Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
The Tri-Gate will transistors definitely help. I didn't mean that lowering the TDP would cause the CPUs to be slower than their predecessors ;) What it can cause, however, is that the performance upgrade will be smaller than what it would have been if the TDPs stayed the same.

Agreed:)

Im itching for a new mATX build, and 1155 has wonderful bang for buck (as the 2600k bests the 980x in the majority of gaming BM)... But 22nm + trigate transistors, with 6 physical cores (+HT) have me very excited for IB. :D

I think other than MBA no other lines up sells well with LV/ULV CPUs - i may be wrong here...

My daily laptop is an ASUS thin and light with a ULV C2D...I like it very much. (though like you, am not sure as to the sales figures of it:p)

Though, Id like to replace it with a discounted 13inch C2D+320m MBA, once the SB/TB refresh hits.
 
Last edited by a moderator:
yeah, that is going to change based on Intel statement, $300 and above price for LV/ULV CPUs are not going to cut for $600 - $800 laptop range.

I think other than MBA no other lines up sells well with LV/ULV CPUs - i may be wrong here...

LV/ULV CPUs aren't very popular but their price might be one reason. The biggest market for them is ultraportables which in most cases can be considered as luxury items (you are only paying for the small, thin form factor). It looks like Intel wants to change this though and I can easily see why. I love my MBA and I'm surprised that there aren't any real competitors to it. I would love to see more MBA alike PCs in e.g. 800$ price range.

What i referring was the desktop Quad Core 65 watts CPUs.

They still cost more, although it looks like the premium has gotten smaller (I remembered incorrectly, I thought they still cost like +50$ more but looks like it's only 20$ or so).
 
FUD...

Mobile ARM CPU is more in 0.5-1 watts range... I think everyone at Intel are scrambling there plan facing the rocket fast rise of ARM cpu. For sure ARM are not powerfull as the x86 chip, but ARM cpu are within 10x mark more efficient than Intel CPU.

I think Intel plan to put x86 into mobile phone and tablet are unrealistic face to ARM cpu. Any efficiency gain Intel could get on x86 cpu by reducing the fabrication process, ARM will do even better with the same trick.

You do know that Intel makes some ARM CPU and they a really not scared of them.
Intel knows the future of CPU better than most companies and the biggest problem in terms of CPU is what is known as the power wall. you can only dump so much heat. 130W and 45W is about the limits for desktops and laptops so going beyond that is hard. going low voltage and getting that to work gives you a lot more room to grow in terms of power and heat dumping later on.

Also the other problem is everything ease but battery power has been growing exponentially. Batteries tech has more or less grown linearly and most of our gains in terms of battery life have been from reducing the power demands not putting in larger batteries.
 
You do know that Intel makes some ARM CPU and they a really not scared of them.
Intel knows the future of CPU better than most companies and the biggest problem in terms of CPU is what is known as the power wall. you can only dump so much heat. 130W and 45W is about the limits for desktops and laptops so going beyond that is hard. going low voltage and getting that to work gives you a lot more room to grow in terms of power and heat dumping later on.

Also the other problem is everything ease but battery power has been growing exponentially. Batteries tech has more or less grown linearly and most of our gains in terms of battery life have been from reducing the power demands not putting in larger batteries.

Intel has their Xscale ARM before sold the whole division to Marvell few years ago, Intel doesn't make ARM cpu anymore.

My point was, Intel target power is 10-15 watts while ARM is less than 1 watts.

I'm sure ARM will not take over Intel in Desktop space anytime soon, but the opposite is still true. I still wonder who will won the next cpu war: slim down a fat architecture or beef up a slime design.
 
There will come a time (sooner than you think) when all that intensive work will *not* demand a so-called "higher-end" processor, or (and more likely), that those high-end processors will require a fraction of the power they require today. Looking it what the iPad 2 is capable of today, it's pretty astounding.

It is possible that the iPad 2 is only capable of doing what it does because it is "new" technology.
Apps written from scratch running on a new OS.

Much less bloat in the code, less of a memory hog.
If desktop applications could be written with that efficiency, most of us wouldn't need more powerful chips.
 
Wattage ratings for CPUs are not power ratings but TDP ratings for OEMs to build appropriate cooling solutions.

They are all we got and pretty much all we need. Idle power usages have gone down every year but the TDP often affects the idle usage too. The TDP determines the suitability of a certain chip. While MBA could run a 130W when it's idling, the CPU would shut itself down when actually doing something since the cooling isn't appropriate.

As the mainstream CPUs are now 35W, that means you can't build a small, thin laptop and put one of those in it without heat issues. Clearly, Intel wants reduce the footprint of laptops and the only way they can do that is to produce more efficient CPUs with lower TDP.

Intel has their Xscale ARM before sold the whole division to Marvell few years ago, Intel doesn't make ARM cpu anymore.

My point was, Intel target power is 10-15 watts while ARM is less than 1 watts.

I'm sure ARM will not take over Intel in Desktop space anytime soon, but the opposite is still true. I still wonder who will won the next cpu war: slim down a fat architecture or beef up a slime design.

I remember reading an article about ARM vs Intel what stated that the possible issue with ARM is that power consumption and performance don't scale up evenly. ARM seems to work great in ~1W areas but its performance might be horrible when you start increasing the frequency and core count and thus the TDP (i.e. it does not scale up. E.g. you double the clock speed but your TDP becomes 10 times as big). Especially if the architecture is designed for 1W areas.
 
Last edited:
It is possible that the iPad 2 is only capable of doing what it does because it is "new" technology.
Apps written from scratch running on a new OS.

Much less bloat in the code, less of a memory hog.
If desktop applications could be written with that efficiency, most of us wouldn't need more powerful chips.
I have been thinking the same thing for a while now. Constraints on the hardware side will force the people writing software to be more inventive and efficient.

Having (virtually) unlimited resources in the desktop realm doesn't exactly give you a lot of incentive to reinvent the wheel every time you pump out a new update. Things tend to bloat instead of get rebuilt efficiently. Is there any reason why I should have to download 500MB updates to MS Office every time I open up a Word document on my iMac? Hardly seems efficient to me, and clearly not a viable option on iOS.
 
There is going to be a market for machines that converge laptop and tablet forms. I don't know if it will do well in the consumer market or not, but manufacturer in the next few years is going to try combine the basic ease and form of a tablet with the usability and power of a laptop, and I wouldn't be surprised to see an Intel processor as a big selling point.

The iPad and its competitors are very limited in their productive capability, mostly handling consumptive tasks. You wouldn't dream of real video editing on them. Adobe released some photoshop companion programs, but having actual Photoshop on the iPad is a silly idea. Right now.

It wouldn't surprise me to see Dell or or someone like that produce a Windows Tablet with an Intel processor that actually has the horsepower to do this stuff. Strategically, Intel can do itself enormous favors if it gets a foot into that market.
 
I remember reading an article about ARM vs Intel what stated that the possible issue with ARM is that power consumption and performance don't scale up evenly. ARM seems to work great in ~1W areas but its performance might be horrible when you start increasing the frequency and core count and thus the TDP (i.e. it does not scale up. E.g. you double the clock speed but your TDP becomes 10 times as big). Especially if the architecture is designed for 1W areas.

I've read about the ARM since it's first use in the Newton. and in my understanding, the ARM is a pure RISC design, a very small core built with efficiency in mind. They don't have branch prediction and deep execution pipe like x86 processor, limiting their effective power in desktop environment. It's like comparing a regular 3L V6 engine with a 1.6 turbo V4 running at 11,000 RPM, both could achieve about the same HP. But the V6 can be push more ahead burning fuel and the V4 will have better fuel efficiency at low speed. While ARM is already push to it's limit, core multiplication and expending the base design of ARM can obliterate those limit in near future.

The interesting part come from Intel, saying right now ARM mobile CPU is growing twice as fast as the Moore Law.
 
Then and than

You would think that a "professional" writer would know the difference between "then" and "than".
 
Intel is feeling the heat from all of these rumors of Apple switching to ARM in future notebooks, in my opinion, apple should do both, switch to ARM and have Intel and/or AMD fabricate them.
 
I just hope the entire line isn't dumbed down... some of us still need the higher-end processors for intensive work.

That said, I wouldn't mind having a low-power machine as a secondary on the go (no, not a phone or tablet).

I wouldn't worry about that. They still make mainframes. High-end laptops will just become vanity and more expensive.

FUD...

Mobile ARM CPU is more in 0.5-1 watts range... I think everyone at Intel are scrambling there plan facing the rocket fast rise of ARM cpu. For sure ARM are not powerfull as the x86 chip, but ARM cpu are within 10x mark more efficient than Intel CPU.

I think Intel plan to put x86 into mobile phone and tablet are unrealistic face to ARM cpu. Any efficiency gain Intel could get on x86 cpu by reducing the fabrication process, ARM will do even better with the same trick.

On what basis do you say they are 10x more efficient? They are slow and hence burn less power and energy. There is nothing inherently inefficient about x86. ARM ISA is equally bloated.

Intel's fabrication edge is not easy to duplicate either. ARM can't use the same tricks in the same time frame. Intel has dedicated fabs tailored for their chips while ARMs built in these shared fabs like TSMC. TSMC stays 2 generations behind Intel in fab technology (and for good, fundamental economic reasons).

There will come a time (sooner than you think) when all that intensive work will *not* demand a so-called "higher-end" processor, or (and more likely), that those high-end processors will require a fraction of the power they require today. Looking it what the iPad 2 is capable of today, it's pretty astounding.

Really don't understand what you mean. Are you saying work will become less intensive, or processors will become faster+more-energy efficient? or are you saying software will become multi-threaded allowing it to leverage multiple energy-efficient cores to get performance, making it both fast and less energy?
 
Last edited by a moderator:
Most won't admit it, but Apple shaped the product road-map of most of those computer computer that's still relevant today. that includes Google, Sony, HTC, Samsung, Motorola etc...
 
.
TOO LATE. Apple is already developing their own custom ARM chips for future notebooks.

I would not count on that. Apple is long ways from delivering a CPU fast enough to run a big OS. Also, as much as I dislike windows, I think they will have to figure out how to run windows on it. Being able to run windows in parallels and bootcamps was one of the biggest reasons Macbooks took off when they Apple switched to Intel cpus.

It is possible that the iPad 2 is only capable of doing what it does because it is "new" technology.
Apps written from scratch running on a new OS.

Much less bloat in the code, less of a memory hog.
If desktop applications could be written with that efficiency, most of us wouldn't need more powerful chips.

Interesting thoughts. For some apps, we always will, e.g., photoshop is already pretty optimized but it still needs a lot of power. Video editing, number crunching for analysis, workstation stuff, DSS systems, are all examples of things that will continue to need processing power.

I would totally agree however that my mom would not need a lot of power if it was not for the bloated, poorly written code.
 
Last edited by a moderator:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.