It is shocking it has taken Intel so long to get on the band wagon. I think the faster they start making 3rd party ARM designs, the better Apple products will get.
You give your internal partners 12 months to right the ship, then you shop for outside demand for your capacity.
Remember the world was coming out of a recession for the last 3 years... hard to predict x86 sales, and you don't just 'slap together a fab,' and with Intel's chip design 5 years out, manufacturing has to commit to deliver to to that as well. The last thing Intel wanted was the inability to supply the world with x86 chips and let AMD back in the game. And once it looks like x86 demand has slacked, giving you capacity... it's not like Intel can just 'roll-in' a custom fab intake process (the legal stuff probably takes a year... you don't want Intel chip design teams lunching with the manufacturing discussing what altera just sent over via courier in the morning.).
It's less about ARM, and more about being a generic 'fab for hire.'
It is shocking it has taken Intel so long to get on the band wagon. I think the faster they start making 3rd party ARM designs, the better Apple products will get.
What is shocking? Intel has had an ARM architectural license for more than a decade. They inherited StrongARM from DEC and then designed their own ARM chip called Xscale before selling it to Marvell in 2006. In fact if anything, they were on the bandwagon way before most others were. The only thing shocking is that people are completely unaware of this.
There may be more in play than simply Intel becoming an ARM manufacturer and/or foundry.
Several of us have suggested the possibility of some of Apple's next computers containing both x86 and ARM chips.
ARM is a RISC architecture which lends itself to parallel processing (among other things).
The original 8086 architecture is CISC. When Intel evolved that to the x86 architecture, they combined a proprietary front end that translates CISC instructions to RISC instructions for execution.
As I understand it, the front end consumes a lot of power and generates a lot of heat.
So, maybe they could make a hybrid chip or a separate front-end CISC chip that interfaces the ARM chip at high-speed. The CISC chip could be used when needed to execute (translate) legacy RISC code, and idle (or off) when executing ARM code. Or, the RISC code could be translated when the code is loaded then turned off while the code is running.
It would seem that a 64-bit ARM architecture, like the A7, could facilitate the above process -- especially if it requires more RAM.
The real news should be the fact that these are not only 64 bit chips, but that Intel is opening up it's 14nm process rather than sticking 3rd parties with the 20.
Intel has already been known to manufacture FPGA's for Altera which already made use of Arm cores. It isn't surprising in the slightest that the mere advanced to full fledged chips happened.
Want to earn money & help your family without going anywhere.........I make $60h - $92h...how? I'm working online now its ur turn to make money at home......................Buzz55.?om
Never use a big word unless it a) is more apt than a small one, and b) has the intended meaning.
You're welcome!
I was thinking the same thing when I read that. If ARM has a stranglehold on mobile CPU design, how is Intel manufacturing ARM chips going to relinquish that?
This could address the yield issues TSMC is having. I like the idea of starting with making the SOC for the AppleTV. Maybe we could see the AppleTV made in the USA. I would also love to see Apple make a server that used ARM chips. They could beta test it in their own server farm and then promote the energy savings.
<<The news came as a surprise to many, given Intel's uphill battle to relinquish ARM's stranglehold on the mobile marketplace. >>
The sentence makes sense. The news is surprising because Intel is doing a 180 in strategy. They were trying to relinquish ARM's stranglehold. That strategy failed. So if you can't beat 'em, join 'em.
Never use a big word unless it a) is more apt than a small one, and b) has the intended meaning.
You're welcome!
And don't forget to throw in several French expressions in the articles, a practice which seems to be fashionable among journalists writing for English language magazines and newspapers.
This is somewhat surprising because Intel's executives alluded to a willingness to be a foundry as long as it involved Intel's own architecture. This piece about Stacey Smith echoes similar comments made by then CEO Paul Otellini to that effect.
Since then Apple has moved to production of their own SoC designs, relations with Samsung have further soured, and Intel has continued to see little penetration of their Atom SoC's. As a couple other readers pointed out, I believe Intel is now hedging their bets as to the future of computing. They are likely trying to straddle the fence as ultra low power, "good enough", devices displace traditional computing platforms for general purpose consumption. Thus they will not collapse under their own weight should they fail to deliver a mobile SoC that is price competitive and superior performance per watt on general purpose tasks. If they can't win over the mobile device (smartphones, tablets, watches, etc) markets and convince them to use Intel Architecture then generating revenue as an extremely innovative foundry is better than receiving nothing.
With this revelation I would not be surprised to see Intel secure an agreement with Apple to produce Apple designed ARM chips. This would be a fairly large shift in the mobile computing landscape and I think Intel recognizes that. There is another aspect to consider which could also be driving Intel down this path. Consider Microsoft's direction with Windows 8 which is arguably a schizophrenic operating system trying to straddle the fence as well. With this move Intel will begin to gain more experience with producing ARM based SoC's. Would it be a shock to see an Intel chip that pairs an ARM SoC with Intel Architecture. Imagine a chip that could sleep all of the IA64 cores while running the New UI apps on a Windows 9 machine; While I do acknowledge the complexity of such a suggestion, I wouldn't be surprised to see such a convergence.
And don't forget to throw in several French expressions in the articles, a practice which seems to be fashionable among journalists writing for English language magazines and newspapers.
Would it be a shock to see an Intel chip that pairs an ARM SoC with Intel Architecture. Imagine a chip that could sleep all of the IA64 cores while running the New UI apps on a Windows 9 machine; While I do acknowledge the complexity of such a suggestion, I wouldn't be surprised to see such a convergence.
Actually yes that would be shocking.
While IA64 was decently successful in the HPC market, it failed miserably to generate any interesting in the general PC market where X86 Ruled.
While Intel said that it is still committed to Itanium(it did become profitable), for intel to not consider using X86+ARM and jump straight to IA64+ARM would be incredibly out of left field.
X86 beats ARM without trying and will continue to do so for a considerable amount of time.
I believe you may have ignored the qualifier in his quotation of the article which specifically identifies the mobile market. Within that market I would disagree with your statement. I will quantify it with this quote from Forbes, "ARM, which doesn’t manufacture chips but rather licenses their blueprints to other chip makers, is present on more than 95% of the world’s smartphones...". I do think there will still be a place for x86 based systems for a considerable amount of time; however I also believe there is acceleration within the shift away from traditional PC's toward tablets and smartphones.
Thank you for the correction. I intended to mean x86-64, not the Itanium Architecture.
Was about to say...heh
Either way, yes, Intel combining ARM IP with it's own X86-64 interpretation doesn't seem that impossible I would agree(After all, AMD is already doing so for embedded applications at least). Though part of me would like IA64 to get an ARM pairing, it's not very good as far as I see, being that ARM already does parallel very well, and Itanium is basically a giant parallel chip. That, and Itanium is expensive enough on it's own/X86-64 is everywhere else mostly.
Comments
It is shocking it has taken Intel so long to get on the band wagon. I think the faster they start making 3rd party ARM designs, the better Apple products will get.
You give your internal partners 12 months to right the ship, then you shop for outside demand for your capacity.
Remember the world was coming out of a recession for the last 3 years... hard to predict x86 sales, and you don't just 'slap together a fab,' and with Intel's chip design 5 years out, manufacturing has to commit to deliver to to that as well. The last thing Intel wanted was the inability to supply the world with x86 chips and let AMD back in the game. And once it looks like x86 demand has slacked, giving you capacity... it's not like Intel can just 'roll-in' a custom fab intake process (the legal stuff probably takes a year... you don't want Intel chip design teams lunching with the manufacturing discussing what altera just sent over via courier in the morning.).
It's less about ARM, and more about being a generic 'fab for hire.'
It is shocking it has taken Intel so long to get on the band wagon. I think the faster they start making 3rd party ARM designs, the better Apple products will get.
What is shocking? Intel has had an ARM architectural license for more than a decade. They inherited StrongARM from DEC and then designed their own ARM chip called Xscale before selling it to Marvell in 2006. In fact if anything, they were on the bandwagon way before most others were. The only thing shocking is that people are completely unaware of this.
Several of us have suggested the possibility of some of Apple's next computers containing both x86 and ARM chips.
ARM is a RISC architecture which lends itself to parallel processing (among other things).
The original 8086 architecture is CISC. When Intel evolved that to the x86 architecture, they combined a proprietary front end that translates CISC instructions to RISC instructions for execution.
As I understand it, the front end consumes a lot of power and generates a lot of heat.
So, maybe they could make a hybrid chip or a separate front-end CISC chip that interfaces the ARM chip at high-speed. The CISC chip could be used when needed to execute (translate) legacy RISC code, and idle (or off) when executing ARM code. Or, the RISC code could be translated when the code is loaded then turned off while the code is running.
It would seem that a 64-bit ARM architecture, like the A7, could facilitate the above process -- especially if it requires more RAM.
because the criminals are about to get their comeuppance
The real news should be the fact that these are not only 64 bit chips, but that Intel is opening up it's 14nm process rather than sticking 3rd parties with the 20.
Intel has already been known to manufacture FPGA's for Altera which already made use of Arm cores. It isn't surprising in the slightest that the mere advanced to full fledged chips happened.
HP to Ship 64-Bit ARM-Based Moonshot Server in 2014
http://www.eweek.com/servers/hp-to-ship-64-bit-arm-based-moonshot-server-in-2014.html
and this:
New Blockbuster Memory Chip may Kick A7 into another Orbit
http://www.patentlyapple.com/patently-apple/2013/10/new-blockbuster-memory-chip-may-kick-a7-into-another-orbit.html
Want to earn money & help your family without going anywhere.........I make $60h - $92h...how? I'm working online now its ur turn to make money at home...................... Buzz55.?om
Just open Home tab nd make money
I was thinking the same thing when I read that. If ARM has a stranglehold on mobile CPU design, how is Intel manufacturing ARM chips going to relinquish that?
I like the idea of starting with making the SOC for the AppleTV.
Maybe we could see the AppleTV made in the USA.
I would also love to see Apple make a server that used ARM chips. They could beta test it in their own server farm and then promote the energy savings.
The sentence makes sense.
The news is surprising because Intel is doing a 180 in strategy.
They were trying to relinquish ARM's stranglehold.
That strategy failed.
So if you can't beat 'em, join 'em.
X86 beats ARM without trying and will continue to do so for a considerable amount of time.
> battle to relinquish ARM's stranglehold <
Never use a big word unless it a) is more apt than a small one, and b) has the intended meaning.
You're welcome!
And don't forget to throw in several French expressions in the articles, a practice which seems to be fashionable among journalists writing for English language magazines and newspapers.
This is somewhat surprising because Intel's executives alluded to a willingness to be a foundry as long as it involved Intel's own architecture. This piece about Stacey Smith echoes similar comments made by then CEO Paul Otellini to that effect.
http://arstechnica.com/apple/2011/05/intel-exec-throws-fuel-on-the-intel-to-fab-apple-chips-rumor-fire/
http://allthingsd.com/20110607/arm-twisting-intel-to-fab-chips-for-apple/
Since then Apple has moved to production of their own SoC designs, relations with Samsung have further soured, and Intel has continued to see little penetration of their Atom SoC's. As a couple other readers pointed out, I believe Intel is now hedging their bets as to the future of computing. They are likely trying to straddle the fence as ultra low power, "good enough", devices displace traditional computing platforms for general purpose consumption. Thus they will not collapse under their own weight should they fail to deliver a mobile SoC that is price competitive and superior performance per watt on general purpose tasks. If they can't win over the mobile device (smartphones, tablets, watches, etc) markets and convince them to use Intel Architecture then generating revenue as an extremely innovative foundry is better than receiving nothing.
With this revelation I would not be surprised to see Intel secure an agreement with Apple to produce Apple designed ARM chips. This would be a fairly large shift in the mobile computing landscape and I think Intel recognizes that. There is another aspect to consider which could also be driving Intel down this path. Consider Microsoft's direction with Windows 8 which is arguably a schizophrenic operating system trying to straddle the fence as well. With this move Intel will begin to gain more experience with producing ARM based SoC's. Would it be a shock to see an Intel chip that pairs an ARM SoC with Intel Architecture. Imagine a chip that could sleep all of the IA64 cores while running the New UI apps on a Windows 9 machine; While I do acknowledge the complexity of such a suggestion, I wouldn't be surprised to see such a convergence.
And don't forget to throw in several French expressions in the articles, a practice which seems to be fashionable among journalists writing for English language magazines and newspapers.
That is so passé...
Would it be a shock to see an Intel chip that pairs an ARM SoC with Intel Architecture. Imagine a chip that could sleep all of the IA64 cores while running the New UI apps on a Windows 9 machine; While I do acknowledge the complexity of such a suggestion, I wouldn't be surprised to see such a convergence.
Actually yes that would be shocking.
While IA64 was decently successful in the HPC market, it failed miserably to generate any interesting in the general PC market where X86 Ruled.
While Intel said that it is still committed to Itanium(it did become profitable), for intel to not consider using X86+ARM and jump straight to IA64+ARM would be incredibly out of left field.
X86 beats ARM without trying and will continue to do so for a considerable amount of time.
I believe you may have ignored the qualifier in his quotation of the article which specifically identifies the mobile market. Within that market I would disagree with your statement. I will quantify it with this quote from Forbes, "ARM, which doesn’t manufacture chips but rather licenses their blueprints to other chip makers, is present on more than 95% of the world’s smartphones...". I do think there will still be a place for x86 based systems for a considerable amount of time; however I also believe there is acceleration within the shift away from traditional PC's toward tablets and smartphones.
http://www.forbes.com/sites/parmyolson/2013/09/20/as-gadgets-shrink-arm-still-reigns-as-processor-king/
Thank you for the correction. I intended to mean x86-64, not the Itanium Architecture.
Thank you for the correction. I intended to mean x86-64, not the Itanium Architecture.
Was about to say...heh
Either way, yes, Intel combining ARM IP with it's own X86-64 interpretation doesn't seem that impossible I would agree(After all, AMD is already doing so for embedded applications at least). Though part of me would like IA64 to get an ARM pairing, it's not very good as far as I see, being that ARM already does parallel very well, and Itanium is basically a giant parallel chip. That, and Itanium is expensive enough on it's own/X86-64 is everywhere else mostly.