Apple's future iPhones may leverage ARM's v7 Cortex designs
With Apple's first proprietary system-on-a-chip designs still rumored to be under active development, the iPhone maker may turn to ARM's standard v7 Cortex designs for multi-touch devices it plans to launch this year before building off the platform next year.
A recent job posting by Apple dug up by MacRumors seeks and iPhone developer with extensive knowledge of ARM's embedded processors, including their NEON vector unit.
As the rumor site points out, NEON is an extended instruction set similar to Intel's SSE or PowerPC's AltiVec which can accelerate multimedia applications. It's also said to be the marketing term for the most recent version of those extensions specifically for the ARMv7 Cortex processor, which is the successor to the ARMv6 chip currently employed in the second-generation of both the iPhone and iPod touch.
One theory behind Apple's apparent interest in NEON is that the extensions may help facilitate a number of multimedia-intensive operations due to start turning up in the iPhone this year, such as video recording, video processing, and rudimentary video editing.
Recently, the Wall Street Journal cited people familiar with Apple's ongoing efforts to develop its own embedded processors as saying those designs are unlikely to materialize until sometime next year. Still, mounting evidence suggest that this year's iPhone and iPod touch will sport a materially distinct architecture from their predecessors.
Given that Apple's latest job listing is dated this month and third-generation iPhones are already believed to be nearing production, it's possible the company's plan is to adopt Cortex chips based off ARM's reference designs this year, before making its own proprietary additions to those schematics for devices that will launch next year.
A recent job posting by Apple dug up by MacRumors seeks and iPhone developer with extensive knowledge of ARM's embedded processors, including their NEON vector unit.
As the rumor site points out, NEON is an extended instruction set similar to Intel's SSE or PowerPC's AltiVec which can accelerate multimedia applications. It's also said to be the marketing term for the most recent version of those extensions specifically for the ARMv7 Cortex processor, which is the successor to the ARMv6 chip currently employed in the second-generation of both the iPhone and iPod touch.
One theory behind Apple's apparent interest in NEON is that the extensions may help facilitate a number of multimedia-intensive operations due to start turning up in the iPhone this year, such as video recording, video processing, and rudimentary video editing.
Recently, the Wall Street Journal cited people familiar with Apple's ongoing efforts to develop its own embedded processors as saying those designs are unlikely to materialize until sometime next year. Still, mounting evidence suggest that this year's iPhone and iPod touch will sport a materially distinct architecture from their predecessors.
Given that Apple's latest job listing is dated this month and third-generation iPhones are already believed to be nearing production, it's possible the company's plan is to adopt Cortex chips based off ARM's reference designs this year, before making its own proprietary additions to those schematics for devices that will launch next year.
Comments
Ok, on a serious note...
As they say to the end of the article - it isn't likely that the new iPhone will have this technology. Still... its feasible, considering the turnover time of the iPod touch, that they may be able to pioneer this technology in that.
After all, the first iPod Touch was an afterthought after the iPhone's success was it not, with a very short run-up time?
Also considering the rampup of CPU in the iPod Touch, and Nike + inclusion, perhaps we are seeing a trend where the iPod touch will start as more of a "road - test" of technologies that will make their way into the iPhone, but tested on a simpler platform in the iPod - a platform that specifically lacks other more phone-relevant technologies, like GPS and cellular radios & basebands, and also lacks many of the hardware-specific software complexities of the iPhone.
Spelling error: "turing up in the iPhone"...
That's right, I'm predicting multiple devices...
Second post, but this time for a reason.
Spelling error: "turing up in the iPhone"...
Heh heh, must have been a Freudian slip..... http://en.wikipedia.org/wiki/Alan_Turing
I do believe Apple would be looking at Cortex chips for a beefier ARM based product due to ship this year. This would allow them to lead right into their own custom designs in 2010 with version 2 product.
Right now off the shelf ARM parts are fine and give Apple time to complete their design process with PA Semi.
As for the iPhone this year would appear to be handled by a faster ARM 6 product, faster Imagination graphics and updated wireless product and other tech.
A driver for faster Cortex processing would, IMO, be driven by HD video out options and more NAND memory which would allow for larger phones for video recording/playback.
First post
Ok, on a serious note...
As they say to the end of the article - it isn't likely that the new iPhone will have this technology. Still... its feasible, considering the turnover time of the iPod touch, that they may be able to pioneer this technology in that.
After all, the first iPod Touch was an afterthought after the iPhone's success was it not, with a very short run-up time?
Also considering the rampup of CPU in the iPod Touch, and Nike + inclusion, perhaps we are seeing a trend where the iPod touch will start as more of a "road - test" of technologies that will make their way into the iPhone, but tested on a simpler platform in the iPod - a platform that specifically lacks other more phone-relevant technologies, like GPS and cellular radios & basebands, and also lacks many of the hardware-specific software complexities of the iPhone.
I think the earliest we may see ARMv7 is with the Touch. After all, Apple is pitting it a super gaming console. We know it has a faster clocked CPU and they have pushed the games pretty hard. I don?t think that will stop anytime soon.
I would not be surprised is the upcoming event also showed some new HW attachments for these handhelds. Once the D-pads comes I think we?ll see an even bigger congregation of major game developers working their way on iPhone OS X.
Given that Apple's latest job listing is dated this month and third-generation iPhones are already believed to be nearing production, it's possible the company's plan is to adopt Cortex chips based off ARM's reference designs this year, before making its own proprietary additions to those schematics for devices that will launch next year.
[ View this article at AppleInsider.com ]
If Apple is like any company I've worked at, it's highly likely that they already have such expertise in-house and have already worked the key issues, and made the key decisions and designs. Most hiring is for new (non-executive) people who step in behind the leaders to continue, fine-tune, and expand the work already being done, not for these new people to start up work in a new area, especially as those leaders are looking to move on to the next challenge and the next products.
Thus, I believe these June-July iPhones will already use the Cortex v7 chips.
I think the earliest we may see ARMv7 is with the Touch. After all, Apple is pitting it a super gaming console. We know it has a faster clocked CPU and they have pushed the games pretty hard. I don’t think that will stop anytime soon.
I would not be surprised is the upcoming event also showed some new HW attachments for these handhelds. Once the D-pads comes I think we’ll see an even bigger congregation of major game developers working their way on iPhone OS X.
I will be shocked if Apple is not using a Cortex based processor most likely an A8 with a Power VR graphics core. If they hope to compete with the PRE on power consumption/ Performance then they need the A8. The A8 design is available from Samsung if a PA Semi version is not already being built. http://www.samsung.com/global/busine...ure_200902.pdf
I will be shocked if Apple is not using a Cortex based processor most likely an A8 with a Power VR graphics core. If they hope to compete with the PRE on power consumption/ Performance then they need the A8. The A8 design is available from Samsung if a PA Semi version is not already being built. http://www.samsung.com/global/busine...ure_200902.pdf
1) I was going to ask you if it was used in any product yet, which usually needs to happen before Apple with their excessive needs pre model type, but I decided to look it up first. The Palm Pre and a whole slow of other are using the ARM v7 Cortex-A8 already or will use it shortly. That is good news for this being the base of the next time.
2) Any chance that they could use Cortex-A9 or some other chip instead?
3) Why do you think it’ll be PowerVR over Nvidia’s Tegra?
PS: Welcome to the AI forums and keep up the informative posts.
1)
3) Why do you think it?ll be PowerVR over Nvidia?s Tegra?
Apple buys 3.6% of Imagination
The Tegra is a nice chip but Apple has been investing in Imagination (who also was on the steering group for OpenCL)
Apple buys 3.6% of Imagination
The Tegra is a nice chip but Apple has been investing in Imagination (who also was on the steering group for OpenCL)
Even with Apple?s new relationship with Nvidia that seems pretty conclusive.
1) I was going to ask you if it was used in any product yet, which usually needs to happen before Apple with their excessive needs pre model type, but I decided to look it up first. The Palm Pre and a whole slow of other are using the ARM v7 Cortex-A8 already or will use it shortly. That is good news for this being the base of the next time.
2) Any chance that they could use Cortex-A9 or some other chip instead?
3) Why do you think it’ll be PowerVR over Nvidia’s Tegra?
Two reasons for the PowerVR graphics Apple invested in a stock offering from Imagination buying 3.6% of the company which I think was a cash infusion to keep things rolling and Imagination announced that Apple was using their IP.
Good question on the A9. The design was made available in Oct 07 and it is normally 2 years before it would appear in an end user device. Apple bought PA Semi in April 08 which would have given them about a year of design time assuming the production for Iphone 3rd version was in the April 09 for the first run and then May for the main run. This would allow test/assembly for a late June launch. The timeline for a A9 would be very aggressive, but since Apple controls the design they could probably trim a couple months from the normal timeline because they don't have all the back and forth when dealing with an external customer and no need to market ect. I would love to see the A9 as it would tie in nicely with the Snow Leopard Open CL and Grand Central Work.
Anyone have any ideas what Apple wanted with PA Semi? Surely, the most common answer is using their staff?s ARM knowledge to create better chips with ARM CPUs, but what specifically could they do with it that isn?t already done? Make it considerably smaller and more compact, like Apple?s Mac MoBos?
I think that's the million dollar question.
Custom design is nice but the expense of chip design pretty much means a custom design has got to be significantly better than "off the shelf"
Is Apple going to be able to add value features that no one else can ape (albeit at lower quality) with turnkey SoC?
I know not enough about design nor can I even postulate on what unique features Apple could add. I guess that's what is going to make the mobile market for Apple exciting again. We're back to that "anything can happen" with Apple products.
Let the speculation begin!
Anyone have any ideas what Apple wanted with PA Semi? Surely, the most common answer is using their staff?s ARM knowledge to create better chips with ARM CPUs, but what specifically could they do with it that isn?t already done? Make it considerably smaller and more compact, like Apple?s Mac MoBos?
PA SEMI's expertise was performance per watt. From the beginning, P.A. focused on developing a family of products that is extensible and scalable. This required a lot of upfront investment and design time to ensure that their efforts would be reusable. They hope that by extensively reusing their IP, that they can tape out and productize their designs much more rapidly; the stated goal was a 3 month tape out, once the first core had been developed. This shorter development time would allow them to address more markets than would be otherwise possible. The PWRficient family incorporates three elements:
The PA6T processor core
The CONEXIUM interchange
The ENVOI I/O system and offload engines
Moreover, each element can be scaled in a variety of dimensions to accommodate various markets and price points.
I'm just hoping. Because really the software doesn't matter, there's an app store. I'm sure if apple doesn't release it, some third party will find a way. BTW if we all use the same app store account only one of us would have to buy the apps
The PA6T processor core
The CONEXIUM interchange
The ENVOI I/O system and offload engines
Moreover, each element can be scaled in a variety of dimensions to accommodate various markets and price points.
I think Apple's going to have to do just that. It only makes sense that Apple looks at their current lineup and says "We own PA Semi and have licenses for Imagination graphics, ARM architectural license (rumored). Why are we buying Marvel chips for the Airport Extreme and Time Capsule? Why are we using Pentium M and Nvidia graphics for the Apple TV? Etc.
Many of Apple's current non Macintosh products can be run using their own chip design to help amortize the cost of development and add features.
I expect an Apple to field a storage device or home server appliance eventually. Today we see NAS powered by Freescale, ARM and Atom chips deliver great performance for the home. I see an Apple designed chip being used here effectively as well.
I'm starting to think with all the clamoring about what the iPhone (software) can and can not do. That maybe it's been a big blinder to the HW side of things. The thing about it is, it could just get faster. You know, like computers, we don't worry about software, we want faster processors, more ram, faster buss speeds, etc. Also, with the iPhone being how old now, 4? I'm sure even battery tech as improved so much that they could boost speed above 800mhz and see no significant loss in battery usage. That would make for some much more powerful apps. You know someday I'll have CAD software for my iPod
I'm just hoping. Because really the software doesn't matter, there's an app store. I'm sure if apple doesn't release it, some third party will find a way. BTW if we all use the same app store account only one of us would have to buy the apps
The iPhone is almost 2 years old since the first one was sold. The CPU is the same from the original model and it is underclocked from 620Mhz to about 400MHz. ARMv7 starts at 600MHz so we?ll see if they use a faster CPU that they can underclock to save power, so they can use it next year and so they up the clock speed as they make the software more power efficient. They have a history of this and it seemed to work out well.
There is newer battery tech out but I am not sure how well tested it is, the Lithium Polymer Ion batteries are still very common for these devices. I hope they do use a better battery tech but I?m not expecting it.
I expect an Apple to field a storage device or home server appliance eventually. Today we see NAS powered by Freescale, ARM and Atom chips deliver great performance for the home. I see an Apple designed chip being used here effectively as well.
I see a higher consumer market for a home NAS than I do for a tablet device. A real NAS from Apple would be great.
I'm starting to think with all the clamoring about what the iPhone (software) can and can not do. That maybe it's been a big blinder to the HW side of things. The thing about it is, it could just get faster. You know, like computers, we don't worry about software, we want faster processors, more ram, faster buss speeds, etc. Also, with the iPhone being how old now, 4? I'm sure even battery tech as improved so much that they could boost speed above 800mhz and see no significant loss in battery usage. That would make for some much more powerful apps. You know someday I'll have CAD software for my iPod
I'm just hoping. Because really the software doesn't matter, there's an app store. I'm sure if apple doesn't release it, some third party will find a way. BTW if we all use the same app store account only one of us would have to buy the apps
Actually battery technology moves very slow. You can expect about a 4% improvement per year. Compared to a doubling every two years for Silicon ICs. In my mind the key is less the battery and more on the ICs and display that use the battery. My guess is if you compare a Iphone 3G running the current software to the next generation Iphone running the Iphone OS 3.0 software you will have noticeable speed and battery life improvements. It won't be just a few percent. Anyone who has used the 4.0 beta of Safari with the Java Script performance improvements via the Web Kit Squirrel Fish knows that the next version of Safari on the Iphone will be much faster.