Has anyone determined whether the A5X is built on 28nm (TSMC) / 32nm (Samsung) lithography, or is Apple still sourcing 45nm chips from Samsung? With the revelation that the LTE baseband is built on a 45nm process ( http://www.anandtech.com/show/5661/the-new-ipad-4g-contains-qualcomms-mdm9600 ) , I'm getting a bit worried. A 45nm A5X would probably be a deal-breaker for me considering that the 28/32nm process represents a full node shrink and will produce chips with significantly lower power consumption at higher clocks.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
A5X: 28/32nm or 45nm?
- Thread starter frick
- Start date
- Sort by reaction score
No one knows until it is released and being examined. Apple never publishes any information about these stuff. We don't even know whether its CPU is Coretx A9 or A15 (the latter is more likely if it wants to remain competitive in terms of spec for another year).
I also find it silly that they put so much emphasis on the so caled quad-core graphic. Kind of like trying to confuse the general public into thinking (when compared other quad-core devices) - "Oh, the iPad is quad-core too". Graphic chips in general has a lot more cores than CPU, and that's why using GPU for general computing is a lot faster when doing heavily parallelized work. But counting the number of cores in GPU is idiotic, the difference between different GPU architectures is a lot more than the difference between CPU architectures.
From the news today, it sounds like the clocks are unchanged at 1 GHz. This does not bode well -- makes me think the chip is still at 45nm (unless all the power budget went to the new display and the two extra GPU cores).
Still standing by my statement that a 45nm A5X is a deal-killer. If this is the case, the next iPad will probably be getting a Cortex A15-based SoC (new architecture) *AND* a process shrink to 32nm or 28nm, depending on whether Apple sources the chips from Samsung or TSMC. This iPad would hypothetically be quite a bit more powerful (and power-efficient) than a 45nm A5X-based iPad 3; and I'm not going plunk down $5-600 until I have something that will last me at least 3-5 years.
Still standing by my statement that a 45nm A5X is a deal-killer. If this is the case, the next iPad will probably be getting a Cortex A15-based SoC (new architecture) *AND* a process shrink to 32nm or 28nm, depending on whether Apple sources the chips from Samsung or TSMC. This iPad would hypothetically be quite a bit more powerful (and power-efficient) than a 45nm A5X-based iPad 3; and I'm not going plunk down $5-600 until I have something that will last me at least 3-5 years.
You do realize there's always a 'latest' technology that bests the previous one, correct? If you're waiting for an iPad that won't be obsoleted in two years, you'll be waiting a long, long time. You must not have been paying attention to Apple over the past five years.
Why would you care? If it does what you need and the battery lasts long enough why wait another year? While the next iPad will no doubt be a bit faster, and possibly more efficient power wise, they are probably more likely to use that extra savings in a smaller battery, and adding something else. Or making it smaller again.
Why could a spec be a deal breaker? If they beefed up the battery to compensate for power consumption, who cares what process the chip was made on? It's the most powerful iPad made, with the best screen, and the same battery life as it's predecessor. What difference does it make if you disagree w/ the engineering decisions made on the chip? Yes, the battery life could theoretically be longer if they used a different process. But they never claimed to be increasing battery life. Apple focuses on end user experience and overall performance, they do not chase specs on particular components just for the heck of it.
Even the iPad 1 can last 3 years since it's already been 2 years and it's still useable. It all depends on usage. Even the most powerful computer may not last 3 years depending what it's used for.
Two problems with a 45nm A5X / larger battery:
1) Weight / size of the iPad (self-explanatory)
2) Likely shortened battery lifetime. If the A5X is 45nm, it will be big (~160 mm^2) and will likely get quite hot under load. Hot iPad innards = significantly shortened Li-Ion battery life.
Additionally (and this is my personal bias), the 45nm process is just old. I bought a 40nm Radeon 5870 in 2009. Now it's 2012, and buying a mobile device (where power usage/battery life are paramount) built on a 3-year-old process just seems ridiculous. Die shrinks are coming about slower and slower nowadays (it's taken quite awhile for the 40nm -> 28nm transition; Intel is supposedly having significant problems with its 22nm process), so if you're concerned about product longevity, I'm thinking the optimal time to buy a mobile device is on the leading edge of one of these die shrinks.
The CPU is still 45nm. Just because a chip is produced at a smaller manufacturing process doesn't mean lower power consumption or better power efficiency. There are multiple factors to battery efficiency and the 2012 iPad will have the same battery life with a slightly larger battery and weight.
Regarding the quad-core graphics, there was a need for faster rendering on the new Retina screen which required a better PowerVR chip. At the time of sourcing, the quad-core graphics processor drew less power and still gave adequate performance for the Retina display.
Purchasing a product based on specs will normally get you into trouble. Why not purchase a product because it will perform for you faithfully over the next few years. Product specs don't have any bearing on performance besides being components of performance. $500-600 is also not a whole lot to pay for the new iPad IMO. We're looking at technology that is supposedly worth a lot more on the consumer markets.
Regarding the quad-core graphics, there was a need for faster rendering on the new Retina screen which required a better PowerVR chip. At the time of sourcing, the quad-core graphics processor drew less power and still gave adequate performance for the Retina display.
Purchasing a product based on specs will normally get you into trouble. Why not purchase a product because it will perform for you faithfully over the next few years. Product specs don't have any bearing on performance besides being components of performance. $500-600 is also not a whole lot to pay for the new iPad IMO. We're looking at technology that is supposedly worth a lot more on the consumer markets.
I just "plunked down" $500 for an iPad and I will have it on Day 1 (Friday). Then I will use the CRAP out of it (gently of course and in a case the whole time
) for the next 10-11 months and then I will sell it for $400 MINIMUM come February 1st. Then I will "survive" (I don't know how I'll survive without a piece of aluminum /cry /tears /end of world rant) for about one month ipadless and put in the same preorder for the "Newer New iPad".If you are bad at math, that means I paid $100 to have the latest and greatest technology for eleven months instead of saving $100 like you and spending my time complaining on message boards. The consideration between 42nm and 28nm is MEANINGLESS if the user experience is excellent. But if you can find people that use the new iPad and are saying to themselves and others, "This iPad is OK but if it had a smaller die shrink on the cpu it would be GRRRRRREAT!" then you may have something to worry about....
If $100 for almost a full year of use is TOO ESSSSSPENSIVE for you concerning a piece of tech, you shouldn't even be considering buying an iPad, fact.
It appears you may be buying a tablet for the wrong reasons. You may be better satisfied with the "specs" of a MacBook Pro.
No, it's definitely not A15.
Based on technology advancements, and Apple's suppliers, they haven't lined up advancements necessary for the A15. Benchmarks from this morning corroborate this.
Apps will not be released on the App store that don't perform well, because they are being designed with iPad2/3 in mind. You would be fine.
----------
My home PC is a Core2/4GB system I built in 2006, ~5.5 years old now, still doing all I need from it and more. It's completely feasible. Most family and college buddies, all hardware and software engineers are mostly using Core2's, a few i5's. Core2's on desks around office. There's hardly a justification for more outside of synthetic benchmarks.who would keep an ipad for 3-5 years?
iPhone 4, I can't fathom having any reason to 'upgrade' it. CPU is powerful, GPU is powerful, retina IPS is brilliant. For web, maps, social networks, sms, it's already overkill. It's pretty much as big as a phone can get and still be easily used one handed. There has yet to be an iPhone 5/6 rumor that sounds luring, most sound worse.
----------
Have you noticed the new A5X has a nice fancy metal heatspreader on it? Rather than the old ceramic packaging. Likely to help bond it thermally to the metal rear chassis. Yea... there's a reason for that ;-)
The only thing I find nice about the next iPhone is the LTE. Right now 3G is still too slow to eliminate having a voice plan and just using Google Voice or its alternatives like VOIP, as one's primary phone. But if LTE is going to be as fast as it promises to be, VOIP and Google Voice might become reliable enough to only have a phone with a data plan. I'd consider that sufficient reason to upgrade my iPhone 4.
who would keep an ipad for 3-5 years?
Any person without an unlimited cash supply.
True. Generation changes of fundamental technologies are more important than incremental changes at product level. Couple of examples:
arm6 to arm7 from ip3g to ip3gs; 3gs still being sold.
Retina display in ip4.
Good point. Would like to hear more about it.
Sent from my cm_tenderloin using Tapatalk 2 Beta-2
CPU Power Consumption not relevant
The power consumption of the CPU is not really relevant. The screen and the wireless subsystems (4G and Wifi) when active are much bigger power hogs.
The power consumption of the CPU is not really relevant. The screen and the wireless subsystems (4G and Wifi) when active are much bigger power hogs.
Gah, it's 45nm:
http://www.chipworks.com/en/technic...ns/2012/03/the-new-ipad-a-closer-look-inside/
Will definitely be returning once the hubbub dies down at the local Apple Store. Someone else should be able to get one at least!
http://www.chipworks.com/en/technic...ns/2012/03/the-new-ipad-a-closer-look-inside/
Will definitely be returning once the hubbub dies down at the local Apple Store. Someone else should be able to get one at least!