Interesting Atom vs ARM lecture

[thecursedfly]

Read here:
http://www.groupsrv.com/computers/post-2746546.html



In the light of recent discussions, I thought this article would be of
interest:

http://www.linuxdevices.com/news/NS5492118276.html

See also: http://laptoping.com/intel-atom-benchmark.html which
shows Atom/Silverthorne is less than half the speed of a Celeron-M...

So the quad core ARM Cortex is 2x smaller, uses 1.5x less power and is twice
as fast according to the competition's benchmark results? Impressive.

Indeed. Intel's benchmarks claim that a 1.6GHz Atom is only 4-6.5 times faster
than a single-issue 400MHz ARM11. Cortex-A8 and A9 are twice as fast as
ARM11, so by Intel's own admission the latest ARM cores are 1.25-2x as fast as
Atom at the same frequency... (which is what I predicted a while ago). If you
compare the SIMD capabilities then ARM's advantage increases even further.

So with a quad core Cortex-A9 at 800MHz you would get 2.5-4x the performance
of a single core Atom at 1.6GHz while still using less power...

How much cheaper is the ARM?

ARM's are definitely much cheaper, but it is hard to compare like with like. Eg.
OMAP is a "single-chip smart phone". It includes everything you need, a memory
controller, LCD controller, graphics accelerator, sound, CMOS camera interace,
on-board memory, etc. Samsung sells similar phone chips with many MBytes
of flash and DRAM in the same package.

The Atom is needs additional chips for all of this, which increases the cost, board
size and power consumption.

I think Intel are really sticking their necks out by making statements like:

"Thus, the average power consumption of Atom is lower" -
http://techon.nikkeibp.co.jp/english/NEWS_EN/20080403/149992/

That's pure marketing spin and then some...

Atom's lowest idle power is 100mW!!! - this is in the deepest sleep mode doing
absolutely nothing! That drains most batteries rather quickly... Compare this with
the idle power of ARMs which is typically well below 1mW.



I had no idea that ARM CPUs are that good... o.O
If it weren't for the fact that they require to recompile/adapt every OS and application, ARMs would be perfect for netbooks..
Maybe Maemo can soon get in the netbook market too? :)

[SD69]

Your links are from 2008. Comparisons from now on have to be made with the next generation Atom chip - Moorestown. Of course, it has to be considered if Moorestown can catch up to the head start that Cortex has...

[thecursedfly]

thx for noticing that, I didn't check the date and don't know that much about latest CPU technologies... :)
still found it interesting..
if you find some up to date comparison, maybe between processors with the same clock speed (if there are any, since x86 processors have higher clock speed since some years ago), feel free to post the links, and thx in advance ;)

[allnameswereout]

I'd say we either compare existing devices or admit we are speculating about future products. If the latter, still must be done carefully.

Moorestown is the Intel Corporation's code name for successor to the Menlow platform designed for mobile Internet devices. Current proposals call for it to use one tenth the idle power of Silverthorne and to hit the market in 2009-2010.

Moorestown consists of a system on a chip, code-named “Lincroft”, which integrates the 45nm processor, graphics, memory controller and video encode/decode onto a single chip and an I/O hub codenamed “Langwell”, which supports a range of I/O ports to connect with wireless, storage, and display components in addition to incorporating several board level functionshttp://en.wikipedia.org/wiki/Moorestown_%28CPU%29

Also funny article:
http://en.wikipedia.org/wiki/Mobile_Internet_device

[khalid]

If it weren't for the fact that they require to recompile/adapt every OS and application, ARMs would be perfect for netbooks..
Maybe Maemo can soon get in the netbook market too? :)

There is some movement in that direction. Always Innovating touchbook (http://www.alwaysinnovating.com/touchbook/) is based on ARM OMAP3 and there is a project in Maemo Garage (http://maemo-beagle.garage.maemo.org/) to port maemo to beagleboard. Touchbook is also based upon beagleboard.

[attila77]

I'd say we either compare existing devices or admit we are speculating about future products. If the latter, still must be done carefully.

Right, if we let in Moorestown in the picture, then it's not A8 timeframe any more, but rather A9, to the likes of OMAP4 and Osprey, and that's going to be quite a fight as the two families finally fight for/on comparable devices.

[speculatrix]

http://www.embeddedtechjournal.com/articles_2009/20090922_arm.htm

has a few comments on this topic too

[eiffel]

...the A8 has about a 2x advantage over the Atom N270. In other words, it delivers about the same performance at half the clock speed. And one-fifth the power.

"Not a bad tradeoff, as long as you don’t need x86 compatibility."

[asys3]

I talked to an Intel representive at Cebit this year and we talked about the future of Atom or x86 devices vs. ARM based ones.

He said nothing is better than true x86 experience - alle these emulations cost performance and are not that what people want - especially they don't want anything else on an ARM plattform.

Well I said what about the power consumption?

He said we'll see very economic Intel processor already next year...

OK - we'll see.

I think though Intel has advantages in performance the consume too much power.
On the other hand ARM comes from the economic side and improves performance,

We'll see who is first in the middle ;)

I'll talk to the Intel guy next year at CeBit again ;)

asys3

[GeneralAntilles]

He said nothing is better than true x86 experience - alle these emulations cost performance and are not that what people want - especially they don't want anything else on an ARM plattform.


So, basically, he's a Windows guy? ;)

[Laughing Man]

Wow.. I have alot more I need to learn about processors. I thought speeds on an x86 platform could be compared to ARM directly (1:1). Apparantly not though.. So it's more like 1Ghz on x86 = something like 600 Mhz on ARM?

[GeneralAntilles]

Wow.. I have alot more I need to learn about processors. I thought speeds on an x86 platform could be compared to ARM directly (1:1). Apparantly not though.. So it's more like 1Ghz on x86 = something like 600 Mhz on ARM?

A 1:0.6 ratio is no more useful than a 1:1 ratio. It depends entirely on the individual processors being considered and the tasks being run on them.

[allnameswereout]

Wow.. I have alot more I need to learn about processors. I thought speeds on an x86 platform could be compared to ARM directly (1:1). Apparantly not though.. So it's more like 1Ghz on x86 = something like 600 Mhz on ARM?No way! See: Megahertz myth (http://en.wikipedia.org/wiki/Megahertz_myth). Its only valid if you compare the very same architecture and revision (including stepping), but you're comparing a RISC with a CISC. In general, MHz compare is worthless and an Intel marketing tool, and this purely false advertising (as done in compares versus Alpha, MIPS, and POWER) and is why AMD quit mentioning MHz/GHz. Besides, it also matters what you're calculating, e.g. fp vs int.

There are other ways to calculate, such as MIPS, but the only way to figure it out is 1) user experiences (although these are not objective they give an estimate/indication; problem is different software is used, and its as perceived; not really a benchmark) 2) benchmarks. Basically the only valid option. Period.

[TA-t3]

Well said!
Acked-by: TA-t3