Apple's Plans for P.A. Semi? Not Interested in the Chips?

[pilotError]

Actually.. As a founding member of the PowerPC alliance, Apple "owns" the instruction set (together with Freescale, IBM and Power.org). P. A. Semi only have a broad license from IBM.

I don't think Apple is buying PA for their IP portfolio and great engineers. I don't think they bought them for their products either. I think it's a mix of everything. PA is a great small company with great minds, great portfolio, great product, great promise and a bright future. They are just starting to make money and Apple might snatch them up while they are cheap. It's a sort of venture investment. They might use them as they use FileMaker Inc.. i.e. in a not so obvious way, and not in their core business. Sure they might use their processors in future Apple TVs or AirPort stations but I think PA will be a quite autonomous entity within Apple. Perhaps they will use some talent, patents and some product, but the main think is to use it as a pure investment, like they did with ARM once.. Sure, they used ARM chips in Newton but Apple earned a lot more money selling the stock at a later date.

One thing is certain.. PA is a really good company. If Apple wanted to use them for making custom made chips for any application using any architecture, PA is probably one of the best companies in the business. Certainly compared to their size.

I would venture to guess that Apple is frustrated by not being able to get what they want from off the shelf components and doesn't have enough leverage in certain industries to get folks to do things for them.

It's has to be tough to differentiate yourself in a commodity market. Apple did that with the iPhone, but that business is dog eat dog. China Bob can go and source the same hardware and copy the interface and build a reasonable knockoff in a few months. It's a matter of time before Nokia and the rest are caught up in the smart phone business.

The guys from PA have worked on a bunch of the premier chips of the past decade (or two). Obviously their experience with power management has to be very attractive to Apple. I see this as being able to develop their vision without everyone else in the industry being able to buy it 2 months later. Being the owner of that IP certainly offsets the ability of everyone else to copy it.

Does anyone know if PA Semi does RF work?

 

[kornyboy]

Wirelessly posted (iPhone: Mozilla/5.0 (iPhone; U; CPU like Mac OS X; en) AppleWebKit/420.1 (KHTML, like Gecko) Version/3.0 Mobile/4A102 Safari/419.3)

I'd love to see something out of this, but I doubt we will see anything noticeable...

So no biggie imho...

I agree. Especially if the acquisition was motivated by intelectual property and engineering talent. Here's to hoping though.

 

[kornyboy]

Wirelessly posted (iPhone: Mozilla/5.0 (iPhone; U; CPU like Mac OS X; en) AppleWebKit/420.1 (KHTML, like Gecko) Version/3.0 Mobile/4A102 Safari/419.3)

I think that it would be silly of Apple to purchase a company that has a contract with the DOD and drop it. I think they may keep that customer but I don't suspect that they would let everyone know about it. DOD stuff is held with pretty high security which could mean tight lips for people at Apple.

 

[crap freakboy]

skynet?

we may laugh now...

 

[farmboy]

It's a bit more complicated than that.

If PA Semi is a defense contractor, all Apple has to do, if they buy them, is to fulfill the contract. Basically, Apple just has to make sure the project(s) PA Semi's team is/are are currently working on continue until the end of the contract. Defense contractors get bought out all the time in the middle of contracts.

Some of these contracts are book length, and are designed so that you don't get to just stop production. If these components were indeed "process critical", most DOD contracts contain specific provisions to assure production until DOD determines product usefulness is expired. There is also a period for "technology transfer" to allow DOD to hook up with another vendor while production continues. Apple would assume those terms as well when purchasing the company. If necessary, Apple would spin off a division to handle DOD needs under the contract.

 

[RogueWarrior65]

Maybe Apple is quietly developing mil-spec Macs. The Macbook Toughbook comes to mind.

 

[morespce54]

...DOD stuff is held with pretty high security which could mean tight lips for people at Apple.

Which would be a real turnover in Apple (Steve) PR strategy...

 

[macerroneous]

Figured it out

As others have pointed out, this is about the diverging bridges: dual operating systems.

Mac OSX has taken off largely due to the fact that it runs on Intel machines so that insecure switchers can still run Windows. I believe Mac market share would not have budged were this not true. Mac OSX needs x86 to grow.

Touch OSX has taken off because of the incomparable elegance of the complete hardware/software package. Have you EVER heard a complaint that Windows Mobile won't run on iPhone? Now the incredible developer response to the SDK will cement this platform as the de-facto mobile OS. It doesn't use x86, (although it could, if Steve wants). Instead, Steve will use this to slash MSFT's throat! Now he's got a product that EVERYBODY wants and it CAN'T run Windows! AND nobody wants it to run Windows.

We will see a gradual migration of the Touch OS upward into larger and larger devices with a huge, well-nurtured developer base, and no-one will ask for MSFT compatibility! Steve has secured the fastest, coolest chip for all of his mobile, Windows Incompatible products. Expect this chip, or at least the fruit of this chip design team, to run Apple's new Tablet in the very near future. Who knows, over the next few years the Touch OSX line may overtake the Mac OSX (and Vista?). Time to don protective gear and watch Ballmer's head explode.

 

[morespce54]

Maybe Apple is quietly developing mil-spec Macs. The Macbook Toughbook comes to mind.

Weren't they selling the Power Mac as a military grade computer once?

 

[Axemantitan]

Which would be a real turn-over in Apple (Steve) PR strategy...

Mmmm, apple turnover…

 

[penhead]

No easy answer

Look; these guys are processor design gurus. They have already designed an astonishingly great low power, network centric cpu that is turning heads absolutely everywhere in the business. This processor and the technology behind it can run the OS X kernel.

Buying them is obviously not about using their products, but on using their talents and building on their experience to produce something great in the future. Jonathan Ives biggest accomplishment before joining Apple was designing toilets; this is probably a much, much better fit for Apples next, great thing than hiring Ive ever was.

The really interesting part about this is the points it gives us on what that might be, and in what direction Apple is headed. So I offer these extremely pointed and oversimplified points:

A. Apple wants to bring much more processing power to portable devices (iphone platform)
A2. The only reason for Apple to do it itself is because it is frustrated with the lack of progress in this area

B. Apple is looking to port a lot more of OS X and its Apps to portable platforms
B2. owning the platform processor makes this a lot easier

In a broad, strategic outlook, this deal makes a lot of sense.

No more time to extrapolate -- maybe later

 

[freediverdude]

I'm surprised nobody has mentioned a Stealthbook yet.

 

[macerroneous]

Let me extrapolate

Steve showed us his Touch OS, and we saw it was good. And he showed us his SDK, and we saw it was good. And soon he shall descend from on high and deliver unto us the tablets, and we shall see that they are so good that we shall be filled with the power of hardware lust and touchscreen envy. Thus spake the Steve: "Yea I sayeth unto you: my elegant hardware shall never be defiled by the corruption known as Windows, nor shall the purity of my Touch OS taste the forbidden hardware of the cloner." And to maketh it be so, the Steve brought his chip design in-house, and kept his Touch OS incompatible with x86 machines. Not to mention his holy patent protection.

 

[iPoodOverZune]

Wow, so many ideas here! Apple is good at making people's head spin.

Think of it this way: Tostitos is the only one that can make a chip shaped like a scoop. Now we get to see what Apple will do with the new smart processor guy to make the ULTIMATE chip.

More signs of growing pains: Over 200,000 iPhone application developers require considerable technical assistance. Intel is expanding multi-core processors from quads to 12's and beyond. Making compatible applications optimized for this and supporting the many Mac developers attempting to do the same means Apple needs a significantly larger staff. Enter PA Semi with 150 highly skilled employees pushing the envelope in power efficient multi-processor computers......
....Say hello to better technical support, on the phone, on-line, and within its own applications and hardware for the Macintosh platform and mobile devices such as the iPhone.

....
2.) This chip manufacturer owns the power instruction set. Apple may want control of the instruction set to control phased development/elimination of the chips? I don't know what advantage that would have...

3.) As far as talent and IPR goes, Apple probably made a good move to better understand low power and multicore processing for mobile devices. Maybe a greater alliance with Intel in the future? The manufacturer's site indicates the chip materials are very green friendly, maybe that has some benefit?

To quote the referenced EETimes article, "P.A. Semi customers were told the acquiring company was not interested in the startup's products or road map, but is buying the company for its intellectual property and engineering talent."

This makes perfect sense to me. Apple has proven their ability to switch architectures on the fly without non-techie customers so much as noticing the change. I agree with earlier posters, this acquisition is to develop true mobile devices. Smaller than MBA so you can carry it with you everywhere, bigger than iPhone so you can use it for real work without squinting. I honestly can't wait to see it, but I hope that this locking down of the developers isn't an indication of the future direction. They used the mobile phone network as an excuse for the lock-down on the iPhone and people are buying into it... not a good sign. If that's the real reason, why isn't the iPod touch an open platform?!? Let's hope a true mobile device is treated more like a computer in that respect than a mobile phone.

.........
As for the iPhone, as i mentioned above, PA Semi does not make low-power chips for small devices like smartphones or iPods. These processors use 50X more power than the iPhone does currently. Phone processors are measured in MILLIWATTS, not WATTS......
The iPhone's ARM11 processor is by NO MEANS top of the line and current. ARM's new Cortex-A8 and Cortex-A9 cores offer 2-3X the performance of the ARM11 with the same or lower power use. These chips are available as dual-core and scale to over 1Ghz, way more than adequate for any future iPhone. And these Cotex cores are the CURRENTLY available generation.
I agree with what you are saying, and it's not as much of a stretch as some may think that they could apply their knowledge to chips in Apple's iPhone or other platforms. Apparently, members of their team have worked on Intel's Itanium and ADM's Opterons.

Actually.. As a founding member of the PowerPC alliance, Apple "owns" the instruction set (together with Freescale, IBM and Power.org). P. A. Semi only have a broad license from IBM.
I don't think Apple is buying PA for their IP portfolio and great engineers. I don't think they bought them for their products either. I think it's a mix of everything. PA is a great small company with great minds, great portfolio, great product, great promise and a bright future. They are just starting to make money and Apple might snatch them up while they are cheap. It's a sort of venture investment. They might use them as they use FileMaker Inc.. i.e. in a not so obvious way, and not in their core business. Sure they might use their processors in future Apple TVs or AirPort stations but I think PA will be a quite autonomous entity within Apple. Perhaps they will use some talent, patents and some product, but the main think is to use it as a pure investment, like they did with ARM once.. Sure, they used ARM chips in Newton but Apple earned a lot more money selling the stock at a later date.
One thing is certain.. PA is a really good company. If Apple wanted to use them for making custom made chips for any application using any architecture, PA is probably one of the best companies in the business. Certainly compared to their size.

I would venture to guess that Apple is frustrated by not being able to get what they want from off the shelf components and doesn't have enough leverage in certain industries to get folks to do things for them......
The guys from PA have worked on a bunch of the premier chips of the past decade (or two). Obviously their experience with power management has to be very attractive to Apple. I see this as being able to develop their vision without everyone else in the industry being able to buy it 2 months later. Being the owner of that IP certainly offsets the ability of everyone else to copy it.
Does anyone know if PA Semi does RF work?

As others have pointed out, this is about the diverging bridges: dual operating systems.

Mac OSX has taken off largely due to the fact that it runs on Intel machines so that insecure switchers can still run Windows. I believe Mac market share would not have budged were this not true. Mac OSX needs x86 to grow.

Touch OSX has taken off because of the incomparable elegance of the complete hardware/software package. Have you EVER heard a complaint that Windows Mobile won't run on iPhone? Now the incredible developer response to the SDK will cement this platform as the de-facto mobile OS. It doesn't use x86, (although it could, if Steve wants). Instead, Steve will use this to slash MSFT's throat! Now he's got a product that EVERYBODY wants and it CAN'T run Windows! AND nobody wants it to run Windows.

We will see a gradual migration of the Touch OS upward into larger and larger devices with a huge, well-nurtured developer base, and no-one will ask for MSFT compatibility! Steve has secured the fastest, coolest chip for all of his mobile, Windows Incompatible products. Expect this chip, or at least the fruit of this chip design team, to run Apple's new Tablet in the very near future. Who knows, over the next few years the Touch OSX line may overtake the Mac OSX (and Vista?). Time to don protective gear and watch Ballmer's head explode.

Look; these guys are processor design gurus. They have already designed an astonishingly great low power, network centric cpu that is turning heads absolutely everywhere in the business. This processor and the technology behind it can run the OS X kernel.
Buying them is obviously not about using their products, but on using their talents and building on their experience to produce something great in the future. Jonathan Ives biggest accomplishment before joining Apple was designing toilets; this is probably a much, much better fit for Apples next, great thing than hiring Ive ever was.
The really interesting part about this is the points it gives us on what that might be, and in what direction Apple is headed. So I offer these extremely pointed and oversimplified points:

A. Apple wants to bring much more processing power to portable devices (iphone platform)
A2. The only reason for Apple to do it itself is because it is frustrated with the lack of progress in this area

B. Apple is looking to port a lot more of OS X and its Apps to portable platforms
B2. owning the platform processor makes this a lot easier

In a broad, strategic outlook, this deal makes a lot of sense.

No more time to extrapolate -- maybe later

Another view:

It has happened often that Apple uses their expertise/technology buy-outs/patents in a different way than many or most people would have guessed, sometimes completely stumping them. Therefore, it wont be a stretch that Intel OS X ain't going anywhere or back to PPC again nor are their product lines to take a radically different paths, but for some other reasons. Arstechnica had provided some perspective before; read about LLVM here

http://arstechnica.com/reviews/os/mac-os-x-10-5.ars/11
http://video.google.com/videoplay?docid=6189170937161128523
http://llvm.org/

This buy-out could possibly be an integration for some other purpose

Disclaimer: I am not an expert and the above is presented only for information that I came across elsewhere.

 

[hbchad]

New mil spec hardware

Apple as a defense contractor. iRaq meet iBomb. Part of iKill 08.

Scary.

 

[Oneness]

I'm surprised nobody has mentioned a Stealthbook yet.

I already have one, er, had one. I seem to have missplaced it. It is quite stealthy.

 

[mactastic1971]

I am not going to speculate why Apple aquired PA SEMI, however I have had the pleasure of working with these guys recently and they are indeed, incredibly high functioning compared to some of the other engineering dross in the valley. I'm sure whatever they turn they're collective talents to, it will only be a good thing for Mr Jobs!

 

[Sun Baked]

Wow, so many ideas here! Apple is good at making people's head spin.

Well as long as it doesn't spin like Linda Blair's head we will be fine.

 

[GroundLoop]

Some of these contracts are book length, and are designed so that you don't get to just stop production. If these components were indeed "process critical", most DOD contracts contain specific provisions to assure production until DOD determines product usefulness is expired. There is also a period for "technology transfer" to allow DOD to hook up with another vendor while production continues. Apple would assume those terms as well when purchasing the company. If necessary, Apple would spin off a division to handle DOD needs under the contract.

While this is true, if there are a truly "fab-less" corporation, then their work is entirely done. They can hand the designs over to any manufacturer whenever the DoD needs materials for spares and repairs. The only thing that PA would have to do is continue the sourcing efforts.

The real question is how many of their DoD contracts are still in the design phase?

Hickman

 

[aeryck]

At one point Apple owned most of ARM . This isn't exactly new ground for them to own a chip designer. They slowly sold off their ownership and made pounds of money. It also makes no sense to just stop a money making proposition such as PA Semi- unless you plan to make more money doing something else.

 

[-hh]

Biggd, I sware if you don't smarten up your next computer will be a Hackintosh

Well, having a "magic chip" inside the Intel-Apple box that gets used for something important (like video) would certainly put a quick kabosh on all of the Hackintosh .. and the threat from Pystar .. regardless of what happens legally with EULAs.

1.) On the DoD contracting topic, it could be that Apple wants some experience with THE largest consumer in the world. US federal government spends more than anyone else, Apple may want part of that pie. It is also good to note that the defense contracts just need to be completed as well, and Apple may not want any DoD experience.

Or it could be that Apple wouldn't mind if they got the DoD's attention in regards to how the Feds seem to have been ignoring the law (Federal Acquisition Regulations), which says that all sole-source procurements have to have a signed justification: gosh, where's the one for buying MS-Windows?.

One thing is certain.. PA is a really good company. If Apple wanted to use them for making custom made chips for any application using any architecture, PA is probably one of the best companies in the business. Certainly compared to their size.

Imagine what such smart guys can do if given the challenge of developing a chip-based "security dongle" to circumvent Hackintosh threats.

I think that it would be silly of Apple to purchase a company that has a contract with the DOD and drop it...

Some of these contracts are book length, and are designed so that you don't get to just stop production. If these components were indeed "process critical", most DOD contracts contain specific provisions to assure production until DOD determines product usefulness is expired. There is also a period for "technology transfer" to allow DOD to hook up with another vendor while production continues. Apple would assume those terms as well when purchasing the company. If necessary, Apple would spin off a division to handle DOD needs under the contract.

Correct. On this part, the Press was way off of base. Besides, once the chip is designed and its in the foundry, what work is there really for PA Semi to do except be the middleman on the contracts and reap the profits? Why would one want to shoot a cash cow?

Time to don protective gear and watch Ballmer's head explode.

Maybe that video will be the hidden Easter Egg in Mac Office 2010?


-hh

 

[Analog Kid]

As others have pointed out, this is about the diverging bridges: dual operating systems.

Mac OSX has taken off largely due to the fact that it runs on Intel machines so that insecure switchers can still run Windows. I believe Mac market share would not have budged were this not true. Mac OSX needs x86 to grow.

Touch OSX has taken off because of the incomparable elegance of the complete hardware/software package. Have you EVER heard a complaint that Windows Mobile won't run on iPhone? Now the incredible developer response to the SDK will cement this platform as the de-facto mobile OS. It doesn't use x86, (although it could, if Steve wants). Instead, Steve will use this to slash MSFT's throat! Now he's got a product that EVERYBODY wants and it CAN'T run Windows! AND nobody wants it to run Windows.

We will see a gradual migration of the Touch OS upward into larger and larger devices with a huge, well-nurtured developer base, and no-one will ask for MSFT compatibility! Steve has secured the fastest, coolest chip for all of his mobile, Windows Incompatible products. Expect this chip, or at least the fruit of this chip design team, to run Apple's new Tablet in the very near future. Who knows, over the next few years the Touch OSX line may overtake the Mac OSX (and Vista?). Time to don protective gear and watch Ballmer's head explode.

I read through almost all of your post fully expecting to reply with a "not everything is about Microsoft" response-- but somewhere in the middle there I think you converted me. I think it was the bit about having an incompatible platform that's gaining traction anyway.

It also makes no sense to just stop a money making proposition such as PA Semi- unless you plan to make more money doing something else.

Whatever tiny profits PA was making (and they had to be small if the company just sold for $250mil), Apple could easily lose in other businesses just by diluting their focus. Apple isn't, and never has been, a conglomerate.

 

[jouster]

Maybe Apple is quietly developing mil-spec Macs. The Macbook Toughbook comes to mind.

Why would they need a relatively small processor design company for that?

 

[ichi]

seen one?

I'm surprised nobody has mentioned a Stealthbook yet.

riiiight, i thought not.

 

[mdriftmeyer]

Think larger networking iron ala data switches, multimedia streaming, embedded devices for the Health Industries, Federal industries, etc.

Raise your hand if you worked at Apple or NeXT.

The uses of these chips for real-time HD decode/encoding and much more come to mind as well.

The fact a single chip can manage the following now it's clear these folks can make inroads for Apple into many markets they currently don't even have product placement:

The CONEXIUMTM coherent crossbar is an on-chip fabric that interconnects the two 64-bit superscalar CPUs, two DDR2 memory controllers, a dual-ported 2MB L2 cache, and the ENVOITM I/O subsystem to deliver on-chip symmetric multiprocessing with coherent I/O.

Dynamic power management on the PWRficient 1682M results in a worst-case total power dissipation of only 25W with both CPUs running at 2GHz and memory and I/O interfaces active at their maximum rate. Typical power dissipation ranges from 5W in portable applications to 13W in high-performance applications that require 10 Gigabit Ethernet interfacing; in power-saving modes, power dissipation drops to around 1W (typical).

The ENVOI I/O subsystem provides 24 configurable SERDES lanes for high-speed serial I/O, which may be used for PCI Express, XAUI, or SGMII interfacing in a wide range of configurations.

There are 8 PCI Express engines, supporting link widths of 1, 2, 4, 8, and 16 lanes for general peripheral connection with up to 4GB/s bandwidth per engine. The two XAUI (10 Gigabit Ethernet) and four SGMII (10/100/1000Mbit Ethernet) protocol engines each feature packet processing, including line-rate packet filtering, VLAN flow control, and TCP/IP acceleration. The two XAUI links can be optionally used as SGMII links, enabling a total of 6 SGMII links.

ENVOI includes a multichannel DMA engine to perform block copies from memory to memory, memory to device, or device to device. Optionally, computation and transformation functions may be applied to the data as it is copied; these offload functions include bulk encryption and the computation of hash, CRC, and XOR values. ENVOI includes a coherent I/O cache and adaptive prefetch unit to improve system performance.

The system controller includes power and CPU-frequency management circuits, watchdog and regular timers, an interrupt controller compatible with the OpenPIC standard, two UARTs, three SMBus channels, and a debug controller. The local/boot bus supports LPC and SPI NAND Flash, CompactFlashTM (including TrueIDE mode) and general multiplexed address/data protocols.

PWRficient provides server-class reliability. All internal SRAMs that hold modified data are protected with single-bit correct, double-bit detect ECC, and CONEXIUM transactions are parity protected. The memory controller supports a combined ECC/CRC to correct single-bit errors and detect double-bit errors and single chip failures.