IBM Plans Collaboration On Power Architecture

TheInternet writes "According to CNET News, IBM has made a series of announcements regarding the opening-up of the Power chip architecture. The story lacks technical details, but apparently, IBM is going to divulge more information about Power/PowerPC, and expects collaboration from the industry on the future of the chip. Nick Donofrio is quoted as saying: 'We will free electronics manufacturers from the limitations of proprietary microprocessor architectures', and Red Hat and Sony are two companies listed as taking part. Power5 was also shown, as was the Blue Gene/L supercomputer, using 32 500MHz processors to achieve 128 gigaflops."

Comments

ATX PowerPC

[niko9]

Another interesting
link at the Inquirer .
Seems IBM is courting third party mobo makers to make PowerPC boards.

Their emracing linux and opening up their hardware platform. Sound Like
their getting their troops in line for THE desktop battle.

I, for one, would love to be running Debian Linux on a ATX PowerPC board. Of
course, they would have to sell enough of them to get the price down.

Good luck to 'em.

Re:ATX PowerPC

[momerath2003]

"mobo"? Is that slang for motherboarder?

Re:ATX PowerPC

[niko9]

Sorry. Here's the correct link

Re:ATX PowerPC

[stephenisu]

Yes 'mobo' is short for motherboard. Not motherbox. Cause if you interested in you your mothers box, you would be a 'mofo'

Re:ATX PowerPC

[stephenisu]

This will be great if it does start a good desktop competition. People might start writing REAL portable code, and we can finally kill x86. Seriously, x86 is going to stop scaling properly.

Re:ATX PowerPC

[homer_ca]

The PowerPC ATX motherboard has been one of the longest running vaporware products on slashdotters' wishlists. Well, not exactly vaporware. They did exist, but the problem was that the motherboard itself cost as much as a complete x86 system. It would be nice if they managed to sell them for a decent price this time. Might even be able to run MacOnLinux, but if that ever happened, Apple would definitely complain.

Re:ATX PowerPC

[Phishcast]

Seriously, x86 is going to stop scaling properly.

I remember hearing this long before x86 was as fast as it is today.

There's also no way we'll ever be able to push more than 9600bps through our dialup modems...

Re:ATX PowerPC

[Endive4Ever]

I picked up an IBM RS/6000 PREP box (a model 7248) at auction about a year and a half ago. It was a nice desktop box, with PCI and ISA slots, integrated S3 video and ethernet, and used PS/2 mouse and keyboard, standard VGA, etc. It was essentially a PC with a PowerPC processor. These boxes are fairly common and easily obtained at low cost on eBay (I paid $15 for mine at auction). PREP stands for PowerPC REference Platoform, and yes, it was capable of running NT, AIX, Linux, and NetBSD. I ran all four of them at various times while fooling with it, then installed AIX on it and sold it on eBay a few months ago.

It wasn't an ATX footprint motherboard, but about the same size and dimension, in a nice PS/2-ish IBM case.

If you want to experiment with hardware like this, search eBay for RS/6000. The 7248 boxes aren't rip-roaring speedy but they're nice enough for what you'll pay (under $50).

Re:ATX PowerPC

[Anonymous Coward]

GOING to? I'd argue that it already has. We're getting chips with higher and higher clock speeds, which is great for performance, but not so great for the electricity bill. I'd be far happier with the PC market if they'd stop ratcheting up the performance of the systems and focused instead on knocking a few dozen watts off the power needed to run the damn things.

We've a P4-based system at home, and it doesn't take long for it to make the room nice and warm. Great in winter, but not so good in summer...

Re:ATX PowerPC

[afidel]

Uh, why?
x86 as it has been implemented since the origional Pentium actually works pretty well. You use fat, complex ops which use little space in icache and convert them to svelte, fast micro-ops for fast cores. True x86 is register starved, but that's why x86-64 added a bunch of registers and cleaned up the parts of the architecture which were truely bad. The death of x86 has been predicted for several decades and I just don't see it. People have too much invested in non-open code to just dump the architecture without a REALLY good reason.

Re:ATX PowerPC

[larry bagina]

Having interest in your father's package makes you a 'homo'

Unless you're a girl. In that case, you'd be a "resident of Arkansas".

Re:ATX PowerPC

[drinkypoo]

Certainly there are enough geeks to buy them to make them profitable for one or two manufacturers, even selling them at PC-like prices. It would be interesting to find out what the G5 would cost as a commodity part. It would also set the stage for another amiga accelerator :) But seriously folks, linux users would buy the things like crazy. Gentoo on a commodity G5 is my ideal of the ideal computer, assuming I'm not tied to any given ISA like I am now, what with my tendency to do pc gaming.

Re:ATX PowerPC

[Anonymous Coward]

Right- the ceiling may have been a good deal higher than it was thought to be, but dialup still hit a ceiling with 56k. People may have been wrong in thinking x86 wouldn't get to speeds it achieved with Pentium II, but it's still showing its limits more and more.

Re:ATX PowerPC

[drinkypoo]

The fact is that the competitors to x86 which do not encompass it (as Athlon has done) have not managed to be as fast as x86. Their processors simply do not have the performance any more and when they do manage to bang out a chip faster than an x86 it ends up being insanely expensive.

There ARE people making low-power computers, such as with transmeta processors. Of course, most of those are laptops. But certainly not all of them! The real issue is that even a PC processor which consumes a lot of power generally still consumes less power than your average incandescent light bulb and it's hard to sell a CPU based on power consumption at that point. (Boy am I ready for wide-spectrum LED lighting to get here at a reasonable price.)

I have an Athlon XP with two hard drives and a bunch of big fans, I'm clearly not worried about power consumption. I also have a 19" monitor, and my PC sound system is a stereo. But in the bedroom I have an i-Opener, so there's room even in my household for both conspicuous consumption and efficient operation. So it would be nice if things were lower power. But in general it's hard to get people excited about low-power computing, they want powerful stuff. And I personally want expandability and most of the low power processors are not only slow but usually installed only in tiny computers or on boards with 1 PCI slot, or what have you. For a terminal (like my i-Opener) that's okay, but for any other purpose it doesn't get the job done.

Re:ATX PowerPC

[Hoser McMoose]

If you look at SPEC CPU scores (about the only widely used cross-platform CPU tests), 10 years ago Alpha had x86 beat solid while SPARC, PA-RISC and MIPS were doing pretty well.

Now if you look at SPEC scores x86 has the two fastest CINT scores out there with Athlon64/Opteron and P4EE/Xeon. Those two chips are also two of the top 4 chips when it comes to CFP scores, with only the IBM Power4 and Intel Itanium2 being ahead. Alpha is no longer competitive, SPARC is getting it's butt whipped and MIPS has totally failed on the high-end and PA-RISC is on life support.

All those people predicting x86's demise are clearly out of touch with reality. x86 is not only continuing to do well, but it's doing BETTER now than it ever used to!

Re:ATX PowerPC

[Seehund]

Well, not exactly vaporware. They did exist, but the problem was that the motherboard itself cost as much as a complete x86 system.

Speaking of which, here is Momentum's 970FX evaluation mobo. Only $4500 for a basic evaluation system. :)

There's of course the more humanely priced Pegasos II for 500 Euros, but it's rather underperforming for 2004 and compared to those complete x86 systems you speak of. I've heard Marvell, who supply the northbridge for the PegII, will sell a "Discovery III" northbridge for the PPC970, so things might improve. Mai Logic, who made the failed Articia S northbridge and Teron motherboards (a.k.a. "AmigaOne", "Dragon", "Boxer" and whatnot), are supposedly planning an "Articia I" (IIRC) northbridge for the 970. I guess it's likely that they'll make a Teron based on this as well.

Re:ATX PowerPC

[stilwebm]

This highlights a fact that the previous poster missed. The architectures of the Pentium 4 and Itanium families are vastly different from the "x86 architecture" in the previous generations of chips. The x86 instruction set is there and the chip presents itself in a backwords compatible way to the system. The innards are vastly evolved, however. A similar analogy is the leap from the Pentium Pro/II/III architecture or K6 architecture to the Athlon architecture.

SPARC, MIPS, and PA-RISC have had relatively minor architecture changes over the same time period. The IBM Power chips have had much better evolutionary gains.

Re:ATX PowerPC

[Anonymous Coward]

Plus in fact we *AREN'T* getting chips with higher and higher clock speeds. Intel took the better part of a year to get from 3.2GHz to 3.4GHz.

Re:ATX PowerPC

[Too Much Noise]

Actually, Alpha is being actively killed by HP as it would have wiped the floor with their new poster-child, Itanium. If I remember correctly, the more exact formulation was 'discontinued R&D, only one crippled architectural upgrade'.

As of SPARC, it looks to me like SPARC is kicking SPARC's ass - as in Fujitsu vs. Sun versions.

Finally, look at the difference between Itanium and x86/amd64-class CPUs. In theory EPIC is all fine and dandy, but ... why does it need 6MB of level 3 cache to show it? Makes one wonder how a comparable cache level would affect, say, a Xeon machine (given the obvious improvement that the extra cache brought to the P4EE CPUs). A similar story would hold for Power (see for instance this result, with 128MB off-chip L3 cache).

There might not even be a 'better way' to design a general-purpose CPU. Everybody has to optimize for something. Remember Intel bolting MMX then SSE/SSE2/SSE3 on x86 only because there was a heavy demand for it? And now, given the success of amd64, adding that as well to the Pentium4-class CPUs? x86 is not standing still. But that's the same for all the 'still alive' platforms.

Re:ATX PowerPC

[Anonymous Coward]

motherfox?

Re:ATX PowerPC

[megabeck42]

I'm fairly certain that if Alpha or MIPS had the same amount of research dollars and money put behind it, those scores would be considerably higher. The x86 architecture is not the enabling factor that gets those high scores; rather, its preventing even higher scores.

Re:ATX PowerPC

[Hoser McMoose]

Actually, Alpha is being actively killed by HP as it would have wiped the floor with their new poster-child, Itanium

Alpha has been in getting beaten to death in a long, slow and agonizing process over the past 7 years or so. This dates back to before Digital was bought out by Compaq and Intel (Intel bought a fab and a lot of Alpha technology in an odd legal settlement just before the Compaq deal... presumably the deal was made at Compaq's request). The merge with HP and the Itanium was just the final nail in the coffin for the long-suffering Alpha architecture.

In theory EPIC is all fine and dandy, but ... why does it need 6MB of level 3 cache to show it?

Why? Because it's an in-order VLIW machine. The whole idea of VLIW is to simplify the processor core to make it simple, easy and consume very little power while making room for lots of cache and memory bandwidth. Same basic argument that was made for RISC over CISC years ago. Of course, some of this became a little screwed up because the Itanium isn't really simple or easy and it consumes a LOT of power.

There might not even be a 'better way' to design a general-purpose CPU.

Well, what we'ver really seen recently is that the ISA doesn't seem to matter that much. Just take a look at the 4 companies producing x86 chips these days. We have Intel, who's P4 design breaks down x86 instructions into micro-ops that are sent through a long pipeline. We have AMD who takes x86 instructions and repackages them as macro-ops and sends them through one of several somewhat shorter pipelines. We have VIA (using the old IDT/Centaur "Winchip" design) who are doing a pretty simple CISC chip in a fairly classic way (think: Pentium on steriods.. though not too many 'roids). And finally we have Transmeta doing a hybrid software/hybrid translation of x86 instructions into VLIW instructions. That's 4 different companies using 4 rather different methods of designing microprocessors, all using the same instruction set, 5 if you count the Intel Pentium-M (though that is kinda-sorta of halfway in between the P4 and the K7/K8 design).

In the end, AMD's Athlon processor probably has a lot more in common with the IBM PowerPC 970 than it does with Intel's P4 when you look at the internals. On the outside though, the P4 and the Athlon are basically the same (same ISA, very comperable performance and power consumption).

Anyway, as you mentioned, x86 is not really standing still. The combination of SSE(1/2/3) replacing the old stack-based FPU and x86-64/AMD64 adding 64-bit addressing and doubling the number of GPRs removes two of the last major architectural limitations of x86. There really just isn't that much left that's "wrong" with the architecture for people to complain about... not that that's ever stopped people from complaining!

Re:ATX PowerPC

[phoenix_rizzen]

This is exactly the path that Intel is taking. The next generation of CPUs will be based on the Pentium-M architecture (which is nothing more than a Pentium-!!! with a few P4 technology bits thrown in). The CPUs run slower (1.7 GHz currently) and a hell of a lot cooler, but perform just as well as a P4 1000 MHz faster.

There have been at least 2 stories about this on Slashdot and Ars Technica. Check'm out.

Re:ATX PowerPC

[phoenix_rizzen]

If PREP is based on Plato, then what's the name for the platform based on Socrates? :)

Re:ATX PowerPC

[Berzelius]

I totally agree with you. I don't want a 5000 MHz Pentium (!) 4. Get me a 10 Watt computer with decent performance (~ 1GHz P3) and I would happily pay the same ammount of money for it. New uses I have: my one web and file server without the ridiculous energy bill. Currently I find it too expensive to leave my computer online all day.

Re:ATX PowerPC

[FewClues]

I actually wrote and National Semiconductor published an application note "proving" 1200 bps was the limit for a serial stream down a copper line.

Re:ATX PowerPC

[Anonymous Coward]

CHRP.

Why do you ask?

Re:ATX PowerPC

[Pharmboy]

I actually wrote and National Semiconductor published an application note "proving" 1200 bps was the limit for a serial stream down a copper line.

Then why should we listen to you now? ;)

Re:ATX PowerPC

[Pharmboy]

There might not even be a 'better way' to design a general-purpose CPU.

Transmeta has the right idea, IMHO. Design the hardware to be a simple engine, and use code to emulate hardware. (they call it Morph Code) Makes it much smaller, you can virtually 'upgrade' the cpu with code, and since it uses so much less power, you can blade the living crap out of it.

Although they have not implimented this technique to its fullest, they have a history of good coders working for them (Linus Torvald, for example) and they do have products out there.

But this design would allow a single cpu to be used for different purposes, since so much of the hardware is emulated in code.

Re:ATX PowerPC

[Pharmboy]

More importantly, we are not getting computers with that great a performance jump. Part of this is feature bloat in Windows and Linux, and part is feature bloat in the chips.

Try installing Windows 95 on a P3@1.0Ghz and compare to XP on a P4/3.0Ghz.

I just ran Siege to test my web servers (a 2.6 P4 and a dual p3-1.0, both with 1.5gb ram) The 2xp3 blew the doors off the p4, 3 times faster. Then again, a 1.0ghz p3 IS faster than a 1.4ghz p4 for everything but multimedia, from my experience.

So we are getting more and more features, that use more and more power, and cause more and more bloat, to the end that the average computer isn't much faster than it was 8 years ago for most everyday tasks.

Huh?

[Mr. Sketch]

Hey, this looks like a legit article. What's up with that? But it's April 1st according to Slashdot, since it's on GMT. Maybe that will be /.s April Fools joke, no April Fools articles?

Re:Huh?

[Trejkaz]

Umm, the article posted right before this one was an April Fools article. But it's still only one April Fools article from an entire page full of April 1 posts.

Re:Huh?

[mcc]

Maybe that will be /.s April Fools joke, no April Fools articles?

What a lovely dream.

Re:Huh?

[Jesrad]

This one is legit, the same was posted yesterday on various other websites, i.e. : on macbidouille.com (sorry, french article).

Excellent!

[Tyrdium]

So, does this mean we might see some good PowerPC emulators coming out? I remember reading that one of the main problems with writing them was the fact that the chip was quite closed, so one essentially had to reverse-engineer the entire instruction set.

Re:Excellent!

[Raindance]

As the POWER arcitecture is in reverse-ENDIAN order from the x86 arictecture, and to my knowledge, the x86 cannot switch order on the fly, I believe...

Such an emulator would necessarily be dog-slow compared to the real thing.

Comments from folks who understand more of the Arstechnica processor guides than I?

Re:Excellent!

[homer_ca]

This article explains the difficulty of emulating the PowerPC instruction set:

The 68K processor is a 32-bit processor, but its instructions are 16-bit. This means that there are at most 65,536 items in the instruction matrix that need to be translated to the X86's instruction set. This creates a relatively small loop to run through for emulation. This small number of translations also requires little memory and can even reside in the cache for optimal emulation performance.

....

Emulating the PowerPC is a far more difficult proposition, as it is a true 32-bit processor. This means that millions of instructions need to be translated. This larger 'hash' table explains why memory performance is so crucial in emulation. Moreover, such a large matrix running in the emulation loop can stymie performance.

Re:Excellent!

[idiot900]

As the POWER arcitecture is in reverse-ENDIAN order from the x86 arictecture, and to my knowledge, the x86 cannot switch order on the fly, I believe...

Such an emulator would necessarily be dog-slow compared to the real thing.

Keep in mind that this is only a constant cost, and only for reads and writes to things outside the processor (most commonly RAM). Once a value is in a register, you can leave it in the host endianness. Certainly there is a speed hit for every access, but you take a bigger hit in other things. For example: emulating the MMU, doing the math for every virtual memory access. Maybe you could leverage the host MMU for this in some way, but then good luck writing emulator code portable across architectures.

Re:Excellent!

[Endive4Ever]

Naw. It would just have to be emulated on a nicer chip, say an UltraSparc, and not on an Intel.

Re:Excellent!

[generationxyu]

The biggest problem, AFAIK, is that emulating a RISC chip on a CISC chip doesn't work very well. RISC chips usually have fewer instructions, but individual instructions are done in fewer cycles. If, for instance, one could map every instruction on the PPC to an instruction on the x86, those instructions would take longer to execute.

Re:Excellent!

[bsd4me]

My databooks are MIA at the moment or I would give more details, but the PPC ISA has versions of MOVE that do the byte swapping automagically.

Re:Excellent!

[Profane MuthaFucka]

Sounds to me like a reporter wrote some notes, then went back to them a few days later and didn't understand what they wrote.

I doubt that anyone would write the main loop of an emulator the way it was described.

Re:Excellent!

[Pharmboy]

actually the powerpc, as of the power4 gen and forward (read: 970 or Apple's new chip) is 64 bit. It can run 32 bit code in native mode, but its actually a 64 bit chip.

Hard work ahead

[EmbeddedJanitor]

Perhaps PPC can make a go of the desktop and server market, but it will be hard work to displace x86.

In mobile space everyone is just going with ARM and PPC lower poser devices don't seem to be going anywhere useful.

Re:Hard work ahead

[Teckla]

Perhaps PPC can make a go of the desktop and server market, but it will be hard work to displace x86.

PPC is already doing well in the desktop and server markets. Macs (desktop) use PowerPC CPUs, and many IBM servers use Power CPUs.

-Teckla

Re:Hard work ahead

[EmbeddedJanitor]

Mac has, what, sub-5% of the desktop and you call that well? Please don't take my pulse doctor!

One place where PPC does seem to be doing OK is in video processing/DTV set top box.

Re:Hard work ahead

[Anonymous Coward]

Oh, and the embedded market, which is 8 times the size of the desktop market.

Stupid x86 fanboys not only think that AMD Athlon is the biggest thing out there, but that it's not possible that any other architecture could be faster. Now, or ever.

Re:Hard work ahead

[Anonymous Coward]

Sorry, not only is the embedded market not using PPC chips, but they seem to prefer ARM and pre-x86 (8086/88) chips.

The only thing worse than an x86 fanboy is a Mac zealot. Overpriced hardware will never be successful. Get over it.

Re:Hard work ahead

[thisgooroo]

well, motorola (or their chip spinoff) is selling their PPCs to the embedded market

Re:Hard work ahead

[Hoser McMoose]

PowerPC chips are most definitely used in the embedded market! What the heck do you think powers Cisco routers? Or for that matter, the hardware diagnostics on Sun's Opteron-based V20z server? Hell, those two rovers doing their thing on Mars are using embedded PowerPC processors.

IBM's 4xx line of PowerPC chips are very much targetted at the embedded market, as are Motorola's MPC500 series. They are aiming rather higher than the areas where ARM competes, which in turn is already aiming towards the higher end of the embedded space, but it's still the embedded market. Stuff like Motorola 6805s and PICs make up the bulk of the volume in the embedded market, but PowerPC takes up it's chunk of the dollars from this market.

Another Link

[anzha]

From Tom's Hardware: IBM's processor plans: Build your own microchip.

OK my first thought - Open CPU

[baryon351]

Anyone else think this on the first read?

Does this sound like a possible good contender for a general purpose replacement for x86 as an "Open CPU" that would work well with F/OSS apps? One that can't be tied down with DRM in such a way that only large megacorps (I won't name them, except to call them "Microsoft" and "Intel". Oh hell I named them whoops) can end up defining what may or may not run?

I've read the general slashdot crowd clamoring for something like this that's free from central control. Does this look likely to you? Would it be a benefit if it did come about?

Re:OK my first thought - Open CPU

[FlamerPope]

Does this sound like a possible good contender for a general purpose replacement for x86 as an "Open CPU" that would work well with F/OSS apps? One that can't be tied down with DRM in such a way that only large megacorps (I won't name them, except to call them "Microsoft" and "Intel". Oh hell I named them whoops) can end up defining what may or may not run?

I'm not so sure about that. The article said that, "...IBM will still continue to control many of the most important elements of the Power architecture, such as its instruction set...". I bet that if IBM wanted to include some sort of hook for DRM, they'd still be able to do so regardless of the objection of potential collaborators.

Re:OK my first thought - Open CPU

[Trejkaz]

Last I heard, DRM wasn't going to be implemented on the CPU, but on some kind of coprocessor. In theory if the PowerPC architecture is opened, then Intel can start making them, and they can make PowerPC systems which support DRM.

Re:OK my first thought - Open CPU

[slash-tard]

As far as I know x86 chips already work well with F/OSS apps.

The instruction set is also well documented, you can even get books for free from AMD and Intel. If you had enough money Im sure AMD would make you a custom Athlon 64 but why bother?

The problem is making chips costs money. You need to design it, test via software, fab it, really test it, and repeat until you get what you want.

My first though was trying to lure some of the other "RISC" vendors like Sun, SGI, and Fujitsu (sparc), into considering the architecture. If they pool R&D it could help all of them.

Re:OK my first thought - Open CPU

[42forty-two42]

I can see it now - Gentoo PPC. They ship you a chip fab, and you spend the next few hours doping your CPU. Of course, the system supports prebuilt CPUs, but nobody actually uses them.

Re:OK my first thought - Open CPU

[Goalie_Ca]

You got it all wrong. First they send you a really big FPGA device and then let you spend hours compiling the VHDL.

Re:OK my first thought - Open CPU

[Anonymous Coward]

If DRM gets that bad, you can switch to an OS portable enough to run on non-DRM systems.

And if it's a GPLed CPU that you want, we've got those too.

Re:OK my first thought - Open CPU

[badriram]

But, I also see Microsoft being on the bandwagon if PPC does get popular. Which would also mean that we could install Windows on a mac, maybe that would also force apple to release OS X on x86 boxes.

Re:OK my first thought - Open CPU

[zhenlin]

OpenCores

Open CPUs for all!

Apple

[cuban321]

Does apple have any say in this at all? What's to stop people from building custom Macs?

I for one was just thinking about how I wish PowerPC was more open. This will give us an alternate platform to work with in case DRM/MS does kill x86.

Re:Apple

[\\]

The license that comes with OS X (and other Apple software as far as I know) says that you cannot install the software on non-Apple hardware. Probably doesn't worry them at all as far as corporate customers.

Re:Apple

[Anonymous Coward]

Who cares? Why should Apple's fear of hardware competition keep everyone from taking advantage of a better, faster platform to run Linux on?

Re:Apple

[tievape]

Does apple have any say in this at all? What's to stop people from building custom Macs?

Macs have custom ROM chips on their motherboards. So you would need to track down an Apple motherboard in order to build your own. Hard to do but not impossible

Re:Apple

[squiggleslash]

That's not been the case for a little over a decade. Macs have been running OpenFirmware since the second generation of PPC Macs.

To handle this, Mac OS has generally included a "ROM" file which contained the part of Mac OS that had previously been stored on the ROM.

Mac OS and Mac OS X does do some checks before allowing itself to be installed on a machine, making sure it really is being installed on a Mac. I believe though that all it does ultimately is query OpenFirmware.

Re:Apple

[BlowChunx]

This hasn't been true for a while, and differentiates the "Old World" macs from the "New World" macs...

Re:Apple

[squiggleslash]

Not exactly. "Old World" Macs have OpenFirmware too, an older, more limited, version. The Mac clones are all "Old World" and do not have the ROMs.

Re:Apple

[ScottEllsworth]

The work in making a clone or custom mac is not in the design of the cpu, but in getting all the rest of the hardware integrated with the OS. Add in the thrill of device drivers, and potentially having to reverse engineer any custom Apple HW, and you get a pretty much clone-free market. Opening up the cpu will not change that.

It does mean that the architecture might be used in more places than it now is. Next generation video games might move to the Power architecture if they see a benefit. Similarly, good Linux offerings will make scientists consider clustering PPC boxes rather than x86 boxes for high performance numerical computing.

If these events happen, then new support, and new understanding of the architecture will hit the streets in a way that will benefit Apple. For example, if serious money goes into optimizing gcc by someone other than Apple, releases of OS X will get better without Apple effort. Similarly, more money from other hardware purchasers gives IBM incentive to advance the architecture.

Scott

Re:Apple

[JayPee]

Interestingly, there are ways ways to install OS X on unsupported machines. Granted, these are simply legacy Mac's but perhaps with more tweaking one could eventually install OS X on those killer dual 1.4GHz G5 prototypes from Momentum Computing.

Re:Apple

[zhenlin]

PowerPC is quite open. The Mac architecture is mostly open. The difference is, for the most part, there are few suppliers of PowerPC/Mac hardware.

Then again... I think the show-stopper for custom Mac builders will be Apple's patents. (and proprietary I/O controllers)

Re:Apple

[phoenix_rizzen]

Sony's PlayStation 3 will use 3 CPUs based on the POWER/PowerPC architecture. Not PowerPC CPUs directly, but similar.

Nintendo has also made a few public nods toward IBM's POWER/PowerPC for its next-gen console. However, since there are no real, public plans for their next-gen console, this isn't really worth noting.

PowerPC is also used quite often in the embedded world due to its low heat output and low power requirements.

hmm

[User 956]

This fits with IBM's vision for spreading the 970. There's two groups: "Pervasive" and "Deep." IBM uses "pervasive" to describe a wide range wired and wireless devices powered by the 970 chips, (i.e. p2p sharing of naked petrified natalie portman pictures). "Deep" computing describes IBM's high performance technical computing products, like Blue Gene.

Opening the architecture swings the door for pervasive market penetration, indeed.

That is a great idea!!!

[Omega1045]

Now if they would just open up software in a similar way, we would have a real movement on our hands!

/. tricked you guys

[Ignorant Aardvark]

128 Gigaflops? April's Fools!

(Hey, it's started already, just look at that pigeon story).

Re:/. tricked you guys

[Anonymous Coward]

128 gigaflops from a supercomputer with 32 chips that have multiple cores.

A believable article with factual material being posted on Slashdot.

Oh my God, it is a joke!

Re:/. tricked you guys

[imsabbel]

Hm. 128GFLOPS/32CPUS*500MHZ=8 fp ops per second or 4 MACs per second. Not impossible for a bigger processor with low clockrate that focusses more on ipc.

Re:/. tricked you guys

[Anonymous Coward]

heh, that's nothing, go look up project OrCon (Organic Control) or project pigeon. 3 pigeons (for redundancy) strapped to the head of a missile trained to home in on tanks.

128 gigaflops

[Neil Blender]

WoooHooo! So someone is finally going to try to unseat Microsoft for the most failures!

G5 Hardware Specs

[morcheeba]

I'm still waiting on IBM to release it's PowerPC 970 (aka G5) hardware specs so that I can see what its high-speed bus looks like. The only thing publically available now is some fairly-broad "powerpc family" software arch documents - no electrical specs or 970-specific info.

Re:G5 Hardware Specs

[Wesley Felter]

Maybe they'll release the ApplePI docs right after Intel releases the Pentium 4 and Itanium 2 FSB protocol docs...

Re:G5 Hardware Specs

[Hoser McMoose]

Uhh.. you mean like these Pentium 4 and Itanium 2 docs?

Intel has top-notch documentation, far and away the best in the industry. The only company that comes close is AMD. IBM's public documentation for their processors is absolutely abysmal in comparison. Maybe they ahve good documentation buried somewhere in the company, but they sure don't like sharing it with anyone.

As it stands now, PowerPC is no where near as "open" as x86 is, IBM has a LONG way to go.

Re:G5 Hardware Specs

[Anonymous Coward]

And I'd like to learn more about the high speed FSB on the Pentium 4. Anyone know where that signal-level spec is located ?

Re:G5 Hardware Specs

[Wesley Felter]

Granted, Intel has better documentation.

I was just pointing out that IBM isn't the only one keeping their bus protocols a secret.

motives?

[ArbitraryConstant]

There are sound business reasons for all the big-headline stuff IBM's been doing lately, but I think they're doing it because it's the biggest, best, and cheapest PR anyone's ever heard of.

Sure they're positioning POWER/PowerPC as the only architechture that can challenge x86 in a meaningful way. Sure if they release enough of the firmware and stuff it'll probably be better for some open source stuff than x86 is. Sure they're a services company and this will put them on the back end of even more stuff.

But honestly. Everyone loves them. That's what they really need if they want to entrench themselves everywhere.

Re:motives?

[curious.corn]

Firmware? Like openfirmaware? Mods... you've been trolled... and let me feed it too... it's ATI/nVIDIA that don't make enough OF video cards... (and these you pay through your nose... ask any Apple customer)...

CHRP?

[j_cavera]

Is this the same thing/similar to what they did in the mid 90's with the "Common Hardware Reference Platform"? The idea was a common set of (PowerPC-based) chips, support chips, firmware, drivers, etc. covering everything (in theory) from desktops to embedded systems. What? You don't remember that? Worked well, didn't it....

Re:CHRP?

[Anonymous Coward]

It worked so well that all their own rs/6000's and pseries machines from 43p-150 and up is completely CHRP based, and now to a large extent the iSeries share the hardware platform.

Re:CHRP?

[suzerain]

The death (or...lack of birth) os CHRP was as much to do with Apple as anyone else. See, the idea was that the Mac would have been the biggest OS for the platform, and it was going to be the logical evolution of Apple's somewhat closed cloning requirements, at the time.

But, when Steve Jobs came back, he killed Mac cloning and, by extension, the CHRP platform. I still believe it could have evolved as a credible competitor to the Wintel architecture (MS even had NT ported to the architecture, but they, too yanked support before it could get out of the gate).

double edged sword?

[ronchie02]

This strikes me as odd. They are one of the founders of the TCPA, yet they want collaboration and the like from others? I dunno, something it that doesn't seem right :/

Re:double edged sword?

[SenseiLeNoir]

The TCPA should NOT be compared to Palladium, and other so called "trusted computing initiatives" which are really there to hadcuff you.

There has already been a huge debate on this, and IBM have created a very large rebuttal against many of the complaints.

http://www.research.ibm.com/gsal/tcpa/tcpa_rebut ta l.pdf

TCPA is a security system designed to store senstive keys in a area of the computer. This has nothing to do with DRM etc...

Re:double edged sword?

[Anonymous Coward]

I dunno, something it that doesn't seem right :/

*sigh* Standard issue Slahsdoid paranoia again...

Why don't we replace the :/ "doubtful" smiley with /. meaning "ass-backwards and one-eyed".

Hey, I'm not trolling. TCPA has to do with passwords and crypto, not Digital Restrictions Management...

Dusting off some old tech...

[Johnathon_Dough]

I wonder if Microsoft has kept that old NT version which runs on PowerPC in anyways up to date?
I know this is mostly aimed at embedded devices, but with how aggressively IBM is pushing the Power architecture, and now with this, is intel going to slowly fall behind?

If MS were to release it's server line for the Power5 or somesuch, how quickly would intel scramble to stay in Microsofts' good graces?

Besides, I would love to be able to dual-boot my powerbook back and forth between multiple OS' without emulators etc.

Re:Dusting off some old tech...

[generationxyu]

That's not too difficult.

Re:Dusting off some old tech...

[Hoser McMoose]

I wonder if Microsoft has kept that old NT version which runs on PowerPC in anyways up to date?

Rumor has it that the first XBox2 development kits ran on Macs (PowerPC) running a custom port of WinNT.

I know this is mostly aimed at embedded devices

Don't forget that WinCE has supported PowerPC chips for ages. It's not like Microsoft is incapable of supporting PPC, there's just never been any demand for it on the desktop or server side.

If MS were to release it's server line for the Power5 or somesuch, how quickly would intel scramble to stay in Microsofts' good graces?

Considering that Microsoft would probably only get about 2 customers for such an operating system, I don't think Intel would be too worried. People who buy IBM Power5 systems are looking at the seriously high-end. They are looking for a complete package of hardware, software and support, so they are not going to go off and install Windows on the thing!

Re:Dusting off some old tech...

[tillemetry]

Isn't the new XBox emulator running old NT code on Mac PPC chips? The XBox2 is supposed to have a number of G5 chips if memory serves. Think it does...

http://www.theinquirer.net/?article=14407

It seems like...

[Cheo]

IBM is actually trying to work with the Open Source community, at least to be thought of as a "nice guy". It would be very interresting to find out their real plans. Does anybody smell "conspiracy" here?

Re:It seems like...

[Daniel_Staal]

I think at this point IBM has no main conspiracy.

IBM sells two things: Hardware and Support. Open Source doesn't hurt either. In fact, it makes it more likely IBM can sell Support. (And may help sell hardware, especially if IBM provide better developer-level support, or Open Source can help out entrenched opposition...)

No conspiracy. Supporting Open Source makes IBM money. Nothing more.

Hmmmmmm

[warlockgs]

I, for one, welcome our OpenPowerPC overlords.

No, seriously, I think this is a great step. When we get to control the functionality and content of our silicon, and contribute to the specs, I think a LOT of creative people will come forward and throw out some truly awe-inspiring ideas. Look what happened with Linux, *BSD, countless GNU projects. The list goes on, people. I think this could be a stepping stone towards getting some really new chip technology on a roll.

Lets just hope this is a sincere effort on IBM's behalf.

No not really

[Anonymous Coward]

Umm, the instruction set is documented and everything. There's this PDF you can download.

The problems with emulators have to do with RISC vs CISC differences and register-rich vs register-poor architectures. I have to go, so I'm not going to go into the details here, but the general idea is this: for the specific case of emulation, it's easier to write an emulator if your host architecture is more RISCy than your emulated architecture, and it's easier to write an emulator if your host architecture has more registers than your emulated architecture.

The PowerPC has a very cluttered instruction set, but it still basically follows RISC as a philosophy-- you're still in a situation where instructions from other architectures have mostly instructions that can be broken down efficiently into a series of PowerPC instructions. Which means that efficiently assembling series' of PowerPC instructions into single instructions while emulating on more CISCy platforms is kind of hard. The PowerPC also has a whole lot of registers, and they're all general purpose so you can't play neat optimization tricks as easily as you can when emulating the Intel x86. Meanwhile, the architecture you probably want to do this emulation on-- x86-- is shit for registers.

The PPC emulation problem has to do with unfortunate conflicts between design philosophies and emulation perverse cases more than anything else.

-- Super Ugly Ultraman

Re:No not really

[afidel]

The P4's and the 970's fetch and decode pipeline phases are similar in one very important respect: both processors break down instructions in their native ISA's format into a smaller, simpler format for use inside the CPU. The P4 breaks down each x86 CISC instruction into smaller micro-ops (or "uops"), which more or less resemble the instructions on a RISC machine. Most x86 instructions decode into 2 or 3 uops, but some of the longer, more complex and rarely used instructions decode into many more uops. The 970 breaks its instructions down into what it calls "IOPs", presumably short for "internal operations". Like uops on the P4, it is these IOPs that are actually executed out-of-order by the 970's execution core. And also like uops, cracking instructions down into multiple, more atomic and more strictly defined IOPs can help the back end squeeze out some extra instruction-level parallelism (ILP) by giving it more freedom to schedule code.
link

It's true that many of the more common PPC instructions break down to only a single iop vs 2-3 uops for the P4, but the PPC is pretty far from what I would classify as a RISC chip.

Re:No not really

[Anonymous Coward]

I know. Err, or rather, the REALLY low-level chip design is a little beyond my scope, but I was aware PPC had kind of left the RISC area. That's why I tried to emphasize the important thing isn't that the PPC is RISC-- it's just that the PPC is more RISC than the P4. I think one can say that makes a notable difference, no?

-- Super Ugly Ultraman

Really good thinking

[Anonymous Coward]

The PPC doesn't seem to be particularly taking off just on its own-- either in terms of performance or in terms of public adoption-- so IBM tries to make it a public platform of sorts, hoping that will give the PPC the kick it needs to really grow.

Add to this the fact that it looks like-- uncertainly, but it looks like-- all three of the major video game consoles in the next generation (GC2, PS3 and XBox) will be using some workalike of the PPC chip, and it looks like the PPC architecture may just be to the next 10-20 years what the MIPS architecture was to the last...

-- Super Ugly Ultraman

Re:Really good thinking

[thisgooroo]

The PPC doesn't seem to be particularly taking off just on its own-- either in terms of performance or in terms of public adoption

not in the desktop market, but PPC (also ARM and MIPS) outsells x86 in the embedded market and is a pretty strong force on servers.

the problem with the desktop market is/was that when the PPC came out, the x86 was already established as the dominant platform in an environment where closed source applications got distributed in binary form only, which makes it pretty hard to get marketshare, since programs either won't run at all or extremely slow (on an emulator). and since the marketshare is so small it's almost impossible to get anyone to port his/her programs to the new architecture. combine that with microsoft's tendency and ability to sabotage such ports (e.g. by buying the company and dropping the port "for lack of demand" as they have done several times when OS/2 started to gain some popularity), and you can understand why its hard for a new architecture to take off (could that be the reason for IBM getting behind open source?)

in terms of performance

until pretty recently (until intel started to riscify the x86) RISC style CPU were performance leaders, and the G5 seems to hold up pretty good.

all three of the major video game consoles in the next generation (GC2, PS3 and XBox) will be using some workalike of the PPC chip

just an example of the embedded market

Good for Power5

[philthedrill]

This is an important step, at least for the Power5. It's immensely complex, and I think feedback from collaborators such as OS people is important when they (IBM) ask themselves if a design decision makes sense. For example, SMT adds 24% to the die area for each core (see here). Compare that with Intel's HyperThreading, which adds little area but is still complicated to verify. Getting feedback and involving other groups can help determine if design decisions/features are worthwhile.

Re:Good for Power5

[wsloand]

SMT doesn't necessarily add 24%, that's the entire jump from Power4 to Power5. I'm guessing that SMT is not the only new feature.

Re:Good for Power5

[Anonymous Coward]

Are you sure? From the Hotchips slide:

• Simultaneous multithreaded (SMT) core
• • Up to 2 virtual processors per real processor
  • 24% area growth per core for SMT
  • Natural extension to POWER4 design

Details are sketchy... but yes, SMT is probably not the only new feature.

Oh great...

[HaeMaker]

So when does SCO sue because AIX runs on PowerPC so they must be releasing SCO Intellectual Property...

why would you use PowerPC

[Anonymous Coward]

x86 and windows are established platforms and people and industry have put lot of time and effort in adopting them. why do you guys are interested in destroying this enormous value created with hardwork over time and replace with totally unproven vaporware system which only the geeks and use?

Even on April 1st, you don't play such cruel jokes.

Re:why would you use PowerPC

[GrahamCox]

Sails and oars are established propulsion methods that shipbuilders since the year dot have put a lot of time and effort in adopting them [sic]. Why are you guys with the damn screw propellors and steam engines interested in destroying this enormous value created with hard work over time and replace it with a totally unproven propulsion system that only the big shipping lines can use?

Re:why would you use PowerPC

[thisgooroo]

shame on the inventors of the automobile to destroy the market that whipmakers, saddle makers, and buggy manufacturers created with their hard work

btw, what makes you think that the PPC is vaporware? ever heard of servers and embedded systems? why do you think intel had to license ARM?

What we really need in an Open Architecture

[randall_burns]

IBM does seem to be trying here-though I doubt the PowerPC architecture is really a huge money maker for them though. Still, simply having a base of acceptence and a reasonably licensed design is only part of the equation here.

The underlying problem with major chip architectures is that they require a cast of thousands to design and implement. I aware of only one exception to that rule: Chuck Moore's Forth Chips. Chuck has acheived a lot in that area working either alone or with a few others.

The sheer complexity of modern chip design seems likely to be a problem as we try to build smaller, lower power machines-or work CPU's into micro-machine and eventually nanotechnology devices.

Re:What we really need in an Open Architecture

[Wesley Felter]

If you care about performance or power, you need to be using the latest 90nm fab processes. If you are using 90nm, a mask set costs almost $1M. If you are going to spend $1M just on manufacturing set-up costs, you might as well spend a few million designing the chip.

PowerPC vs. Power

[Anonymous Coward]

Could someone explain the difference between PowerPC and Power chips? Or, perhaps, how they are the same. A good link to info would be fine too.

Re:PowerPC vs. Power

[Anonymous Coward]

Could someone explain the difference between PowerPC and Power chips?

PowerPC chips have some apple specific optimizations in them.

Re:PowerPC vs. Power

[Hoser McMoose]

PowerPC is an instruction set that IBM uses for all of their current processors. POWER is two things: first it's an OLD ISA that they used to use long ago, and second it is a marketing name that IBM uses for their high-end PowerPC processors.

Despite popular belief, the "Power4" and "Power5" processors do NOT use the POWER ISA, they are PowerPC chips. Same as the PowerPC 970 chips used in Apple's new Macs and same as the PowerPC 405 used in the Nintendo Gamecube and Cisco routers.

Motorola also produces chips that use the PowerPC architecture, among many other ISAs.

Re:PowerPC vs. Power

[Anonymous Coward]

Nonsense.

Re:PowerPC vs. Power

[salimma]

Funny, I thought the ISA is still called POWER, not PowerPC? PowerPCs implement the POWER ISA, although until the introduction of the PPC 970 they were limited to a 32-bit subset of the full ISA.

Re:PowerPC vs. Power

[thisgooroo]

brief history: IBM developped the power architecture in the late 1980s. the first system that came out with it was the RS/6000 (1990?). at that time the CPU was a 3 chip set. around the same time, apple came to realize that the 68000 was running out of steam, and they had a look at the then available architectures and came to the conclusion that the power architecture was their best bet, so they approached IBM. at that time they had a very cozy relationship with motorola, so they insisted on motorola being part of the deal. the first result of this cooperation was a revised architecture, basically the IBM architecture with some modifications base on input from motorola and apple (i suspect this was about the first time that IBM accepted that other people understood CPU architecture as well).

the powerpc is an implementation of this architecture on a single chip. unless i am grossly mistaken, the power cpus on IBM's high end systems were multichip, at least until very recently

business is business

[randall_burns]

The way software engineers make money is continually showing a higher price/performance ratio. Microsoft and Intel are two big monopolies that eat at the pocket of every single software engineer. Replacing the WinTel monopoly with something truly open architecture is the type of thing that will be necessary to jump-start IT--which in the US is starting to become a declining industry. We need to think about how to produce $50 PC's--and just open sourcing the OS, CPU and memory design is a big step in that direction.

Re:business is business

[salesgeek]

Replacing the WinTel monopoly with something truly open architecture is the type of thing that will be necessary to jump-start IT--which in the US is starting to become a declining industry. We need to think about how to produce $50 PC's--and just open sourcing the OS, CPU and memory design is a big step in that direction.

Agreed. IBM is doing something that has needed to be done for a long time - they are quietly delivering the replacement for the 80's era application platforms. Interesting that this time IBM has commoditized the processor and os in a way that would prevent a competitor from hijacking control of the archetecture the way MS and Intel did.

Re:business is business

[randall_burns]

Well, there are _still_ some bottlenecks here. It looks like IBM isn't really Open Sourcing the chip design-just trying to license it on reasonable terms. I suspect they'll have to go further to make this _really_ take off(i.e. have a design that is available with no royalty charges). If there is a truly open source design that is really cheap, I suspect we'll see this combined with the market Sun is developing at Walmart to imply $100 PC's.

Now, I think this goes beyond a 80's application delivery platform. With $100 PC's, we can start to seriously look at things like a PC on every students desk in elementary schools-and a PC as an interface to just about every machine in creation

Re:business is business

[badriram]

yes and a completely open architecture would feed these engineers how??

Re:business is business

[salesgeek]

I think this goes beyond a 80's application delivery platform. With $100 PC's, we can start to seriously look at things like a PC on every students desk in elementary schools-and a PC as an interface to just about every machine in creation

Exactly. The 80s era platform (PC/MAC/Workstation) is too expensive to do anything other than be a general purpose device. IBM is gunning for a platform that can be applied to low computing power devices to high end super computers. Learn the archetecture once, reuse. Add custom functionality as needed. It's not Open Source though - it's a new level of customization that allows engineers to build devices that use a simmilar software environment and leverages that to drive down the costs associated with writing the code that drives advanced or additional functionality. IBM being a driver allows less concern for submarine patents and other IP hijacking a la Rambus, SCO and opportrepreneuers.

I'm sure the other microprocessor manufacturers will be rolling out simmilar plans...

The next ten years will be very interesting...

Re:business is business

[randall_burns]

If computer prices drop, then the total costs associated with hiring engineers go down. That means that certain projects become viable that aren't now and markets expand.

I think a very good case can be made that the introduction of Linux was the key factor in making the internet bloom as a business tool. Expensive AT&T Unix couldn't make that happen-inexpensive Linux could.

My contention is that the Microsoft/Intel monopolies are at this point a major drain on the salaries of software engineers world-wide--because they actively tend to keep computing more expensive than it would be otherwise.

Those monopolies may even be a threat to world peace. Really, cheap computing means things like even relatively poor countries can protect their borders by placing secure web cams everywhere along their periphery. That would in some cases remove any doubt of who the agressive nation is in the case of border disputes. I'm giving this mainly as an example of what we can do with really, really inexpensive computing.

I was there

[deadline]

I attended the event. It is pretty big news. There was a lot of interesting presentations and it is really an astounding direction for IBM. As one person from IBM put it,

IBM has seen how well the Open Source/Community model has worked for Linux. Now they believe that it will benefit the deployment of POWER derived technology.

The details are a little sketchy at this point, but Wladawsky-Berger basically said this is of the same magnitude as the decision to embrace Linux.

I think I heard the word "community" in almost every other sentence. I truly believe IBM "gets it" and is moving forward in bold direction. The people I talked to afterward were credible and excited.

There will be a longer story on ClusterWorld tomorrow. (sign up for three free issues of the magazine as well)

I saw the small "Blue Gene" system. Very cool both performance wise and thermally (32 CPUs in a table top box). I also saw the new Power blade server. Nice.

Re:I was there

[Khalid]

I taught it was an April fool ! is this fo real ?

Re:I was there

[aeoo]

I agree. When I saw this news, it totally blew me away! I can't believe there are not already most posts on this topic. IBM is basically reinventing the hardware industry with this move. This might have the same far reaching long term impact as when RMS started coding his first GNU program. I'm sure back at the time, when GNU was only a few programs, it was probably nothing and not even newsworthy, but look at what happened later. I think this is a key decision of the quivalent magnitude.

Re:I was there

[drinkypoo]

IBM is really not opening their core though in the sense of open software. Also they are only opening the 400 series PowerPC core. The chip that seems like the very best of them is the 440GX, which is a 466-600MHz chip. So this is just IBM licensing their core more widely as if they were MIPS or something. That's great, since their cores are probably better anyway :) Frankly I'm more excited about the prospect of affordable ATX PowerPC boards. It would be nice to see it actually happen :)

Only a Partial Blue Gene/L

[Betelgeuse on Ice]

That Blue Gene/L machine being shown is only a small part of the full machine they are building for LLNL. When its complete, IBM estimates that it will run at 360 TFLOPS, at a fraction of the size and power consumption of the current #1 supercomputer. Even if they miss the mark by 50% it represents a fairly significant leap in processing and power consumption. And hey, since it will only occupy 64 racks, you can just about fit one in your garage! (Nuclear reactor to power it not included...)

Re:Only a Partial Blue Gene/L

[bhima]

But the whole point of Blue Gene/L is that you would not need the Reactor!

fp

[Anonymous Coward]

first post, ballbags!

Re:fp

[Anonymous Coward]

WTF is this ball bags think going around

Re:Please READ THE WHOLE POST before modding down

[Anonymous Coward]

FTFFAQ!

But you won't find the thing about the numbers not adding up. That's because slashdot is written in Pearl and Pearl does this weird rounding thing with numbers that noone quite understands. I think it can only use numbers as low as 10 and can only increment by 10.

What About DB2 ?

[tealover]

When they open up DB2, then they'll get my attention. Collaboration on processor design is not new at all.

Horse's mouth

[samoverton]

straight from the

Another OSS advantage

[gr8_phk]

You can always recompile for whatever architecture you want. No waiting for some monopolist to decide if/when to do the porting.

no

[Anonymous Coward]

It doesn't appear that way at all.

It's a cash grab. If cash means DRM, it will have DM.

Triple boot

[Neo-Rio-101]

With that Power chip, I can forsee a future where I can build a PC that can triple boot between Windows, MacOS, and Linux!

The next question I'd have to ask myself is what possible gain I could actually get from doing that...

Re:Triple boot

[thisgooroo]

not very likely as long as windows runs only on intel cpus.

Re:Triple boot

[Exitthree]

Use Windows for web serving, Linux for Office applications, and Mac OS for games!

Re:Triple boot

[Andy_R]

...and AMD cpus, and Transmeta, and Cyrix, and Windows is included in virtual PC for Macs with IBM cpus and Motorola cpus, and there's that cut down thing for PDAs and Phones, which I think runs on ARM chips, and Hitachis and...

Re:Triple boot

[sean23007]

I know I dual-boot my webserver ... you all should too!

Very Smart Move by IBM

[Constantin]

As you all know, the costs of developing new chip architectures is escalating. However, once designed, chips can be replicated at relatively low cost (at least by comptent fabs like IBMs). So, to maximize profits despite the high upfront costs, what is one to do?

Design a kick-ass chip, sign up a lot of partners to establish street credibility, maintain processor improvement momentum, deliver chips on time, then sell as many chips as possible, of course! AMD performed in some, but certainly not all of these aspects, hence their current standing in the chip industry. Don't even get me started on the slow train wreck called Motorola.

The power architecture was always meant to be flexible, ranging from the $10,000+ quad-core uber-chip Power5's on down... So it's only logical that we will find stripped-down versions of the Power5 architecture in everything from Apple Desktops to next-gen consoles from Sony and MS.

As I see it, this is a great PR step by IBM to get some mindshare from the growing Linux camp. When you combine the incredible performance, lower prices, etc. of the 970 architecture, folks like Intel will have to take notice sooner or later, particularly when it hits their most profitable processor lines. However, Apple may not be happy to face competition in a market segment that it has had to itself for now.

As for MS and their PowerPC line of NT or whatever, who cares. If they need to make the switch, they'll find a way. In the meantime, it's the Linux/Unix folks who'll benefit the most from no longer being squeezed between SPARC and XEON pricing.

Re:Very Smart Move by IBM

[News for nerds]

>So it's only logical that we will find
>stripped-down versions of the Power5
>architecture in everything from Apple Desktops
>to next-gen consoles from Sony and MS.

For Sony's case in PS3, the Cell architecture is vastly different from PowerPC, though IBM will help it in its production or perhaps in design work with Sony and Toshiba. You should've referred to Next Nintendo console and next MS console as Power5 architecture machine.

IBM -- How many human tears for a GFLOP or a MIP?

[turtleshadow]

All this talk of opening up the Power 5 architecture just means to me that IBM is exiting the hard side of the manufacturing business and entering into being a designer and integrator; all the while and letting the poisonous drudge work to be done in various countries where public & workers rights are not well enforced.

IBM has moved assembly of Thinkpad , Netvista/Adaptiva (really Sanmima-SCI), and mid end Servers to outside the US to be a player. Now comes the egress of high end chips design work and assembly/integration.

Eventually IBM won't manufacture supercomputers in the USA at all but will design them, have some subcontractor build them then import them under the IBM name and sell them to the US GOV.

How many broken backs, destroyed environments, and tears for a cheap GFLOP or a MIP?

Nothing but a bunch of marketing BS

[Hoser McMoose]

My first thought is that this is nothing but a bunch of marketing BS, no substance at all.

Seriously, read the article, just what is IBM opening up? Answer: nothing that everyone else isn't already opening up.

The instruction set is still controlled by IBM, and while you are free to make your own PPC chips, it's not like that's anything new. Everyone is free to make their own SPARC chips as well, and from the looks of things SPARC has fewer restrictions than what IBM is proposing.

IBM will still license you the core, but that's hardly any different from what a half-dozen other chip markets will do using a wide variety of architectures.

So what does this buy you over x86? It's not like the x86 architecture is somehow *closed*. The ISA is fully documented and there are at least 4 companies producing x86 processors at the moment, possibly more if you look at the embedded space.

To me it sounds like nothing but a bunch of marketing BS trying to jump on some sort of open source bandwagon.

Reminds me of SPARC

[Ungrounded Lightning]

Is this the same thing/similar to what they did in the mid 90's with the "Common Hardware Reference Platform"? The idea was a common set of (PowerPC-based) chips, support chips, firmware, drivers, etc. covering everything (in theory) from desktops to embedded systems. What? You don't remember that? Worked well, didn't it.

Reminds me of Sun's SPARC archetecture. This was supposed to be used in everything from the top-end processors to little embedded thingies, using different performance silicon but a common instruction set. And it was supposed to be open.

You don't hear much about that either, do you?

Re:Reminds me of SPARC

[Anonymous Coward]

From http://www.openbsd.org/lyrics.html#33

"[...] This song is an allegory of the recent difficulties we went through dealing with Sun, who refused our request for documentation about their UltraSPARC III processors. We want documentation, because these are the fastest processors with a per-page eXecute bit in the MMU, needed to fully support our new W^X security feature. In the meantime, the AMD Hammer has come onto the scene, and this processor supports an eXecute bit in 64-bit mode. And it is going to be faster... [...]"

Also on http://www.deadly.org (somewhere)

Re:Reminds me of SPARC

[HanzoSpam]

Reminds me of Sun's SPARC archetecture. This was supposed to be used in everything from the top-end processors to little embedded thingies, using different performance silicon but a common instruction set. And it was supposed to be open.

You don't hear much about that either, do you?

The difference here is that Power/PowerPC is already being used in everything from the top-end processors to little embedded thingies. IBM has a pre-established market with an interest in their technology. Other than a few half-hearted SPARC based Sun clones, Sun never had anything besides high hopes for the SPARC technology.

No, not at all.

[Bensmum]

Everything that you send over the network is correct endian too. Because x86 is backwards, it has dedicated byte-swapping instructions so its quite fast to go from little endian to big and the other way around.

Lets hope they actually mean it, unlike sun.

[Bensmum]

Sparc is supposed to be an open architecture already, but that's clearly not the case. You need to keep in mind that "open" is a buzzword now, it could just be IBM looking for a way to get more attention.

Re:Lets hope they actually mean it, unlike sun.

[h8sg8s]

Clearly you don't understand the use of "open" in the SPARC case. It's arguably the most open CPU standard out there. Everyone is pissed off at Sun for this perceived lack of "openness" when in fact they practically invented the term (see SPARC, OpenOffice, Java, etc, etc, etc...) Let's see IBM open source DB2 or AIX before you start chunking stones at Sun.

Re:Lets hope they actually mean it, unlike sun.

[Bensmum]

No, its most certainly is not. Sparc is not open, you cannot get documentation on it without signing an NDA. Just because other companies who have been forced to sign NDAs can make sparc chips, doesn't make sparc open. And what does DB2 or AIX have to do with anything? When did I say IBM was open? I said open is a buzzword and IBM has been taking advantage of that fact lately.

Re:1 april

[Anonymous Coward]

72%

Sun was here

[Mr. Piddle]

Before people get too excited about this big "development", remember that SPARC was a completely open architecture since something like 1987. Sun and Fujitsu manufacture their chips independently, and there are free SPARC designs downloadable over the WWW. IIRC, the only licensing cost is if you want to use the "SPARC" logo for branding and marketing.

Check www.sparc.org for the rest.

Why would'd you use PowerPC?

[bsd4me]

I'm glad to see that people think that desktops are the only computers that people use...

I come from the embedded world. I don't know any engineer who would put an x86 in anything. On the other hand, I have put PPC in a lot of places, and there has been a trenendous amount of work to make this a stable, robust platform.

Endian-ness is NOT a problem.

[CaptainPinko]

As the POWER arcitecture is in reverse-ENDIAN

This is incorrect, see:

"PowerPC is open-endian, supporting both big-endian *AND* little-endian memory models" IBM

I found that link a long time ago from this obscure website, though you may have heard of it: /.

I suggest you check it, it's not a bad place if only there were fewer trolls and more people RTA. ;) Oh, btw apparently PowerPC is already already open:

"Aside from compatibility, one of the best things about the PowerPC architecture is that it is open: it specifies an instruction set architecture (ISA) that allows anyone to design and fabricate PowerPC-compatible processors; and source code for software modules developed in support of PowerPC is freely available." (same source)

Sometimes you don't get what you pay for

[randall_burns]

Moore has been able to place a _lot_ of cpu's on a single chip because his underlying design is so dang simple-and it is something that one(albeit VERY smart) guy can create/understand.

In the current state of the art, that may not be that big of a deal, but eventually, we are going to need to go beyond these huge, complicated facilities if we want to keep on delivering a better cost/performance ratio.

From what I can see(and I don't claim to be a chip design expert) Moore's stuff has potentially huge advantages when it comes to power consumption and miniturization. When you are trying to make really, really small designs(say stuff to fit on a robot the size of a fly), I suspect Moore's designs will be there _long_ before PowerPC will.

i'd like someone with authority to read this

[dalutong]

Some with authority please answer a few questions:

1) will this mean consumer products being developed to compete with x86 products?

2) who do you think will support this first? Will the major distributers jump on or will it be the little guys?

3) what will be the price of a open ppc system?

4) not that i'm interested... but might this allow mac clones?

5) for the developers, would you support PPC? I use debian so PPC is supported (whether this is a good or bad thing...) but it is usually behind.

6) what would having an emerging platform like ppc do wrt DRM? Could it break down the company's support DRMs plans?

actually, i could go on forever. i'll stop there.

Re:i'd like someone with authority to read this

[Gizzmonic]

By the authority vested in me by the great state of Texas, I will answer:

1)No

2)Everyone will

3)More than you could afford.

4)No

5)No

6)No

Re:i'd like someone with authority to read this

[Anonymous Coward]

I'm the manager of a seven-eleven. Here are the answers to your questions:

1. Nobody cares enough about x86 to bother competing with it. These chips will go into stuff that people big on x86 won't be interested in anyway.

2. Distributors don't jump on things, they park forklifts on them. They also park forklifts on little guys. That leaves support for these chips to come from the manufacturers of products with embedded CPUs.

3. For information on the price of an open PPC system, contact an open PPC system manufacturer. Check the phone book.

4. Not that I'm interested, but no. Mac clones are impossible because they're incompatible with Windows, period.

5. Of course we intend to support PPC for the developers. That's what the announcement is about. IBM is supporting PPC for developers in unprecedented ways. That debian PPC is "behind" is both a gross misconception and utterly irrelevant.

6. An open platform would enable the development of DRM that isn't tied to any one manufacturer's proprietary system. Unfortunately, the whole point of DRM is to lock end users into a proprietary system. So what would a new platform do to DRM? Nothing significant.

6. What the hell is "company's support DRMs plans?"

Re:i'd like someone with authority to read this

[Anonymous Coward]

In answer to all your questions: I am a member of your local authorities and i wish you'd stopped earlier. Now STFU, junior! Instead, come take a look at our Alpha factory. It is so much c00ler than this IBM thingy.

Re:i'd like someone with authority to read this

[Anonymous Coward]

1-- A mac is a consumer product right... as will the PS3 and XBox2... and the Xbox1 does have an intel chip right... this means IBM would like more people to have access to their work.

2-- The little guys will want to carve out a new market... the big guys will follow if they see a profit.

3-- The little guys will make it so they need to make a profit... so higher prices... sometime after CHARP IBM released an open sourced PPC-ATX motherboard design... it was designed to run g3 chips I *think*... anyway I only saw one company selling those boards... something like $500 no chip... no box

4-- NO, you don't have the right ROMs and what not... a mac just isn't a cute looking box, and a poweer pc chip... the OS does a lot of checking to make sure apple did make the motherboard.

5-- developers do support the ppc, and no lag really exist. the only lag you see is from the debian fokes.

6-- PPC isn't an emerging platform... and it can't do a single thing about a companys DRM plans.

Pricing Policy of IBM and Motorola

[bstadil]

The problem is not the Mobo makers and designers. The problem is the pricing policy that IBM and Motorola has/had.

The Chipset and CPU's were under the Telecom Divisions where they are used to very high margins and close to Zero price-elasticity as the equipments goes into areas where performance and reliability is paramount. (What does an extra $200 / CPU add to a $100K switch.

The price for an IBM Northbridge in 1K is around $85 compare this to Itnel Chipset that can be had for $9+-. The PowerPC itself was for a long time only available using Bumpchip technology maning you needed a very expensive socket or had to solder the cpu to the board.

In summary IBM and Moto was not interested in initial low volume low profit market.

Compare this to TI where you can buy DSP's in small quantities for almost the same as 100K price. TI understand they market needs to be developed and the pricing strategy needs to make the innovators job doable.

Re:1 april

[Anonymous Coward]

Macs don't make you gay, Mac-envy makes you gay. Kinda like how nancyboys always want to be Hollywood starlets.

SPARC is already open

[IGnatius T Foobar]

The SPARC is already an open architecture processor. It's been that way for years. Sun was the big player behind it, and certainly the best known, but the SPARC design is the closest thing there is to an "open source cpu." There's even a non-Sun organization (SPARC International) they spun off to act as a steward for the standard.

SPARC processors are made by Sun, Texas Instruments, Hitachi, and others. There's a history of all the chips made on their web site.

Dunno why they're too blind to see that this would be as good an idea for Java.

Re:SPARC is already open

[Anonymous Coward]

Here's what a Sun worker had to say about it.

Re:SPARC is already open

[Mr. Piddle]

Dunno why they're too blind to see that this would be as good an idea for Java.

??? Java Community Process ???

Java is very comparible in openness to SPARC. All the Java specs are there just as they are for SPARC. Compliance tests are there just as they are for SPARC.

Anyone is free to make their own JVM, class libraries, and Java compiler just as anyone is free to make their own SPARC core.

I see practically no difference between Sun's strategies regarding SPARC and Java. The OSS community should really be able to appreciate this.

No discussion of self-reconfiguring machines???

[Toe, The]

Oh, it's nothing big. IBM, the company that nano-wrote its name in atoms, is now creating self-evolving machines.

Does no one care? Is this not the biggest news to come along since, well, the beginning of organic life?

Did you read this? Look at the second-to-last paragraph:

"...IBM is working on future Power chips that can physically reconfigure themselves -- adding memory or accelerators, for example -- to optimize performance or power utilization for a specific application."

Re:No discussion of self-reconfiguring machines???

[Anonymous Coward]

It isn't sexy... if IBM added any sense of open source to it then the /.eers would be having a wet dream of it... i've been saying it for years... get a /. fan excited you need to do one of three things... use the word open, talk about games, or talk about free... so the best story would be about a free open source game console thats built on a self-reconfiguring PPC machine with hot swapable ram... and processors.

oh irony

[neko9]

on news.com.com page with article about IBM opening up its Power line is Intel Itanium 2 ad where is mentioned Threefold performance improvement over RISC...

i want desktop Power PC with Linux preinstalled. screw x86. its a dead end.

Assuming, of course

[Sycraft-fu]

That no assembly optimisation was used. If it was you have to either hope it's also there for your platform, or that they provided higher level code. If they did provide higher level code, you have to hope that it still performs up to spec.

While for a great many apps, performance isn't very relivant (the systems are faster than they need to be anyhow) there are still plenty where time is critical and low level optimisation can help. Games would be a good example.

Re:Assuming, of course

[TheRaven64]

Not to mention the amount of code around that assumes sizeof(int) == size(int*), and that both of these are 4 (not an issue with PowerPC, unless you use a 970, but will be soon). Oh, and of course all of the coders who think that you can get the least significant byte of an integer by simply casting the pointer to a char*, ignoring the fact that other architectures use a different byte order.

The advantage of open source is that you can fix these problems yourself (or pay someone to), you don't have to wait for the manufacturer to get around to supporting your platform of choice.

IBM, I love you

[Weaselmancer]

First open source, now open hardware. I'm...pleasantly stunned. Go Blue!

Weaselmancer

PS: At the Risk of -1 Redundant, this is a great move. I'm in embedded design, and I've discovered a few things that wound up in errata sheets later on. If I had been working on an open chip like this, I'd have worked out a fix and contributed it back to the project.

Sure beats skimming errata sheets endlessly and knowing there's nothing you can do to fix things.

Re:IBM -- How many human tears for a GFLOP or a MI

[Anonymous Coward]

and do YOU know how many people, trees, and large mamalian sea-going creatures sacrificed themselves to create that tin-foil hat you use to protect your mind from the evil capitolist brain beams?

good move...

[qtothemax]

This seems like a smart buisness move by IBM that will end up benefiting everyone. IBM probably believes that they will controll most of the manufacturing, because the only company with the fab capacity to make many is Intel, and i doubt they'd want to jeopardize thier probably higher profit margin pentium market share by making a power chip. At the same time, small fabs will be able to make the chips and keep the price reasonable. Also, i'd have to assume that IBM expects to provide most of the service and support for the chips. So the result will be a better designed chip because more people are working on it, oppritunities for small fabs to make this excelent chip, and profits for IBM because if the chip design is improved, market share will rise, with IBM reaping a good part of the profits. Bottom line in this optomistic scenario is a cheap, high preformance chip, and added profits for everyone but Intel. Everyone wins.

The problem the PowerPC has on the desktop

[MemoryDragon]

You can sum it easily up.
No early adopters who can build their own PowerPC based PCs. The PowerPC and related boards are no commodity hardware you can get for affordable prices.
Many early adopters love to build their own computers, they basically are locked out.
That basically means no early adopters, no long term mass market. What is left is only niche markets like Apple, who is not too unhappy not to have commodity hardware in their machine (high prices)

PowerPC is strong in many areas but as long as you cant get PowerPC ATX boards and processors at the same prices as their x86 counterparts, and the DRM lock in on the x86 side is not very strong yet, people still will buy x86.

Re:The problem the PowerPC has on the desktop

[BiggerIsBetter]

Right on. I've tried getting hold of PPC gear, and nobody I've talked to locally will even bother to sell it to me. You can get "development kits" from manufactures for embedded stuff if you're prepared to pay $5000 or so for 6 year old technology. I think there's a couple of $1000 PPC boards targeted at AmigaOS and Linux too, but I've never even seen a 3rd party verify they even work worth a damn.

There was a northbridge company (MAI?) that held promise to bring boards to the masses but that idea seems to have fizzled too, with their "evaluation" boards costing $3900 for 1. Nobody seem to be serious about volume sales, and setting a realistic market price for what is essentially a slow-ass PC platform. It's always developer or evaluation or sample product.

These days I don't really give a damn. x86 works fine for me, and I can get low power x86 stuff like Mini/Nano-ITX cheaply too. Any advantage of PPC is lost on me these days. PowerPC is nice, but they missed the boat. IMO, going "Open" isn't enough for IBM to reach even the geek masses. They need salable product on the shelf (or website) and they need it to be competitively priced. Anything less is just cheap talk.

Errr, no. And no.

[Jesrad]

The Power and PowerPC architectures are not big-endian or little-endian, they can run in both modes. All you need for that is a set of instructions that lets you read and write data to and from memory in both endian modes, and these architectures have them.

That, and the x86 has dedicated instructions for reversing the byte order.

The battle is in the living room

[Anonymous Coward]

The battlefield is moving to the living room. Increasingly, the consoles are going to usurp more and more tasks that are currently performed on the PC. I'm not saying that people are gonna be running spreadsheets on their PS3, but for email, IM, VOIP, surfing etc the consoles will increasingly become the device of choice. Gaming is a social activity especially online gaming. I think its inevitable that the consoles will be relied upon to deliver those functions which enhance and support the social aspect of gaming. And given IBMs wins with all three major console makers, they're well ahead in this battle.

Eh - does not compute

[DABANSHEE]

From what I remember, IBM released to the public domain their CHRP/PReP motherboard & PPC Northbridge designs a couple of years ago, leading to a number of Asian manufacturers releasing complete PPC referance boards with prices approaching the PowerPC Mac clones days. For example I think the think the MIA referance board evolved from IBM releasing their referance board to the public domain.

Anyway even if I'm wrong about the Northbridges (I'm just going by what I can recall from the Slashdot thread, etc that was on this subject a couple of years ago), if the platform has been opened to the public domain, there's nothing stopping others from making Northbridges. Plus there's no requirement to go through the expence & time to reverse engineer them. Afterall MAI's Articia PowerPC northbridge does exist (& apparently the bugs are now getting sorted)

Anyway what's stopping people from using Moto or Apple northbridges, let alone MAI's new revisions? Afterall I think Eyetech's Amiga.One boards are simply just MAI referance board designs. Terra Soft also uses MAI referance board designs in some of their products too. & of course there's the Genesi Pegasos II, apparently the cheapest new ATX PowerPC mainboard available today (embedded boards & diverted Apple spares boards are cheaper but niether are ATX) which also has Genesi's new [I]"April2"[/I] chip that they claim fixes MAI's [I]Articia[/I] northbridge bugs.

Mind you, fully assembled Pegasos G4 systems do work out about 15% dearer than undiscounted Apple sytems in Oz, plus they come with a slower G4 chip, so yes they are more expensive. But I think this comes down to the lack of X86 style economies of scale & the lack of X86 style hardware vender competition, more than the cost of some IBM Northbridge. I've also heard it is possible to install Mac OS9/X on the Eyetech & Pegasos boards, maybe via Linux though