What's Behind LG's Sweet New Phone?
by Anand Lal Shimpi on July 13, 2009 12:00 AM EST- Posted in
- Anand
I saw this over at Engadget today, it's a video of what many are assuming will be the next-generation Chocolate by LG.
Two things stuck out about the phone; first, its form factor. It's unusually long for a smartphone, but I do remember at least one more example of an unusually long smartphone:
The phone above my iPhone in this pic is actually a mockup of a Moorestown based device that Intel has been showing off for a while. Could the new LG phone be the first Moorestown based smartphone?
The answer is unfortunately, no. Despite the similarities in design to the Moorestown mockup and despite the ridiculously fast UI, there's no Moorestown in this phone. The video lists the phone as coming out in 2009 and Moorestown isn't arriving until 2010. Chances are that this new LG phone is using a SoC based on ARM's Cortex A8. While I don't think it's impossible for the Cortex A8 to deliver such a smooth UI, it is just far more believable that Moorestown could do it.
The second thing that stuck out is obviously the phone's awkward shape. The extremely long phone has a 4" screen, something called an "Active Flash UI" , touch screen keyboard and an active camera UI that lets you smoothly apply filters to your viewfinder before you take your photo (hooray).
Intel and LG do have a good working relationship, but the fruits of their collaboration won't be seen with this phone. That being said, if the video isn't a blatant lie, that's one sweet looking phone.
Facebook
Twitter
RSS
37 Comments
View All Comments
doncerdo - Tuesday, July 14, 2009 - link
Anand why is it that you compare Moorestown with ARM? Even more important, why is it you can affirm that Moorestown is faster than ARM? Aren't you comparing apples with oranges? Last I knew ARM was many times "faster" based on MIPS and Moorestown faster according to specs.Besides last time I heard x86 instruction set was actually pretty inneficient, that's why people thought RISC was destined to kill x86 CISC chips. With making CPUs smaller and making them superscalar and out of order, the x86 instruction set issue wasn't wasn't much of a problem. So isn't Moorestown closer in the way it behaves to the old clunky CISC chips of old?
Talking about the UI based on marketing materials is also a pretty novice mistake. Finally flash interface on phones...hmmm if you look outside the iPhone universe you'd see many feature phones use them.
Given you know so much about CPUs in general and knowing that you know about this industry more than anyone, while you make some of the best tech articles on the planet, I was actuallly kind of surprised by this blog entry.
Anand Lal Shimpi - Tuesday, July 14, 2009 - link
Have a read through my Atom architecture piece and compare it to the Cortex A8 coverage in my recent iPhone 3GS review to see where Moorestown's strengths come from:http://www.anandtech.com/showdoc.aspx?i=3276">http://www.anandtech.com/showdoc.aspx?i=3276
x86 is a penalty at such small chip sizes, but Moorestown will/does have the performance advantage. Where Intel has difficulty competing is in power requirements - ARM is still the victor there and that's what fundamentally matters in these smartphones. Moorestown isn't small/low power enough to fit into something like the iPhone, you'll only see it in larger devices.
Take care,
Anand
IntelUser2000 - Tuesday, July 14, 2009 - link
No Anand is right. According to specintbase(not threaded) benchmark on the Atom presentation...Atom 1.1Ghz is 50% faster than Cortex A8 at 1GHz. Hyperthreading can extend that by additional 38%.
According to ARM, the Cortex A8 delivers 2.0 DMIPS/MHz, making it 2000 DMIPS at 1GHz: http://www.arm.com/products/CPUs/ARM_Cortex-A8.htm...">http://www.arm.com/products/CPUs/ARM_Cortex-A8.htm...
Atom can achieve 4000 DMIPS at 1.6GHz, which equals to 2500 DMIPS/MHz.
With Moorestown, performance will go up again.
IntelUser2000 - Tuesday, July 14, 2009 - link
Ok let's see this RISC vs. x86 argument again. It's the additional transistors required on x86 for decoders that gives the x86 the "disadvantage".On the Pentium III core with 0.25um process technology, the 3-wide instruction decoder took 9.2mm2. Even considering each process generation(0.25 to 0.18 to 0.13 and so on) reduction only results to a 70% of the original size(0.7x), the 3-wide decoder would take 1.5mm2 at 45nm on the Atom.
The core on the Atom is approximately 9mm2, meaning the decoder portion probably takes 10% of the core(assuming 2-issue takes 1mm2).
That 10%, can easily be mitigated by better design and process technology. Even if they are equal, Atom will be only disadvantaged by 10%, or in other words, within competitive margins.
andrewaggb - Thursday, July 16, 2009 - link
RISC vs x86 has been visited lots of times. In theory RISC is much cleaner, prettier, more consistent, easier to design/scale/debug etc. But the trend for years has been to keep adding new instructions, mmx, sse1-4 etc, and powerpc has been doing that as well. This blurs the line bit. And CISC has merits too. It's a trade off, more complex cpu with CISC but smaller programs or bigger programs with a simpler cpu with RISC? With clock speeds being artificially limited for power/heat reasons and cache making up the majority of the cpu die I'm not sure if the merits of RISC really matter. And for all it's merits, designs like Itanium still struggle to show why they are better than a good x86 design.deputc26 - Tuesday, July 14, 2009 - link
Could this be a tegra based device? The fast graphical UI would be possible on it no?dabbigj - Monday, July 13, 2009 - link
I had my hands on it during while shooting a commercial, i really liked the interface and it is quick although not as quick as in the video. The earliest phone was from january and the latest one was made in may and the january one was constanly bluescreening.All in all i thought it was nice and i really like the profile of the phone and the interface although it was a bit confusing during the first few minutes.
Anand Lal Shimpi - Tuesday, July 14, 2009 - link
Drop me an email and tell me more :)Take care,
Anand
T2k - Wednesday, July 15, 2009 - link
You should contact SE and request a Satio (IIRC it was formerly known as Idou) for a shootout - SE is working on a few big thing for the last quarter and 1Q2010 including this Satio and a high-end Android-based unit.http://www.sonyericsson.com/cws/products/mobilepho...">http://www.sonyericsson.com/cws/product...ephones/...
anandtech02148 - Monday, July 13, 2009 - link
with every month we hear and see a new smartphone coming out and the year long waits until it pass by the 3 american cellular cartel's approval. it feels as though the cellular are the new Microsoft whom can barely provide decent 3G. Gotta give it to the hardware makers for trying to be innovative in the US market. with that said, i like the Htc Hero's ability to render adobe Flash.