• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Kabini Rumors

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Intel17 said:
I love how AMD loyalists like to think that AMD, on a 28nm process, will be able to pack in as much performance in a given power envelope as Intel will on a 22nm process for a low power design.
.
Click to expand...

There was a thread about the intel node and it was closed
because intel afficionados complained that it was trolling ..

Anyway , the guy was right , Intel s 22nm is actualy 26nm ,
so dont expect a big advantage due to their process.

For thoses who are skeptic , just look at SB transistor density
at 32nm and compare to IB.

Despite most of the added transistor budget has been devoted
to the IGPU , wich should allow for even better density scaling
we can see that transistor density increased by only 60% , far
from the theorical 100% , and this match with an actual 26nm node.....
 
Abwx said:
Anyway , the guy was right , Intel s 22nm is actualy 26nm ,
so dont expect a big advantage due to their process.
Click to expand...
I beg to differ: Expect a big advantage, not due to 26nm, but due to finfets. These things are from the future 😀

monstercameron said:
how much info is there on the next gen atoms? and how can you draw that conclusion?
Click to expand...
THey will also use out of order execution, like any other modern architecture does, besides the old Atoms, which use the slow in-order-execution. It will be definitly an interesting competition, I am especially interested in the die size. The old Atoms were rather huge compared to AMD's Bobcat architecture.
 
Last edited:
Finfet provide some advantage only at lower frequencies ,
dont be fooled by Intel s slide that showed 37% better
perfs than.....32nm planar!!

Should they display a comparison with 22/26nm planar
that their slide would show way lower advantage ,
particularly at average and high frequencies.

Edit : The finfets are decades old devices in the
discrete transistors and high frequency analog ICs
market , where they are called multigate mosfet transistors,
so much for being from the future...
 
Last edited:
Abwx said:
Finfet provide some advantage only at lower frequencies ,
dont be fooled by Intel s slide that showed 37% better
perfs than.....32nm planar!!

Should they display a comparison with 22/26nm planar
that their slide would show way lower advantage ,
particularly at average and high frequencies.
Click to expand...
Well the topic here is Atom, isn't it? Guess at which clocks these chip will run ...
 
Dont know for the frequency but anywhere above 0.8V wont
show dramaticaly lower TDP than the same product manufactured
with TSMC s bulk process.
 
erunion said:
People are leaning on its GPU performance increase as its differentiator, but they must not understand just how puny its predecessors' performance was. E-350 is not even in the same league as a mere HD 4000.
Click to expand...

E-350 was on HD2000 level (the slower Sandy Bridge GPU variant). HD4000 is way faster. I expect a big GPU advantage for Temash against Bay Trail, at least a 2x advantage over Bay Trail. On the CPU side it probably will lose badly to Silvermont.

zlatan said:
In a performance demanding game the HD Graphics 4000 won't run in 1+ GHz, just only 350 MHz.
Click to expand...

This is plain wrong.
 
E-350 is not even in the same league as a mere HD 4000.
Click to expand...
Your right the 17w Ivy Bridge CPU inside surface Pro, is about
2x or 3x as fast as the E350.

The E-350 is a brazo's model and a 40nm chip, from Jan 4th 2011. (Motherboard & chip = ~100$)
The i5-3317U is a Ivy bridge model and 22nm chip, from april 23rd 2012. (costs ~225$)

The jump from 40nm to 22nm is a big one.

Hopefully AMD can do better at 28nm than they can at 40nm.
 
Last edited:
mikk said:
E-350 was on HD2000 level (the slower Sandy Bridge GPU variant). HD4000 is way faster. I expect a big GPU advantage for Temash against Bay Trail, at least a 2x advantage over Bay Trail. On the CPU side it probably will lose badly to Silvermont.
Click to expand...

Given how little we know about Silvermont, I don't feel especially confident predicting its performance.
 
NTMBK said:
Given how little we know about Silvermont, I don't feel especially confident predicting its performance.
Click to expand...


Silvermont tablet is targeted for 2 Ghz while Jaguar for Tablet only 1,0 Ghz at a lower TDP. Furthermore Silvermont makes a bigger jump in IPC than Jaguar. This is all I need to know.
 
Abwx said:
Likely 12W TDP for the 2GHZ model , is that what you call lower TDP.?.
Click to expand...

The slide dissagrees.

intel_atom_valleyview_baytrail_5.jpg
 
Abwx said:
It says that the lower power parts will be about 5W ,
frequencies being 1.2-1.4Ghz according to THG.

The low power is attained thanks to subpar GPU as well...
Click to expand...


Bay Trail Tablet is targeted for 3W or less at 2.1 Ghz. Don't confuse desktop Silvermont with tablet Silvermont.
 
beginner99 said:
As you can see in above chart brazos graphics pretty mich sucked and was way worse than HD 4000. And Haswell mobile graphics will be a lot faster than HD 4000. So no Kabini will lose also in that area and will be demolished in CPU performance.
Click to expand...

You're correct on the CPU front but the gap has been seriously narrowed. On GPU, only Intel the company believes this.
 
Pretty early to be declaring what will be viable. At the very least it's about time the brazos class of product was refreshed by AMD.
 
Vesku said:
Pretty early to be declaring what will be viable. At the very least it's about time the brazos class of product was refreshed by AMD.
Click to expand...

Indeed. The canning of the 28nm Bobcats has to be one of my biggest disappointments from AMD of recent years (yes, even higher than Bulldozer, before you all pile in 😛 ). That chip could have made for some very nice ultraportables, and maybe staved off the demise of the netbook- if GloFo had delivered on their promises.
 
Arkadrel said:
Your right the 17w Ivy Bridge CPU inside surface Pro, is about
2x or 3x as fast as the E350.

The E-350 is a brazo's model and a 40nm chip, from Jan 4th 2011. (Motherboard & chip = ~100$)
The i5-3317U is a Ivy bridge model and 22nm chip, from april 23rd 2012. (costs ~225$)

The jump from 40nm to 22nm is a big one.

Hopefully AMD can do better at 28nm than they can at 40nm.
Click to expand...

no doubt kabini will improve over Brazos. I don't doubt that it will double the gpu performance. The question is to what effect?

AMD is pushing the idea of gaming tablets. but with less GPU performance than hd4000 or trinity, will consumers see them that way? Or will they see them as still being the same budget product that Brazos is?
 
True that it has been the most "stable" AMD plateform ever , heck ,
even the frequency was barely increased , by a miserable 10%
or so....
 
erunion said:
no doubt kabini will improve over Brazos. I don't doubt that it will double the gpu performance. The question is to what effect?

AMD is pushing the idea of gaming tablets. but with less GPU performance than hd4000 or trinity, will consumers see them that way? Or will they see them as still being the same budget product that Brazos is?
Click to expand...

temash is their tablet chip and while it will be slower(by how much is yet to be seen) it will have much better battery life in comparison to and i5 ulv w/ hd4000. which seems more appealing to customers?

a)$499 atom clovertrail tablet, weak cpu & gpu(or valleyview), 3hr battery under full load and 8hrs normal usage

b)$499 temash tablet, better than atom cpu and gpu, 2hr battery under full load and 6hrs normal usage

c)$999 i5 tablet, better than temash cpu and gpu, <1hr battery under full load and 4hrs normal usage
 
From S/A reading their thread there about Kabini.

The dual core (1ghz) Temash is supposedly ~3-3.5w TPD
With a typical power usage of ~1w.

The quad cores (Temash) are supposed to clocked higher and have a ~8w TPD.
These will have 2 CU GCN.

Both will have: Dedicated video encode/decode pipeline, southbridge on chip, ARM cores? for TrustZone.


Crazy we re gonna start seeing x86 at 1watt typical power loads.
Even if that might just be 2x 1ghz cores.
 
Last edited:
You must log in or register to reply here.

TRENDING THREADS

Back
Top