Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Wirelessly posted (Mozilla/5.0 (iPod; U; CPU iPhone OS 4_3_3 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8J2 Safari/6533.18.5)

dlewis23 said:
Why would they want to run anything that fast? Apple is mainly interested in battery life with passable performance. The fact that they go shun Flash on the platform illustrates that.

Why wouldn't they want anything to run faster? Shunning flash does illustrate that. With anything the ipad 2 shows that Apple is very much so interested in performance. With in a year they go from an A8 to an A9. The next logical step is an A15. It gives them more performance and better battery life.

Also if they do go with a screen thats 2048x1536 they are going to need more performance much more then can be squeezed out of a Cortex A9.

If they want to take more away from the Desktop/Laptop they are going to need more performance.

I think the main reason would be gaming.
 
:mad:

Your avatar made me punch my screen...:(

Tee-Hee :D

Which could mean:

* there'll be no refresh of the iPad2 later this year as a new iPad2 SKU (2s).

Or it could mean:

* Apple will not release an iPad3 later this year with a vastly more upgraded 4 to follow mid-next year.

Your premise is so vague you're almost certain to be right, even if you're actually wrong.



The analysts have taught me well. ;)
 
There is still a good chance of an upgraded iPad2 with enhanced resolution and perhaps more memory.

Nope, there is no chance of this, never was. This is just a recycled rumor from when Gruber was breathing his own farts earlier this year.

It's senseless, but great link bait so it keeps getting repeated.

Doubled resolution is going to wait for A6 to make sure there is no slow downs when pushing 4x the pixels.
 
re 2012 release

oh cmon tsmc 2012 is too long to wait for an a6 - get your fab and test floor moving

"Don't Let Me Wait Too Long..." - george harrison
 
Gaming...........

It's driven the industry for the past two decades and forced tech to push hard every year.

Apple has never been there for gaming, but now, for once Apple has it's one chance to become THE portable gaming champion.

If they don't do it, then others will.
 
Wirelessly posted (Mozilla/5.0 (iPod; U; CPU iPhone OS 4_3_3 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8J2 Safari/6533.18.5)

Piggie said:
Gaming...........

It's driven the industry for the past two decades and forced tech to push hard every year.

Apple has never been there for gaming, but now, for once Apple has it's one chance to become THE portable gaming champion.

If they don't do it, then others will.

Exactly as I said earlier. :D
 
Wirelessly posted (Mozilla/5.0 (iPod; U; CPU iPhone OS 4_3_3 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8J2 Safari/6533.18.5)

Popeye206 said:
Darn... no iPhone 5 for me then. Why be stuck with an A5 if an A6 is coming? :p

Yes... I'm being sarcastic.

That's probably a good thing. :rolleyes:
 
I'd like to see one of these integrated into a MBP. Two chipsets, two OS's running simultaneously, one as an overlay... working off the same data between them, OSX base, with iOS as on-screen keyboards/programmable control surfaces or an overlay...

Perhaps that's a little sophisticated for where they're currently at. ...but I wonder how long it'll take them to get there.



Right now they can't even get iOS to keep track of a video file. :confused:
 
Last edited:
Please let it not be so! TSMC never gets their crap together in new node shrinks. Just look at the 40nm shrink. It took them 1.5 years to get it sorta right and at the same time harmed ATI GPU sales, due to low yields and constrained chip supplies.

What do you thunk would happen with Apple? Same thing, even worse constrained supplies.
 
I doubt it. If it were an A15 design, it would likely be first on market, by several months.

It will be a 28nm die shrink of A5 that focuses on power savings to compensate for the LTE chip it is likely to include. Then A7 will be a cortex A15 design.

I think it will go a bit beyond that. The A5 is class leading in using the 543MP2 GPU. It simply pounds the Tegra II into the pavement (as well as the A5 having the SIMD NEON instruction set and the Tegra II does not) and will be even with the Tegra III.

So I expect the A6 to be a Cortex A9 (perhaps a still a dual core) but with a 544MP4 providing about 3.5X the graphics performance. Pushing a Retina display would need about 4X the performance boost on the GPU to see a straight across shift in performance.

I would love to see the Rouge series but that will not be in the cards. The 543 core was released on Jan 2009 and started showing up March 2011. The 544 core was released Jun 2010 making a March 2012 showing about right.

Likewise, the 554 core was just released 8 months back so I think it is a little hopeful thinking to see that core included.
 
You think... you have no true inside information one way or the other.

Neither did Gruber when he started this rumor in February. He was just making it up.

I do have logic on my side as Apple takes 6 months to ramp up production and the iPad is fairly new product, that they would dramatically change it again after only half a year when they only just finally caught up with demand for iPad 2 is nonsense.

It will be challenging enough to make sure they can get iPad 3 with the A6 necessary for 4x times pixel pushing power out the door in early 2012.

Not only do I think there is zero chance of Retina iPad this year, I think there is good chance it will be late next year trying to get all the pieces to fall into place.
 
i feel this is what is going into the iPad 3. the A6 is likely to be quad core and with the rumored display resolution i think they will need a more powerful chip to drive it. the A5 is by no means a slouch as its scores are very impressive but a resolution of that magnitude needs powerful silicon. anyone else disagree?
 
I think it will go a bit beyond that. The A5 is class leading in using the 543MP2 GPU. It simply pounds the Tegra II into the pavement (as well as the A5 having the SIMD NEON instruction set and the Tegra II does not) and will be even with the Tegra III.

The Tegra line is not the only point of comparison. Snapdragon, OMAP and Exynos line all feature the SIMD execution units featured in the reference design of the A8/A9 ARM7 cores.

So I expect the A6 to be a Cortex A9 (perhaps a still a dual core) but with a 544MP4 providing about 3.5X the graphics performance. Pushing a Retina display would need about 4X the performance boost on the GPU to see a straight across shift in performance.

It's already been expected that the jump to 543MP2 was to prepare for Retina displays on the iPad. So, either that was the case, or they could have just stuck with the 535 on the A5 and then jumped to a 54xMPx core or straight to the 6xx series on the A7. It's my opinion that they increased it just for the sake of performance. And being comfortable with that performance, they'll focus on power consumption on their next iteration by doing a process shrink of their current design.

I would love to see the Rouge series but that will not be in the cards. The 543 core was released on Jan 2009 and started showing up March 2011. The 544 core was released Jun 2010 making a March 2012 showing about right.

544 core is only slightly incremental over the 543, so I don't see a big push towards that one.

Likewise, the 554 core was just released 8 months back so I think it is a little hopeful thinking to see that core included.

I suppose it's possible, but Apple does seem to want to stay on bleeding edge when they upgrade their SoC. Given that March 2013 would be right time for a 6xx core to appear, it wouldn't surprise me at all for A6 to be power optimized A5 with LTE added to the phone and then A7 pops up as a dual/quad A15 config with 6xx cores.

Also, the 554 core is essentially two 544 cores per 1 554 core. It's just a way to cut the overhead in half for higher core counts. Doesn't really bring anything new to the table in terms of new silicon. My understanding is that this isn't far off for the 544 compared to the 543 either, the 544 chief advantages simply being standard compliance rather than architecture advantages. I can't find literature to back that up, but I seem to recall reading it somewhere.
 
Last edited:
Quad-core? :cool:

you know, I am not so sure about that. Although totally possible. The second gen i5's and i7's on the macbook air and pro were dual-core when the first gen's were quad core. From what I here quad core can eat more power.

Like I said it is possible, but not necessary since they seem to switch dual to quad to back to dual, for increased performance.
 
Apple could hire TSMC. And buy AMD outright.

Apple likes to have multiple suppliers for their components. That way there is no single point of failure, and Apple can start bidding wars between suppliers. As in "Brand X can make these parts for 52.3 cents each in lots of 1 million. If you can't match that, we'll take our business elsewhere..." So the TSMC deal makes perfect sense.

There is, of course, the small matter of legal action against Samsung. Which may or may not affect Apple and Samsung's business relationship.

So what about Intel? Apple supposedly served notice to Intel, warning them that if they didn't reduce the power consumption of their mobile x86 SoCs that they'd go elsewhere. Intel's response was to spew $300 million into the Wintel community in an attempt to bribe them into copying the MacBook Air. Irrational, short-sighted, and reeking of desperation. Much better to spend the $300 actually improving their own product and/or process.

Then what will Apple do if Intel ignores their demands? Who could they go to? Well, there's always AMD. They are the second largest maker of x86-compatible chips and they bought ATI in 2006. And guess what. Their market cap is only $4.6 billion. Apple could either acquire a majority of their stock or buy them outright. Then Apple could use their engineering talent to massage the AMD chips for lower power consumption.

Apple acquiring AMD is a long shot. Especially considering that it is also possible that Apple could be planning to migrate some or all of their Mac lines to ARM-based chips. That would take years, considering how long it took for Adobe to migrate their professional suites from OS 9 to OS X and Cocoa. But I'm sure Apple could sell millions of MacBook Airs even if they don't run Adobe bloatware.
 
http://www.imgtec.com/corporate/newsdetail.asp?NewsID=602

PowerVR with Ray Tracing would be very, very cool, albeit were a few generation away from seeing it on mobile platforms.

And don't expect Apple to blaze any trails there. They may add a core with ray-tracing functionality, but they'd never force it on developers unless it was the heir apparent graphics technology. They are all about making it as easy for the devs as possible.

Apple likes to have multiple suppliers for their components. That way there is no single point of failure, and Apple can start bidding wars between suppliers. As in "Brand X can make these parts for 52.3 cents each in lots of 1 million. If you can't match that, we'll take our business elsewhere..." So the TSMC deal makes perfect sense.

There is, of course, the small matter of legal action against Samsung. Which may or may not affect Apple and Samsung's business relationship.

So what about Intel? Apple supposedly served notice to Intel, warning them that if they didn't reduce the power consumption of their mobile x86 SoCs that they'd go elsewhere. Intel's response was to spew $300 million into the Wintel community in an attempt to bribe them into copying the MacBook Air. Irrational, short-sighted, and reeking of desperation. Much better to spend the $300 actually improving their own product and/or process.

Then what will Apple do if Intel ignores their demands? Who could they go to? Well, there's always AMD. They are the second largest maker of x86-compatible chips and they bought ATI in 2006. And guess what. Their market cap is only $4.6 billion. Apple could either acquire a majority of their stock or buy them outright. Then Apple could use their engineering talent to massage the AMD chips for lower power consumption.

Apple acquiring AMD is a long shot. Especially considering that it is also possible that Apple could be planning to migrate some or all of their Mac lines to ARM-based chips. That would take years, considering how long it took for Adobe to migrate their professional suites from OS 9 to OS X and Cocoa. But I'm sure Apple could sell millions of MacBook Airs even if they don't run Adobe bloatware.

You can't switch suppliers for a SoC willy-nilly. Designs have to be tuned to the reference libraries of their fabs and it makes more sense to stick with a supplier to understand their process and make the best chips possible.

Apple would never acquire AMD. While they are happy to use their chips, acquiring them would mean no one else could use them, which is a bad thing. Their own products could fall off of the market's top offerings and Apple would be stuck with silicon and engineers they don't want. If they source from them like all of their other component makers, then they share the same risks the market does, and they are free to move between suppliers in the same fashion other companies do. It makes much more sense to license your partners' technologies and help invest in their business so that you share their success; success you helped to create in the first place.

That's precisely what Apple has done. They own about 10% of ImgTec, who supplies their GPUs, and they are a direct licensee of the ARM cores. They have access to the core technology and are free to modify it as they see it, yet they retain the market's same suppliers and simply beat them by out-engineering them without the fear of acquiring themselves into a corner by owning too much in house IP.
 
The Tegra line is not the only point of comparison. Snapdragon, OMAP and Exynos line all feature the SIMD execution units featured in the reference design of the A8/A9 ARM7 cores.

But in the tablet race, the Tegra II is the only line of comparison because Honeycomb was built around it and Android is the only real competitor so far in this field. Comparable OMAPs (to even the A5) are still several months out. The Adreno 300 GPUs (Qualcomm) also look really good but may be a bit latter.

It's already been expected that the jump to 543MP2 was to prepare for Retina displays on the iPad. So, either that was the case, or they could have just stuck with the 535 on the A5 and then jumped to a 54xMPx core or straight to the 6xx series on the A7.

I don't think so. Apple recognizes the single biggest draw to iOS right now is gaming. The 543MP2 gave a HUGE boost in gaming performance. With Sony's NGP sporting a quad core A9 design with the 543MP4, Apple knows what it has to shoot for to compete on what it sees as its primary point of competition. Mobile gaming.

It's my opinion that they increased it just for the sake of performance. And being comfortable with that performance, they'll focus on power consumption on their next iteration by doing a process shrink of their current design.

Apple increased it because the 535 was pushed really hard on the iPhone 4 and the iPad as it was. I am amazed at how much faster many graphics operations (like fills and blits) are on the iPhone 3G S when compared to the iPhone 4. The 543MP2 was a requirement to maintain the mobile gaming edge. If Apple does jump to a Retina display, they will need to offset all those darn pixels with something. Heading to a 28nm will really help the power thing so they may be able to increase performance while maintaining the same power envelope.

544 core is only slightly incremental over the 543, so I don't see a big push towards that one.

Agreed. They might just go quad on the 543 and maintain dual on the Cortex A9. The 554 is the next logical choice. Double the pipes and get about 100% more performance similar to simply adding additional cores.

I suppose it's possible, but Apple does seem to want to stay on bleeding edge when they upgrade their SoC. Given that March 2013 would be right time for a 6xx core to appear, it wouldn't surprise me at all for A6 to be power optimized A5 with LTE added to the phone and then A7 pops up as a dual/quad A15 config with 6xx cores.

Also, the 554 core is essentially two 544 cores per 1 554 core. It's just a way to cut the overhead in half for higher core counts. Doesn't really bring anything new to the table in terms of new silicon. My understanding is that this isn't far off for the 544 compared to the 543 either, the 544 chief advantages simply being standard compliance rather than architecture advantages. I can't find literature to back that up, but I seem to recall reading it somewhere.

544 adds full Direct X compliance but no updates to OpenGL.
 
While not really a creative logo, it is just a silicon wafer with dies scored in it and a few of the dies colored black. It's OK; but not very innovative or even interesting. But I've seen much worse logos.

Thanks for clarifying. I thought it was a crossword puzzle. :D
 
Why wouldn't they want anything to run faster? Shunning flash does illustrate that. With anything the ipad 2 shows that Apple is very much so interested in performance. With in a year they go from an A8 to an A9. The next logical step is an A15. It gives them more performance and better battery life.

Also if they do go with a screen thats 2048x1536 they are going to need more performance much more then can be squeezed out of a Cortex A9.

If they want to take more away from the Desktop/Laptop they are going to need more performance.

They don't need 2GHz worth of performance. Running anything that fast means it's much less efficient per cycle. There is always a trade off between speed and battery life. Doubling the screen resolution makes almost no difference to the the speed of anything. Rendering takes a relatively small amount of power and its performance is mostly affected by the graphics processor. Even video would probably be easier to render into a larger screen.

As for "take more away from the Desktop/Laptop", you can't do real work on tablets, so this really isn't happening.
 
But in the tablet race, the Tegra II is the only line of comparison because Honeycomb was built around it and Android is the only real competitor so far in this field. Comparable OMAPs (to even the A5) are still several months out. The Adreno 300 GPUs (Qualcomm) also look really good but may be a bit latter.

Not really. Qualcomm has silicon out there (TouchPad), Exynos is out there (Samsung Galaxy Tab 10.1) as well as the OMAP 4430 (playbook, droid 3, upcoming bionic). Just because they don't same a comparable GPU doesn't mean they aren't comparable SoCs, especially in the areas of CPU performance. GPU performance is a subset of a SoC's function.

I don't think so. Apple recognizes the single biggest draw to iOS right now is gaming. The 543MP2 gave a HUGE boost in gaming performance. With Sony's NGP sporting a quad core A9 design with the 543MP4, Apple knows what it has to shoot for to compete on what it sees as its primary point of competition. Mobile gaming.

You think gaming is what is winning Apple the mobile war? That's an aspect of it, but far from the only, much less primary. The draw of it is the design, ecosystem, app selection etc.

While the iPhone competes with the DS, 3DS and PSP variants, they have overlapping market segments. Neither is a subset of the other. What they are doing is leeching casual gamers from those platforms and adding new players to the market. There will always be a contingent of hardcore gamers that demand physical buttons and other amenities that only a dedicated device will offer. Thus, while good graphics are essential for those platforms, they are only a nicety on the mobile phone platforms. Game selection will trump AAA titles (at least for now). For instance, what big graphical title can you name for iOS? Most people might manage Infinity Blade. But for most people it's cut the rope, angry birds, fruit ninja, etc. Small time wasters with undemanding graphics.

On top of that, the mobile phone platforms will necessary lag the gaming platforms because they'll have more levels of abstraction between them and the hardware. They'll have to have a lot more than a matching 543MP4+ core to match the Vita's performance.

Finally, people don't follow mobile phone platforms for "killer app" games like they do dedicated consoles. For instance, Mario is a huge draw for Nintendo systems, whereas Uncharted will move some Vitas (me included here). No one buys a mobile phone thinking "Oooh, this game is coming out so I'll buy that." A phone means much more to a person and those users also have a reasonable expectation of every game they could want being available on their platform. That certainly makes the most sense for app developers, so until mobile phone developers see enough of a market to justify acquiring exclusives on mobile phone games, games will be a feature and not a purchase decider.

Apple increased it because the 535 was pushed really hard on the iPhone 4 and the iPad as it was. I am amazed at how much faster many graphics operations (like fills and blits) are on the iPhone 3G S when compared to the iPhone 4. The 543MP2 was a requirement to maintain the mobile gaming edge. If Apple does jump to a Retina display, they will need to offset all those darn pixels with something. Heading to a 28nm will really help the power thing so they may be able to increase performance while maintaining the same power envelope.

It's not clear that Apple has prioritized the gaming edge. They certainly want to stay on top in terms of the ARM core they use, but we don't even know for sure they want to go retina on the iPad, so their intent there remains to be seen.

544 adds full Direct X compliance but no updates to OpenGL.

That's what I recall reading, but the wikipedia page showed identical DirectX 9c compliance and I couldn't be bothered to look further into it.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.