have one computer that runs faster than 1GHz and most run sub 500MHz. I have several P1 class machines and have run KDE 3.2 on one without problem. I can see how a lab at a premier University would rather opt for the multihead solution from a noise, heat and performance standpoint. The rest of us can get buy on other people's discards.
Sooner or later, someone is going to make a deb package and the minor increase in effort will be obliterated for everyone.
Friends don't help friends install M$ junk.
We used to run whole research departments developing mathematical modelling, computational physics programmes on a single DEC VAX 11/750 with 8 MB of main memory and like 80 MB of hard disk space. It was so underutilised that astrophysics would rent out time on the darn thing to geophysics and chemistry.
For what you are talking about to be comparable to what the article is talking about (did you read it?) everybody in the department would have had to been huddled around the VAX with keyboards and monitors plugged into it directly. The whole point of this hack is to allow the school to avoid the price of terminals.
Even if you insist on running a pointy-clicky GUI, with X10 we used to run dozens of graphics terminals off of one VAX
Of course: and any idiot could set up such a system today as well. But the difference in price between cheap PCs and terminals is small enough that most people choose to just go with the extra PCs.
This article just proves that the net progress of computing is actually backwards because the computers certainly are getting bigger/faster/better more slowly than the intelligence and creativity of the users -- now they all need a GUI just to edit text and compile programs.
Yes, computing is totally moving backwards. Now there are hundreds of millions of old people, children, women and business people who use them on a daily basis, tweaking fonts in documents directly without asking a geek living in the basement to tweak their template or tell them the "code" they have to type in. Computing was so much better back in the days of VAXes when those cretins just stayed away from the computer.
The typical UTS terminal connected to Unisys 1100 and 2200 mainframes in the 80's and 90's did all text and screen editing (line/character insertions and deletions) locally and could protect, justify, and even selectively enforce input in defined character fields on the screen without communicating with the host box at all.
That's why such terminals are called "block mode" terminals, and why they behave so differently from host-dependent terminals like the VT1xx.
Some of the later UTS terminals also did graphics and I think one could even use a mouse with MAPPER on the mainframe under some configurations.
Mainframe/UNIX Bit Twiddler and long time Windows/Linux Hobbyist.
The Theorem Theorem: If If, Then Then.
To have multi-users upon a single system would need to establish the environment's restrictions; which is near impossible because the verry nature of USB devices is Plug'N'Play which is impossible to limit to any specific setting and is constrantly changing on the USB's identification chain.
It's not quite a big a problem as it appears since many modern onboard USB interfaces present as multiple root hubs.
...in Australia. Not only am I not from BYOND or Litha, I don't even recognise the names. (-:
Think of me as Joe Random Linux Consultant. I've been into computers since just before the IBM PC existed, and into Linux since you had to install Slackware from floppies.
Got time? Spend some of it coding or testing
404 Company Goes Public
I called it the Biputer, and it's just a custom video card with OS hooks.
I was also looking into using white LEDs to make a flashlight-sized projector...two of my ideas in a single week.
Even if you insist on running a pointy-clicky GUI, with X10 we used to run dozens of graphics terminals off of one VAX
Yes, but the X server was running on the terminals. The X apps running on the VAX only sent X primitives down the wire to the terminals, and the terminal bore the burdon of rendering stuff on the screen and processing kb/mouse input, turning them into events to send down the wire back to the app running on the VAX.
This article is about plugging the monitors/kbs/mices directly into the box with the X primitives processed on that box.
Think about the difference for a while before you reply to this.
You make the mistake of thinking you can educate the fundamental stupidity out of people. You can't.
"Criminal monopoly"? "redundant copy of Windoze 2000"? Har, har. How cute.
Looks like these folks run 10 stations per standard pc:
http://www.userful.com/products/1-box
Windows can support multiple monitors. Many of the computers in the UT Austin business school are hooked up to 2 20'' LCD screens. This is a feature that comes with most NVidia cards. However, you can only use one user at a time. (1 large desktop or two small ones).
Remember this is a low cost solution.
The mobo has AGP and 3 PCI slots, that's it.
Sure you could run many more serial terms off the same box at the same time but a serial term still cost more than a cheap video card and a usb hub. both require a monitor, keyboard and mouse. I would not be surprised to find out that four serial termals alone cost more than this setup.
The fact that people want GUIs kind of puts serial terms out of this market anyway.
not so hard
Anyone remember Citrix metaframe.
After that MS changed their user license, & later bought Citrix or something. Meaning one can only do the multiple network terminal thing on one Windows license if one uses a version of NT3.51 or NT4 that was purchased before the law suit.
Oh? That's interesting.
Feel free to start hacking away at it... I'm sure the guys at RealVNC would be overjoyed to include it, since that would turn a free GPL'd program into a replacement for Citrix, with it's multi-million-dollar license fees.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Environmentally this is a real savings. The biggest electrical power consumer in a computer is typically the processor. While you won't save 75% on the electricity you probably would save about 50%. And that's a big deal in a lot of places. Especially places where you have to install backup generators because the power grid is so unreliable.
The maintenance is also a lot easier. 1/4 the software installs and upgrades. And another big cost saving is desk space. Desk space is one of the most expensive types of real estate and one CPU is a lot better than 4.
The biggest downside would be if anyone wanted to futz with the machine. Turning it off would affect 4 people. I'd say that's the biggest weakness of Windows in this situation - install/reboot cycling wouldn't work at all with multiple users.
or maybe "-1 Factually Incorrect" to be less flamey about it... but so often do i see posts which are just blatanty wrong about their facts, and they maybe not be "trolls" or "flamebait".
tasty electronic music vittles
And we had access to supercomputers at Argonne, NCAR, LANL, LBL and Cornell over the ARPAnet. in the freaking early 1980's...
AND we produced beautifully typeset scholarly papers...
I find it ironic that this is modded interesting as there is nothing less interesting than listening to someone brag.
Yes, z/OS is apparently certified as a UNIX(tm), but its history is certainly not the same as Unix-like OSes, and neither is its feature set.
:-)
Also, mainframes like the Unisys A-series and 2200-series boxes use MCP and OS2200, and those are about as different from a "Unix" as one can get and still be running in text-mode.
Are you talking about things like Sun E10k's and the like (which might fit a loose definition of mainframe)?
I'm mainly just curious.
Mainframe/UNIX Bit Twiddler and long time Windows/Linux Hobbyist.
The Theorem Theorem: If If, Then Then.