[Arm-netbook] Good netbook based on Cortex-A9

Gordan Bobic gordan at bobich.net
Mon Jul 30 21:53:34 BST 2012


On 30/07/2012 20:54, Alejandro Mery wrote:
> On 30 July 2012 21:36, Gordan Bobic<gordan at bobich.net>  wrote:
>> On 07/30/2012 08:32 PM, Alejandro Mery wrote:
>>> On 30 July 2012 21:25, Gordan Bobic<gordan at bobich.net>   wrote:
>>>> On 07/30/2012 05:29 PM, lkcl luke wrote:
>>>>> On Mon, Jul 30, 2012 at 4:49 PM, Alejandro Mery<amery at geeks.cl>    wrote:
>>>>>> On 30 July 2012 17:44, Gordan Bobic<gordan at bobich.net>    wrote:
>>>>>>> On 07/30/2012 03:19 PM, lkcl luke wrote:
>>>>>>>>      ok - so that's good to know.  it'd be helpful to get a list of part
>>>>>>>> numbers known to be interchangeable.  can you recall what they are?
>>>>>>>
>>>>>>> Sorry, I haven't really kept track of part numbers, over and above what
>>>>>>> I wrote up in the relevant articles which I linked previously.
>>>>>>
>>>>>> I think the goal of luke's question is to find a pair of compatible
>>>>>>> =720p and>=1080p panels so we can have to models of the same eoma68
>>>>>> laptop
>>>>>
>>>>>     ideally, yes.  although it is really up to you guys: what do *you*
>>>>> want to see in a laptop?  the I/O Board can be made reasonably
>>>>> flexible, if the right matching cases are found and enough flying
>>>>> leads put on it, it might be possible to use the same I/O Board for 2
>>>>> different laptops, especially if they have a CD-ROM hole and that's
>>>>> used as the EOMA-68 CPU Card slot as well as having several of the
>>>>> connectors on it (just ignoring all other holes).
>>>>>
>>>>>     also as previously mentioned it should not be hard to do an LCD
>>>>> Daughter Board, with an FPC over to it with about 32 pins or so:
>>>>> that'd be enough to transfer the RGB/TTL signals, some power and a PWM
>>>>> signal (for backlight control).  the daughter board in the 1080p case
>>>>> could have a dual LVDS IC on it, and for the 720p a single LVDS IC.
>>>>>
>>>>>     i'd really like the single LVDS IC version to cover at least two 10in
>>>>> LCD panels - 1024x600 *shudder*, 1200x720, maybe 1388x768 - but we
>>>>> need to find matching parts that all have the same connector and the
>>>>> same LCD Backlight design and voltages.
>>>>
>>>> The problem is that while 1024 and 1280 panels are physically the same
>>>> size, the only 1366 panel isn't. I'd pretty much ignore the 1366
>>>> possibility alltogether. There is no choice of panels, and if you make
>>>> the chassis fit it, it won't fit the other panels properly.
>>>>
>>>> The problem is also lack of distinction. Many competing products have
>>>> 1024 or 1280 panels. Only one has 1920x1200, and that's, as mentioned,
>>>> 16:10, so you're stuck with a chice of 1 again because hardly anyone
>>>> makes 10" 16:10 panels. But if you're going to make 1920x1200 laptops,
>>>> why bother with a lower res option? People will be buying it for the
>>>> high res screen. For 1280 there are already better options than the A10,
>>>> including the AC100 and the Genesi.
>>>
>>> I assume the reason to care about>1000p is cost ;-)
>>
>> If you're concerned about cost you'll be buying an AC100 anyway.
>>
>>> but an
>>> ass-kicking open source friendly 10" 1920x1200 laptop will catch far
>>> more attention as there are mostly no competitors (yet).
>>
>> That depends. How OS frieldy is it really? Upstreamed kernel support for
>> everything? Substantially cheaper than the competition with 5x the raw
>> CPU performance with all other things being equal?
>
> well, the point of eoma68 is that the laptop part will remain after
> you upgrade to a newer/better CPU card while the competitor will
> totally dish support once the next generation comes out

Sure, but the benefits of that depend on how often you are planning to 
upgrade. Personally, I tend to not upgrade my machines until I have a 
VERY good reason to do so. This involves either hardware failure or a 
very substantial performance boost.

For example, the hardware I use daily:

1) Desktop: Core2 class, currently 4.5 yerars old. No plans to upgrade. 
Core i just doesn't provide enough improvement, and as much as I would 
love one of those EVGA SR2 boards for 12-16 cores worth of compiling 
goodness, I don't see a real, tangible _need_ for it.

1.1) Monitor: IBM T221. 3840x2400, 2005 vintage. I don't see that being 
beaten any time soon. I'm vaguely pondering getting another 1-2 of them 
as and when they are available just to mothball them in case my current 
one blows up in the next decade or two until something higher res 
becomes available. As you can tell from that statement, I don't expect 
it to actually get improved upon any time soon, so no upgrade.

2) Laptop: IBM ThinkPad T60, upgraded to the fastest Core 2 CPU it can 
handle (2.33GHz), 3GB of RAM (can't see more than 3GB even with 3GB 
installed, sadly), and that awesome 2048x1536 15" screen. Intel SSD. I 
am reasonably tempted to finally replace this 2006 vintage machine with 
the new MacBook Pro, purely because of the 2880x1800 screen (I find the 
ThinkPad's performance is ample for just about anything).

3) Netbook: Toshiba AC100, upgraded with 1280x720 screen, USB 
SuperTalent RC8 SSD fitted internally, improved cooling to facilitate 
the OC from 1GHz to 1.4GHz. Open to upgrade potential, but the only 
thing that measures up on that front is the Asus Transformer Prime HD - 
but not for £500+.

Anyway, the point I'm making with all this is that if something is 
decently good to begin with, it's likely to last 6+ years before 
something offers a reasonably compelling reason to upgrade. The cycle is 
shorter with emerging technologies (e.g. ARM netbooks, higher resolution 
10" TFT panels and emerging display technologies like PixelQi), but even 
then, the upgrade cycle is unlikely to be shorter than 2 years. So the 
real question is whether having an open upgrade path is really worth 
that much of a premium on something like a laptop. I guess it depends on 
how much you'd save every few years by not replacing the chassis itself, 
but my understanding is that the chassis itself is not a substantial 
part of the cost of a laptop. You might as well just make a box that 
allows plugging in an EOMA card into a Motorola Lapdock if you want to 
cut a few corners. :)

Gordan



More information about the arm-netbook mailing list