Slashdot Mirror
Cheap PPC Linux Machines From IBM
ksheff writes "According to this story, IBM is planning on introducing low-end SMP servers and deskside machines based on the PPC970. The machines would be able to run Linux and AIX. A 4-way machine is expected to cost less than $3500! IBM expects a 20x increase in the number of PPC Linux servers by 2006."
Unless I'm missing something, this could definately serve as a linux workstation. The power of the new G5 with linux, what could be better?
Now if I only had a spare $3,500 to spend on it...
In linux libertas
We just heard on Slashdot that the new 3 billion plant wasn't living up to expectations, so IBM has to capitalize on this oppurtunity.
This is also a good thing for the mac community because now the G5 will get a lot more "work" done on it, because IBM will have to compete with other 64-bit manufacturers on a broader stage than just the Mac arena.
Error 407 - No creative sig found
Looks like Intel's competition is going to be coming more and more from IBM, not AMD... ...which is good news. An IA64 vs PowerPC battle benefits everyone - if one wins out over the other, nobody's stuck with old, inferior technology. Personally, I'd be perfectly happy running my software on either architecture.
;)
If x86-64 were to succeed, on the other hand, we'd have an iron clad guarantee that server rooms would always be at least 5 degrees warmer than the outside world
This isn't meant for you running Mac OS X on it. Its meant for you to run Linux on. Besides, I'd guess Mac On Linux will probably run fine. With a bit of hacking, you could probably run OS X, if you really felt like it.
I really doubt that people wanting to run Mac OS X are the targeted group here. It is, as IBM says it is, for servers and Linux desktops.
Slashdot is a waste of time. I enjoy wasting time.
Unless it's prices significantly lower than Apple's offerings, I wouldn't bet on it as a workstation. MacOSX already offers a great kernel with an even better GUI, and right now I wouldn't put money on Linux against that for a work desktop.
The server market, on the other hand will definitely get a great boost. Cheap PPC970 and 64-bit = heaven for databases, web, and app servers.
That's fine. You probably don't want or need ECC memory, the ability to compile and run 64-bit userland software, IBM's excellent VisualAge C/C++/Fortran compilers (xlc/xlC/xlf90) or a rackmount case. And that's fine - you're more than welcome to buy an Apple. I've preordered one myself! But if IBM delivers on this promise, guess what I'll be recommending to my boss that we have in the machine room. The same G5 I'll have on my desk? Not exactly. :)
Come on moderator. Explain why I'm wrong instead of slapping me with a mod point.
1: In order for PCI stuff to work with this platform, you need firmware for PPC. Guess what? The multitude of X86 cheap stuff doesnt work on these platforms. You probably pay 3-6 times what you'd normally pay for NICS and GFX cards. Apple does this all the time.
2: SMP's nice. So is PPC. But how much will you actually save if you went to this versus a new 1 or 2 Athlon setup?
3: Yes, firmware inconsistancy (86 or PPC) is a problem, but WHERE do you go for parts you need NOW? I can go to WalMart, or ripoff computer store and buy parts I need now. With this, it's atleast 1-5 days for parts. Not a good idea.
4: In my statement about Beowulf beating this, What's the cost/performance of 4 Athlon 1.5GHz with 1 gig of ram each (on 100MBps) versus one of these? I bet the name of "server" raises that cost atleast 1000$.
5: What about power consumption issues? Last I've seen the G5's, they gobbled power faster than an overclocked Athlon. If they're "always on", might want to consider power adjustements to that beo'.
Okay, last time IBM redefined the PC, they fscked up royally. But maybe this time they can successfully redefine what it means to be IBM-compatible, with a machine that rivals the Macintosh, aimed at the Linux PC market.
If these machines can be coaxed into running Darwin, maybe there will be some limited amount of binary compatibility with OS X - and people could run programs on both boxes. Compatibility is a good thing, but who says IBM has to be PC-compatible? Besides, these days the Apples are more IBM than your average PC.
I say this can only be a Good Thing.
-uso..
Dreams, dreams, don't doubt dreams, dreaming children's dreaming dreams. Sailor Moon SS
If Intel and MS proceed to only allowing signed software this could provide a nice escape path for Linux users
the PPC970 is a 64 bit cpu...
How can you compair a Athlon to a G5?
Yes If you want a cheap dual smp rig then Athlons are fine. But if you want a powerfull workstation, backed by a huge company, THis is a pretty cool idea.
im sure since IBM is the creator of a G5 they are going to be able to support them better than anything else they sell. Companies are going to see this and they will realize that these machines will be quick and easy for IBM to repair.
Oh Sure these companies can call Bob's computer warehouse, but we all know how much better IBMs support will be....
If I worked for a Multimillion dollar company I wouldnt want the computer repair guy going to walmart to pick up spare parts for their new server/workstation.....
keanmarine.com
These servers are for real sysadmins, and to run real server applications, not a GUI.
Less is more !
However, if the IBM Machine is geared towards business server use it's going to have an ATI 8MB hard mounted video card with no AGP slot.
Why put an ATI Radeon 9800 Pro in a machine that is going to show a login: prompt at best?
As a rock-in-roll Physicist once said, No matter where you go, there you are.
Uh, sorry to piss on your cornflakes, but Linux is already faster than OS X. Besides, I think that IBM knows a thing or two about systems optimization, I think that if they can't get performance-related PPC patches into the kernel proper, they'll just fork and release their own sets of patches. Remember, you're dealing with IBM here, and they don't fuck around when they release a system.
Marxism is the opiate of dumbasses
is anyone buying itanium chips? I think most of intel's fire in the server market comes from Xeon sales. The Opteron competes against the Xeon, not the itanium.
autopr0n is like, down and stuff.
Of that I have no doubt, IBM's reputation precedes them. I'm skeptical about their reasoning behind marketing Linux on ppc though. I think that's a flakey business move. There is a much more powerful OS available for that processor than Linux ppc. I think it's just a move based on the hype of the word Linux. I hate to make it sound so stupid but...
IBM already shoulders the enormous design costs of POWER4 for their high-end pSeries unix boxes. The tweaks necessary to make the PPC 970 for Apple have already been done at Apple's behest. It costs IBM very little additional R&D money to make low-end servers based on a chip they already design and manufacture for other reasons.
This makes PPC the only competitor to x86 in the commodity server space, except Sun, but Sun's product lineup grows more stale and outclassed by the day. Using IBM's compiler the 2GHz PPC970 performs approximately equivalently to a 2.8GHz p4 using icc, which is far beyond the performance offered by the in-order execution (!!) 1.05 GHz UltraSparc iii.
Having an alternative to x86 in the server space is desirable, because PPC will always have better heat dissipation and power consumption at a given level of performance. These are important considerations especially in the blade server market. In addition these are 64-bit boxes which will allow going beyond the 4GB memory barrier without using the "segmented memory" hack of the 36-bit memory addressed Xeons.
In short, this could work.
Not at the moment. While IBM can offer an extremely powerful and cheap PPC server with Linux, I seriously doubt that they can challenge Apple or MS for desktop/laptop marketshare with Linux. Yes, Linux has become more mainstream on the business sides but widespsread consumer adoption on desktops or laptops is years away.
What is keeping most consumers from using Linux on the laptop?: Office. While OpenOffice and StarOffice may be compatible with MS Office in most respects, those consumers willing to change are only now starting to make the migration. For those not interested, they will have to be dragged kicking and screaming as people in general abhor change.
Ironically, this is the same problem that MS faced years ago with Excel. In Excel's beginnings, most people used Lotus 1-2-3 as their spreadsheet. Now most people used Excel. What changed? 1) Consumers started using GUI interfaces and slowly changed their ways to work in a GUI environment. Coupled with Lotus' slow, slow creation of a GUI version of 1-2-3, it meant doom for 1-2-3. 2) MS made it easier for the transition by making many parts of Excel like 1-2-3. They even copied many of the shortcut keystrokes to make the migration easier.
Well, there's spam egg sausage and spam, that's not got much spam in it.
A computer is a tool, not a home, it's not a fashion statement. OS X gets this right. Trivial time-wasters like themes--while they may keep you from getting bored--really don't have much practical value.
/. stuff, volunteering to run some free community network centers/labs
That's bull. Mac OS X only helps "just getting work done" if you're functionally computer illiterate.
I'm a creative pro (supposedly Mac's main market) and yet I do all my photo processing (which is extensive) in Linux.
Why? Becuse it's about 100x faster in Linux due to the degree that I have been able to optimize my workflow:
1) Focus-follows-mouse, always shunned by non-Unix systems and now even by Unix systems (OS X, GNOME) saves endless point-and-click strokes (find titlebar, click to focus) when you have dozens of image windows open. Each one of these is a savings of several seconds. If you're performing hundreds or thousands of manipulations on a single task in multiple windows, that adds up to hours saved, not just minutes, on focus policy alone.
2) Fast cut/paste. Here again, the reviled behavior of X (highlight with left button, move to another window that focuses automatically, middle click where you want it to paste) saves incredible amounts of time versus the OS X or Windows behaviors (highlight with left button, hit CTRL-C, click on titlebar of destination window, click where you want to place cursor for paste, hit CTRL-V). The combination of focus-follows-mouse and keystroke-free copy/paste here again saves hours, not just minutes, when performing reptitious tasks.
3) Floating windows are my call. Once again I can keep GIMP tool windows, layer/channel dialogs, a kcalc, my conferencing window and others on top at my discretion, rather than always having to hunt down and raise some windows (by clicking on a taskbar or a dock) that I know I will need over and over again or being stuck with others on top that I don't want there and that just take up screen real estate. And when I am done with them, I can release them from forceed raise behavior.
4) Ability to turn of automatic raise when windows receive a click (done by combining focus follows mouse + titlebar-raise-only). I can have one window partially obscuring another and be working (inputting) in the "lower" (partially obscured) window while referring to one or more upper windows that partially obscure it. No need to "raise this one, look, raise that one, work, raise this one, look some more, raise that one, work some more, oh hell, just make a hardcopy, hmm, where shall we set the hardcopy..."
6) Scriptability/rapid application development. Yes, the dreaded command line shell. Many of my most intense post-production tasks (i.e. laying out posters with their captions, borders, copyright notices, anti-aliasing, interpolating to proper sizes, etc.) are database driven and processed through command line tools like ImageMagick. This allows me to do things like "makeposter 20x16 img_2525.crw" and in a single pass have the image automatically fetched from archive, converted from Canon raw, edited, captioned, matted, etc. according to a list of edits and captions I've saved ahead of time for images in my database, then sent to post-production (i.e. output). Don't tell me that there is a "makeposter" command in Mac OS X that will automatically query my database of images and perform these tasks for me, or that Apple will be willing to write me one.
[Perhaps AppleScript is capable of this stuff, perhaps not... I don't know AppleScript. But I will happily refuse to buy arguements that as well as my system works for me, I should switch to Mac OS X simply because AppleScript just "gets it right" or is "just more elegant" as scripting languages go. You'll have to give me real benefits, not techno-spiritual ones.]
7) The X-factor. I take pictures and I write prose. Those are the things I do for a living. I have other things that I do as hobbies (i.e. the
STOP . AMERICA . NOW
For the longest time I have been listing to the faint mumblings of the once so-called PowerPC Reference Platform.
/Dread
I for one, would be seriously interested in another platform besides Intel (wintel) where the hardware specs are as open as possible.
- Slower clocked RISC chips seem still to outperform the Intel line, although RISC perhaps is an outdated acronym.
- Couple this with a basically organically grown Intel MoBo based PC, with archaic Need To Have's like the floppy disk.
- Add to this a brilliant company (Apple) constantly being marginalized by having an 'incompatible' plaform.
(Apple does not want apple clones, but they sure would want to profit from the techno push the PC plaform gives to hardware)
It could end up in a faster evolving hardware platform, where software (think linux) and hardware (like Prep) evolve at the same speed (think Edsgar Dijkstra)
I, however, would not hold my breath.
Wrong. I don't know where you got this myth but it is, indeed, a myth. That's your point 3 as well - completely misguided and misinformed.
Then the Beowulf comments, now those are really clueless. Obviously you don't understand what a Beowulf cluster is. It's a protocol for building a distributed supercomputer using multiple linux boxes. You could make a Beowulf cluster with these, if you wanted to, but talking about the performance of a PPC970 versus that of a Beowulf cluster in general is simply nonsense. You're just horribly confused, or trolling.
The power consumption statement definately makes me lean towards trolling. That's marvelously clueless, totally reversing the actual relationships. The PPC lines run very cool compared to Intel and AMDs offerings, but you claim the opposite.
So yes, you definately deserve the modslapping, and another one as well. If you don't know something that's no shame, but if you don't have a clue and start spouting off whatever comes into your head pretending to be an expert, that is shameful indeed.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
32 bits gives you 4 GB of total address space. This is not enough to map a large database file alongside program and library code and variable space. So the program would have to swap different pieces of the file in and out of memory, in a non-transparent way. This is really no better than explicit read()/write(). 64 bits provides about 4 billion times as much memory space as 32 bits. Even if the machine does not have nearly this much physical RAM, the address space can be used by the VM to do file mapping and other fun stuff.
Sometimes I despair of Slashdot. Here's IBM offering us a quad processor system at a price we can afford and we go off maundering about Mac OS X. This is not about Mac OS X. It's about a quad processor machine that you can afford to put under your desk. Isn't anyone else excited about that?
I'm old enough to remember when discussions on Slashdot were well informed.