I'll likely get one of these. My main computer is a 15" rMBP (with the discrete GPU) that's only a year old. I also have an iPad (3rd gen) with keyboard for taking notes in meetings, reading, browsing on the couch, and light gaming, but it's getting old. My work often requires that I have a second computer in order to test changes to our IT infrastructure and I travel more than just a little bit, so I have an 11" 2012 Air as well.
The new MacBook looks like a wonderful replacement for the Air that may even obviate the need for a new iPad. It looks like the perfect travel computer and if its performance is on par with my Air, I'd have no complaints. It has a much nicer display, is lighter, and has a better battery life, so using on a plane would be great too.
This machine may not be for everyone, but it's just what I was waiting for.
I'll likely get one of these. My main computer is a 15" rMBP (with the discrete GPU) that's only a year old. I also have an iPad (3rd gen) with keyboard for taking notes in meetings, reading, browsing on the couch, and light gaming, but it's getting old. My work often requires that I have a second computer in order to test changes to our IT infrastructure and I travel more than just a little bit, so I have an 11" 2012 Air as well.
The new MacBook looks like a wonderful replacement for the Air that may even obviate the need for a new iPad. It looks like the perfect travel computer and if its performance is on par with my Air, I'd have no complaints. It has a much nicer display, is lighter, and has a better battery life, so using on a plane would be great too.
This machine may not be for everyone, but it's just what I was waiting for.
1) So far it's the one I'm getting for my niche use as a Windows machine, but since I can't buy it today I might just wait to aee what other USB-C-powered Windows notebooks come out by this Summer. if the prices are under $800 I may forego having a great 12" MacBook. With the high price of Core-M under $800 may not be likely.
2) What sort of tests do you perform with a Mac on your network. IOW, are using Windows, do you have Mac apps, or is this purely general browser and defailt network tools diagnosistics that are mostly OS agnostic?
2) What sort of tests do you perform with a Mac on your network. IOW, are using Windows, do you have Mac apps, or is this purely general browser and defailt network tools diagnosistics that are mostly OS agnostic?
I manage(d) a small (4000 user) university network and am the IT security person for said university. I find it immensely useful to have a computer to test the effects of changes that I make to the network, or security appliances and such, without having to disconnect from my management sessions. It's mainly platform neutral stuff like web browsing and access to various IT services from different networks.
If you type ipad into the search and then 5300, you can see the scores between the Intel HD 5300 in the Macbook and last year's iPad. There were rumors that Apple tested ARM out in a laptop and given the single USB C port, this design looks like it was made for an ARM chip. They might have designed it, shown it to Intel and asked if they could build a chip to match so that they didn't have to deal with the headache of software compatibility.
There are advantages going either route. ARM is cheaper, a bit faster and already known to work smoothly over long periods in the iPad but Intel has more features and more software compatibility. Shaving off $200-300 going with ARM would have let them increase their marketshare. I don't think people really want to deal with software incompatibility again on the desktop side, that affected the Surface RT.
Intel has shown here that you couldn't get a crazy advantage out of ARM like quad processors in the same power threshold as a single Intel chip. In the same price threshold you could but the price premium isn't such a big deal when you get the compatibility.
Yes it does. I've (just) read your comments, but other than stating geekbench 3 isn't suited as a benchmark, no argument can be found.
Geekbench 3 is suited as a cross platform benchmark and the numbers speak for themselves: A8X 1.5 GHz 4488 and 1.1 GHz Core M 4267.
Corrected for the GHz Core M is faster but only slightly so and I suspect the low clock rate is for a reason (Core M dissipates a lot it seems).
Again, no it does not.
Regarding Geekbench, ARMv8 has its own instructions for SHA1/SHA2 that are ~4x faster than C++ that is regularly used for ARMv7 and x86 in the benchmarks. This can be looked at as an optimization in favor of ARMv8.
Some examples in Geekbench 3 (data for the A8 comes from jfpoole's Geekbench account. jfpoole is a founder and developer with Geekbench):
Apple A8 (32-bit/ARMv7 mode uses C++):
SHA1 scores: 1324 singe core / 2662 multi core
Apple A8 (64-bit mode uses ARMv8's instruction set):
SHA1 scores: 4542 singe core / 8784 multi core
Intel 5Y71 - ASUS T300 Chi (32-bit mode):
SHA1 scores: 3151 singe core / 5310 multi core
Intel 5Y71 - ASUS T300 Chi (64-bit mode):
SHA1 scores: 3520 singe core / 5822 multi core
I would also suggest you read my other posts as I highlighted the performance of the 5Y71 in the ASUS T300 Chi, a proper implementation of Core M.
If you type ipad into the search and then 5300, you can see the scores between the Intel HD 5300 in the Macbook and last year's iPad.
That GFXBench 3.0 benchmark for OpenGL ES 3.0 or equivalent is not to be trusted.
According to that same benchmark:
A8x (iPad Air 2) - 32.1 Fps
HD 5000 (Windows) - 29.6 Fps
HD 5000 (OS X) - 26.2 Fps
HD 4400 (Windows) - 25.7 Fps
HD 5300 (Windows) - 24.9 Fps
This is clearly nonsense, the iPad Air 2's GPU is not more powerful (or more capable) than the HD 5000 in the 2014 MacBook Air. Just because the benchmark has results for ARM and x86 devices for DirectX 9 or OpenGL ES 3.0 features, does not mean it should be treated as a valid source for comparison between the two.
That GFXBench 3.0 benchmark for OpenGL ES 3.0 or equivalent is not to be trusted.
According to that same benchmark:
A8x (iPad Air 2) - 32.1 Fps
HD 5000 (Windows) - 29.6 Fps
HD 5000 (OS X) - 26.2 Fps
<span style="line-height:1.4em;">HD 4400 (Windows) - 25.7 Fps</span>
HD 5300 (Windows) - 24.9 Fps
This is clearly nonsense, the iPad Air 2's GPU is not more powerful (or more capable) than the HD 5000 in the 2014 MacBook Air. Just because the benchmark has results for ARM and x86 devices for DirectX 9 or OpenGL ES 3.0 features, does not mean it should be treated as a valid source for comparison between the two.
That's true, sometimes higher quality tests show a larger difference. There's another test here with 3DMark:
[VIDEO]
The iPad Air 2 at 6:00 got 31k for the graphics score and the HD5000 on notebookcheck gets 65k:
That's last year's iPad though, they could double graphics speed in this year's iPad where Skylake vs Broadwell would only jump 20-30%. It's still OpenGL ES but the tests are designed to be comparable.
The main point is that the 2015 fanless iPad ARM chips will be in league with the 2015 fanless Intel chips for both CPU and GPU performance. That shouldn't come as a surprise.
They haven't said anything about a preorder, but it goes on sale on the 10th.
That is the pre-order date. What we don't know is when on that date the pre-order begins. I seem to recall the iPhone 6 pre-order starting at midnight PT, but don't quote me on that.
That's last year's iPad though, they could double graphics speed in this year's iPad where Skylake vs Broadwell would only jump 20-30%. It's still OpenGL ES but the tests are designed to be comparable.
The main point is that the 2015 fanless iPad ARM chips will be in league with the 2015 fanless Intel chips for both CPU and GPU performance. That shouldn't come as a surprise.
And this is a prime example of the misuse of benchmarks to figure out the performance difference between different processors. The iPad 3 uses an A5X which is not supported by the Metal API, so one test is going through Metal (a low latency - almost hardware level access) and one is going through a high-level API. Don't know if the 3D has 2D benchmarking, but unless you are using games or another 3D based application.
You are testing:
- The CPU across architectures (architecture can change even within generations of the same CPU family)
- Operating system libraries, operating system and power utility based drivers that may ramp up and ramp down based on power usage profiles which could differ in hardware implementations (how well the processor is heat-synced), how well the power utility drivers are implemented (change in if when power is ramped up or down to save battery power vs computing power). Even how well an environment was optimized for the benchmark itself. i.e. this is much more a factor when benchmarking a portable device than a desktop device (except for the OS libraries in general).
- Trying to test CPU performance itself which may have little to do with overall performance.
This benchmark is going to indicate that the iPad Air 2 is a much better platform for gaming as long as the operating system is up to date and uses the Metal API to bypass layers within the operating system to give you more direct hardware access for 3D based games.
Basically an abuse of benchmarks in this case since in this thread we are really discussing a device that is not really meant as a gaming platform.
This applies in different ways to different benchmarks. Take the benchmark across architectures with a big pinch of salt.
That is the pre-order date. What we don't know is when on that date the pre-order begins. I seem to recall the iPhone 6 pre-order starting at midnight PT, but don't quote me on that.
You sure? I could have sworn that they said "available" during the presentation. I tried to double check on the website, but it just says "coming soon".
You sure? I could have sworn that they said "available" during the presentation. I tried to double check on the website, but it just says "coming soon".
The watch was "preorder" on 10, "available" 24.
Mea culpa. I foolishly assumed it was about ?Watch.
You are correct, there is no stated date for the MacBook.
That specific processor has a little more "configurability" to it based on manufacturer's implementations.
If you look at the specs the specifics in regards to "Configurable TDP-up Frequency" "Configurable TDP-up" (Watts) are the range that the "base frequency" can be "adjusted" to. So when you see the Apple 1.3Ghz option, this is the same processor - just that Apple has implemented it as a 1.3Ghz option (with > 4.5W and < 6.0W) consumption (slight more drain on the battery). Apple could have even gone up to 1.4Ghz/6W option but I am guessing Apple either did not think it was a good implementation to use based on battery or heat and how much the case acts as a heat-sync. (It can also be configured down to 600Mhz/3.5W - which Apple probably thinks would provide a noticeable performance hit).
That's last year's iPad though, they could double graphics speed in this year's iPad where Skylake vs Broadwell would only jump 20-30%. It's still OpenGL ES but the tests are designed to be comparable.
The main point is that the 2015 fanless iPad ARM chips will be in league with the 2015 fanless Intel chips for both CPU and GPU performance. That shouldn't come as a surprise.
I'm not sure about that. Imagination Technologies stated that the PowerVR 7XT series, launching later this year, will only offer up to 60% improvement clock for clock, cluster for cluster. It is said that Intel has been focusing a considerable amount on improving graphics performance with Skylake's graphics offering up to a 50% improvement (unconfirmed). Skylake will support full feature DirectX 12_0 (current chips that support DX11 will not utilize all DX12 features). PowerVR 7XT offers improvements with options to offer DirectX 11.1 features (tessellation and geometry shaders). If anything, Apple's 2015 iPad chips will be closer to Intel's 2014 chips (still good in my opinion).
That aside, the advantages from Intel's offerings go beyond what benchmarks such as Geekbench or 3Dmark Ice Storm can show us. Of course, Intel charges a considerable premium for their parts.
Overall, I'm quite excited about the levels of performance and feature support we're seeing in mobile. The A9 chips should see a nice boost this year (improvements such as PVR 7XT and LPDDR4). It brings us closer to convergence across a multitude of devices with mostly performance being the limiting factor.
You sure? I could have sworn that they said "available" during the presentation. I tried to double check on the website, but it just says "coming soon".
The watch was "preorder" on 10, "available" 24.
Mea culpa. I foolishly assumed it was about ?Watch.
You are correct, there is no stated date for the MacBook.
Another sad sign of the new Apple.
Announce but don't give dates. Too close to Microsoft's way of doing things.
in the official PR statement sent to journalist Apple said the Macbook will be available
on april 10 via Online Stores and Retailer shops.
Source: german press statement
Preise und Verfügbarkeit
Das völlig neue MacBook wird ab Freitag 10. April über den Apple Online Store, die Apple Retail Stores und über ausgewählte autorisierte Apple Händlern ausgeliefert.
Comments
I'll likely get one of these. My main computer is a 15" rMBP (with the discrete GPU) that's only a year old. I also have an iPad (3rd gen) with keyboard for taking notes in meetings, reading, browsing on the couch, and light gaming, but it's getting old. My work often requires that I have a second computer in order to test changes to our IT infrastructure and I travel more than just a little bit, so I have an 11" 2012 Air as well.
The new MacBook looks like a wonderful replacement for the Air that may even obviate the need for a new iPad. It looks like the perfect travel computer and if its performance is on par with my Air, I'd have no complaints. It has a much nicer display, is lighter, and has a better battery life, so using on a plane would be great too.
This machine may not be for everyone, but it's just what I was waiting for.
1) So far it's the one I'm getting for my niche use as a Windows machine, but since I can't buy it today I might just wait to aee what other USB-C-powered Windows notebooks come out by this Summer. if the prices are under $800 I may forego having a great 12" MacBook. With the high price of Core-M under $800 may not be likely.
2) What sort of tests do you perform with a Mac on your network. IOW, are using Windows, do you have Mac apps, or is this purely general browser and defailt network tools diagnosistics that are mostly OS agnostic?
2) What sort of tests do you perform with a Mac on your network. IOW, are using Windows, do you have Mac apps, or is this purely general browser and defailt network tools diagnosistics that are mostly OS agnostic?
I manage(d) a small (4000 user) university network and am the IT security person for said university. I find it immensely useful to have a computer to test the effects of changes that I make to the network, or security appliances and such, without having to disconnect from my management sessions. It's mainly platform neutral stuff like web browsing and access to various IT services from different networks.
The graphics score of the A8X is higher too:
http://gfxbench.com/result.jsp?benchmark=gfx30
If you type ipad into the search and then 5300, you can see the scores between the Intel HD 5300 in the Macbook and last year's iPad. There were rumors that Apple tested ARM out in a laptop and given the single USB C port, this design looks like it was made for an ARM chip. They might have designed it, shown it to Intel and asked if they could build a chip to match so that they didn't have to deal with the headache of software compatibility.
There are advantages going either route. ARM is cheaper, a bit faster and already known to work smoothly over long periods in the iPad but Intel has more features and more software compatibility. Shaving off $200-300 going with ARM would have let them increase their marketshare. I don't think people really want to deal with software incompatibility again on the desktop side, that affected the Surface RT.
Intel has shown here that you couldn't get a crazy advantage out of ARM like quad processors in the same power threshold as a single Intel chip. In the same price threshold you could but the price premium isn't such a big deal when you get the compatibility.
Yes it does. I've (just) read your comments, but other than stating geekbench 3 isn't suited as a benchmark, no argument can be found.
Geekbench 3 is suited as a cross platform benchmark and the numbers speak for themselves: A8X 1.5 GHz 4488 and 1.1 GHz Core M 4267.
Corrected for the GHz Core M is faster but only slightly so and I suspect the low clock rate is for a reason (Core M dissipates a lot it seems).
Again, no it does not.
Regarding Geekbench, ARMv8 has its own instructions for SHA1/SHA2 that are ~4x faster than C++ that is regularly used for ARMv7 and x86 in the benchmarks. This can be looked at as an optimization in favor of ARMv8.
Some examples in Geekbench 3 (data for the A8 comes from jfpoole's Geekbench account. jfpoole is a founder and developer with Geekbench):
Apple A8 (32-bit/ARMv7 mode uses C++):
SHA1 scores: 1324 singe core / 2662 multi core
Apple A8 (64-bit mode uses ARMv8's instruction set):
SHA1 scores: 4542 singe core / 8784 multi core
Intel 5Y71 - ASUS T300 Chi (32-bit mode):
SHA1 scores: 3151 singe core / 5310 multi core
Intel 5Y71 - ASUS T300 Chi (64-bit mode):
SHA1 scores: 3520 singe core / 5822 multi core
I would also suggest you read my other posts as I highlighted the performance of the 5Y71 in the ASUS T300 Chi, a proper implementation of Core M.
Apple A8x
Single-core: 1808
Multi-core: 4529
Intel 5Y71
Single-core: 2944
Multi-core: 5680
As for other, non-Geekbench benchmarks?
3DMark 11
iPad Air 2 (A8x)
Ice Storm: 21650
Cloud Gate: N/A
Sky Diver: N/A
Fire Strike: N/A
T300 Chi (5Y71)
Ice Storm: 44803
Cloud Gate: 4784
Sky Diver: 2440
Fire Strike: 633
The graphics score of the A8X is higher too:
http://gfxbench.com/result.jsp?benchmark=gfx30
If you type ipad into the search and then 5300, you can see the scores between the Intel HD 5300 in the Macbook and last year's iPad.
That GFXBench 3.0 benchmark for OpenGL ES 3.0 or equivalent is not to be trusted.
According to that same benchmark:
A8x (iPad Air 2) - 32.1 Fps
HD 5000 (Windows) - 29.6 Fps
HD 5000 (OS X) - 26.2 Fps
HD 4400 (Windows) - 25.7 Fps
HD 5300 (Windows) - 24.9 Fps
This is clearly nonsense, the iPad Air 2's GPU is not more powerful (or more capable) than the HD 5000 in the 2014 MacBook Air. Just because the benchmark has results for ARM and x86 devices for DirectX 9 or OpenGL ES 3.0 features, does not mean it should be treated as a valid source for comparison between the two.
That's true, sometimes higher quality tests show a larger difference. There's another test here with 3DMark:
[VIDEO]
The iPad Air 2 at 6:00 got 31k for the graphics score and the HD5000 on notebookcheck gets 65k:
http://www.notebookcheck.net/Intel-HD-Graphics-5000.91978.0.html
That's last year's iPad though, they could double graphics speed in this year's iPad where Skylake vs Broadwell would only jump 20-30%. It's still OpenGL ES but the tests are designed to be comparable.
The main point is that the 2015 fanless iPad ARM chips will be in league with the 2015 fanless Intel chips for both CPU and GPU performance. That shouldn't come as a surprise.
This page seems so indicate that it supports virtualization, or am I reading it incorrectly? (no sarcasm, genuine curiosity)
http://ark.intel.com/products/84672/Intel-Core-M-5Y71-Processor-4M-Cache-up-to-2_90-GHz
nice find - anyone know when Apple are opening up pre-order?
They haven't said anything about a preorder, but it goes on sale on the 10th.
That is the pre-order date. What we don't know is when on that date the pre-order begins. I seem to recall the iPhone 6 pre-order starting at midnight PT, but don't quote me on that.
That's true, sometimes higher quality tests show a larger difference. There's another test here with 3DMark:
The iPad Air 2 at 6:00 got 31k for the graphics score and the HD5000 on notebookcheck gets 65k:
http://www.notebookcheck.net/Intel-HD-Graphics-5000.91978.0.html
That's last year's iPad though, they could double graphics speed in this year's iPad where Skylake vs Broadwell would only jump 20-30%. It's still OpenGL ES but the tests are designed to be comparable.
The main point is that the 2015 fanless iPad ARM chips will be in league with the 2015 fanless Intel chips for both CPU and GPU performance. That shouldn't come as a surprise.
And this is a prime example of the misuse of benchmarks to figure out the performance difference between different processors. The iPad 3 uses an A5X which is not supported by the Metal API, so one test is going through Metal (a low latency - almost hardware level access) and one is going through a high-level API. Don't know if the 3D has 2D benchmarking, but unless you are using games or another 3D based application.
You are testing:
- The CPU across architectures (architecture can change even within generations of the same CPU family)
- Operating system libraries, operating system and power utility based drivers that may ramp up and ramp down based on power usage profiles which could differ in hardware implementations (how well the processor is heat-synced), how well the power utility drivers are implemented (change in if when power is ramped up or down to save battery power vs computing power). Even how well an environment was optimized for the benchmark itself. i.e. this is much more a factor when benchmarking a portable device than a desktop device (except for the OS libraries in general).
- Trying to test CPU performance itself which may have little to do with overall performance.
This benchmark is going to indicate that the iPad Air 2 is a much better platform for gaming as long as the operating system is up to date and uses the Metal API to bypass layers within the operating system to give you more direct hardware access for 3D based games.
Basically an abuse of benchmarks in this case since in this thread we are really discussing a device that is not really meant as a gaming platform.
This applies in different ways to different benchmarks. Take the benchmark across architectures with a big pinch of salt.
You sure? I could have sworn that they said "available" during the presentation. I tried to double check on the website, but it just says "coming soon".
The watch was "preorder" on 10, "available" 24.
Mea culpa. I foolishly assumed it was about ?Watch.
You are correct, there is no stated date for the MacBook.
5Y71 (MacBook top model):
- Base frequency: 1.2 GHz
- Max turbo frequency: 2.9 GHz
- Graphics base frequency: 300 MHz
- Graphics max frequency: 900 MHz
That specific processor has a little more "configurability" to it based on manufacturer's implementations.
If you look at the specs the specifics in regards to "Configurable TDP-up Frequency" "Configurable TDP-up" (Watts) are the range that the "base frequency" can be "adjusted" to. So when you see the Apple 1.3Ghz option, this is the same processor - just that Apple has implemented it as a 1.3Ghz option (with > 4.5W and < 6.0W) consumption (slight more drain on the battery). Apple could have even gone up to 1.4Ghz/6W option but I am guessing Apple either did not think it was a good implementation to use based on battery or heat and how much the case acts as a heat-sync. (It can also be configured down to 600Mhz/3.5W - which Apple probably thinks would provide a noticeable performance hit).
That's true, sometimes higher quality tests show a larger difference. There's another test here with 3DMark:
The iPad Air 2 at 6:00 got 31k for the graphics score and the HD5000 on notebookcheck gets 65k:
http://www.notebookcheck.net/Intel-HD-Graphics-5000.91978.0.html
That's last year's iPad though, they could double graphics speed in this year's iPad where Skylake vs Broadwell would only jump 20-30%. It's still OpenGL ES but the tests are designed to be comparable.
The main point is that the 2015 fanless iPad ARM chips will be in league with the 2015 fanless Intel chips for both CPU and GPU performance. That shouldn't come as a surprise.
I'm not sure about that. Imagination Technologies stated that the PowerVR 7XT series, launching later this year, will only offer up to 60% improvement clock for clock, cluster for cluster. It is said that Intel has been focusing a considerable amount on improving graphics performance with Skylake's graphics offering up to a 50% improvement (unconfirmed). Skylake will support full feature DirectX 12_0 (current chips that support DX11 will not utilize all DX12 features). PowerVR 7XT offers improvements with options to offer DirectX 11.1 features (tessellation and geometry shaders). If anything, Apple's 2015 iPad chips will be closer to Intel's 2014 chips (still good in my opinion).
That aside, the advantages from Intel's offerings go beyond what benchmarks such as Geekbench or 3Dmark Ice Storm can show us. Of course, Intel charges a considerable premium for their parts.
Overall, I'm quite excited about the levels of performance and feature support we're seeing in mobile. The A9 chips should see a nice boost this year (improvements such as PVR 7XT and LPDDR4). It brings us closer to convergence across a multitude of devices with mostly performance being the limiting factor.
Another sad sign of the new Apple.
Announce but don't give dates. Too close to Microsoft's way of doing things.
in the official PR statement sent to journalist Apple said the Macbook will be available
on april 10 via Online Stores and Retailer shops.
Source: german press statement
Preise und Verfügbarkeit
Das völlig neue MacBook wird ab Freitag 10. April über den Apple Online Store, die Apple Retail Stores und über ausgewählte autorisierte Apple Händlern ausgeliefert.
Not in the UK.
Here, it's 'Coming Soon', much like iTunes Radio.
So we can look forward to it arriving in 2018 or later, if ever.