[AppleInsider] Apple may abandon Intel for its Macs starting with post-Broadwell

[TreVader]

You do realize you can't simply take Geekbench scores and start making conclusions about performance on different architectures right? Otherwise Samsung's new Exynos 7420 smartphone SoC would be faster than a 15W Core i5 5200U (Broadwell-U) in MT tasks, and it most certainly isn't.

Also where does the 1/10 peak power draw statement comes from? Please enlighten us, I'm curious.

It isn't? Do you have both and would like to compare them?


Don't just assume something is faster because intel makes it. That's what got you guys into this mess of using windows in the first place. If intel was so awesome why aren't any of us using "intel inside" phones? Because they aren't awesome at all.

Geekbench isn't the only thing I can compare, A8X competes very well with 3Ghz intel CPUs in browser benchmarks too. In real world use it's also just as fast as intel.


Intel's favorite thing to do is to say "well, that benchmark isn't accurate, but look at our benchmarks right here!". Nobody beleives them anymore, because they have finally overpromised and underdelivered so much that Apple is through with them for Macbook Air.

 

[Sweepr]

It isn't? Do you have both and would like to compare them?

I have an Exynos 5433 phone and while it is very fast it doesn't match a Core i5 4200U in MT tasks like Geekbench suggests.

Geekbench isn't the only thing I can compare, A8X competes very well with 3Ghz intel CPUs in browser benchmarks too. In real world use it's also just as fast as intel.

Try running some real software instead of that crap called iOS and you'll see how responsive those chips are compared to Broadwell-U and why Apple is not ditching Intel anytime soon.

 

[TreVader]

Try running some real software instead of that crap called iOS and you'll see how responsive those chips are compared to Broadwell-U and why Apple is not ditching Intel anytime soon.

So, your solution is for me to run your terrible software, which will then slow my device down to the point where I will see that intel is the fastest?



:\



I think we will have to just agree to disagree. It won't be long until people see who was right.

 

[Brunnis]

It isn't? Do you have both and would like to compare them?


Don't just assume something is faster because intel makes it. That's what got you guys into this mess of using windows in the first place. If intel was so awesome why aren't any of us using "intel inside" phones? Because they aren't awesome at all.

Geekbench isn't the only thing I can compare, A8X competes very well with 3Ghz intel CPUs in browser benchmarks too. In real world use it's also just as fast as intel.


Intel's favorite thing to do is to say "well, that benchmark isn't accurate, but look at our benchmarks right here!". Nobody beleives them anymore, because they have finally overpromised and underdelivered so much that Apple is through with them for Macbook Air.

Can you post links that compare the performance of the A8X vs Haswell? I'm really interested in this and would appreciate to know where you've seen/read it. Or have you run tests yourself?

A couple of days ago, I read this article on Anandtech: http://anandtech.com/show/8928/acer-chromebook-13-1080p-with-tegra-k1

It contains several browser benchmark results for the low-end Celeron 2955U (1.4GHz dual core, 2MB cache, no HT, no turbo). The results paint a very different picture from the one you have, with the Intel chip beating the A8X quite severely in the heavier tests despite lacking a core. What do you think is the reason for these anomalous results?

 

[TreVader]

Can you post links that compare the performance of the A8X vs Haswell? I'm really interested in this and would appreciate to know where you've seen/read it. Or have you run tests yourself?

A couple of days ago, I read this article on Anandtech: http://anandtech.com/show/8928/acer-chromebook-13-1080p-with-tegra-k1

It contains several browser benchmark results for the low-end Celeron 2955U (1.4GHz dual core, 2MB cache, no HT, no turbo). The results paint a very different picture from the one you have, with the Intel chip beating the A8X quite severely in the heavier tests despite lacking a core. What do you think is the reason for these anomalous results?

There are several threads that contain that information, and it is common knowledge amongst people on this forum.


Off the top of my head, the A8X scores something like 1800/4500 Mutli/Single. Intel's top of the line Core M scores around 1900/3800 in the Yoga 3 Pro.


You can search A8, A8X, or A7 and you will find three separate threads where this is discussed. I am sure you already are aware of the scores. He cited some unknown Exynos chip (I've never heard of it and I am well versed in mobile SoCs) and provided no evidence. I don't need to provide evidence for something like A8 or A8X because it is assumed you would know what everyone else already does before you came in this thread.

Edit: Missed your question. The benchmarks you cite are WebXpert and Mozilla Kraken, which are anomalies. In the bottom of the page, the Air 2 beats every single thing by a large margin in the last benchmark, and continues to dominate every other page.

 

[Essence_of_War]

Intel's favorite thing to do is to say "well, that benchmark isn't accurate, but look at our benchmarks right here!"

The benchmarks you cite are WebXpert and Mozilla Kraken, which are anomalies.

Really?

 

[Brunnis]

There are several threads that contain that information, and it is common knowledge amongst people on this forum.


Off the top of my head, the A8X scores something like 1800/4500 Mutli/Single. Intel's top of the line Core M scores around 1900/3800 in the Yoga 3 Pro.


You can search A8, A8X, or A7 and you will find three separate threads where this is discussed. I am sure you already are aware of the scores. He cited some unknown Exynos chip (I've never heard of it and I am well versed in mobile SoCs) and provided no evidence. I don't need to provide evidence for something like A8 or A8X because it is assumed you would know what everyone else already does before you came in this thread.

No, I'm not really aware of any comprehensive tests that show the A8X besting Haswell/Broadwell. But thanks for being so pleasant, as always.

Anyway, Anandtech's results are quite opposite of what you're claiming in regards to browser performance. That's why I thought it would be interesting to hear your views on it.

Edit: Missed your question. Tegra K1 is a buggy mess, and anybody who has used it will tell you that. It does perform well in a few benchmarks, but if there is any company that is BUILT around fudging benchmarks, it's nvidia. They have been doing this for decades. Tegra K1 was a flop just like every other tegra, and winning a couple browser benchmarks while losing all others isn't all that impressive. On top of that, K1 runs extremely hot and sucks power like crazy. Or did you actually buy a Nexus 9, and can take pics to show how awesome it is? I've never met somebody who didn't return it!

I wasn't interested in the K1 results in the link I provided. The article provides results for the low-clocked Acer C720 with a Haswell Celeron 2955U. The C720 results compared to the Ipad Air 2 results were what I thought was interesting.

 

[TreVader]

I wasn't interested in the K1 results in the link I provided. The article provides results for the low-clocked Acer C720 with a Haswell Celeron 2955U. The C720 results compared to the Ipad Air 2 results were what I thought was interesting.

I misread the question (or only read the beginning, clicked the link, thought you meant K1).


The two that apple doesn't win handedly (against cited ARM and x86-64 android) are outliers. Apple does not necessarily win even a majority of benchmarks against Core M, my point is that it provides adequate performance and competes, using a FAR smaller TDP. Anybody who has held an ipad air 2 will tell you it does not heat up to more than luke warm even under heavy load. It also lasts 10-11hrs on a tiny battery that wouldn't give you more than a couple hours on a Core M device.

 

[dawheat]

I'd certainly not buy another Macbook Pro if it didn't have a i7. I'm able to use it as my primary machine as it can handle everything my desktop can. Last I recall, it has something like twice the single core performance and three times the multicire performance of the A8X.

It's easy to decry Intel if you're going to cherry pick SKUs, but they are still far and away the best at the high end. So if Intel is terrible, then everyone else is a disaster.

Even in the most optimistic case, the A-series is years away from the powerhouse Intel chips available now. And until Apple can cut an entire model line over, I doubt they'll be as foolish to introduce consumer and operations complexity by having different SKUs of the same generation be incompatible. That would be a nightmare.

Lastly being able to boot into Windows is still close to mandatory for corporate machines . It's just the way things are unfortunately

Edit - my point is let's not get carried away in calling Intel names. Core M is kind of disappointing but that's not yet where their strength is.

 

[Nothingness]

The benchmarks you cite are WebXpert and Mozilla Kraken, which are anomalies.

Really?

WebXPRT was developed with Intel, so definitely suspicious. Kraken is a different story...

 

[Enigmoid]

MS just reported record revenue on Surface. There are rumors that say MS is ditching windows on ARM. If they go all out x86, I think they will eat a chunk out of ipad/mac sales

I will throw in my two cents and say from what I have observed in the student and professional environment the surface pro is immensely popular.

 

[xthetenth]

Edit: Missed your question. The benchmarks you cite are WebXpert and Mozilla Kraken, which are anomalies. In the bottom of the page, the Air 2 beats every single thing by a large margin in the last benchmark, and continues to dominate every other page.

So other than sunspider, which is notorious for the performance optimizations it receives and is being compared cross-platform, you're arguing that the Air 2's screen quality means its processor is better than a random chromebook's processor.

And the surface pro is a vastly superior device to anything with a fruit on it, and Apple really needs to step up its game, its laptops are dated compared to windows offerings.

 

[III-V]

BDW-Y is the same die as the GT2 BDW-Us, as far as I know.

Core M is supposedly on its own process flavor (there's 14nm SoC, 14nm HP, and 14nm... Core M).

I will throw in my two cents and say from what I have observed in the student and professional environment the surface pro is immensely popular.

It only started with the Surface Pro 3. The first and second were duds. Used to work at a campus store.

 

[justanotherboy]

He cited some unknown Exynos chip (I've never heard of it and I am well versed in mobile SoCs) and provided no evidence.

Are you serious?

 

[III-V]

Are you serious?

Yeah, he is. He's been given numerous infractions for his crackpot theories, however this subforum is no longer moderated; thus we're subjected to his alternate realities on anything regarding Apple.

 

[elemein]

All CPUs have instruction fetch unit or they couldn't load instructions from memory to feed the decoders :biggrin:

Joke aside these diagrams are just missing a lot of information, all of the CPU's mentioned have an instruction cache, a branch prediction unit and most likely a form of hardware prefetch.

Yeah I figured that much xD Thanks! Seems Apple and Qualcomm were just being a little tight lipped.

Also, unrelated entirely, but does anyoen know why the ICache is sometimes depicted AFTER the instruction fetch unit?

 

[OBLAMA2009]

man, when i imagine macs without intel, im thinking of something as bad as those arm chromebooks. what a joke those are man

 

[TreVader]

Yeah, he is. He's been given numerous infractions for his crackpot theories, however this subforum is no longer moderated; thus we're subjected to his alternate realities on anything regarding Apple.

At least you're honest and can see that I'm infracted for my opinions on Intel, not my attitude.


It's tough to be the only one to see the elephant in the room. It will all be worth it when apple releases this thing and all my predictions come to fruition.

 

[tential]

At least you're honest and can see that I'm infracted for my opinions on Intel, not my attitude.


It's tough to be the only one to see the elephant in the room. It will all be worth it when apple releases this thing and all my predictions come to fruition.

No, your infracted because you can't post in a civilized manner.

As for you being the only one to see the elephant in the room? Have you ever considered that maybe, if you're the ONLY person seeing it, you're delusional?

 

[TreVader]

No, your infracted because you can't post in a civilized manner.

As for you being the only one to see the elephant in the room? Have you ever considered that maybe, if you're the ONLY person seeing it, you're delusional?

I post in the EXACT same manner the intel crew posts in. I only give back what I get from you guys, the difference is that suddenly you have somebody you don't agree with, and you don't like that!


Neither does any of the rest of the people here who constantly state that it's impossible for intel to be beat in anything, despite them getting beat multiple times before from the lowely likes of AMD.


You guys cannot handle criticism. You freak out, and then whine to the mods when I insult you back. I don't have time to report all your attacks, I have a life, but you guys all seem to find the time to report even stuff that is totally inoccuous.

 

[tential]

I post in the EXACT same manner the intel crew posts in. I only give back what I get from you guys, the difference is that suddenly you have somebody you don't agree with, and you don't like that!


Neither does any of the rest of the people here who constantly state that it's impossible for intel to be beat in anything, despite them getting beat multiple times before from the lowely likes of AMD.


You guys cannot handle criticism. You freak out, and then whine to the mods when I insult you back. I don't have time to report all your attacks, I have a life, but you guys all seem to find the time to report even stuff that is totally inoccuous.

When you think everyone is against you, chances are you're doing something to deserve it.
Being a very strong headed person myself, I've had to learn that the hard way many times.

Many people you're accusing to be "intel fanboys" are actually not. You just simply don't always make the best arguments for yourself. Even when you do have a decent argument at times, you present it in such a preposterous manner as to cause people to just laugh at you and take nothing you say seriously.

We could make an insane laugh reel from comments you've posted just in this thread alone.

You need to tone down some of your more outlandish statements if you want people to take you seriously.

 

[ninaholic37]

Off the top of my head, the A8X scores something like 1800/4500 Mutli/Single. Intel's top of the line Core M scores around 1900/3800 in the Yoga 3 Pro.

I read reviews that Yoga 3 Pro with Core M does not perform very well either. Could be another reason why we don't see Macbook Air with it yet, maybe they will wait for the Skylake version or something else. I imagine it would take a long time to scale A8X up to 15W (Broadwell-U levels) well though. Has anything with ARM ever tried this yet? What is the highest TDP ARM chip version out there / in testing phase at the moment? Surely someone has pondered the idea of how this could work or be engineered, you would think. :hmm:

 

[SAAA]

I read reviews that Yoga 3 Pro with Core M does not perform very well either. Could be another reason why we don't see Macbook Air with it yet, maybe they will wait for the Skylake version or something else. I imagine it would take a long time to scale A8X up to 15W (Broadwell-U levels) well though. Has anything with ARM ever tried this yet? What is the highest TDP ARM chip version out there / in testing phase at the moment? Surely someone has pondered the idea of how this could work or be engineered, you would think. :hmm:

I second these toughts: is there anyone who even tried to build a 15W or+ ARM soc for mobile (laptops, mini pcs)? Be it a 2-8 cores I don't mind but just that specific TDP and consequently higher clocks.
Or are they really limited to sub 3GHz clocks with the underlying process and architectures?

If that's the case then well I see a problem and I'm sure it's not going to be solved by 16nm finfet either: we all know what happened when Intel switched to 22nm at the max clock speeds... they decreased, while at any other point they only got more power efficient.

So beside maybe Apple who seemingly developed a high IPC core the rest of the crowd should still be limited by low max clocks and consequently following the route of more cores. 10,12 who's next?

 

[Qwertilot]

Entirely possible in principle - they're building arm based server chips you know

The standard arm cores don't aim there.

Whether anyone is doing 15w laptopish chips it in earnest is probably something you'd have to ask Apple Maybe Nvidia could hit 15w semi sensibly for some sort of micro console by really cranking up their iGPU.

 

[trivik12]

Isn't IPad Pro confirmed already for H1 this year. May be it will come up with desktop mode with better Multi-tasking ablities. Current iphone/ipad are not comparable to mac as they are constrained big time. You cannot open 30 tabs on your ipad like what people do on their mac.

I hope Apple releases a machine with A8x comparable to macbook air in feature capabilities and then compare their performances.