ARM-Based Windows 8 Notebooks in Mid-2013?

Status
Not open for further replies.
how small and fast will the arm socs be? intel already has 22 nm chips on the way, amd's just got their 32 nm chips out. 2013 is a little late for debut. intel can have a mature ultrabook market by the time arm brings out a notebook.
and if it can't play crysis...
 

croc4

Distinguished
Sep 3, 2010
12
0
18,510
Why?, ARM based CPU's fail compared to x86 versions in the performance arena, and we all know how badly the 'netbooks' did, very slow even doing basic tasks, so adding an even less powered CPU is going to work 'better'?, I'm sure the performance will be better than todays versions, but still won't be close to low end x86 cpus. I see fail written all over this
 

amk-aka-Phantom

Distinguished
Mar 10, 2011
3,004
0
20,860
[citation][nom]techguy378[/nom]Windows performs much better on an ARM processor than it does on an x86 processor.[/citation]

Windows does NOT perform on an ARM processor yet, troll, it's x86-only.
 

zanny

Distinguished
Jul 18, 2008
214
0
18,680
[citation][nom]de5_roy[/nom]how small and fast will the arm socs be? intel already has 22 nm chips on the way, amd's just got their 32 nm chips out. 2013 is a little late for debut. intel can have a mature ultrabook market by the time arm brings out a notebook.and if it can't play crysis...[/citation]

Kal'el Tegra chips are 45nm, the Wayne series after that in early 2012 will be 28nm, and I expect ~20nm by 2013. They use TSMC for both gpus and socs, while AMD is using TSMC for the radeon 7000 gpus that are on their way.

[citation][nom]croc4[/nom]Why?, ARM based CPU's fail compared to x86 versions in the performance arena, and we all know how badly the 'netbooks' did, very slow even doing basic tasks, so adding an even less powered CPU is going to work 'better'?, I'm sure the performance will be better than todays versions, but still won't be close to low end x86 cpus. I see fail written all over this[/citation]

ARM processors are extremely power efficient, and once you have 4 cores at 2ghz each performance on consumer grade laptops becomes a non issue. I expect by 2015 for there to be a general market of $300 - 500 ARM windows laptops with battery life in the neighborhood of a day under load, and the $1k + market will be dominated by skylake by Intel at 16nm.

The ARM assembly language is just fundementally better than x86 because it isn't backwards compatable all the way to the original i386 instruction set. Sandy Bridge E is a great example. 2.3 billion transistors, but in the end a ton of those are spent supporting legacy instructions that no longer make sense with current designs of cpus and hardware in general, but are still there because they keep the zombie beast of x86 alive for so long. Its not like didn't try to fix it - they made Itanium as a better replacement. Only problem was the problem the windows ARM machines will have - if you cant run x86 software, you are dead from inception. One advantage for ARM will be a somewhat reasonable portability between iOS and Android to ARM, and Linux already runs on ARM as do many linux apps, GCC compiles to ARM, etc. The OSS community tools work on ARM already, whereas for a long time after Itanium came out they didn't, but that is irrelevant for average joe consumer.

One other thing to consider - how many general purpose consumers use software anymore? Everything is web based. You can get the entire rage of general purpose software through google docs and the prepackaged bloatware MS will throw in.

What is hilarious to me is how much this much be costing M$. They have to rewrite everything in the NT kernel for ARM, port .net and everything else they have made for a decade, and expect it to sell well enough to justify that huge investment.
 

hpglow

Distinguished
Sep 14, 2011
11
0
18,510
[citation][nom]de5_roy[/nom]how small and fast will the arm socs be? intel already has 22 nm chips on the way, amd's just got their 32 nm chips out. 2013 is a little late for debut. intel can have a mature ultrabook market by the time arm brings out a notebook.and if it can't play crysis...[/citation]
At 40nm many ARM processors consume less than 1W. Nothing intel has comes anywhere close to that. If these SOC makers were to utilize a smaller node they would bring that figure down quite a bit. Performance is another arena entirely, however, given the fact that the ARM processors would not have to worry about backwards compatibility they will likely not be as slow as people think.
 

kartu

Distinguished
Mar 3, 2009
959
0
18,980
[citation][nom]croc4[/nom]Why?, ARM based CPU's fail compared to x86 versions in the performance arena, and we all know how badly the 'netbooks' did, very slow even doing basic tasks, so adding an even less powered CPU is going to work 'better'?, I'm sure the performance will be better than todays versions, but still won't be close to low end x86 cpus. I see fail written all over this[/citation]

I was surprise to discover that while they "fail" they are not it's not orders of magnitude behind. 800Mhz A8 was up to two times slower than 1Ghz Atom in tests:
http://iltsarnews.blogspot.com/2010/04/arm-vs-x86-low-power-vs-performance.html

Poiwer usage was 1 : 2 to 1 : 3 and that with Arm board having no power saving features.


Note that most users need "fast enough" CPU.
 

billybobser

Distinguished
Aug 25, 2011
432
0
18,790
x86 is the junk.


If intel broke away from it, we'd have alot more progress. AMD just competes in Intel's market for price point.

It's about time someone moved computers forward.


(staying on x86 is like staying on IE5)
 
[citation][nom]billybobser[/nom]x86 is the junk. If intel broke away from it, we'd have alot more progress. AMD just competes in Intel's market for price point.It's about time someone moved computers forward.(staying on x86 is like staying on IE5)[/citation]

just because your phone performs good with it doesn't mean a PC will

x86 is junk? give me a break
 

croc4

Distinguished
Sep 3, 2010
12
0
18,510
Well, I'm not going to hold my breath, I realize arms are better power wise, I was not arguing that. And as far as "general software" or "fast enough" CPU's that's where I disagree. My wife has a netbook (atom based) and for all intent and purposes its fast enough and the battery life is pretty good, but what does she do when I'm not using my laptop?, she uses mine because hers is 'too slow' she says and all she is doing is face booking and emailing / surfing the web, all of which the atom can do, but not "fast" enough apparently, so toss in a lower performing CPU and I doubt the end user experience is going to be 'better'. Maybe I'm wrong.........
 
G

Guest

Guest
"At 40nm many ARM processors consume less than 1W. Nothing intel has comes anywhere close to that. If these SOC makers were to utilize a smaller node they would bring that figure down quite a bit. Performance is another arena entirely, however, given the fact that the ARM processors would not have to worry about backwards compatibility they will likely not be as slow as people think. "

None that can do any kind of performance like Kal-El,etc. By 2013, Intel will have Silvermont ATOMS. These are SOC ATOMS redeveloped to compete with the low power ARMs. Power consumption will be very similar between the two but the ATOM will have FAR greater performance. Would you really buy an ARM laptop over a Silvermont ATOM laptop? The ATOM laptop will run all of your same software BUT will have much greater performance at the same low cost. ARM is a failure on the PC platform in my opinion. This is a pretty big gamble by those suppliers.
 

allenpan

Distinguished
May 4, 2001
72
0
18,635
port application from CE to windows or windows to CE is relatively easy, as long as on .Net framework, it works on both CE and Windows

also back in win2K, there is windows support for RISC CPU, even back to windows NT
 
I just think its going to be a really uphill battle to convince people Windows, even with metro, is going to be, or at least, give the perception of being "fresh, hip, and new".

The masses have spoken.

Performance isn't the main factor when considering gadgets these days.

There's exceptions to everything.
 

dalethepcman

Distinguished
Jul 1, 2010
1,636
0
19,860
I can see Intel creating SOC's similar to kal-el in the future. Something like an ivy bridge 8+2 core on a single chip, with the +2 (simple) cores being low power and running whenever the system is just sitting there and all other cores in a sleep state, then when the simple cores hits 40-50% cpu light up the real cores and burn through any cpu intensive tasks then go back to sleep when the work is done.
 

tanjo

Distinguished
Sep 24, 2011
272
1
18,810
Let's wait for the first ARM powered Win8 notebook and see if it is acceptable (unlike Atom... though I highly doubt it). Otherwise, Haswell (w/Thunderbolt) for utrabooks.
 

twstd1

Distinguished
Nov 26, 2008
104
0
18,710
I understand everybody worrying about the hardware not being able to run windows 8 effectively but I don't believe the hardware is going to be the problem here. It's whether or not MS can get the code to run right on this hardware. I can't see putting the blame on a quad-core chip running @ 2 to 2.5 GHz. That chip is more than powerful enough to do the job. I think it's going to come down to how well MS works it's magic on the functionality of the code running on an ARM based chip.
 
[citation][nom]Zanny[/nom]Kal'el Tegra chips are 45nm, the Wayne series after that in early 2012 will be 28nm, and I expect ~20nm by 2013. They use TSMC for both gpus and socs, while AMD is using TSMC for the radeon 7000 gpus that are on their way.ARM processors are extremely power efficient, and once you have 4 cores at 2ghz each performance on consumer grade laptops becomes a non issue. I expect by 2015 for there to be a general market of $300 - 500 ARM windows laptops with battery life in the neighborhood of a day under load, and the $1k + market will be dominated by skylake by Intel at 16nm. The ARM assembly language is just fundementally better than x86 because it isn't backwards compatable all the way to the original i386 instruction set. Sandy Bridge E is a great example. 2.3 billion transistors, but in the end a ton of those are spent supporting legacy instructions that no longer make sense with current designs of cpus and hardware in general, but are still there because they keep the zombie beast of x86 alive for so long. Its not like didn't try to fix it - they made Itanium as a better replacement. Only problem was the problem the windows ARM machines will have - if you cant run x86 software, you are dead from inception. One advantage for ARM will be a somewhat reasonable portability between iOS and Android to ARM, and Linux already runs on ARM as do many linux apps, GCC compiles to ARM, etc. The OSS community tools work on ARM already, whereas for a long time after Itanium came out they didn't, but that is irrelevant for average joe consumer.One other thing to consider - how many general purpose consumers use software anymore? Everything is web based. You can get the entire rage of general purpose software through google docs and the prepackaged bloatware MS will throw in.What is hilarious to me is how much this much be costing M$. They have to rewrite everything in the NT kernel for ARM, port .net and everything else they have made for a decade, and expect it to sell well enough to justify that huge investment.[/citation]

No, just no.

Do not compare ARM vs x86, the ISA's are too different to be directly compared. And don't even attempt to go into the utter disaster that was IA64. Itanium was NOT a better instruction set, Intel did not try to save the world from x86, it was the exact opposite. They were trying to lock out third party manufacturers from producing CPU's capable of running MS Windows. IA64 sucked, nobody liked it and thus it never gained any traction. In the heavy RISC world you have SPARC and PPC, and both of those are better ISA's the IA64.

x86 is not a bad instruction set. It works very well for what it was designed to do, run single process's at high speed. We've just moved onto bigger and better micro-processing design's, of course trying to move the entire world to a different architecture isn't going to work without an evolutionary growth (EMT64 vs IA64) vs a disruptive revolutionary one.
 



*Shudder*

And you realize that Atom's are the absolute slowest low power x86 CPU you can get. AMD and Via both produce faster CPU's that drink approximately the same amounts of power (once the whole platform is taken together). Atom's are deliberately crippled worse then the Celerons where, to prevent them from competing with Intel's budget processors.
 
Status
Not open for further replies.