Yeah, but it should be doable in 28nm. That said, I don't think the tablet market is so clearly defined as to have precise power targets, yet. I mean, is there a definite rule that says you can't make a tablet based on a 3W SoC? If it idles at 500mW, is that still too much?
The AMD Execution Thread [2007 - 2017]
- Thread starter overclocked_enthusiasm
- Start date
-
- Tags
- amd
- Status
Yeah, but it should be doable in 28nm. That said, I don't think the tablet market is so clearly defined as to have precise power targets, yet. I mean, is there a definite rule that says you can't make a tablet based on a 3W SoC? If it idles at 500mW, is that still too much?
Those rules are defined by limits of battery tech, the practical need to make them thin and avoid passive heatsinks.
What I mean is that the tablet market isn't clearly defined in terms of customer expectations, yet. Does a tablet have to have 10 or 20 hours of battery life, or is just 6 hours OK? Does it have to be thinner than 15mm, or is it still fine if it's 20mm thick, or even a bit more? Does it have to be under 800g, or is 1.2kg OK, provided it offers something more? That something could be performance, x86… hell, even DX11 is a nice checkbox feature (clearly nothing more than that in a tablet).
Bobcat most likely can't squeeze into an iPad, perhaps not even in 28nm, but do all tablets have to be as small, thin and light as the iPad? What about battery life?
The two power envelopes of the Bobcat SKUs is entirely defined by their target markets: Net tops and ultra portable lap tops.
I can't see any reason why AMD couldn't launch a 4W BC clocked at 900MHz (30% lower). At 4W TDP, typical power would be much less and completely dwarfed by the power consumption of a 9" screen.
They probably couldn't shoe horn one into Galaxy S Tab size (400grams, 14.8 WH battery), but iPad-size? No problem.
Make the chassis in aiuminium and use it as a heat sink.
Cheers
I can't see any reason why AMD couldn't launch a 4W BC clocked at 900MHz (30% lower). At 4W TDP, typical power would be much less and completely dwarfed by the power consumption of a 9" screen.
They probably couldn't shoe horn one into Galaxy S Tab size (400grams, 14.8 WH battery), but iPad-size? No problem.
Make the chassis in aiuminium and use it as a heat sink.
Cheers
Yes, if you want to have a saleworthy product, compared to the halo product.
iPad is a leader by far in battery life.
That will make it, in all probability, too hot to hold.
Probably not. Not to mention that AMD is very far from getting it's packaging tech right to put ram/flash and SoC in one package.
Bobcat is capable to use 4GB 1333 MHz DDR3. Aple A4 is probably using mobile LPDRAMs and something around 200 MHz and 2*32bit. Thats actualy a quite good reason for some serious power differences.(with data movement from and to CPU at those speeds and the memory itself)
So on mobile devices ARM power consumption looks impresive if u compare just bare theoretical CPU performance and frequency but the reality could be different in more demanding aplications.
Silent_Buddha
Legend
Sure, and the iPad also leaves a lot of sales on the table because it doesn't meet the needs of many potential tablet users.
Even though the HP Slate was a flawed tablet marketed to the business segment, it sold out immediately. Same goes for the ExoPC slate which was sold out for 2 months.
Both offered 4-5 hours of battery life (~3 if watching HD video). Both targetted different markets, Slate for business and ExoPC for home and home media including 1080p playback and output through HDMI (enough for an on the go lightweight HTCP).
Both still flawed due to the pathetic performance of the Atom CPU among other things. But offering quite a few capabilities that the iPad completely lacks.
Including iPad, there are still alrge consumer segments who's demands aren't met by those offerings.
Asus for example is coming out with a 12.1" slate that will be quite nice for digital artists as it has a large screen size (the 8.9" screen on the HP is entirely too small and too low res. for this) combined with an active digitizer. And to cater to its niche it'll also feature a CULV processor although Sandy Bridge is also being considered. Battery life drops to 3-3.5 hours, but it's a worthwile trade off.
And of course a plethora of slate designs from 7" up to the massive 12.1" Asus.
IMO, if Apple were smart they would also look into introducing a higher performance slate using OSX.
Battery life is certainly very important, especially for on the go media consumption. But that's still just a part of the potential slate market. Other market segments are going to require features and abilities that a power sipper like the iPad just can't and won't ever be able to fill. Active Digitizer (never buying another Capacitive touch only slate, /shudder) is going to consume power. More memory (2 gigs in the above HP and ExoPC) is going to consume more power. More performant CPUs are going to consume more power. More performant GPUs are going to consume more power. A more capable OS is going to consume more power. Etc.
The beauty is that at least in the Android and Windows slate markets, there will eventually be options for everyone who desires a slate. And the iPad will probably remain the king of battery life. But honestly that's a non-issue for me and many others. It's a consideration but depending on our needs, may not be in even the top 5 requirements for a slate.
Regards,
SB
The HP had a production volume of 5000 units.
When you speak about devices selling out, and using that as an indication of success, you need to bear in mind volumes. A device such as the HP Slate was always going to sell to a certain category of people, whose job it is to check if it could be a suitable device for their organization. It's a bit like the Dual GPU cards that ATI and nVidia produce - how many of those ever find their way outside the people in the industry, be it development, QC, publication/review or retail? The Galaxy Tab is pretty much the only non-Apple tablet that has seen any success at all.
In related news, Samsung reckons they will produce four times (!) the SoC wafers for Apple in 2011, vs 2010. 20000 wafers vs. 5000. The volumes are staggering, considering a wafer is just over 70000mm2 and an A4 processor is 53mm2.
Last edited by a moderator:
Silent_Buddha
Legend
Sure, but it's still notable as HP and many media sites were doubtful HP would be able to sell even that many claiming there was zero demand for a Windows based slate.
HP were forced to not only recommence production but also to ramp up production after that in an attempt to meet ongoing demand. It still took them a few months to get production up to a high enough level to successfully meet demand for the device. Note that I'm not implying anywhere near iPad level of demand. The device has a far too limited target market for that. Only that demand for the device far exceeded anything HP, analysts or the media predicted.
BTW - while late to the party AMD's Brazos platform compares quite favorably to Atom (Pinetrail or Ion 2) with greater performance and lower power consumption. It should remain faster than Intel's upcoming Oaktrail platform, but unknown how it will compare in power consumption.
Regards,
SB
HP were forced to not only recommence production but also to ramp up production after that in an attempt to meet ongoing demand. It still took them a few months to get production up to a high enough level to successfully meet demand for the device. Note that I'm not implying anywhere near iPad level of demand. The device has a far too limited target market for that. Only that demand for the device far exceeded anything HP, analysts or the media predicted.
BTW - while late to the party AMD's Brazos platform compares quite favorably to Atom (Pinetrail or Ion 2) with greater performance and lower power consumption. It should remain faster than Intel's upcoming Oaktrail platform, but unknown how it will compare in power consumption.
Regards,
SB
9W is too much for a tablet.
Apparently, there's a 5W variant: http://news.softpedia.com/news/AMD-Showcases-5W-Version-of-Ontario-C-50-APU-181254.shtml
Silent_Buddha
Legend
It all depends on the use case and battery used. Current Pinetrail based tablets can achieve 4-5 hours using rather smallish batteries. Oaktrail and Brazos are expected to expand that by 1-2 hours at least.
For a media only device both Atom and Bobcat are probably overkill and thus the battery life hurts them.
However, Arm based tablets are completely and totally insufficient for me and many others that require more utility from our tablets. Thus while battery life isn't as good as a less performant CPU with a greatly limited and/or castrated OS, it's fantastic since there is nothing better at the moment.
Once Windows 8 for ARM arrives, I'll revisit the battery life angle. Or if Apple can muster up the courage to launch an OSX based tablet that will run on ARM.
I suspect, I may still find the performance quite lacking however for my use cases. Heck, I and some others are already looking ahead to Sandy Bridge based tablets for our needs even though battery life would then drop to around 3-4 hours.
Battery life doesn't mean much if it's used in a vacuum with no reference to performance levels and capabilities of a device, IMO. But I'll agree for a media consumption only device, limited OS systems with lower performance CPUs are mostly fine and Atom/Bobcat aren't going to make much of a dent there.
However, trying to limit the slate market in that way is shortsighted when there's a market there for more powerful and capable slates.
Regards,
SB
I'd be curious to know about your supposed use cases. The iPad is my primary device for home and the sole device for travel. The MBP is too heavy to lug through airports and hotels when an iPad gets the job done faster and the battery lasts so much longer. Thanks to iTap RDP, I've even been able to do code reviews and make check-ins remotely, while iSSH lets me fool around with my cloud instances. Then there are the games - Chaos Rings, Secret of Mana, Infinity Blade...good stuff all the way around.
Silent_Buddha
Legend
Number one is being able to run anything that I may want or need at any given time without having to go through RDP in order to do it.
Easy and fast transfers to and from the device. Available use of external media. HD video output, preferably over HDMI (iPad is extremely limited with a maximum output of 1024x768), capable of outputing a display of more than just video and certain defined applications. Which again gets back to not being able to use any available application I need without the use of RDP which may not be able to output via the VGA, component or composite output of the iPad.
I imagine Android slates are going to be far more useful in this area, but will again be far more limited than a full OS. And that's above and beyond how locked down they might be depending on manufacturer.
And even then there's performance implications with certain apps. Especially if RDP is unavailable or undesireable.
And without an active digitizer, pen input is absolutely craptacular on a capacitive touch display which makes handwritten notation unbearable and is absolutely useless when it comes to any serious inking or sketching. 2 areas where a slate device would presumably excel.
Regards,
SB
Easy and fast transfers to and from the device. Available use of external media. HD video output, preferably over HDMI (iPad is extremely limited with a maximum output of 1024x768), capable of outputing a display of more than just video and certain defined applications. Which again gets back to not being able to use any available application I need without the use of RDP which may not be able to output via the VGA, component or composite output of the iPad.
I imagine Android slates are going to be far more useful in this area, but will again be far more limited than a full OS. And that's above and beyond how locked down they might be depending on manufacturer.
And even then there's performance implications with certain apps. Especially if RDP is unavailable or undesireable.
And without an active digitizer, pen input is absolutely craptacular on a capacitive touch display which makes handwritten notation unbearable and is absolutely useless when it comes to any serious inking or sketching. 2 areas where a slate device would presumably excel.
Regards,
SB
Is it likely that the iPad 2 will get an active digitiser? Even if its a 3G model only feature?
Silent_Buddha
Legend
I would hope so, but I'm doubtful myself. Apple view it primarily as a media consumption device, as such capacitive touch is really all you need. The iPad is definitely a really well executed product in that respect, although I think they messed up by not having an HDMI out or supporting any sort of external media. HDMI I can understand however, as the iPad may not be able to playback 1080p content (it's currently limited to 720p video sources IIRC). But even then 720p over HDMI would still have been fantastic for a media device, IMO.
It's admirable that companies have been able to wrangle business and developement uses from it, but it's still not the primary target so an active digitizer for pen input may be a battery life and cost sacrifice that Apple will be unwilling to make.
Regards,
SB
It's admirable that companies have been able to wrangle business and developement uses from it, but it's still not the primary target so an active digitizer for pen input may be a battery life and cost sacrifice that Apple will be unwilling to make.
Regards,
SB
Silent_Buddha
Legend
Sure for something with as small of a screen as a smartphone you're going to be limited in what you could do with a pen anyways.
Move up to a proper slate with a decent sized screen and you'll find there's a lot of people asking for pen input. Especially if it's a Windows based slate as then handwriting speed is orders of magnitude faster than any form of onscreen keyboard/keypad and less likely to have errors in writing/typing.
And that doesn't even get into all the people that have been waiting for a slate device for art related sketching and inking, yet another demographic that wouldn't have wanted a pen based smartphone as the format is just too small.
As well slates become fantastic devices for electronic note taking in classes or meetings. But only if you have an active digitizer pen.
Just look at how fast the capacitive touch pen market is growing as people try to find a way to draw or take notes on an iPad. Only to find out that the experience with a capacitive touch pen is really REALLY bad.
Regards,
SB
Move up to a proper slate with a decent sized screen and you'll find there's a lot of people asking for pen input. Especially if it's a Windows based slate as then handwriting speed is orders of magnitude faster than any form of onscreen keyboard/keypad and less likely to have errors in writing/typing.
And that doesn't even get into all the people that have been waiting for a slate device for art related sketching and inking, yet another demographic that wouldn't have wanted a pen based smartphone as the format is just too small.
As well slates become fantastic devices for electronic note taking in classes or meetings. But only if you have an active digitizer pen.
Just look at how fast the capacitive touch pen market is growing as people try to find a way to draw or take notes on an iPad. Only to find out that the experience with a capacitive touch pen is really REALLY bad.
Regards,
SB
itsmydamnation
Veteran
i wonder if AMD started to increasing there Margins and shipping numbers with bobcat/bulldozer/llano if we could see much more agressive pricing of GPU's come RV9XX. if the rest of AMD performed better then the graphics department could be more agressive and i think in general everyone agrees that at a comprable SKU to SKU comparision between AMD and NV AMD has more room to move.
from what i hear around the net AMD are selling very good numbers of bobcat based APU's and there is a market that no one seems to talk much about and in terms of volume it could potencially dwaf all the other markets we talk about and it seems perfect for bobcat APU's and that the developing world.
the selling potencial in places like china/india/brazil etc must be massive
its good for AMD in lots of ways
1. its on bluk ( cheaper then SOI)
2. its really small
3. bobcat offers enough single thread performace over ATOM to make operating systems bearable
4. it would be easy for them to release high freq/higher TDP APU's if needed into just these markets probably at almost 0 aditional cost.
5. TSMC seems to have heaps of 40nm capacity these days
6. because its being made at TSMC they can transistion to a new node when it makes sense not just because they need to rip the old gear out of the FAB to put the new stuff in.
is there even enough room for the pins in a straight 28nm shink of Zacate?
to me it seems like if Intel wanted to compete in this market it wouldn't be profitable for them it would just be a case of market share.
im feeling a bit bullish towards amds future.
from what i hear around the net AMD are selling very good numbers of bobcat based APU's and there is a market that no one seems to talk much about and in terms of volume it could potencially dwaf all the other markets we talk about and it seems perfect for bobcat APU's and that the developing world.
the selling potencial in places like china/india/brazil etc must be massive
its good for AMD in lots of ways
1. its on bluk ( cheaper then SOI)
2. its really small
3. bobcat offers enough single thread performace over ATOM to make operating systems bearable
4. it would be easy for them to release high freq/higher TDP APU's if needed into just these markets probably at almost 0 aditional cost.
5. TSMC seems to have heaps of 40nm capacity these days
6. because its being made at TSMC they can transistion to a new node when it makes sense not just because they need to rip the old gear out of the FAB to put the new stuff in.
is there even enough room for the pins in a straight 28nm shink of Zacate?
to me it seems like if Intel wanted to compete in this market it wouldn't be profitable for them it would just be a case of market share.
im feeling a bit bullish towards amds future.
- Status
