Richland & Kabini rumours

[itsmydamnation]

And they are in the exact same category as Atom.

A Core based Celeron would run in circles around any.

hardly, Celeron will have no turbo, no AVX, poor GPU support and poor GPU performance, Celeron will also have higher platform costs.

Celeron will loose in multi-threaded workloads and might even lose in single thread in situations where AVX is used. let not talk about gaming performance particular with non blockbuster titles. It is hardly as clear cut as you make out.

SHOCK HORROR SURPRISE?!?!?!@!@!@@#12 :whiste:

 

[NTMBK]

hardly, Celeron will have no turbo, no AVX, poor GPU support and poor GPU performance, Celeron will also have higher platform costs.

Celeron will loose in multi-threaded workloads and might even lose in single thread in situations where AVX is used. let not talk about gaming performance particular with non blockbuster titles. It is hardly as clear cut as you make out.

SHOCK HORROR SURPRISE?!?!?!@!@!@@#12 :whiste:

He was comparing Brazos, not Kabini- in which case yes, the low end Celerons definitely outclass it. Take a look at these benchmarks: http://www.tomshardware.com/reviews/nettop-performance-review,3211-5.html Of course, it's not a vast gap- the gap between the Celeron and the i3 is usually much larger. So its very conceivable that Kabini will close the gap with the low end Celerons, and potentially even the Pentiums. But the i3s will probably remain out of its reach.

 

[NTMBK]

And they are in the exact same category as Atom.

A Core based Celeron would run in circles around any.

No arguments on that from me! There's a reason I got my girlfriend a nettop with an i3 in it, not a Bobcat. But there are plenty of people out there who are less discerning, and just want something faster than a 10 year old single core P4, in a smaller box.

EDIT: Just for comparison, to give some idea why these systems might seem appealing to your average Joe:

Rest of benchmarks here. And for reference, my parents were using a 1.8Ghz Pentium 4 until they replaced it 3 months ago. Going from that to even a Bobcat, with Windows 7 and a modern hard drive, would feel like a big step forward. (The i5 that they actually got was far superior of course- given how long they held onto their last PC, I didn't want them stuck with Bobcat for a decade!)

 

[frozentundra123456]

so explain bobcat then, theres 1 billion first world people, 2-3 billion second world people. There is a huge market that kabini can sell into.

Bobcat was placed in desktops and full sized notebooks, both of which it is barely adequate for, if at all, because the OEMs want to maximize profits and sell at a very low price. Unfortunately many people just see the low price and dont know how limited the performance is.

Kabini may be more suited for this kind of use, but I will reserve judgement until the chip actually turns out.

 

[krumme]

I disagree here. At the top of the market bracket Kabini faces Celeron/Pentium, and if AMD charges too much Intel will be more than happy to raise fab utilization in order to supply customers there.

At the bottom, AMD must be cost competitive against ARM solutions. Nobody will go for anemic performance if it costs too much because ARM already anemic performance with better power efficiency and reasonable costs. And I'm not even counting Silvermont in Q4.

The inelastic component of Kabini demand is *very small*:

- Hardware that needs to be x86 compatible but needs more power than Clovertrail+ and less costly than Haswell.

As I said, Kabini faces a much tougher environment than Brazos faced when it arrived. I doubt that AMD can repeat Brazos success with Kabini. AMD management f*ed up big time by canning Brazos shrink, but what can we expect from those guys?

40% margin is super for most companies. And temash/bobcat is fit for it cost wice doing the ROI exercise. Its made for the market of tomorrow, unilike the rest of the AMD big CPU/APU lineup that is just junk business.

I bet your analysis of temash / kabini for 2013 will prove wrong. If there is a product that can get windows 8 of the ground for tablets 2013 its exactly temash. Temash is a key here, and if i was RR i would take the phone, call MS ceo and put some heavy pressure on - and hopefulle he have already done that. Thats his job, get into the fight and get his hands dirty. Otellinini would have done that long ago.

And yes, right now, Intel is selectively dumping/giving good offers for ulv core to the oem, to take the temash/kabini up front. Standard tactics. But how long will it last?

If RR knows his job, he would have played this the brutal way, and only let 2-3 OEM get access to Temash the first ½ year. Exclusive rights. Negotiate, and punish the others. Punish. Punish. Punish. And even before starting the negotiations, he/AMD should start by proclaiming "officially" that Dell will not get AMD processors for excactly the next 5 years (the official excuse will ofcourse be something else). And then stand by it.

If RR cant do it, let the board hire a Russian. He will always meet with 3 heavy armoured guards. Thats a good signal that this is serious business.

 

[Olikan]

Bobcat was placed in desktops and full sized notebooks, both of which it is barely adequate for, if at all, because the OEMs want to maximize profits and sell at a very low price. Unfortunately many people just see the low price and dont know how limited the performance is.

people bought a cheap laptop, with low performance.... where is the problem?

 

[krumme]

people bought a cheap laptop, with low performance.... where is the problem?

Bobcat is a fine balance with a cheap 5400rpm hd. The problems is high-def netflix and a few other situations. Plays 1080p youtube fine today. If they consumers would pay for some extra, it would probably be better higher res sreen and/or better battery life

As temash can get into no fan teritory, i think we will even see a lot of consumers prefer that together with slow 64gb ssd. Cheap, small and doesnt make a sound.

It will move the entire notebook segment to smaller size. 25w will end up in a lot of dirt cheap desktops.

For an OEM this is the perfect product for your profit in these low margin times. You can scale it to 90% of your consumer market if you want to. It lower the total cost of implementation across all portfolios when you can use the same platform. Its a far more scalable platform than bobcat, that really only was fit for 18w. Its a brilliant opportunity for the OEM to save cost and introduce new tablet products.

 

[monstercameron]

Bobcat is a fine balance with a cheap 5400rpm hd. The problems is high-def netflix and a few other situations. Plays 1080p youtube fine today. If they consumers would pay for some extra, it would probably be better higher res sreen and/or better battery life

hd netflix and medium quality hulu aren't problems(or shouldn't). netflix modern ui/metro app plays just fine on my c-50 based netbook and hulu has been fixing their web player medium quality so work fine. Most other flash videos work fine too.

 

[krumme]

hd netflix and medium quality hulu aren't problems(or shouldn't). netflix modern ui/metro app plays just fine on my c-50 based netbook and hulu has been fixing their web player medium quality so work fine. Most other flash videos work fine too.

Okey, its obviously about using the hardware, and it seems what matters is fine now because drivers and programs have matured.

Then nothing stands in the way for consumers embracing even the low performance variants. My guess is, they take the low power to the more performance oriented variants.

 

[pablo87]

I disagree here. At the top of the market bracket Kabini faces Celeron/Pentium, and if AMD charges too much Intel will be more than happy to raise fab utilization in order to supply customers there.

At the bottom, AMD must be cost competitive against ARM solutions. Nobody will go for anemic performance if it costs too much because ARM already anemic performance with better power efficiency and reasonable costs. And I'm not even counting Silvermont in Q4.

The inelastic component of Kabini demand is *very small*:

- Hardware that needs to be x86 compatible but needs more power than Clovertrail+ and less costly than Haswell.

As I said, Kabini faces a much tougher environment than Brazos faced when it arrived. I doubt that AMD can repeat Brazos success with Kabini. AMD management f*ed up big time by canning Brazos shrink, but what can we expect from those guys?

What I was referring to in terms of profitability is Brazos/kabini is a lot lower overall cost than trinity/Richland yet they sell for similar prices (Kabini with its integrated I/O and more appealing vectors might garner higher ASPs), so its a lot more profitable.

In terms of elasticity, history has shown this to be the case, the reasons are varied and I won't get into it.

Kabini is a much stronger competitor than Brazos though:

- much faster, more competitive IPC
- fch I/o is integrated and overall power is lower - these are very good differentiators and help lower costs at a time when memory and HDD costs are rising.
- its a quad core, intel's meet comp stuff is still dual core
- its the architecture in the 2 main gaming consoles and its going to be in tablets as well as notebooks so it will have appeal in and of itself
- its GPU is current and can cross-fire - Intel has nothing comparative generally, and certainly not in the $40 price range.

canning the brazos die shrink was dumb but that's going to be water under the bridge very soon.

 

[nehalem256]

canning the brazos die shrink was dumb but that's going to be water under the bridge very soon.

They didn't can it because they wanted to. They canned it because GF failed to deliver a 28nm process to build it on. And their wasn't time (or it wasn't profitable) to port the design over to TSMC.

 

[Third_Eye]

I disagree here. At the top of the market bracket Kabini faces Celeron/Pentium, and if AMD charges too much Intel will be more than happy to raise fab utilization in order to supply customers there.

At the bottom, AMD must be cost competitive against ARM solutions. Nobody will go for anemic performance if it costs too much because ARM already anemic performance with better power efficiency and reasonable costs. And I'm not even counting Silvermont in Q4.

The inelastic component of Kabini demand is *very small*:

- Hardware that needs to be x86 compatible but needs more power than Clovertrail+ and less costly than Haswell.

I am more excited about Temash than Kabini. I explained in detail why in the below link in SA forums.

http://semiaccurate.com/forums/showpost.php?p=178465&postcount=247

Brazos was a success as far as the markets it targeted and ROIs. The problem was, by the time Brazos was released, its initial target market "Netbook" started to dwindle. Brazos single handedly exposed Atom's weakness..
It is a surprise that Brazos was still a success when given the "finger" by the market.

The cancellation of Wichita and Krishna by AMD due to GF 28nm screw-up is the one that hurt AMD really bad. So Brazos 2 (which is nothing but Brazos 1 with a better southbridge) had to fill the gap and enter the cheap notebook scene where it was easily bettered by the Celerons.

So in Temash we have a CPU core which is equivalent/better (cinebench) to a Sandy i3 core and running at a far lower power budget and 30% lesser die-size made in a process that has matured > 18 months and for a market where AMD's penetration is close to zero.

As I said, Kabini faces a much tougher environment than Brazos faced when it arrived. I doubt that AMD can repeat Brazos success with Kabini. AMD management f*ed up big time by canning Brazos shrink, but what can we expect from those guys?

The big difference is that Jaguar cores already have
a) 1 guaranteed new market in PS4
b) 1 likely new one with XBox720.

So it has opened 2 new markets for AMD for the architecture which are different form the "PC" based market that AMD has been focussing all along. So Kabini might face a glut in the traditional PC market but might end up in the Console/Gaming market.....

The tablet market is something that is more friendly to GPU over CPU. So AMD will have a huge advantage over Intel CloverTrail +.

I am a little disappointed that AMD is positioning Temash only as a Windows 8 product and does not encourage adoption of Android or other OSs.
http://semiaccurate.com/forums/showpost.php?p=178474&postcount=253

Also in case of Temash, AMD has already partnered with all 3 big Taiwaneese ODMs Quanta, Compal as well as Wistron to provide reference Tablet form factors.

https://www.youtube.com/watch?v=8N7L...PClrQ&index=17 - Quanta
https://www.youtube.com/watch?v=OjBn...PClrQ&index=18 - Wistron
https://www.youtube.com/watch?v=b1kg...PClrQ&index=20 - Compal

So AMD has done the groundwork for OEMs to just pick and choose rather than design....

 

[krumme]

Well obviously the standard work is done this time. After all AMD is growing up and starting to realize its business is not designing and producing APUs but providing profit for the OEM /Sony. I doubt the cpu designer, Dirk ever understanded - to the bone - the business he was in.

If Intel succeded this time in keeping kabini out of the 13.3 and sub notebook size with its kickback, bonus, and half leagal grease, RR needs to have implanted new balls. If he can not prevent it, with his history, he is absolutely useless. We all know he doesnt really know what a transistor is, but who cares, but he have to deliver right here. Right now. This should be his prime time.

Who cant sell a full windows experience, in a fanless tablet, less than 10mm?
http://www.youtube.com/watch?v=8N7LNHRl2jE

 

[happysmiles]

RR is the bees knees

 

[Idontcare]

If Intel succeded this time in keeping kabini out of the 13.3 and sub notebook size with its kickback, bonus, and half leagal grease, RR needs to have implanted new balls.

What is this mythical unethical Intel to which you refer

Didn't you get the memo, Intel has been named a 2013 World's Most Ethical Company

And since it is there in print, in black and white, they can't possibly be doing anything unethical or shady. Umpossible I say!












Yes I am being facetious. Everyone knows those lists are "pay-to-be-listed" which makes it all the sadder, someone in Intel actually felt the need to pay up and get Intel on the "whitewashed and purified" list. LOL

 

[vampirr]

Intel may seem like moral and ethical company but what they have done numerous times against some companies and mostly AMD is atrocious... AMD has done one or two but compared to Intel, AMD is an angel...

 

[mrmt]

Intel may seem like moral and ethical company but what they have done numerous times against some companies and mostly AMD is atrocious... AMD has done one or two but compared to Intel, AMD is an angel...

John Fruehe, Randy Allen, Mike Houston, Dave Baumann... It's nice to see AMD digging its own pockets to pay those guys to upheld ethical standards in the industry.

 

[vampirr]

Just couple of people on forums... Compared to Intel that used bribery, gave "special" discounts to shops that delayed batch of AMD's CPU's, the cinebench scandal with Intel compiler, breaking antitrust laws in more than 10 countries and I was partially a victim because I got a Pentium 4 and not something from AMD that was better in 2004...

 

[Gideon]

Extracted from this post:
http://semiaccurate.com/forums/showpost.php?p=178465&postcount=247

A Temash A6-1450 in a docked mode at 1.4GHz, has a die-size of 110 mm2 and Cinebench score of 1.39 at a 05W TDP
A Corei3 2367M at its standard speed 1.4GHz, has a die-size of 149 mm2 and Cinebench score of 1.34 at a 17W TDP

Dude, hold your horses... While Temash might very well turn out to be an excellent product, you're definitely getting overhyped!

AMD has confirmed, that when Turbo Dock WILL raise TDP:

http://blogs.amd.com/home/2013/02/25/what-your-tablet-doesnt-want-you-to-know/

In engineering speak, AMD Turbo Dock technology is delivered through “dynamic TDP” that lets the processor automatically adjust its top frequency to the scenario in which it is operating.

This article also confirms it (I have to link it for the third time, as no-one seems to actually read it):

http://www.slashgear.com/amd-turbo-dock-temash-tablet-reference-design-hands-on-24271087/

The differences are clear: docked, the Temash ramps up from 8W to 15W, and that’s enough to see a 40- to 50-percent hike in the Fish Bowl score.

8W seems a bit high, considering we have been hearing about 5.9W a lot, but this also states that we may well have Temashes with as high a TDP:

http://techreport.com/news/24186/new-details-early-benchmark-data-revealed-for-15w-kabini-apu

AMD plans to offer Temash in both sub-5W and 8-9W variants.

So it'll be at Kabini's performance with also Kabini's TDP.

Considering, that Intel will also release a 1.4 Ghz i3 with 13W TDP (+ southbridge) the power draw will actually be at least comparable. Obviously Temash will be a much cheaper alternative to build and buy for the OEMs. It is a SOC and has 1 mem-channel while Intel's 7W SDP i3s START at 200$+. It won't however be magically performing at i3 levels with 3x smaller TDP. Actually in Single Threaded scenarios it'll probably be up to 50% slower than i3 Unless it has a Turbo on top of the Turbo Dock (dawg ), in which case it'll be probably closer to 30% slower.

Don't build your expectations up to an insane level, you're bound to be dissapointed then ...

 

[monstercameron]

Just couple of people on forums... Compared to Intel that used bribery, gave "special" discounts to shops that delayed batch of AMD's CPU's, the cinebench scandal with Intel compiler, breaking antitrust laws in more than 10 countries and I was partially a victim because I got a Pentium 4 and not something from AMD that was better in 2004...

perspective is key.

 

[Olikan]

Some OpenCL benches... kabini vs tamesh... looks to have some bandwidth problems...

 

[monstercameron]

^what benchmarking tools are they using for those results?

 

[Olikan]

^what benchmarking tools are they using for those results?

forgot the link...http://semiaccurate.com/2013/03/08/opencl-performance-of-next-gen-low-power-apus/#.UTn2UIF4_z8

These tests were performed under Windows OS with an OpenCL driver version 1124.2 (SDK 2.8).

there is a low-end Llano vs Kabini in the link, but is kinda point less, since GCN is FAR better than VLIW-5 in compute (arguably, 90% better is still kinda impressive)

 

[Gideon]

Just to prove the point about too high expectations...

I personally hoped that, Temash will offer at least SOME power savings compared to Bobcat.

Seems like a very modest hope right? After all, it's a SOC, built on a smaller process and is heavily optimized to power gate better. We also all remember the Hot Chips talk about faster CC6 and C6 times and about L2 running on half the frequency and being able to be power gated separately, etc, etc ... We also had articles like these:

http://www.eetimes.com/design/eda-d...rocessor-core-with-RTL-clock-gating-analysis-

AMD was able to lower the typical power by approximately 20% while increasing frequency at the given voltage by over 10%.

It turns out that at the same clocks and core-count, Temash will offer NO power-draw advantage whatsoever, over Bobcat! Unless AMD itself posted invalid info.

Here is a slide, from the Bobcat based Z-60 we probably remember:

And here is the information about Temash:

http://www.amd.com/us/products/notebook/tablets/Pages/tablets.aspx#4

Total system power for the reference platform is projected at 2.8 W at idle, 3.7 W during web browsing, 5.3 W during video playback and .07 W during a system S3 “sleep” state. Battery life calculations were derived using a 35Whr battery pack at 98% utilization. The power projections are based on the “Larne” reference system with a configuration including the A4-1200 Dual Core 1.0GHz APU, AMD Radeon™ HD 8180 series graphics, 2GB DDR3-1066 system memory and Microsoft Windows 8. TEM-1

So if it's a typo on AMD's part, it's quite a typo (they specify "Larne" platform 2 times as well as the A4 branding and GPU) and they'd better fix it fast.

So even with really modest expectations you can get burned

 

[frozentundra123456]

Just couple of people on forums... Compared to Intel that used bribery, gave "special" discounts to shops that delayed batch of AMD's CPU's, the cinebench scandal with Intel compiler, breaking antitrust laws in more than 10 countries and I was partially a victim because I got a Pentium 4 and not something from AMD that was better in 2004...

I bought an Athlon in the same time frame. No body forced me to buy a pentium.