Confuscious says, the comment section is not what you think it is... It's not meant to talk about the article, but rather to respond to the first comment.
The A7 is limited to 31 bits of address space (2GB) by iOS 7 currently. While it is technically a 64 bit operating system, Apple found it to be more efficient to limit address space. Many developers, including myself, are hoping to see at least 42 bits of address space in iOS 8. Then we will see real competition with desktop computers.
The comment above is full of myths and misconceptions.
For the real scoop, you might enjoy reading this article:
You're right. It'll be the iPad Pro. Not Luke the MBA or iPad Air. It'll be different platform. Remember the patents of smart enclosure that has concealed led's for keyboard and other features? That all spells iPad pro. Not just a bigger screen.
I am not sure. Apple will not mess with the iPad - way too successful / new / awesome. Though an iPad Pro is a possibility I suspect it is the MBA that is likely to be 'messed' with. Pros are way to conservative when it comes to their toolset (for good reasons) and developing a pro tablet from scratch seems too risky and niche, to me. The aging MBA is ripe for development, however. I like the idea of a new platform, somewhere in between a chrome book, OSX and iOS, with extended battery life, lower price, and I suspect a completely new keyboard (touch with haptic feedback).
A device like this would create a clearer division between the existing pro line and the consumer line (a good thing in my opinion) and be as capable, or close to, the current mba's.
It fits him much better than the analysts I assume you're thinking of.
analyst |?anl-ist| noun
• a person who conducts analysis.
analysis |??nal?sis| noun (pl. analyses |-?s?z| )
• detailed examination of the elements or structure of something.
• the process of separating something into its constituent elements.
This required "independent analysis"? Or, you know, you can simply read the ARMv8-A reference manual for this information.
Say what? All ARMv8 defines is an instruction set. Apple built a processor that runs ArMv8 faster than ARM's own cores can. Nothing in any ARM documentation will tell you how many instructions the A7 can do at once or how deep the pipeline is.
I thought that originally everyone in the smartphone industry said the whole nonsense about the A7 was just marketing BS. However, aren't Qualcomm's high-end Snapdragon chips every bit as powerful as Apple's A7? I'm guessing that all the different companies high-end processors are pretty much in the same ballpark when it comes to processing power. Usually one company never gets that much further than another because most companies are privy to the same technology. It's not like some aliens came down and gave one company some extra knowledge. Samsung's Galaxy S5 processor is basically an off the shelf processor with a slightly higher clock-speed than some competitor's offerings but it's still pretty powerful. Apart from the Galaxy S5 not having a 64-bit processor I'm sure it's up there with the A7 in processing power and graphics capabilities. I'm rather curious how they compare but Apple's A8 will probably make a big leap forward that maybe even Qualcomm or Nvidia will find it hard to keep pace with.
What a load of crap. I knew there was a reason I blocked you, but after seeing this drivel quoted I couldn't let you get away with posting lies.
Qualcomm Snapdragon processors only stay equal with the A7 because a) they have twice the cores and b) are clocked a full 1.0GHz faster. A single core of the A7 is 2.5x as fast as a single Snapdragon 800 core. And 2x as fast as a Samsung Exynos. Samsung and Qualcomm are so far behind it's not even funny. It's actually kind of pathetic.
What I'd give to see the fully working ARM port of OS-X that for sure Apple has in a lab somewhere deep in the bowels of One Infinity Loop.
first born sons are always a fungible assets, as are souls...
Quote:
Originally Posted by unknwntrr
Get me an iPad with keyboard that snaps apps to the side like the surface and I'll replace my Notebook with an iMac and take the iPad on the go! Especially considering the next iPad will have an A8 thats likely going to be comparable with a low/middle-class notebook. I'd love a lighter mobile solution but as long as I need a computer on the go I won't have two computers and therefore have to use the Retina MacBook Pro to be able to do all my work. But on the go the power of an iPad would be more than enough!
I'm surprised there isn't one on the market yet. It would be nice have a lightning connected keyboard and battery pack.
Quote:
Originally Posted by RadarTheKat
Gaming. In AppleTV. High Performance Gaming. With gesture control. Gotta be what Apple is thinking. It's ARM, therefore iOS compatible, where there are more and more immersive games. That's gotta be where they're taking this.
my guess is that apple TV will never be the 'game...' it will be the display of a game running on your High performance game.
Quote:
Originally Posted by WS11
$315 is the price listed on ARK.
A $37 Z3770 (Silvermont) already outperforms Apple's A7, the Z3795 even more-so.
14nm Airmont is the replacement to 22nm Silvermont. Airmont is expected to offer 3x~4x the performance of Silvermont. Airmont is coming around the time an A8 would release, and Airmont's successor (14nm Goldmont) will be out before Apple's A9.
performance = fast.
User experience = quick.
You don't have to be fast to be quick, if you can properly coordinate and anticipate your next move.
The issue is designing an OS and HW set concurrently, as iOS and the A series are at the moment (Apple has tweaked the 1980s and prior era of 'design great hardware and then design an OS on top of it' to 'determine what is the compelling user experience possible, and then design a 'platform' of CPU, ancillary chips, OS, and key apps and services on the system and in the cloud (a simple but compelling point: Apple's implementation of trust is built into the chips and systems at time of build, and almost everything in the OS requiring trust is built upon that bidirectional registration of apple's cert, and the device's cert at time of manufacture).
Intel et al are not optimizing Goldmont for 1 current version of iOS and 2 potential future versions. and no OS is designed specifically for Goldmont. Apple's unique position to put 'end user' requirements into the chip spec for 3-4 implementations allows for performance to be optimized.
I think the end game is that laptops (and the chips they rode in on) become as irrelevant as mainframes are to end-user computing... they're out there but 99.999% of 'computer' users don't use them or don't realize they are using one in the background. Thus, having the fastest chip means nothing if it wasn't designed to meet coordinated set of user needs.
Developers will eventually just develop in a cloud environment with an iOS device as the interface.... a little more sophisticated than Xwindows... although to me it's the ultimate endgame of Project Athena (central distribution of all rights and applications images to end devices... local computing with very little local storage relative to the 'network').
A $37 Z3770 (Silvermont) already outperforms Apple's A7, the Z3795 even more-so.
14nm Airmont is the replacement to 22nm Silvermont. Airmont is expected to offer 3x~4x the performance of Silvermont. Airmont is coming around the time an A8 would release, and Airmont's successor (14nm Goldmont) will be out before Apple's A9.
Where do you get this stuff from?
A single A7 core is actually about 40% faster than a single Z3770 core. The only reason the Z3770 is faster overall than an A7 is because it has 4 cores and is clocked higher (similar to Exynos or Qualcomm). What's impressive about needing double the cores and higher clock to get a measly 20% overall increase in performance? And this is based on raw processor scores. Look at results from actual applications and the A7 looks even better.
And I'm calling BS on the fact that Airmont is 3-4x the performance. Where is your source?
Say what? All ARMv8 defines is an instruction set. Apple built a processor that runs ArMv8 faster than ARM's own cores can. Nothing in any ARM documentation will tell you how many instructions thevA7 can do at once or how deep the pipeline is.
What a load of crap. I knew there was a reason I blocked you, but after seeing this drivel quoted I couldn't let you get away with posting lies.
You're never wrong when you ignore Anandtech's analysis and just listen to your gut assumptions about the A7's performance.
A single A7 core is actually about 40% faster than a single Z3770 core. The only reason the Z3770 is faster overall than an A7 is because it has 4 cores and is clocked higher (similar to Exynos or Qualcomm). What's impressive about needing double the cores and higher clock to get a measly 20% overall increase in performance? And this is based on raw processor scores. Look at results from actual applications and the A7 looks even better.
And I'm calling BS on the fact that Airmont is 3-4x the performance. Where is your source?
The prices are from ARK.
Do keep in mind the Z3770 is being benchmarked while running x86 Windows 8.1, the A7 is running a lightweight iOS 7.x. "Actual applications", such as? Last time I checked Bay Trail-T SoCs are running full desktop applications, the A7 is not.
Mind the "expected". We still need to wait until Computex to get full details on Airmont (Cherry Trail-T), but we have found out some details (VR-Zone).
We don't know what sort of additional advantages will come from the move to an 8th gen Broadwell GPU. As well, if there are any additional architectural changes to the CPU that would allow for even greater performance. The bulk of the improvement is expected to come from the new GPU. There will also be updated API support.
Without getting too far off course, my suggestion was in response to Mac-sochist's idea of an Apple chromebook-like computer that would run an ARM SoC and utilize OS X like software.
An inexpensive Airmont (Cherry Trail-T) solution would offer the performance required to run OS X in its entirety, so there would be no need for Apple to have to fragment their system or spend the time creating/porting OS X applications for ARM. This way it could fall in-line with the rest of the Macbook / OS X line-up.
Everyone is missing the largest part of the puzzle. What is the biggest hole in Apple's entire line up?
It's servers! After Apple canned the XServe, they don't have anything targeting that space. Servers run largely stock software - no need to support gazillions of consumer grade software. A simple recompile inside the Apple labs will give servers all the software they need. Also, servers benefit a lot more from power gains - because they are on 24/7, and working hard almost all the time. No displays, no graphic RAM, no WiFi/3G to worry about. Its all entirely under Apple's control.
I expect a very low power, very small form factor (think smaller than AppleTV) server class machine from Apple. It will be good enough to do handle almost any needs of any one, including handling a large database.
Almost 95% of the databases in the world are less than 1GB in capacity - which means, an AppleTV class computer with a quad core 2 GHz A8, 2GB DDR5, 16GB Flash, Gigabit Ethernet is already very capable of running this database for most workloads. Any performance issue with CPU will be offset by gains from using Flash memory.
Also, there's something very interesting about AppleTV that no one has noticed. There is a LOT of wasted space inside the AppleTV, which makes very little sense, knowing Apple's penchant for making things as compact as possible. However, a quick analysis of the internal space reveals that it is perfect for keeping 4 AA size rechargeable batteries inside!
Can you imagine an AppleTV size server that comes with built in power backup that can keep your server running for 24 hours without power? And costs about $200?
A typical server requires more power for cooling it, than this server will require for running and cooling put together! In terms of upfront costs too, the savings are ridiculously high.
Can you imagine what that will do to the server market? This will be the holy grail of cloud computing - and will allow Apple to thrash Google and Amazon in the cloud space. There is absolutely no way anyone can compete with such a solution from Apple, without completely replacing their entire infrastructure!
Everyone is missing the largest part of the puzzle. What is the biggest hole in Apple's entire line up?
It's servers! After Apple canned the XServe, they don't have anything targeting that space. Servers run largely stock software - no need to support gazillions of consumer grade software. A simple recompile inside the Apple labs will give servers all the software they need. Also, servers benefit a lot more from power gains - because they are on 24/7, and working hard almost all the time. No displays, no graphic RAM, no WiFi/3G to worry about. Its all entirely under Apple's control.
I expect a very low power, very small form factor (think smaller than AppleTV) server class machine from Apple. It will be good enough to do handle almost any needs of any one, including handling a large database.
Almost 95% of the databases in the world are less than 1GB in capacity - which means, an AppleTV class computer with a quad core 2 GHz A8, 2GB DDR5, 16GB Flash, Gigabit Ethernet is already very capable of running this database for most workloads. Any performance issue with CPU will be offset by gains from using Flash memory.
Also, there's something very interesting about AppleTV that no one has noticed. There is a LOT of wasted space inside the AppleTV, which makes very little sense, knowing Apple's penchant for making things as compact as possible. However, a quick analysis of the internal space reveals that it is perfect for keeping 4 AA size rechargeable batteries inside!
Can you imagine an AppleTV size server that comes with built in power backup that can keep your server running for 24 hours without power? And costs about $200?
A typical server requires more power for cooling it, than this server will require for running and cooling put together! In terms of upfront costs too, the savings are ridiculously high.
Can you imagine what that will do to the server market? This will be the holy grail of cloud computing - and will allow Apple to thrash Google and Amazon in the cloud space. There is absolutely no way anyone can compete with such a solution from Apple, without completely replacing their entire infrastructure!
1) I like your thinking and have been wanting an iOS-based Apple Home Server for as long as the iPhone has existed. Before that I wanted a Mac OS X-based Apple Home Server. Windows actually does a pretty good job with their Home Server but it doesn't have the refinement that Apple could bring to the table. For this they have all the parts, they just need to make HW that can support multiple drives. Their Airport Utility would likely be the UI to the settings.
2) ARM blade servers are definitely catching on for reasons you've mentioned but in regards to Xserve I'm not sure Apple will ever want to get back into that market. If their $20 (great for the price) OS X Server option is the level they are falling even further behind Windows and *nix options when they dropped Xserve.
Do keep in mind the Z3770 is being benchmarked while running x86 Windows 8.1, the A7 is running a lightweight iOS 7.x. "Actual applications", such as? Last time I checked Bay Trail-T SoCs are running full desktop applications, the A7 is not.
Mind the "expected". We still need to wait until Computex to get full details on Airmont (Cherry Trail-T), but we have found out some details (VR-Zone).
We don't know what sort of additional advantages will come from the move to an 8th gen Broadwell GPU. As well, if there are any additional architectural changes to the CPU that would allow for even greater performance. The bulk of the improvement is expected to come from the new GPU. There will also be updated API support.
Without getting too far off course, my suggestion was in response to Mac-sochist's idea of an Apple chromebook-like computer that would run an ARM SoC and utilize OS X like software.
An inexpensive Airmont (Cherry Trail-T) solution would offer the performance required to run OS X in its entirety, so there would be no need for Apple to have to fragment their system or spend the time creating/porting OS X applications for ARM. This way it could fall in-line with the rest of the Macbook / OS X line-up.
So what you're saying is that you completely fabricated that 3-4X performance number. I'll go out on a limb and predict the Airmont SoC for mobile (not the one for laptops) will only be 50% faster than the current Z3770. We can come back to this thread in several months when it's out to see who's right (if you're still here).
Apple may used an A series processor in a MacBook Air style laptop that ran iOS but with a mouse pointer GUI. Even if it only ran safari at first you would have access to the web, icloud, iWork for icloud, all they would have to do is creat iLife for icloud and iTunes for icloud. Developers could simply modify their ios apps to work with a pointer.
Anand is a smart guy and I appreciate his detective work (and that of his sources), but I found this paragraph a bit odd:
"The challenge is that at full tilt a pair of Cyclone cores can consume quite a bit of power. So for now, Cyclone's performance is really used to exploit race to sleep and get the device into a low power state as quickly as possible. The other problem I see is that although Cyclone is incredibly forward looking, it launched in devices with only 1GB of RAM. It's very likely that you'll run into memory limits before you hit CPU performance limits if you plan on keeping your device for a long time."
The overall impression from this text is that Cyclone, and the iPhone/iPad, are somehow unbalanced -- that there is a lot of performance potential that isn't being fully realized, and might never be fully realized in existing devices. Yet also contained in this text is the explanation for why Cyclone makes perfect sense: "Cyclone's performance is really used to exploit race to sleep and get the device into a low power state as quickly as possible." Anand makes it sound like this is a second or third order concern, but in the context of a mobile device, it's actually one of the most important concerns. Apple is trying to design a chip that gets its work done quickly, using as little power as possible. Given the physics confronting CPU designers, that means that you'd rather have more transistors running at a lower clock speed than the reverse. And given the nature of the software run on mobile devices, you'd rather have greater instruction level parallelism than thread level parallelism.
So the A7 is the perfect mobile SOC -- low clockspeed, high ILP, low TLP. it is perfectly balanced for its job. The fact that the cyclone core could also be the basis for a very credible desktop CPU is what's secondary here.
It called Apple spending R&D money wisely as usual, the A9 or A10 chips will be in a Mac laptop or Mac desktop in the future.
This is what I have been saying and some people said that it was not possible. I believe Apples next move is put their own processors in to their laptop products. They do not need Intel at this point. I would not be surprise if they did not have a OSX port for their processors already.
Apple does, this is the same company that had OS X running on Intel chips for nearly 10 before switching to them. When Steve Jobs said OS X on the phone in 2007, I don't he was lying most (tech people) didn't want to believe it, just like Apple building their own chips has constantly derided.
Comments
Confucius say: you make name too long.
The A7 is limited to 31 bits of address space (2GB) by iOS 7 currently. While it is technically a 64 bit operating system, Apple found it to be more efficient to limit address space. Many developers, including myself, are hoping to see at least 42 bits of address space in iOS 8. Then we will see real competition with desktop computers.
The comment above is full of myths and misconceptions.
For the real scoop, you might enjoy reading this article:
https://www.mikeash.com/pyblog/friday-qa-2013-09-27-arm64-and-you.html
What in particular are you referring?
A device like this would create a clearer division between the existing pro line and the consumer line (a good thing in my opinion) and be as capable, or close to, the current mba's.
Anand is now an analyst. I've seen it all.
It fits him much better than the analysts I assume you're thinking of.
noun
• a person who conducts analysis.
analysis |??nal?sis|
noun (pl. analyses |-?s?z| )
• detailed examination of the elements or structure of something.
• the process of separating something into its constituent elements.
What a load of crap. I knew there was a reason I blocked you, but after seeing this drivel quoted I couldn't let you get away with posting lies.
Qualcomm Snapdragon processors only stay equal with the A7 because a) they have twice the cores and b) are clocked a full 1.0GHz faster. A single core of the A7 is 2.5x as fast as a single Snapdragon 800 core. And 2x as fast as a Samsung Exynos. Samsung and Qualcomm are so far behind it's not even funny. It's actually kind of pathetic.
What I'd give to see the fully working ARM port of OS-X that for sure Apple has in a lab somewhere deep in the bowels of One Infinity Loop.
first born sons are always a fungible assets, as are souls...
Get me an iPad with keyboard that snaps apps to the side like the surface and I'll replace my Notebook with an iMac and take the iPad on the go! Especially considering the next iPad will have an A8 thats likely going to be comparable with a low/middle-class notebook. I'd love a lighter mobile solution but as long as I need a computer on the go I won't have two computers and therefore have to use the Retina MacBook Pro to be able to do all my work. But on the go the power of an iPad would be more than enough!
I'm surprised there isn't one on the market yet. It would be nice have a lightning connected keyboard and battery pack.
Gaming. In AppleTV. High Performance Gaming. With gesture control. Gotta be what Apple is thinking. It's ARM, therefore iOS compatible, where there are more and more immersive games. That's gotta be where they're taking this.
my guess is that apple TV will never be the 'game...' it will be the display of a game running on your High performance game.
$315 is the price listed on ARK.
A $37 Z3770 (Silvermont) already outperforms Apple's A7, the Z3795 even more-so.
14nm Airmont is the replacement to 22nm Silvermont. Airmont is expected to offer 3x~4x the performance of Silvermont. Airmont is coming around the time an A8 would release, and Airmont's successor (14nm Goldmont) will be out before Apple's A9.
performance = fast.
User experience = quick.
You don't have to be fast to be quick, if you can properly coordinate and anticipate your next move.
The issue is designing an OS and HW set concurrently, as iOS and the A series are at the moment (Apple has tweaked the 1980s and prior era of 'design great hardware and then design an OS on top of it' to 'determine what is the compelling user experience possible, and then design a 'platform' of CPU, ancillary chips, OS, and key apps and services on the system and in the cloud (a simple but compelling point: Apple's implementation of trust is built into the chips and systems at time of build, and almost everything in the OS requiring trust is built upon that bidirectional registration of apple's cert, and the device's cert at time of manufacture).
Intel et al are not optimizing Goldmont for 1 current version of iOS and 2 potential future versions. and no OS is designed specifically for Goldmont. Apple's unique position to put 'end user' requirements into the chip spec for 3-4 implementations allows for performance to be optimized.
I think the end game is that laptops (and the chips they rode in on) become as irrelevant as mainframes are to end-user computing... they're out there but 99.999% of 'computer' users don't use them or don't realize they are using one in the background. Thus, having the fastest chip means nothing if it wasn't designed to meet coordinated set of user needs.
Developers will eventually just develop in a cloud environment with an iOS device as the interface.... a little more sophisticated than Xwindows... although to me it's the ultimate endgame of Project Athena (central distribution of all rights and applications images to end devices... local computing with very little local storage relative to the 'network').
$315 is the price listed on ARK.
A $37 Z3770 (Silvermont) already outperforms Apple's A7, the Z3795 even more-so.
14nm Airmont is the replacement to 22nm Silvermont. Airmont is expected to offer 3x~4x the performance of Silvermont. Airmont is coming around the time an A8 would release, and Airmont's successor (14nm Goldmont) will be out before Apple's A9.
Where do you get this stuff from?
A single A7 core is actually about 40% faster than a single Z3770 core. The only reason the Z3770 is faster overall than an A7 is because it has 4 cores and is clocked higher (similar to Exynos or Qualcomm). What's impressive about needing double the cores and higher clock to get a measly 20% overall increase in performance? And this is based on raw processor scores. Look at results from actual applications and the A7 looks even better.
And I'm calling BS on the fact that Airmont is 3-4x the performance. Where is your source?
You're never wrong when you ignore Anandtech's analysis and just listen to your gut assumptions about the A7's performance.
The minivan of computers. Loved by soccer* moms everywhere.
* Do other countries have a similar term?
Indeed we do, "Soccer Mums". (^_-)
Where do you get this stuff from?
A single A7 core is actually about 40% faster than a single Z3770 core. The only reason the Z3770 is faster overall than an A7 is because it has 4 cores and is clocked higher (similar to Exynos or Qualcomm). What's impressive about needing double the cores and higher clock to get a measly 20% overall increase in performance? And this is based on raw processor scores. Look at results from actual applications and the A7 looks even better.
And I'm calling BS on the fact that Airmont is 3-4x the performance. Where is your source?
The prices are from ARK.
Do keep in mind the Z3770 is being benchmarked while running x86 Windows 8.1, the A7 is running a lightweight iOS 7.x. "Actual applications", such as? Last time I checked Bay Trail-T SoCs are running full desktop applications, the A7 is not.
Mind the "expected". We still need to wait until Computex to get full details on Airmont (Cherry Trail-T), but we have found out some details (VR-Zone).
Z3770 (Silvermont) --> Z47xx (Airmont):
We don't know what sort of additional advantages will come from the move to an 8th gen Broadwell GPU. As well, if there are any additional architectural changes to the CPU that would allow for even greater performance. The bulk of the improvement is expected to come from the new GPU. There will also be updated API support.
Without getting too far off course, my suggestion was in response to Mac-sochist's idea of an Apple chromebook-like computer that would run an ARM SoC and utilize OS X like software.
An inexpensive Airmont (Cherry Trail-T) solution would offer the performance required to run OS X in its entirety, so there would be no need for Apple to have to fragment their system or spend the time creating/porting OS X applications for ARM. This way it could fall in-line with the rest of the Macbook / OS X line-up.
It's servers! After Apple canned the XServe, they don't have anything targeting that space. Servers run largely stock software - no need to support gazillions of consumer grade software. A simple recompile inside the Apple labs will give servers all the software they need. Also, servers benefit a lot more from power gains - because they are on 24/7, and working hard almost all the time. No displays, no graphic RAM, no WiFi/3G to worry about. Its all entirely under Apple's control.
I expect a very low power, very small form factor (think smaller than AppleTV) server class machine from Apple. It will be good enough to do handle almost any needs of any one, including handling a large database.
Almost 95% of the databases in the world are less than 1GB in capacity - which means, an AppleTV class computer with a quad core 2 GHz A8, 2GB DDR5, 16GB Flash, Gigabit Ethernet is already very capable of running this database for most workloads. Any performance issue with CPU will be offset by gains from using Flash memory.
Also, there's something very interesting about AppleTV that no one has noticed. There is a LOT of wasted space inside the AppleTV, which makes very little sense, knowing Apple's penchant for making things as compact as possible. However, a quick analysis of the internal space reveals that it is perfect for keeping 4 AA size rechargeable batteries inside!
Can you imagine an AppleTV size server that comes with built in power backup that can keep your server running for 24 hours without power? And costs about $200?
A typical server requires more power for cooling it, than this server will require for running and cooling put together! In terms of upfront costs too, the savings are ridiculously high.
Can you imagine what that will do to the server market? This will be the holy grail of cloud computing - and will allow Apple to thrash Google and Amazon in the cloud space. There is absolutely no way anyone can compete with such a solution from Apple, without completely replacing their entire infrastructure!
1) I like your thinking and have been wanting an iOS-based Apple Home Server for as long as the iPhone has existed. Before that I wanted a Mac OS X-based Apple Home Server. Windows actually does a pretty good job with their Home Server but it doesn't have the refinement that Apple could bring to the table. For this they have all the parts, they just need to make HW that can support multiple drives. Their Airport Utility would likely be the UI to the settings.
2) ARM blade servers are definitely catching on for reasons you've mentioned but in regards to Xserve I'm not sure Apple will ever want to get back into that market. If their $20 (great for the price) OS X Server option is the level they are falling even further behind Windows and *nix options when they dropped Xserve.
The prices are from ARK.
Do keep in mind the Z3770 is being benchmarked while running x86 Windows 8.1, the A7 is running a lightweight iOS 7.x. "Actual applications", such as? Last time I checked Bay Trail-T SoCs are running full desktop applications, the A7 is not.
Mind the "expected". We still need to wait until Computex to get full details on Airmont (Cherry Trail-T), but we have found out some details (VR-Zone).
Z3770 (Silvermont) --> Z47xx (Airmont):
We don't know what sort of additional advantages will come from the move to an 8th gen Broadwell GPU. As well, if there are any additional architectural changes to the CPU that would allow for even greater performance. The bulk of the improvement is expected to come from the new GPU. There will also be updated API support.
Without getting too far off course, my suggestion was in response to Mac-sochist's idea of an Apple chromebook-like computer that would run an ARM SoC and utilize OS X like software.
An inexpensive Airmont (Cherry Trail-T) solution would offer the performance required to run OS X in its entirety, so there would be no need for Apple to have to fragment their system or spend the time creating/porting OS X applications for ARM. This way it could fall in-line with the rest of the Macbook / OS X line-up.
So what you're saying is that you completely fabricated that 3-4X performance number. I'll go out on a limb and predict the Airmont SoC for mobile (not the one for laptops) will only be 50% faster than the current Z3770. We can come back to this thread in several months when it's out to see who's right (if you're still here).
Remember Apple were running an Intel version of MacOSX since the inception of the OSX.
By copying the structure they may be able to allow windows to run natively as well.
Anand is a smart guy and I appreciate his detective work (and that of his sources), but I found this paragraph a bit odd:
"The challenge is that at full tilt a pair of Cyclone cores can consume quite a bit of power. So for now, Cyclone's performance is really used to exploit race to sleep and get the device into a low power state as quickly as possible. The other problem I see is that although Cyclone is incredibly forward looking, it launched in devices with only 1GB of RAM. It's very likely that you'll run into memory limits before you hit CPU performance limits if you plan on keeping your device for a long time."
The overall impression from this text is that Cyclone, and the iPhone/iPad, are somehow unbalanced -- that there is a lot of performance potential that isn't being fully realized, and might never be fully realized in existing devices. Yet also contained in this text is the explanation for why Cyclone makes perfect sense: "Cyclone's performance is really used to exploit race to sleep and get the device into a low power state as quickly as possible." Anand makes it sound like this is a second or third order concern, but in the context of a mobile device, it's actually one of the most important concerns. Apple is trying to design a chip that gets its work done quickly, using as little power as possible. Given the physics confronting CPU designers, that means that you'd rather have more transistors running at a lower clock speed than the reverse. And given the nature of the software run on mobile devices, you'd rather have greater instruction level parallelism than thread level parallelism.
So the A7 is the perfect mobile SOC -- low clockspeed, high ILP, low TLP. it is perfectly balanced for its job. The fact that the cyclone core could also be the basis for a very credible desktop CPU is what's secondary here.
It called Apple spending R&D money wisely as usual, the A9 or A10 chips will be in a Mac laptop or Mac desktop in the future.
This is what I have been saying and some people said that it was not possible. I believe Apples next move is put their own processors in to their laptop products. They do not need Intel at this point. I would not be surprise if they did not have a OSX port for their processors already.
Apple does, this is the same company that had OS X running on Intel chips for nearly 10 before switching to them. When Steve Jobs said OS X on the phone in 2007, I don't he was lying most (tech people) didn't want to believe it, just like Apple building their own chips has constantly derided.