Haswell to have 10 hours of battery life in tablet, 13 hours docked in real world use

[rgallant]

Once again, too much on performance angle, and too little on the "good enough" factor. Besides ARM isn't remotely close to AMD one bit, the biggest device makers are designing their own chips, and Samsung is fabbing their own on top of that. Why would they switch when most consumers don't even care what inside their phones/tablets? So they can become bitches to Intel like PC OEMs now?

-as stated I 'm a mobile noob - but if good enough is the rule why are people upgrading at much more cost$$$ every 3-6 months ???

 

[LogOver]

Wait for real world reviews. Marketing spin, with no real information is pretty much useless. What size battery? Doing what tasks? Screen size and pixel density? What GPU?

I'm looking forward to Haswell, as it's time for a new desktop. But Haswell in a tablet? Too many unknowns right now.

At least it looks good on the paper. Haswell ULT/ULX doesn't even need to exit power-saving S0i1 state during internet browsing. And it can run Unigine benchmark while consumes just under 8W. Exynos 5 Dual consumes around 10W in game benchmarks.

 

[Imouto]

By that chart the MS Surface RT with a Tegra3 (quite a power hungry chip) and a 31.5 Wh beats the battery life of the MS Surface Pro with an underclocked i5 3317U and a 42 Wh battery.

33% more battery = 33% lower battery life

Please tell me how a Haswell chip can two fold the efficiency of the current Ivy Bridge chips.

Haswell devices lasting 10 hours? Sure, with a backpack for the battery like the MS Surface Pro already needs and still can't do.

 

[blackened23]

Once again, too much on performance angle, and too little on the "good enough" factor. Besides ARM isn't remotely close to AMD one bit, the biggest device makers are designing their own chips, and Samsung is fabbing their own on top of that. Why would they switch when most consumers don't even care what inside their phones/tablets? So they can become bitches to Intel like PC OEMs now?

There is cheap and there is good. ARM SOCs are cheap - again the only reason they're used in tablets is because intel isn't competitive in terms of efficiency. Now that they will be, there is no reason to use ARM SOCs in high end 500$+ devices. Further, I don't consider 200$ tablets to be good. Having used a nexus 7 tablet, it is okay. But you will not find one person who would rather have that over a full blown retina iPad, unless they just hate apple or they can't afford the real thing. Besides which,intel can make an atom SOC on the 14nm node and sell 10$ chips if they want, which will have twice as much battery life as ARM SOCs (if haswell is any indicator).

500$ tablets are a significant market. Sure some people can't afford them - but it has by far the highest profit margins. IMO, intel will get those profits with broadwell. And for those who can't afford the high end chips, i'm sure intel will make ULV low end chips with nearly that much more battery life than both broadwell and ARM SOCs - consider the current battery life of cloverdale. If intel makes a low cost Atom SOC based on the 14nm process, it will be a bloodbath in the low end cheap market that you guys apparently love. You get what you pay for.

 

[LogOver]

Please tell me how a Haswell chip can two fold the efficiency of the current Ivy Bridge chips.

http://www.anandtech.com/show/6355/intels-haswell-architecture/3
http://wccftech.com/intel-haswell-u...ving-features-lynx-point-lp-chipset-detailed/

 

[jpiniero]

500$ tablets are a significant market.

No it's not. The only $500 tablet selling is the iPad normal, and the Mini really has eclipsed it.

 

[blackened23]

No it's not. The only $500 tablet selling is the iPad normal, and the Mini really has eclipsed it.

Like I said, intel will make chips for both the cheap junk crowd, and other chips for the high end market. I don't see a compelling reason to use an ARM SOC for a 500$ tablet if intel has a chip that beats it by every metric - and beats it in performance 20 fold. At this time, ARM SOCs are used because of efficiency and efficiency alone in the 500$+ tablet market.

 

[jpiniero]

I don't see a compelling reason to use an ARM SOC for a 500$ tablet

That's the point - there is no market. It's literally going to take Haswell at Atom prices for Intel to have a shot.

 

[blackened23]

That's the point - there is no market. It's literally going to take Haswell at Atom prices for Intel to have a shot.

Oh okay. If you say so. That sounds like a bold statement with no proof.

Yes, cheap junk moves more volume. Yet, the nexus 10 is continually sold out at the google play store and has sold millions upon millions, while Apple has sold tons of retina ipads. Nobody else has really attempted to compete in the high end market seriously. (in terms of tablets)

 

[ThePolarSky]

...

 

[Imouto]

http://www.anandtech.com/show/6355/intels-haswell-architecture/3
http://wccftech.com/intel-haswell-u...ving-features-lynx-point-lp-chipset-detailed/

Nice read. Still PR and a jump big enough to make it look impossible.

 

[2timer]

There is cheap and there is good. ARM SOCs are cheap - again the only reason they're used in tablets is because intel isn't competitive in terms of efficiency. Now that they will be, there is no reason to use ARM SOCs in high end 500$+ devices. Further, I don't consider 200$ tablets to be good. Having used a nexus 7 tablet, it is okay. But you will not find one person who would rather have that over a full blown retina iPad, unless they just hate apple or they can't afford the real thing. Besides which,intel can make an atom SOC on the 14nm node and sell 10$ chips if they want, which will have twice as much battery life as ARM SOCs (if haswell is any indicator).

500$ tablets are a significant market. Sure some people can't afford them - but it has by far the highest profit margins. IMO, intel will get those profits with broadwell. And for those who can't afford the high end chips, i'm sure intel will make ULV low end chips with nearly that much more battery life than both broadwell and ARM SOCs - consider the current battery life of cloverdale. If intel makes a low cost Atom SOC based on the 14nm process, it will be a bloodbath in the low end cheap market that you guys apparently love. You get what you pay for.

This idea of Intel slaughtering and destroying ARM that you cherish is purely speculation at this point. The average consumer doesn't make purchasing decisions based purely on CPU performance.

ARM tablets sell from $149-249. A Haswell tablet will most likely be twice that. Most people aren't going to spend those bucks on a tablet because they don't need to pay a high price for performance they don't need in the first place

 

[Roland00Address]

By that chart the MS Surface RT with a Tegra3 (quite a power hungry chip) and a 31.5 Wh beats the battery life of the MS Surface Pro with an underclocked i5 3317U and a 42 Wh battery.

33% more battery = 33% lower battery life

Please tell me how a Haswell chip can two fold the efficiency of the current Ivy Bridge chips.

Haswell devices lasting 10 hours? Sure, with a backpack for the battery like the MS Surface Pro already needs and still can't do.

First you need to be comparing the same screens or the comparison is useless. A 1920x1080 screen uses more power than a 1366x768 screen. 1920x1080 is double the amount of pixels as 1366x768. When apple quadrupled the resolution of the ipad 2 with the ipad 3 they also made the battery 70% larger and even then the ipad 2 had 10% better battery life than the ipad 3.

By integrating the southbridge (Platform Controller Hub), as well as picking the rest of the components in the computer wisely you can reduce idle power of the computer substantially and idle power is what affects battery life.

We are going to see a substantial increase in battery life with haswell, who knows if we get the 9 hours in Ultrabooks (mobilemark 2007) that intel is expecting with haswell (currently ultrabooks are required to get 5 hours in mobilemark 2007, intel wants as a goal 7 hours but they only require 5 hours.)

 

[erunion]

Because ARM cost $20 and haswell costs $200

People are willing to pay $200 for PC cpus even though cheap celerons exist.
If Intel can deliver the performance, people will pay for it. ARM chips are commodities, Intel chips aren't. Its that simple.

The ARM hype is so funny to me. ARM chips exist unchallenged in smartphones for a few years and people become committed to the idea that they are unsinkable.

 

[CHADBOGA]

This idea of Intel slaughtering and destroying ARM that you cherish is purely speculation at this point. The average consumer doesn't make purchasing decisions based purely on CPU performance.

ARM tablets sell from $149-249. A Haswell tablet will most likely be twice that. Most people aren't going to spend those bucks on a tablet because they don't need to pay a high price for performance they don't need in the first place

I wouldn't go so far as to say that Intel are going to slaughter ARM, but they are clearly going to rough them up a bit.

Haswell will attack ARM from the top and Bay Trail will attack ARM from the bottom.

I think Bay Trail, rather than Haswell, will do more damage to ARM.

 

[blackened23]

People are willing to pay $200 for PC cpus even though cheap celerons exist.
If Intel can deliver the performance, people will pay for it. ARM chips are commodities, Intel chips aren't. Its that simple.

The ARM hype is so funny to me. ARM chips exist unchallenged in smartphones for a few years and people become committed to the idea that they are unsinkable.

Precisely this. I'm sick of the "Good enough" performance angle. There are a lot of people who want something more than cheap junk, and there are customers who are willing to pay for devices that aren't cheap junk. I'm truly sorry that there are so many cheap junk fans found within this thread, maybe these guys are buying AMD CPUs as well. I am one of those who refuse to buy cheap junk ie 150$ junk android tablets - while some people don't mind buying cheap junk, there are plenty who want more.

Again, intel will have chips for both the low cost and high end segments - and people really need to expand their mindset as to what is possible on a tablet. Haswell will open the doors up in terms of what is possible on a tablet, while simultaneously doing it with a similar battery life similar to what is available on ARM devices. Think of an ultrabook, only that is now possible in a tablet form factor WITH the same battery life as an iPad4. That is an exciting prospect.

I like the iPad, a lot, but there is a lot it can't do by virtue of the ARM SOC found within. If haswell delivers on battery life, this will all change. The capabilities of existing high end tablets will increase a ton.

 

[Bill Brasky]

Like I said, intel will make chips for both the cheap junk crowd, and other chips for the high end market. I don't see a compelling reason to use an ARM SOC for a 500$ tablet if intel has a chip that beats it by every metric - and beats it in performance 20 fold. At this time, ARM SOCs are used because of efficiency and efficiency alone in the 500$+ tablet market.

I do. Apple makes incredible margins selling Ipads with custom IC. Do you really think they'll stop they're R and D and start giving 200 bucks/Ipad to intel? We've already established that Ipads are currently the most popular tablet, and that Ipad users don't really care about performance. What motivation does Apple have to use intel in their tablets?

 

[AtenRa]

I have bought a cheap 8” sub 200.00 euro dual core 1.5GHz Tablet for my wife. I have used it a couple of times while we were for a coffee and I can tell you it was just fine reading Anandtech/browsing and play a few games.

Now, if you need to replace your Laptop with a tablet you will need more performance and a keyboard but I don’t see Haswell tablets bellow the $600-700 mark.

People will not spend that much for a tablet, they could buy a nice 15,6” Laptop + a cheap 8” tablet and have both speding the same or less.

 

[dma0991]

I like the iPad, a lot, but there is a lot it can't do by virtue of the ARM SOC found within. If haswell delivers on battery life, this will all change. The capabilities of existing high end tablets will increase a ton.

It has never occurred to me that it will be able to do more with more performance since its in the good enough category. The problem is a software one, maybe an interfacing(no physical keyboard) one as well, not purely based on hardware performance. Even right now, a mid range tablet could do >80% of what a high end tablet does in terms of usability.

If x86 tablet variants become the norm, the only benefit would be the ability to install any x86 based OS currently available. I feel that is a far more interesting prospect.

 

[Piroko]

-as stated I 'm a mobile noob - but if good enough is the rule why are people upgrading at much more cost$$$ every 3-6 months ???

Imho, for the same reason that quite a lot of people buy new cars as soon as their current ones are 3-4 years old, style obsolescence. Ever older Galaxy S or iphone dramatically lost percieved value as soon as the new one was released, even though an iphone 3GS or Galaxy 1 is still doing just fine for everyday surfing.

Intel may very well be a big player in the tablet market, if - and only if - they manage enough high profile wins. Microsoft might want to push into that market with the Intel wave as well, that could be a nice differentiator.

 

[biostud]

It really depends on whether it will be used in a laptop replacement or just another entertainment tablet. Maybe even apple will make a iPad pro based on haswell. With support for hdmi, mouse and keyboard/docking station pro tablets would be very competitive.

 

[beginner99]

I think haswell will be great and I believe the battery live claims. I mean it's also a battery of battery size. But my X220 laptop with a Sandy bridge has already very long battery life, I never tested it but I think it can reach 10 hours...on the bigger battery pack. Since haswell has like 50x fold lower idle power consumption and tablets have smaller screens, 10 hours doesn't seem that special. However one should never forget that this is only possible with low power settings in windows and that has a huge effect on CPU performance.

But then I also agree with the posters that are dubious of it's success because of pricing. I just don't see a haswell tablet selling for under $700 and people don't really care about CPU performance, the will rather buy the $200 ARM one (or already got one for free from a deal with a provider/carrier which is pretty much standard here nowadays).

 

[lagokc]

Precisely this. I'm sick of the "Good enough" performance angle. There are a lot of people who want something more than cheap junk, and there are customers who are willing to pay for devices that aren't cheap junk. I'm truly sorry that there are so many cheap junk fans found within this thread, maybe these guys are buying AMD CPUs as well. I am one of those who refuse to buy cheap junk ie 150$ junk android tablets - while some people don't mind buying cheap junk, there are plenty who want more.

Again, intel will have chips for both the low cost and high end segments - and people really need to expand their mindset as to what is possible on a tablet. Haswell will open the doors up in terms of what is possible on a tablet, while simultaneously doing it with a similar battery life similar to what is available on ARM devices. Think of an ultrabook, only that is now possible in a tablet form factor WITH the same battery life as an iPad4. That is an exciting prospect.

I like the iPad, a lot, but there is a lot it can't do by virtue of the ARM SOC found within. If haswell delivers on battery life, this will all change. The capabilities of existing high end tablets will increase a ton.

99% of consumers have never heard of x86 and couldn't tell you which is faster between an iPad and an i7. The only way they have of differentiating between quality and 'cheap junk' is how shiny the case it comes in.

 

[Fx1]

There is cheap and there is good. ARM SOCs are cheap - again the only reason they're used in tablets is because intel isn't competitive in terms of efficiency. Now that they will be, there is no reason to use ARM SOCs in high end 500$+ devices. Further, I don't consider 200$ tablets to be good. Having used a nexus 7 tablet, it is okay. But you will not find one person who would rather have that over a full blown retina iPad, unless they just hate apple or they can't afford the real thing. Besides which,intel can make an atom SOC on the 14nm node and sell 10$ chips if they want, which will have twice as much battery life as ARM SOCs (if haswell is any indicator).

500$ tablets are a significant market. Sure some people can't afford them - but it has by far the highest profit margins. IMO, intel will get those profits with broadwell. And for those who can't afford the high end chips, i'm sure intel will make ULV low end chips with nearly that much more battery life than both broadwell and ARM SOCs - consider the current battery life of cloverdale. If intel makes a low cost Atom SOC based on the 14nm process, it will be a bloodbath in the low end cheap market that you guys apparently love. You get what you pay for.

You realise that iPad runs on an ARM Soc?

Do you think Apple will pay intel $100-$200 per CPU for an iPad that they currently pay $10-$20 for?

Atom cant compete today and it wont compete in 5 years time either. Intel is so far behind and works on much higher margins that its never going to be in an iPad.

 

[StrangerGuy]

99% of consumers have never heard of x86 and couldn't tell you which is faster between an iPad and an i7. The only way they have of differentiating between quality and 'cheap junk' is how shiny the case it comes in.

There was a brief time in 2010 where masses genuinely needed more SoC CPU power than a single core A8 to run that slowass OS called Android 2.xx. That boat has long sailed beyond reach out of Intel once ICS/dualcore A9s hit. Just ICS alone transformed browser performance of my 2010 Optimus Black (1GHz A8) from a complete POS to a somewhat acceptable speeds.