Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Thank you very much! Overall not bad...The 6S Plus needs to bring it next year. I am talking A9 Quad Core clocked @ 2.5Ghz. The benchmarks should get close to the 808 and 810 chipset.

Right now Snapdragon is the chipset to beat because of multi-cores used. A8 is way underclocked to compensate battery life.

Lol not at all. The 810 is already beaten by the a8, and it isn't even out yet.
 
Actually concerns me. Despite the hate on these forums, Samsung is really good at what they do. Hopefully no logic board failures down the road.

Huh? TSMC is one of the best and most reliable fans in the world. Samsung is not. Hardly anyone uses samsung as a contract fab. TSMC chips are everywhere you look.

----------

Wow that was fast from Chipworks. Can't wait to see how this compares to Intel. How far are we from the A-series getting into Macbook's and Mac Pro's?

http://www.applenews.zone/2014/09/why-macs-will-get-armd.html

http://www.applenews.zone/2014/09/why-macs-will-get-armd-part-ii.html


http://www.applenews.zone/2014/09/why-macs-will-get-armd-part-iii.html
 
Hardly a news, since rumours about that are around since the beginning.
What Apple has to explain is the reason of the underwhelming performance gain and the use of only 1 Gb of ram in the iPhone 6.
Don't get me wrong, I'm fine with the performance of the new iPhone and I still think it's the best device you can get, but I'm disappointed nevertheless. If you opt for a yearly update policy, and I think it's fine, you have to improve your device more consistently over the year. Performance wise the iPhone 6 is only marginally better than the iPhone 5S. They put all the attention on the new form factor, but the A8 seems to be an "A7S"....

My two cents

----------

I don't care about the technical details only that I have no lag and the phone runs smooth and does what I want.

You are in the wrong thread, then.

----------

Actually concerns me. Despite the hate on these forums, Samsung is really good at what they do. Hopefully no logic board failures down the road.

Tsmc surely isn't worse than Samsung
 
Did you notice it being way too big? :confused: I'm sure that's an hardware issue.

That's highly subjective

----------

I'm sending back my 6 Plus, because apparently, they sent me a wrong product. Everyone keeps saying how it is so huge or even an iPad Mini replacement. Today, I had it in my chest pocket most of the time, hardly noticing that it was there. I can easily hold it in one hand, even use it with one hand for the most common activities (without even using reachability). When I wrap my hand around it, my thumb and middle finger are less than one inch apart (and I do not have giant hands). "Wrapping my hand around it" is not a sentence I have ever used when describing the iPad Mini.

So, Apple, where is the real 6 Plus? Where is that unwieldy tablet phone that people keep talking about? That's what I ordered!
I'm using a Lumia 1520 as daily driver, a phone bigger and heavier than the 6+, and there is nothing wrong about having a big device

----------


Plain false. Intel, Apple, Qualcomm, mediatek, broadcom, Marvell.... all the bigger companies are using TSMC

----------

Two min benchmarks would not truly show this beast's strength, but do something like a demo of a FPS game and you'll see other phone weep after a few minutes... No Throttling (because of low clock) and 20nm is a quasi revolution.

True, but it was the same with A7. I expected something better from A8

----------

Yes, iOS8 is laggy. Strangely, so is Yosemite.

How are the OS's getting more inefficient??

:mad:

First release. Always the same, they are going to improve it

----------

Because it's a $1,000 phone and you should be able to have as many tabs open as you want.

Price has nothing to do with unrealistic expectations

----------

Because it makes it a lot easier to come back to something you found earlier.


My Note 3 has 16 Chrome tabs open right now. :eek:

With 16 Chrome tabs opened, your device is unresponsive....
Ah ok, my bad, a galaxy is unresponsive even with one tab thanks to TouchWiz :D

----------

TSMC FTW MFers!

----------



Wow. I bet that makes you feel extra super special.

And boy, I have my fingers crossed for you. I pray that you aren't one of the "half of Android users" who are at great risk due to the catastrophic privacy bug in Android's browser:

Android Browser flaw a “privacy disaster” for half of Android users
Bug enables malicious sites to grab cookies, passwords from other sites.

http://arstechnica.com/security/2014/09/android-browser-flaw-a-privacy-disaster-for-half-of-android-users/

I pray to the gods of the Internet and everything else online that you aren't affected.
Chrome isn't affected
 
You are joking right? Many of today's top notch GPUs (think nVidia and AMD/ATI) are made by TSMC.

Not just that, but many other things, like core router chips, communications chips, etc..

Their fab is as good as anyone else's, if not better.

The reason you'd use Samsung instead of TSMC is because they have different business models. With samsung, when you make a chip, they build the chip, test it, package it, and sell you that.

With TSMC, you buy the wafer. YOU'RE responsible for testing, packaging, etc.. It takes more skill for that to happen.
 
Thank you very much! Overall not bad...The 6S Plus needs to bring it next year. I am talking A9 Quad Core clocked @ 2.5Ghz. The benchmarks should get close to the 808 and 810 chipset.

Right now Snapdragon is the chipset to beat because of multi-cores used. A8 is way underclocked to compensate battery life.

Shows how much you don't know about chipset construction.
 
Two min benchmarks would not truly show this beast's strength, but do something like a demo of a FPS game and you'll see other phone weep after a few minutes... No Throttling (because of low clock) and 20nm is a quasi revolution.

I saw that in the keynote. They claim sustained performance up to 20 minutes (thats all the chart showed) while "other" dropped off quickly :cool:
 
I want to know if the CPU/GPU is clocked higher in the 6+ model.

Given the extra pixels, one would logically think Apple will have clocked it higher, so that in tests, both device perform the same, despite the 6+ needing to do more work.

This would also explain the need for a larger battery as clocked higher parts would use more power.

I have not seen any speed comparison tests yet.

Someone MUST have done them already to compare both devices?

The Apple A8 is beyond capable of running at 1080p resolution very smoothly, I think Apple knows this pretty well to always have their trademark buttery smooth OS lol, some apps may not be fully compatible with iOS 8 apps yet but bunch of apps are being updated daily now , within one month everything will be almost updated

the GPU in the A8 is a GX6650 from ImaginationTechnologies

Anandtech actually says the GX6650 can compete with the K1 GPU(this is for sure what is in the new phones, support for ATSC)

the Apple A8 offers performance similiar to Nvidia K1 Denver (2.5 ghz 64 bit 7 way superscaler dual core)

http://www.anandtech.com/show/7796/...ounces-new-high-end-mobile-gpu-powervr-gx6650

----------

That's highly subjective

----------


I'm using a Lumia 1520 as daily driver, a phone bigger and heavier than the 6+, and there is nothing wrong about having a big device

----------



Plain false. Intel, Apple, Qualcomm, mediatek, broadcom, Marvell.... all the bigger companies are using TSMC

----------



True, but it was the same with A7. I expected something better from A8

----------



First release. Always the same, they are going to improve it

----------



Price has nothing to do with unrealistic expectations

----------



With 16 Chrome tabs opened, your device is unresponsive....
Ah ok, my bad, a galaxy is unresponsive even with one tab thanks to TouchWiz :D

----------

Chrome isn't affected

about 40-50% of that beautiful "85% market share" Android fanboys glamor about is affected though.
 
Not amateurs by any stretch of the imagination.
Which is why Apple is eyeing 14nm FinFET from Samsung.

20nm does not guarantee less power. Smaller geometries have more power consumption while static (leakage) and depending on the cells used HVT, LVT, etc., you could consume more power if you want to go faster.

Ohmigosh! Someone with technical understanding commenting on this thread. Not sure that is allowed / encouraged on MacRumors! ;)
 
Looking forward to seeing if the efficiency shows in battery life.

No effect on battery life, just thinner.

Case in point: iPhone 6: Larger battery, vastly more efficient SoC, relatively low-res display for the size, yet it's battery life is average compared to similarly sized smartphones.

I'm at the point where I'd like to see Apple create a iPhone 7X focused on durability and battery life. Let's see Ive design an iPhone that doesn't need a case and gets 36-48 hrs battery life. Sapphire Crystal display could obviate the need for scratch protection, so all that's left is to create a chassis that is shock supremely rugged and shock resistant.

I do wonder at times about Ive's design breadth. Everything I've seen from him, including his own personal exhibits, is all based on the same theme. It's one thing to have a personal style, but Ive takes it so far that it seems more like a rut. I'm not convinced he could design the iPhone 6X described above.
 
But Apple is set to go with GloFo next generation, and thus potentially Samsung as well, so that story will short lived.

My understanding is that GF was laughably underprepared for sub-20nm processes until licensing Samsung's FinFET 14nm and ST's FDSOI. Prior to that they had been unable to successfully develop anything in-house at those nodes. I find it difficult to believe that Apple would trust working with them over Samsung 14nm or TSMC 16nm.
 
Will there really only be one generation at the 20-N size? Are they really going to get down to 14-N next year?
I have no idea about this stuff, but it certainly seems impressive.

(a) TSMC will certainly get plenty of value from this fab over many years. Your iPhone (along with everyone else's electronics) has MANY different chips in it. The less demanding chips get fabbed on older processes which are slower and/or use more power and take up a larger area. But each year or two they also get moved down a slot, so you'll find the LCD controller or the audio DAC or whatever will eventually move to this process, halve its power usage, and take up a little less space on the logic board.

(b) The next stop for TSMC is 16nm. [There's a whole political fight going on about the naming of these processes with TSMC et al claiming that Intel labels their nodes (most recently 14nm) incorrectly to look better than anyone else, and Intel replying that a node is defined by a dozen different numbers and there's a perfectly good reason why they can claim their numbers justify 14nm. It's really not worth obsessing about this.]

The more important point is that with 16nm TSMC provides Apple with FinFETs. These are vertical (rather than planar) transistors which take up less area and provide better power/performance. Intel introduced these with Ivy Bridge and they gave them a substantial drop in power at slightly higher performance.
Given that Apple already dropped their power 50% this time round, what most of us are all hoping is that Apple will use FinFETs to give us an A9 that's say 50% faster at the same power --- but we shall see.

All indications so far are that A8 was about the most minimal change possible over the A7 --- very much like say Sandy Bridge to Ivy Bridge --- where the new process was used to drop power dramatically, and to add some more GPU hardware, but with no serious attempt to redesign the CPU. This may be Apple settling into a tick-tock strategy (because it's HARD to design new CPUs!), or it may reflect a one-time thing, namely the aWatch, with Apple's elite chip engineers all taken off the A8 project two or three years ago to design the S1 for that device.
 

Thanks for the links. But these articles seem to be written by someone who is highly biased - if the first paragraph has "Suck it Intel", the objectivity of the author is under question. Chipworks or Anandtech may come up with something more level headed.
 
Sorry but why would you have 6+ tabs open at the same time?

Because with something that big and fast you might instinctively start acting like it is a real computer and not a cell-phone?

How many tabs do you usually have open in Safari on your iMac or MBP?
 
Thanks for the links. But these articles seem to be written by someone who is highly biased - if the first paragraph has "Suck it Intel", the objectivity of the author is under question. Chipworks or Anandtech may come up with something more level headed.

In the second part he says this: :eek:

Conclusion
I have been accused of being an "ARM fanboy" or "biased against x86." This is nonsense. I have no experience at all with ARM and don't have any particular feelings about it. And I was an x86 chip designer for nine years. I simply have an engineering mindset

Why Macs will get ARM'd, part II
 
Actually concerns me. Despite the hate on these forums, Samsung is really good at what they do. Hopefully no logic board failures down the road.
TSMC, or the Taiwan Semiconductor Manufacturing Company, does not have a bad reputation... It's actually got a pretty good reputation and it's been around since the 80's.

Companies that subcontract manufacturing to them include Nvidia (so TMSC chips in Apple products is nothing new), AMD, Marvell, Altera, Xilinx and Qualcomm (i.e if a device has a Snapdragon chip, it's been manufactured by TSMC). Hell, even Intel outsources some work to them.
 
Actually concerns me. Despite the hate on these forums, Samsung is really good at what they do. Hopefully no logic board failures down the road.

TSMC has been doing this longer than Samsung.
 
Thanks for the links. But these articles seem to be written by someone who is highly biased - if the first paragraph has "Suck it Intel", the objectivity of the author is under question. Chipworks or Anandtech may come up with something more level headed.

Taken out of context. The context was that Intel copied AMD's 64-bit ISA, known as AMD64, and renamed it "Intel 64". Of course.

Read the article series -- the author is a long-time specialist who worked on both x86 and RISC chips.

One thing that non-specialists may not appreciate is that, (largely) starting with the products released in 2007 (e.g. Core-2 Duo, Conroe, Penryn, etc) (aside: the previous Pentium Pro/Pentium III were the first halting steps forward), Intel engineers did an unbelievable job (as in -- nobody predicted that they could do nearly this good a job) in working around and optimizing one of the ugliest and most difficult to speed up Instruction Set Architectures in computer history. The fact that they were able to make that ugly duckling fly so well has postponed the day of reckoning for workstations and servers, but, in a cell-phone, the historical baggage that x86 carries matters.

The perceived requirement for ISA compatibility (for consumers -- it never did matter that much to server people), will disappear. Consumers have the App Store on their phones and laptops, and, there are servers out there in the cloud somewhere, and because of the App Store, consumers just won't care about the ISA compatibility issues that Microsoft/Intel were able to exploit so successfully for 20-30 years. All they will care about is access to the set of apps they need.

That is why notebooks like Air are a perfect candidate for ARM, and, why even a consumer at home won't care that their notebook is ARM and their laptop is x86-- their family license will allow them to download the ARM version for the notebook and the x86 version for their laptop. (They already don't care that their Air/MBP have a different architecture from their phone.)

I conclude by noting that AMD has announced an ARM server chip.

http://arstechnica.com/information-technology/2014/08/amd-shows-off-the-guts-of-its-first-arm-chip-for-the-server/

I'm looking forward to an Apple Air based on ARM. It really is a logical next step.
 
Thanks for the links. But these articles seem to be written by someone who is highly biased - if the first paragraph has "Suck it Intel", the objectivity of the author is under question. Chipworks or Anandtech may come up with something more level headed.

Read it in context. I was talking about the name of the architecture. I was proud that I helped invent it at AMD and thus prefer AMD64. Intel would prefer not to use the AMD name for obvious reasons. Find a fact in those articles that is wrong and THEN call me biased.
 
Apple really could, and probably should, buy out TSMC at this point, if they want to do the whole vertical integration thing..

Apple typically applies vertical integration on customer facing elements. Generally, that means hardware integration via software.

By acquiring TSMC (or any other specialized downstream component manufacturer), Apple would be taking on huge oversight responsibilities for... what exactly? What gains are you implying that make this such an obvious decision? How would such an acquisition benefit the end customer?

Considering that Apple already has a design team developing their chips, I really can't think of anything to be gained.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.