The Intel Atom Thread

[Arachnotronic]

Isn't Z3470 the one going in most tablets? Why don't they all just pay the $5 extra and throw in the Z3770 instead? Maybe power consumption / heat / battery life does has something to do with it after all ... ??? :hmm:

*cues twilight zone music*

No. $5 is a big deal when you're talking about $250 tablets with razor thin margins to begin with

 

[TechFan1]

I was looking at the Dell Venue Pro 11, which is supposed to have 10hrs battery life on 30Wh battery. I may get one after I see some reviews.

 

[bullzz]

@ninaholic37 - "Why don't they all just pay the $5 extra .."

OEMs could throw in extra $5 to implement 802.11ac
OEMs could throw in extra $5 to implement bigger batteries in phones/tablets
OEMs could throw in extra $5 to put in quieter fans

they want to cheap out in every possible way. so I dont think Z3770 is at fault here

 

[Roland00Address]

No. $5 is a big deal when you're talking about $250 tablets with razor thin margins to begin with

I am not disagreeing. To add further to the point, also remember a tablet may sell for $250 to the final customer but that is after a middle man such as a store is involved. Those middle man vendors are going to want at least 10% so that tablet may be sold by dell or hp may be sold to the amazons, best buys, wal-marts of the world for $225 of $200 (if the product is not selling and thus the OEMs want to curry favor with the middle man to push product X by offering product X with a higher profit margin.).

The ipad mini 2012 teardown in Nov 2012 was about $188 of BOM cost and about $10 of manufacturing cost.

http://www.isuppli.com/Teardowns/Ne...l-of-Materials-Teardown-Analysis-Reveals.aspx

Most high end tablets with nice screens are going to have a BOM in the $150 to $200 range

Now this is the BOM cost, not the cost for any research and design, warranty, tech support, marketing or software development. With a Windows Table the OEMS are going to be for the most part paying for software development by buying a windows license which itself may be $30. This $30 even if it includes office home and student reduces the profitability for the OEMs.

-------------

In other words the gain or loss of $5 to the BoM is going to be a significant swing in the profitability of the device. Furthermore OEMs can offer the faster chip as a $30 or $50 dollar online order add on where they OEMs will be able to recoup this as a very high profit similar to how they recoup higher nand sizes as high profit.

It is intels best interest to offer two skus for tablets, for by offering a mainstream and a high end sku they can curry favor with the oems. Currying favors with the oems means more design wins.

Just like how apple didn't care if you got an ipad or a mini, just as long as its an apple, intel doesn't care if you get a 3740 or a 3770 just as long as it is an intel and not a qualcomm, a tegra, a samsung, or a rockchip.

 

[cbn]

Lets not forget the really cheap Tablets too.

http://www.zdnet.com/intel-shows-of...-100-tablets-and-wearables-at-idf-7000020495/

For tablets, more than 20 Android and Windows 8 tablets based Intel's new Bay Trail Atom SoC will be on sale by the US holiday season, with prices ranging to below $100, he said.

 

[Nothingness]

you should care though given an SOC is typically half a total tablet/phone TDP. Don't tegra 4/s800 tdp's reach (at load) ~5W-6W? If thats the case the z3770 seems to be a pretty interesting little chip.

Of course I care, but indirectly. My point is that we have no real battery life figures of a whole device. And it's all what matters. If OEM put bad components along with a great CPU, it means they can keep their HW and the great CPU. See my point?

 

[Nothingness]

Some new benchies here: http://www.journaldugeek.com/2013/10/11/intel-atom-bay-trail-bench

So they compare an Intel tablet dev/demo platform against a smartphone that can be bought right now? And still no battery life test.

 

[AtenRa]

People should understand that BayTrail-T has 2W SDP, we still dont know the TDP. I believe it will be close to 4-5W.
Also, it is another thing to have a high 3D Mark graphics score and another to measure real Games. BT-T cannot use high frequency on all 4-cores and its iGPU at the same time. It will not have the same performance in real games as it has in Ice Storm.

 

[mikk]

Look in the datasheet, there is no TDP. We probably won't get an official TDP number but more than 4W is unlikely for the max power draw.

 

[drikkie]

Some hands on at an Intel session:

http://www.theinquirer.net/inquirer/review/2299365/intel-bay-trail-benchmarking-test-and-hands-on

 

[Abwx]

Look in the datasheet, there is no TDP. We probably won't get an official TDP number but more than 4W is unlikely for the max power draw.

You dont get official numbers because they would contradict
what was displayed to the press...

People here seems to not suspect the obvious , that is that
Intel binned a few chips out of dozen thousands ones to get
a few samples that could run at significantly lower voltage
than the average of their production.

Selecting chips that can work with about 20% lower voltage
will reduce the cores power comsumption by 44% , hence
if the chip is basicaly a 7W SoC you ll end with only 4.86W...

Intel didnt publish numbers because they know that the
delivered chips wont be as efficient as the golden samples
they used for their marketing reviews, they simply cant
garantee the specs they displayed with an unknown protocol.

 

[ShintaiDK]

You dont get official numbers because they would contradict
what was displayed to the press...

People here seems to not suspect the obvious , that is that
Intel binned a few chips out of dozen thousands ones to get
a few samples that could run at significantly lower voltage
than the average of their production.

Selecting chips that can work with about 20% lower voltage
will reduce the cores power comsumption by 44% , hence
if the chip is basicaly a 7W SoC you ll end with only 4.86W...

Intel didnt publish numbers because they know that the
delivered chips wont be as efficient as the golden samples
they used for their marketing reviews, they simply cant
garantee the specs they displayed with an unknown protocol.

You keep going on in this thread with your homemade FUD.

 

[Abwx]

You keep going on in this thread with your homemade FUD.

You would expect intel to not use binned chip for marketing
purposes..?..

In short , your own logic is that a logic behaviour is FUD if it hurts intel...

 

[ShintaiDK]

You would expect intel to not use binned chip for marketing
purposes..?..

In short , your own logic is that a logic behaviour is FUD if it hurts intel...

We all know AMD is the best company in the world for you.

But do you really think all the OEMs avoided AMDs tablet chips because they are within the same power envelope as BT? Only to release a massive amount of new BT tablets, while not wanting to touch Temash in tablets like it was infected with the plague. You only see Temash in laptops. Even the MSI Temash tablet got nowhere.

 

[Abwx]

We all know AMD is the best company in the world for you.

Lol , coming from you such an ad hominem is just ironic
but i suppose that it s the result of a lack of argument..

But do you really think all the OEMs avoided AMDs tablet chips because they are within the same power envelope as BT? Only to release a massive amount of new BT tablets, while not wanting to touch Temash in tablets like it was infected with the plague. You only see Temash in laptops. Even the MSI Temash tablet got nowhere.

They were not used as much as expected for the same
old same reasons , guess why.?...TDP..??.Not at all...

And btw , why no official TDP number from Intel.?..

Because the real number is too good and would
surely cause a PR debacle , for sure..

 

[ShintaiDK]

Lol , coming from you such an ad hominem is just ironic
but i suppose that it s the result of a lack of argument..



They were not used as much as expected for the same
old same reasons , guess why.?...TDP..??.Not at all...

And btw , why no official TDP number from Intel.?..

Because the real number is too good and would
surely cause a PR debacle , for sure..

So back to the tinfoil arguments. Blame someone else.

Let me know when you someday get near facts for a change. Until then its impossible to even take your posts somewhat serious.

 

[Sweepr]

They were not used as much as expected for the same
old same reasons , guess why.?...TDP..??.Not at all...

This 3.9W TDP Temash has been in the market for months. Is there any reasonably good 8-10'' based on it out there (or even planned after W8.1 launch)? Sure, there must be some OEM anti-AMD conspiracy, lets just ignore the fact that 2x 1GHz Jaguar cores offer crappy CPU performance compared to BT-T/S800/T4 and less than 1/2 of A4 5000's GPU performance. Based on what we see on laptops I'm not conviced that this chip could offer competitive battery life in a small tablet form factor either.

 

[Arachnotronic]

Lol , coming from you such an ad hominem is just ironic
but i suppose that it s the result of a lack of argument..



They were not used as much as expected for the same
old same reasons , guess why.?...TDP..??.Not at all...

And btw , why no official TDP number from Intel.?..

Because the real number is too good and would
surely cause a PR debacle , for sure..

Bottom line is that Bay Trail is going into super sleek 7-8" tablets, and Temash is powering really cheap, sub-par laptops.

Everything else is, quite frankly, academic.

 

[mrmt]

Bottom line is that Bay Trail is going into super sleek 7-8" tablets, and Temash is powering really cheap, sub-par laptops.

And that's to be expected, no? Jaguar wasn't intended for tablets, it was always a bottom market x86 notebook processor and it fits nicely in the embedded market. Jaguar cannot compete on power consumption with ARM A15, let alone A57 that comes next year, and on top of that it holds no price advantage over the ARM solutions on the market place. That makes it a no-go for the tablet market.

Bobcat and Jaguar were carefuly designed to target the big gap that Intel used to have between their Core and their Atom line up, and guess what, you could only fit that chip on cheap laptops/netbooks.

 

[Abwx]

Once i questionned intel s methodology and their refusal
to publish the TDP specs of their chip it didnt take long
for the usual crowd to push AMD in the debate even if
it was not my point wich is that intel didnt publish said
specs for a reason and early numbers we got from a few
manufacturers all point to the intel "reviews" as being
doctored , that is , the results wont be reproducible
on the average production but nevermind , to keep mumm
will only help proping up the hype , as displayed in this
very thread.

 

[Arachnotronic]

And that's to be expected, no? Jaguar wasn't intended for tablets, it was always a bottom market x86 notebook processor and it fits nicely in the embedded market. Jaguar cannot compete on power consumption with ARM A15, let alone A57 that comes next year, and on top of that it holds no price advantage over the ARM solutions on the market place. That makes it a no-go for the tablet market.

Bobcat and Jaguar were carefuly designed to target the big gap that Intel used to have between their Core and their Atom line up, and guess what, you could only fit that chip on cheap laptops/netbooks.

Bingo. And it's not just power consumption, but power management - Jaguar's lack of turbo also makes it a no go.

 

[Abwx]

Bingo. And it's not just power consumption, but power management - Jaguar's lack of turbo also makes it a no go.

Technicaly speaking a turbo feature is counter productive
on a mobile device running on battery but usefull when
connected to the main.

 

[mrmt]

Bingo. And it's not just power consumption, but power management - Jaguar's lack of turbo also makes it a no go.

The lack of turbo is bad, but the lack of the deep sleep states is even worse.

 

[Nothingness]

Technicaly speaking a turbo feature is counter productive
on a mobile device running on battery but usefull when
connected to the main.

As long as turbo can be turned on/off by software, I don't see why it would be counter-productive.

 

[drikkie]

If you don't have turbo or at least speedstep, you have to find a middle ground for performance, with turbo it can just have low performance and only fire it up (till a certain (bios) constrained max) when needed.

And afaik Bay Trail has the most advanced form of turbo / speedstep, it takes all (sub)parts of the soc into account, cpu per core and gpu (per core?).