IBM Opens Power Microprocessor Architecture

http://bits.blogs.nytimes.com/2014/04/23/ibm-opens-chip-architecture-in-strategy-of-sharing-and-self-interest/

by kjhughes • 12 years ago

160 101 12 years ago

bits.blogs.nytimes.com

I can't actually see anything on Tyan's site about a board. Elsewhere Google says it is going to be cheap. But it would be nice if it was available.

rwmj• 12 years ago

This is not actually "open" as in free and open source software.

https://en.wikipedia.org/wiki/OpenPOWER_Foundation

noahl• 12 years ago

"Open" here seems to mean that other companies can license IBM's core designs to use as part of a larger chip, and also that the instruction set itself will be public (although IBM's implementation will not be).

tormeh3• 12 years ago

That's what we really need: An ISA with a GPL-style copyleft licence. Not that being Intel's competitor is all that profitable, but the x86 market would probably be more diverse if everyone could use the ISA.

Symmetry• 12 years ago

Well, there's OpenRISC[1] and I believe SPARC is too. At least there are open source implementations including Sun's first Niagara.

[1]http://www.openrisc.net/

bri3d• 12 years ago

http://en.wikipedia.org/wiki/LEON

http://space.stackexchange.com/questions/729/why-did-the-esa...

The SPARC ISA is completely unencumbered as far as I know, and there are several copyleft implementations of the architecture that are synthesized into interesting products, often from overseas vendors (for example Navspark: https://www.indiegogo.com/projects/navspark-arduino-compatib... )

rbanffy3• 12 years ago

If we are really talking about free and open-source software, the ISA is irrelevant. Over the years, I've used Emacs on Alpha, SPARC, PA-RISC, POWER, MIPS, ARM and x86 under at least six different operating systems and it works flawlessly on all of them. I never had the chance to confirm that, but I suspect it works very well on IBM zSeries mainframes too.

dman3• 12 years ago

If they want adoption they will have to bring the price of entry level power pc systems down.

Also wonder if its a year or two too late. AMD seems to have picked ARM as their plan B strategy realising that competing with Intel on x86 is hard to do with reasonable margins. Wonder if AMD would have picked POWER if it was available as an alternative a couple of years back.

VonGuard• 12 years ago

No, they still would have picked ARM. Why? Because every mobile device with uptake uses ARM, and Intel has no play there.

dman3• 12 years ago

AMD is not playing in the mobile space with ARM. They realise they are not large enough like Intel to spend their way into the ARM market. ARM is interesting because you have Apple and Qualcomm at the high end, but you also have Mediatek and RockChip at the low end. AMD cant play the "more value for money" in ARM mobile that they do with x86 - because there are already people doing it.

mrweasel• 12 years ago

I honestly thought that the Power architecture was already open. Guess not.

wmf6• 12 years ago

It's been varying degrees of open all along; there was AIM and Exponential in the 90s and then PA Semi and AMCC ten years ago. But marketing/branding/strategy people need to stay employed.

ausjke• 12 years ago

the major power licensees were Freescale and AMCC, the rest is too small to be meaningful, if any. Both Freescale and AMCC are shifting towards ARM and not in the foundation?

With or without them, it's probably too little too late for this architecture, sigh.

MIPS, UltraSparc, and Power, are sadly fading away.

hapless6• 12 years ago

A particularly desperate move from IBM.

This is uncomfortably close to SPARC's story. Sun opened up SPARC designs in 2005. Four years later, the "Rock" project was canceled, marking an end for world-beating SPARC performance.

(Later Sun/Oracle chips have all been based on "Niagara," a low-end chip that didn't even hope to compete on performance. It was intended to be massively multicore and inexpensive, and it was at least one of those things.)

binarycrusader4• 12 years ago

I think your conclusion, "end for world-beating SPARC performance", is perhaps based on out of date information:

http://www.oracle.com/us/solutions/performance-scalability/s...

Also, if you look at the roadmap, you'll see there's even more performance improvements scheduled in the near future:

http://www.oracle.com/us/products/servers-storage/servers/sp...

hapless6• 12 years ago

Oracle's current chips are derived from the massively-multicore Niagara, not from Rock or UltraSPARC III/IV. As a result, you might notice that the benchmarks Oracle brags about are SPECjbb and TPC-C. These are benchmarks that reward integer perf and parallelism, not single-thread or floating-point performance.

On SPEC's general purpose benchmark, SPECcpu, Oracle published only three SPARC results. All three are from Fujitsu chips in re-branded Fujitsu systems. Oracle just doesn't publish results for their in-house chips. I imagine they ran the benchmarks and decided the poor results did not fit their marketing message.

http://www.spec.org/cpu2006/results/cpu2006.html

mindcrime5• 12 years ago

I keep hearing "IBM is opening POWER", but I still don't see any obvious route to dial up Asus or Tyan and order a cheap ATX motherboard that I can slap a POWER processor on, and build a machine to play with. Until that day comes, it's hard for me to get excited about POWER, as much as I might otherwise want to.

wmf6• 12 years ago

In that case you should probably avoid the Open Compute Project, too.

mindcrime5• 12 years ago

I don't really understand this comment. OCP, from what I've seen, is mostly dealing with designs for servers built using x86 and/or ARM processors, of which there are plenty of off the shelf motherboards already available. If people are building (or start building) OCP spec motherboards, so much the better. But with POWER, it's always been just short of impossible to buy an "off the shelf" motherboard with a normal form-factor, that uses normal RAM, etc., and put a box together.

And, IMO, when an architecture is that inaccessible to the hobbyists and experimenters and tinkerers, it really hinders it. I know it definitely diminishes my personal interest, and I want to be an IBM fan in many ways.

hapless6• 12 years ago

SME (Sun MicroElectronics) did exactly this in the 1990s. You could buy a standard ATX motherboard with an UltraSPARC II and PC DIMM slots on it. It was called the Ultra AXmp.

Unfortunately, the internal cost structure was such that the Ultra AXmp cost only a few percent less than a "proprietary" workstation/server from the parent company.

The moral of the story: sticking a high-margin, low-volume chip in a commodity board doesn't put it on a commodity cost basis.

mindcrime5• 12 years ago

Yeah, I was gonna say earlier, but got distracted... it would also help if one could call up IBM (or a reseller) and actually buy a POWER chip - for a reasonable price - to plug into said motherboard.

In the end, it's the entire ecosystem... it's either affordable and accessible or it isn't. It's sad that IBM keeps coming out with pretty cool hardware (Cell, POWER, etc.) that nobody can afford to get convenient access to. :-(

zvrba2• 12 years ago

What can today POWER do that neither Intel nor ARM can? What is its relevance?

mindcrime5• 12 years ago

I'm not sure it can do anything that ARM and x86 can't. But the relevance is that it provides competition to both of those, and competition is good because it keeps things moving forward. And diversity and choice are (mostly) Good Things.

cliffbean• 12 years ago

It's unclear. Diversity helps innovation, but it also adds complexity to everyone who builds on top of it. If there's room for significant innovation, it can be worth the complexity. But if there isn't, and there are many signs that there's not much left to do in the traditional general-purpose ISA space, then we're just living with a bunch of complexity for no benefit.

angersock• 12 years ago

Run a PS3.

mabhatter3• 12 years ago

The performance is in the SYSTEM architecture. Even a mid sized POWER box can address Hundreds of gigs of RAM, Petabytes of direct attached disk and rooms full of tape stored data. You can make an x86 box do that, but it's not really supported at the "plug bits and go" level. You put POWER boxes in place and don't reboot them for years. (Unless you are smart and do your HA/DR tests like a good kid)

For example a Google plays around with racks of disposable X86 boards like candy. When their app becomes "fixed" rather than growing exponentially, they'll want to move to something like POWER because it's DESIGNED to work with dozens of CPUs sharing Petabytes of attached disk easily. Not the silly kludges like Blades, SANS or iSCSI or virtual machines people play with now to hide x86 OS vendor scaling limitations.

zvrba2• 12 years ago

What you're describing sounds more like a combination of better HW (bus, components) and SW (operating system). I agree, IBM delivers superb systems.

However: are there any inherent limitations in the x64 architecture which would make it impossible to achieve the same as with POWER, if you designed it from scratch for that kind of robustness?

anentropic• 12 years ago

any sign of this transition happening at Google?

matthewmacleod2• 12 years ago

For better or worse, the ISA wars are in one sense over. There's relatively little benefit to be gained from using an architecture other than amd64 or ARM at this point; silicon has evolved to the point were the line between RISC and CISC is pretty much nonexistent, save for some minor benefits around code density etc. The benefits of using amd64 are beyond-compare experience of optimisation, commodity pricing, and legacy compatibility; the costs are negligible.

However, I'm somewhat uncomfortable with Intel's increasing and complete near-domination of the "not-low-power" CPU market. I'd love to see other high-end silicon designers start to develop amd64-compatible processors that will encourage competition in the market. AMD's struggling with it, but there's plenty of scope for looking at interesting approaches to increasing performance or reducing power use.

melling• 12 years ago

Intel has dominated for quite some time.

http://en.wikipedia.org/wiki/Intel#Market_share

In all honesty, beyond the cheap CPU market, most people have gone with Intel for a decade(s).

rbanffy3• 12 years ago

Not really.

Intel has been dominating the Windows-compatible desktop computer space and recently made significant progress in the server space but if you lump together mobile and embedded processors and units sold, you'll see is nowhere near a leadership position in overall processor sales.

tkinom• 12 years ago

ARM CPU/SOC shipped in billions of units.

Latest production ARM in Nexus 5, 7 has 4 cores, integrate GPU, VPU (video processing unit.) Run from battery power. IMO, Intel is losing CPU war to ARM.

We are also seeing Qualcomm's 64 bits 8 cores going into the next gen phone soon.

copergi• 12 years ago

You are only looking at a tiny subsection of CPUs, which happen to be the niche intel dominates. CPUs are in a lot more devices than just PCs and laptops.

seanmcdirmid2• 12 years ago

People still forget how ARM just came out of nowhere to dominate the low power space. Makes me wonder if the CPU wars aren't really over yet.

protomyth5• 12 years ago

And I guess a lot of people forget that the PowerPC is in a lot of embedded designs.

rhubarbquid• 12 years ago

ARM didn't "come out of nowhere", it's been used in embedded and mobile systems as a low power CPU for decades now. The earliest mobile use I can find is the Apple Newton in 1993. DEC was marketing their StrongARM product as a low power CPU back in 1996.

ballard• 12 years ago

They're not, never will be. Especially if there were a concerted effort to put together a GNU/Linux like open source hardware stack to get rid of the fear, uncertainty and doubt of what's inside commercial processors and chipsets. Also there needs to be more decapping of commercial chips to see what's actually in them.

willvarfar4• 12 years ago

(Mill mod)

Ahem, sorry to those suffering Mill Fatigue on HN, but any conversation about the ISA wars being over is a red rag to a bull with us! :)

Hope you like the talks on http://millcomputing.com/docs

NOTE: bulls react to movement not colour

dang• 12 years ago

Speaking for myself here – the Mill material is some of the most interesting that's appeared on HN in recent memory and I don't think we're close to being fatigued by it.

matthewmacleod2• 12 years ago

Apologies, I shouldn't have been quite so dismissive.

I mean that the ISA wars are over in the sense of multiple relatively similar architectures competing against one another for negligible gain.

There's still plenty of scope for nontraditional architectures to come in and offer something different. We've already seen that in the more general availability of general-purpose GPU programming. As an architecture nerd, I'm already super-excited by Mill.

pjc502• 12 years ago

Traditional "ISA wars" were about the cost of instruction decode & context-specific functionality. This really is over now: instruction decode costs hardly anything and context specific functionality is rarely used.

As far as I can tell, Mill have realised that the real fight is for cache; the most expensive thing you can do on a modern CPU is a data stall. So the intent is to change the execution model enough to change data access patterns and make prefetch work properly, along with more efficient use of the cache.

The next most expensive thing is a branch mispredict, and Mill attacks that strongly as well.

kabdib• 12 years ago

Agreed. I worked on PPC on the Xbox 360, and rarely had to worry about the underlying ISA, although when you did have to worry about it, things got pretty hairy and ugly, fast (think: DMA, cache coherence and other low-level stuff all ganging up to ruin your day).

You also had to worry about the ISA when you were doing something performance intensive (that's when you start taking advantage of multimedia instructions and so forth; ideally you have abstractions for this, but they'll only take you so far).

userbinator2• 12 years ago

silicon has evolved to the point were the line between RISC and CISC is pretty much nonexistent, save for some minor benefits around code density etc

There's one strategy of increasing performance on a CISC that can't really be done with RISC - making existing complex instructions execute faster. This is particularly attractive since it means all existing software gets a boost without any software-level work. So in some sense, having complex instructions that would not be fast at the time the ISA was designed is like future-proofing. And x86/amd64 still has many places where this sort of optimisation can be done.

I'd love to see other high-end silicon designers start to develop amd64-compatible processors that will encourage competition in the market.

Absolutely. It's unfortunate that many regard the ISA as being too complex, since it is within that complexity that is hidden a wealth of opportunities for optimisation.

leccine• 12 years ago

POWER (Performance Optimization With Enhanced Risc) is an extremely nice architecture compare to the x86. It's firmware system was so ahead of the BIOS disaster (EFi might change this a bit, yet). I remember loading and booting kernels from NFS because you could load the network card drivers[1]. This is 2006, EFI was not known/used widely. The CPU itself has amazing features too.[2]

I guess we are seeing another occurrence of Gresham's law applied to the technical field. "Bad architecture drives out good". In this case bad is cheaper, and this is the only thing we care about after all.

1. Open Firmware allows the system to load platform-independent drivers directly from the PCI card, improving compatibility. http://en.wikipedia.org/wiki/Open_Firmware

2. http://www.spscicomp.org/ScicomP16/presentations/Power7_Perf...

AnthonyMouse• 12 years ago

EFi might change this a bit, yet

As far as I can tell the only reason EFI even exists is NIH syndrome. Intel could have just adopted OpenFirmware for x64, and they still should.

yuhong5• 12 years ago

http://osxbook.com/book/bonus/chapter4/firmware/

TwoBit• 12 years ago

Everything seems like a nice architecture compared to x86. I do love the PowerPC instruction set though.

I'm not a hardware guy, but it always seemed like PowerPC chips underperformed others and ran as hot as hell while doing so.

protomyth5• 12 years ago

I seem to remember an article that pointed out some of the POWER addressing modes made it hard to optimize so it always would run slower than other chips. I would really love to hear what the story actually is.

Someone• 12 years ago

You cannot praise POWER for OpenFirmware. OpenFirmware came out of SUN, and allows for the development of CPU-independent drivers. IIRC, SUN needed that because they shipped workstation variants that either had a 68k CPU or a SPARC CPU.

protomyth5• 12 years ago

Can someone explain Google's involvement?

wmf6• 12 years ago

Sure, the article itself can explain it: "But analysts say the embrace of Power has two crucial advantages for Google. First, the Internet giant builds its own data centers and tweaks the technology in its server computers, and the licensing regime in the Power foundation is hacker-friendly in a way Intel’s handling of its intellectual property is not. The second advantage for Google is negotiating power..."

protomyth5• 12 years ago

I read the article and didn't find their little explanation to be that good, that's why I asked. Why PowerPC? Why not SPARC or ARM? What got them this involved?

frozenport• 12 years ago

Does anybody know the architectural advantages that POWER has over x86?

wmf6• 12 years ago

Depending on what you mean by architecture, there probably aren't many. Maybe reliability and decimal floating point. But IBM's Power server processors tend to be bigger and higher performance than Intel's; e.g. Power8 has four times the threads, four times the cache, and twice the memory bandwidth as Ivy Bridge-EP.

mabhatter3• 12 years ago

This sucks.

It means IBM doesn't want to do HARD engineering anymore. They just want to do sales and handling of stuff for high markup. This is the last really cool thing to go before the finish dismantling that former technical giant.

POWER was never about flops/watt it was about the IBM ecosystem. It's calling in for support on your hardware/software and getting elevated to somebody in the same department that will actually write the fix. It's having actual field engineers that went to the factory and learned how the stuff is put together. It was never cheap... But it was good.

ibm's POWER hardware is the last of the "old iron" lines. X86 stuff just doesn't hold a candle in the "well built" category. My company went from an "open platform" IBM solution back to old fashioned POWER hardware because it was just better built and the "open" platform just resulted in finger pointing between BLADE, SAN, LAN, and OS people... And that was IBM support with their name on the box.. When it's truly mixed support it's going to be very terrible.

It's not like AIX or System i will ever get ported to non-IBM machines for better value. It's not like Apple is gonna dust off OSX and let me build racks of POWER servers either. Betting on "Linux" is a cop out. People use Linux on POWER when they have investments in POWER iron for their mainline business apps.. ROG, COBOL, etc.. And they don't want to pay AIX/enterprise licensing for web/email servers.

This is just the "last call" to show they tried to let the world love them more... And nobody will come. Then they close it down.

DonGateley• 12 years ago

That is very cool. I know that John Cocke who invented the architecture (and was an old drinking buddy of mine) would definitely approve.

It was an outgrowth of his work developing the first optimizing compiler (Fortran) with Fran Allen which earned him a Turing Award. Good on ya, Johh, wherever and with whomever you might be sipping now. :-)

IBM Opens Power Microprocessor Architecture

Related Stories