The new silicon at the heart of Apple's recently-released iPhone 6 and iPhone 6 Plus has increased the company's lead in smartphone application processor performance, according to data released on Monday, with the A8 improving on the already industry-leading 64-bit A7. ...
It HAS a regular 1080 screen. Why is it rendering above that resolution? Why is it downsampling at ALL?
I believe it's something to do with maintaining similar physical sizes of 44 pt x 44 pt squares (which is the minimum recommended target size for touches).
An iPhone 6 Plus is about 2.7 inches wide. With 1080 pixels, that gives the 401 pip value.
That display is also a @3x display. 1080 and 1920 are both divisible by three (360 pt x 640 pt), so let's see what happens with a 44pt square:
Actual pixels of 44pt would be 132 x 132. 132 is about 0.122 the width of the screen or 0.33 inches. That's a decent amount larger than the approx 0.27 inch sides of 44 pt squares on iPhone 6, 5S, 5C, 5, etc.
So apparently to compensate, the iPhone 6 Plus point size is actually 414 pt x 736 pt (1242 x 2208 pixels). A scaling factor is approx 87% (1080 divided by 1242)
If we apply that scaling to the 0.33 inch value, you end up with a 0.287 inch value. Which is closer to the 0.27 inch value of other phones.
Although...
All non-mini iPads had a larger physical size than 0.27 in x 0.27 in for their 44 pt squares. When the mini came out, it had the same resolution as a non mini, yet pixels were smaller. Thankfully when the touch target size was scaled down for the mini, it was still around 0.27 in x 0.27 in.
So across all devices, the physical size of 44pt is a bit varied. Still, in looking only at phone (and iPod touch), they are all very close to the same size.
Apple is playing a very smart game here. They probably have specced out hardware improvements for a decade but are using the software team and the CPU team to eke out what they can now in lower specced machines. Mobile chips can only go so fast or have so many cores or have so much RAM and android machines will reach that limit pretty soon. Apple will stay competitive in benchmarks in that time and then give itself the extra hardware to blast past the competition while Android phones will have no reason to upgrade.
The 3d benchmark does not compute!
Whats up here... With Metal , Scene and Sprite kit one would expect stellar performance !
My suspicion is that the benchmark is somehow not compatible with apples new technology ..
Any insights from those more in the know !?
They can't really test with proprietary technology.
Since many will be upgrading from a 4S or 5, it would have been nice to see the performance increase from those models as well, not just from the 5S. But I guess the purpose of the article was to compare with other vendors.
This is what happen when you make your own processor and do not have to worry about competing interest that qualcom and intel have to worry about. plus when you control the silicon and software you can do things other companies can not do.
This is what happen when you make your own processor and do not have to worry about competing interest that qualcom and intel have to worry about. plus when you control the silicon and software you can do things other companies can not do.
For that exact reason, I also wonder if the tests are not able to measure Apple's chips accurately.
As for cellular speed, I see ping times that are about 2.5x slower with the 6 Plus than the year-old iPad Air (AT&T LTE, 3 bars). This is using Speedtest app. 160+ ms for the 6 Plus, ~60 ms for the iPad Air.
A ping test isn't generally influenced by the local device, but more by the network and endpoint that you are pinging, so not a good test of your device. As well, you should do multiple tests from multiple locations to get an average result.
I believe it's something to do with maintaining similar physical sizes of 44 pt x 44 pt squares (which is the minimum recommended target size for touches).
An iPhone 6 Plus is about 2.7 inches wide. With 1080 pixels, that gives the 401 pip value.
That display is also a @3x display. 1080 and 1920 are both divisible by three (360 pt x 640 pt), so let's see what happens with a 44pt square:
Actual pixels of 44pt would be 132 x 132. 132 is about 0.122 the width of the screen or 0.33 inches. That's a decent amount larger than the approx 0.27 inch sides of 44 pt squares on iPhone 6, 5S, 5C, 5, etc.
So apparently to compensate, the iPhone 6 Plus point size is actually 414 pt x 736 pt (1242 x 2208 pixels). A scaling factor is approx 87% (1080 divided by 1242)
If we apply that scaling to the 0.33 inch value, you end up with a 0.287 inch value. Which is closer to the 0.27 inch value of other phones.
Although...
All non-mini iPads had a larger physical size than 0.27 in x 0.27 in for their 44 pt squares. When the mini came out, it had the same resolution as a non mini, yet pixels were smaller. Thankfully when the touch target size was scaled down for the mini, it was still around 0.27 in x 0.27 in.
So across all devices, the physical size of 44pt is a bit varied. Still, in looking only at phone (and iPod touch), they are all very close to the same size.
I really appreciate all the math, but downscaling still feels like a compromise to me. My guess is they wanted a 1242x2208 display, but couldn't find anyone who could produce the required number at the required quality and required price point.
I wouldn't be surprised if a selling point of the iPhone 6s Plus (or iPhone 7 Plus depending on how long it takes LCD manufacturers to meet Apple's requirements) is a full 3x retina 1242x2208 display.
Although I appreciated the article itself, the article title is disingenuous in light of the severely lackluster GPU performance. GPU matters as much as CPU in many cases, even when you are not playing games.
Download was typically as fast or a little faster with the 6 Plus than the iPad Air. Upload was roughly 50-75% faster with the 6 Plus. Doesn't seem to be LTE-Advanced network here, though.
Quote:
Originally Posted by coolfactor
A ping test isn't generally influenced by the local device, but more by the network and endpoint that you are pinging, so not a good test of your device. As well, you should do multiple tests from multiple locations to get an average result.
Same network, same location, same server, every time, many times over 3 days: a reasonably good test.
Let's see some others' results before dismissing, ok?
Download was typically as fast or a little faster with the 6 Plus than the iPad Air. Upload was roughly 50-75% faster with the 6 Plus. Doesn't seem to be LTE-Advanced network here, though.
Same network, same location, same server, every time, many times over 3 days: a reasonably good test.
Let's see some others' results before dismissing, ok?
Let's see some others' results before concluding, ok? :P
The 3d benchmark does not compute!
Whats up here... With Metal , Scene and Sprite kit one would expect stellar performance !
My suspicion is that the benchmark is somehow not compatible with apples new technology ..
There's only so much optimization you can put into a benchmark for specific hardware, you can't just replace huge chunks of code for one device and call it a fair benchmark. It's a test of the hardware's raw performance, not the software. In a way it gives a false view of the end-user performance but it's designed to compare the same software on different hardware.
Check the bottom of the page, takes you to meetgadgets, check the bottom of the page to see affiliates. They list Samsung phones there for some reason.
phonearena: cough, er it turns out that a device from one of our affiliates was benchmarked with a lower score so we're eh, just going to make a slight adjustment to that so now instead of the LG having 30 minutes more battery life, our custom web script now shows that the Samsung (our affiliate) has well over an hour more battery life.
wouldn't mind seeing these stacked against x86 CPUs
The new Core M is way faster, both CPU and GPU.
But I'm not sure it matters. Core M is also much more expensive than what apple is likely paying for the A8, and most iphone and iPad users are not cpu constrained.
If you're thinking about MacBooks, the A8 is not competitive with Core M.
But apple is gaining fast on Intel. It could be that in another two to three years apple might be making SOCs that really could go in a MacBook air
Comments
The new silicon at the heart of Apple's recently-released iPhone 6 and iPhone 6 Plus has increased the company's lead in smartphone application processor performance, according to data released on Monday, with the A8 improving on the already industry-leading 64-bit A7. ...
It HAS a regular 1080 screen. Why is it rendering above that resolution? Why is it downsampling at ALL?
I believe it's something to do with maintaining similar physical sizes of 44 pt x 44 pt squares (which is the minimum recommended target size for touches).
An iPhone 6 Plus is about 2.7 inches wide. With 1080 pixels, that gives the 401 pip value.
That display is also a @3x display. 1080 and 1920 are both divisible by three (360 pt x 640 pt), so let's see what happens with a 44pt square:
Actual pixels of 44pt would be 132 x 132. 132 is about 0.122 the width of the screen or 0.33 inches. That's a decent amount larger than the approx 0.27 inch sides of 44 pt squares on iPhone 6, 5S, 5C, 5, etc.
So apparently to compensate, the iPhone 6 Plus point size is actually 414 pt x 736 pt (1242 x 2208 pixels). A scaling factor is approx 87% (1080 divided by 1242)
If we apply that scaling to the 0.33 inch value, you end up with a 0.287 inch value. Which is closer to the 0.27 inch value of other phones.
Although...
All non-mini iPads had a larger physical size than 0.27 in x 0.27 in for their 44 pt squares. When the mini came out, it had the same resolution as a non mini, yet pixels were smaller. Thankfully when the touch target size was scaled down for the mini, it was still around 0.27 in x 0.27 in.
So across all devices, the physical size of 44pt is a bit varied. Still, in looking only at phone (and iPod touch), they are all very close to the same size.
They can't really test with proprietary technology.
Since many will be upgrading from a 4S or 5, it would have been nice to see the performance increase from those models as well, not just from the 5S. But I guess the purpose of the article was to compare with other vendors.
For that exact reason, I also wonder if the tests are not able to measure Apple's chips accurately.
It HAS a regular 1080 screen. Why is it rendering above that resolution? Why is it downsampling at ALL?
Gruber has some info on it in his review:
http://daringfireball.net/2014/09/the_iphones_6
And why the higher resolution dimensions are the target for the interface:
http://daringfireball.net/2014/08/larger_iphone_display_conjecture
As for cellular speed, I see ping times that are about 2.5x slower with the 6 Plus than the year-old iPad Air (AT&T LTE, 3 bars). This is using Speedtest app. 160+ ms for the 6 Plus, ~60 ms for the iPad Air.
A ping test isn't generally influenced by the local device, but more by the network and endpoint that you are pinging, so not a good test of your device. As well, you should do multiple tests from multiple locations to get an average result.
I believe it's something to do with maintaining similar physical sizes of 44 pt x 44 pt squares (which is the minimum recommended target size for touches).
An iPhone 6 Plus is about 2.7 inches wide. With 1080 pixels, that gives the 401 pip value.
That display is also a @3x display. 1080 and 1920 are both divisible by three (360 pt x 640 pt), so let's see what happens with a 44pt square:
Actual pixels of 44pt would be 132 x 132. 132 is about 0.122 the width of the screen or 0.33 inches. That's a decent amount larger than the approx 0.27 inch sides of 44 pt squares on iPhone 6, 5S, 5C, 5, etc.
So apparently to compensate, the iPhone 6 Plus point size is actually 414 pt x 736 pt (1242 x 2208 pixels). A scaling factor is approx 87% (1080 divided by 1242)
If we apply that scaling to the 0.33 inch value, you end up with a 0.287 inch value. Which is closer to the 0.27 inch value of other phones.
Although...
All non-mini iPads had a larger physical size than 0.27 in x 0.27 in for their 44 pt squares. When the mini came out, it had the same resolution as a non mini, yet pixels were smaller. Thankfully when the touch target size was scaled down for the mini, it was still around 0.27 in x 0.27 in.
So across all devices, the physical size of 44pt is a bit varied. Still, in looking only at phone (and iPod touch), they are all very close to the same size.
I really appreciate all the math, but downscaling still feels like a compromise to me. My guess is they wanted a 1242x2208 display, but couldn't find anyone who could produce the required number at the required quality and required price point.
I wouldn't be surprised if a selling point of the iPhone 6s Plus (or iPhone 7 Plus depending on how long it takes LCD manufacturers to meet Apple's requirements) is a full 3x retina 1242x2208 display.
what chance for the A8 nfc 4 inch screen
i don't like the larger sizes
would you think that next round we'd have 4 4.7 5.5 all with touch id nfc, etc
and the actual upload and downnload speeds
Download was typically as fast or a little faster with the 6 Plus than the iPad Air. Upload was roughly 50-75% faster with the 6 Plus. Doesn't seem to be LTE-Advanced network here, though.
A ping test isn't generally influenced by the local device, but more by the network and endpoint that you are pinging, so not a good test of your device. As well, you should do multiple tests from multiple locations to get an average result.
Same network, same location, same server, every time, many times over 3 days: a reasonably good test.
Let's see some others' results before dismissing, ok?
Download was typically as fast or a little faster with the 6 Plus than the iPad Air. Upload was roughly 50-75% faster with the 6 Plus. Doesn't seem to be LTE-Advanced network here, though.
Same network, same location, same server, every time, many times over 3 days: a reasonably good test.
Let's see some others' results before dismissing, ok?
Let's see some others' results before concluding, ok? :P
Anybody else notice the 6 Plus charging basically twice as fast when using an iPad 11watt charger? This is nice, considering its much larger battery.
Does the iPhone 6 also charge faster? I don't recall seeing such behavior with the iPhone 5 or 5s.
When is DED going to suss out the facts behind the privacy and security risks of giving keyboard extensions Full Access?
There's only so much optimization you can put into a benchmark for specific hardware, you can't just replace huge chunks of code for one device and call it a fair benchmark. It's a test of the hardware's raw performance, not the software. In a way it gives a false view of the end-user performance but it's designed to compare the same software on different hardware.
I trust Anandtech more:
http://www.phonearena.com/news/Correction-new-battery-life-results-of-G3-and-S5_id57484
Check the bottom of the page, takes you to meetgadgets, check the bottom of the page to see affiliates. They list Samsung phones there for some reason.
phonearena: cough, er it turns out that a device from one of our affiliates was benchmarked with a lower score so we're eh, just going to make a slight adjustment to that so now instead of the LG having 30 minutes more battery life, our custom web script now shows that the Samsung (our affiliate) has well over an hour more battery life.
The new Core M is way faster, both CPU and GPU.
But I'm not sure it matters. Core M is also much more expensive than what apple is likely paying for the A8, and most iphone and iPad users are not cpu constrained.
If you're thinking about MacBooks, the A8 is not competitive with Core M.
But apple is gaining fast on Intel. It could be that in another two to three years apple might be making SOCs that really could go in a MacBook air
Anybody else notice the 6 Plus charging basically twice as fast when using an iPad 11watt charger? This is nice, considering its much larger battery.
Does the iPhone 6 also charge faster? I don't recall seeing such behavior with the iPhone 5 or 5s.
You are far from alone apparently. I am super excited to see this. Wish Apple talked about it or included the 2.1A charger.
http://www.macrumors.com/2014/09/22/iphone-6-and-6-plus-charging/
This should be safe too (without reducing the longevity of the battery).
Let's see some others' results before concluding, ok? :P
Yeah, like fer sure, man. Give us some more results, so we can conclude.