• 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.

Next Gen Exynos laughs at your Tegra 3

Page 9 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Again. Exynos 5250 was never meant to be used in SGS3 or even in a phone. It has nothing to do with time to market, yields or potential losses. All one has to do is read Samsungs official press release.

http://www.samsung.com/global/busine...o?news_id=1267

"Designed specifically for high-end tablets"

"Samsung's Exynos 5250 offers a host of peripheral functions including an embedded image signal processor enabling 8 Megapixel resolution images at 30 frames per second, a full HD 60 frame per second video hardware codec engine for high resolution 1080p video recording and playback, a HDMI 1.4 interface for sharp and crisp multimedia content transmission, along with a diverse scope of embedded booting device interfaces such as SATA, UART, USB and external ports such as USB3.0, eMMC4.5 and eSD3.0."

Samsung should have an advantage over TI and Qualcomm because they design their own chips, they have their own foundry and they are the OEM.

Just because this chip is not coming until Q3 doesnt mean its because Samsung cant produce it before then, perhaps the 2560x1600 screen is not ready until Q3, we dont know for sure

My point is, that 5250 is meant for Q3 but a lower clocked 5230 might come out earlier

Until we see an official announcement, anything is possible. If 4412 was really their next mobile SOC, why the silence from Samsung?
 
sarav said:
I don't know much about Medfield. Is it X86 based? I know Intel is working with Google for Android on x86 processors, but will the apps have to be recompiled or would this not be a requirement as apps run on a VMware?
Click to expand...
Intel's Medfield isn't going to be a match for Krait or next gen Exynos.
Intel won't be competitive until they reach 22nm.
 
lothar said:
Intel's Medfield isn't going to be a match for Krait or next gen Exynos.
Intel won't be competitive until they reach 22nm.
Click to expand...
Evidence?
I believe that 22nm will help with battery life, but why are you saying 32nm vs 32nm won't be competitive?
 
ilkhan said:
Evidence?
I believe that 22nm will help with battery life, but why are you saying 32nm vs 32nm won't be competitive?
Click to expand...
Screen&


"Shipping Smartphones" means those already shipping today.
By the time Medfield comes out, Krait, Samsung's new Exynos SoC, and possibly even Apple will already be out the door.

Medfield maybe competitive against phones shipping today...Won't be against the phones that will be shipping in the next 3-4 months IMO.

I was never comparing 32nm vs 32nm specifically. I'm comparing Intel's 32nm to whatever "xx"nm will be out in the competition by the time Intel release Medfield.
 
Intel's problem isn't die size, no one is rushing to port QSD8250 or OMAP3 to the latest die shrink, Intel's problem is competitive architecture within constraints of the SoC's designed for the portables market. If there is one giant that had all the time, resources and experience needed, it's Intel but the mobile market isn't quite simple to master as both Intel and Nvidia have learned. Even within ARM architecture it's difficult to compete, let alone trying to beat them with a new outside solution.
 
The problem with Intel is that x86 is ancient crap that requires a lot of cruft on the die to maintain backwards compatibility. Intel could make a competitive SoC but it would break twenty years of software which negates Intel's largest advantage.
 
poofyhairguy said:
The problem with Intel is that x86 is ancient crap that requires a lot of cruft on the die to maintain backwards compatibility. Intel could make a competitive SoC but it would break twenty years of software which negates Intel's largest advantage.
Click to expand...

It would be better that Intel cannibalize their own x86 business than for some other company to do it.
 
poofyhairguy said:
The problem with Intel is that x86 is ancient crap that requires a lot of cruft on the die to maintain backwards compatibility. Intel could make a competitive SoC but it would break twenty years of software which negates Intel's largest advantage.
Click to expand...
I have to look this up to get the "exact" details, but I don't think x86 registers take up as much die space as people think it does.
 
sarav said:
Will the Android apps have to be rewritten to work on x86 platforms?

In other news looks like the A6 is going to be quad core A9 based processor(http://www.slashgear.com/quadcore-apple-a6-support-found-in-ios-5-1-beta-06206785/). If Samsung can bring out A15 processor based phones later in the year, it would be a huge win for them as Apple won't update their processors till 2013 (assuming they follow a yearly upgrade pattern).
Click to expand...
This is certainly an interesting development...
So that means Samsung doesn't need to rush Exynos 5250 out the door for their Galaxy S3 like people think they do.
 
lothar said:
I have to look this up to get the "exact" details, but I don't think x86 registers take up as much die space as people think it does.
Click to expand...

Just the fact that ARM was designed for mobiles and x86 wasn't means that ANY x86 backwards compatibility requires a power envelope sacrifice.

Intel and x86 won't be competitive until it can get one or two size generations ahead of ARM, which seems impossible given the rate of ARM growth.
 
poofyhairguy said:
Intel and x86 won't be competitive until it can get one or two size generations ahead of ARM, which seems impossible given the rate of ARM growth.
Click to expand...

what does arm growth have to do with it? whether it's good or not - ARM still is totally reliant on TSMC, GF and samsung right?
 
tommo123 said:
what does arm growth have to do with it? whether it's good or not - ARM still is totally reliant on TSMC, GF and samsung right?
Click to expand...

I meant arm as a platform not ARM as a company or set of designs.

The platform is getting real resources thrown at production capability- enough to stay almost lock-step with Intel in the micron war thanks to companies like TSMC, GF and samsung.
 
manly said:
Aren't Android end user apps written in Java for the Dalvik runtime? They may not even need to be recompiled for x86.
Click to expand...
They are, that is why I am wondering how big of an effort x86 will require.

And for Krait, as far as I can tell this is how the numbering scheme works out as:
gcffS.jpg

The wiki page for snapdragon and Anandtech's charts don't *quite* line up, but thats the main info.
APQ80xxx: No baseband.
MSM82xxx: GSM+CDMA
MSM86xxx: HSPA+
MSM89xxx: HSPA+CDMA+LTE

Third/fourth/fifth(?) digits refer to clock speed and features. I think. I've never seen a decent explanation.
The unhighlighted cells are for part numbers without GPU info. The xx60A chips are the high end dual cores, with the 8660A being the only one without GPU info. There is also an MSM8270 part number that would imply a GSM/CDMA chip but the 70 isn't referenced anywhere else.
 
Last edited:
ilkhan said:
They are, that is why I am wondering how big of an effort x86 will require.
Click to expand...
I'm not an Android developer but Google would just need to expend the effort to port Android to x86 (this may already be largely done with the SDK simulator). It shouldn't affect an app developer much, in theory.
 
manly said:
I'm not an Android developer but Google would just need to expend the effort to port Android to x86 (this may already be largely done with the SDK simulator). It shouldn't affect an app developer much, in theory.
Click to expand...

err, doesn't a port already exist (mostly by intel)? the efforts need to be put in maintaining it..
 
poofyhairguy said:
Just the fact that ARM was designed for mobiles and x86 wasn't means that ANY x86 backwards compatibility requires a power envelope sacrifice.

Intel and x86 won't be competitive until it can get one or two size generations ahead of ARM, which seems impossible given the rate of ARM growth.
Click to expand...

poofyhairguy said:
I meant arm as a platform not ARM as a company or set of designs.

The platform is getting real resources thrown at production capability- enough to stay almost lock-step with Intel in the micron war thanks to companies like TSMC, GF and samsung.
Click to expand...
I wouldn't put much faith in Global Foundries if I were you.
Samsung alone can't fulfill the needs of the entire mobile sector.
TSMC has had misstep in certain process technologies but they are better than Global Foundries by a mile.

There's enough for everyone to go around, Intel included.
 
lothar said:
However, I believe Intel would be a force to be reckoned with at 22nm and beyond.
Nothing different from what I mentioned earlier in this thread.

Intel by far has the best experience in process technology.
Click to expand...

ARM has shown off 14nm testbeds at last ARMCON few months ago but I agree with your assessment, Intel is a force to be reckoned with indeed.
It took them quite a while but it seems like they opened the door to establish themselves in the ultra mobile market much in the same fashion as in the mobile and desktop markets.
 
i remember samsung annonced a 2ghz dual-core smartphone for early 2012 release (most likely the upcoming SGSIII)
http://mashable.com/2011/04/18/samsung-2ghz-dual-core-smartphone-2012/

so if it has 2ghz dual-core processor, why can't it be the exynos 5250 all u guys are talking about?

as far as i know it's also a 2ghz clocked dual-core SoC....also in that same link above, the samsung guy says ''This product will have the data processing capacities of a regular PC''

in the official press release for the new 5250 chip, it says:

"Designed specifically for high-end tablets"

but specifically doesn't translate to ''ONLY'', now does it?
All it means is that it was designed for tablets, not that it can't be used for smartphones...

there may be a chance samsung could put this in the SGSIII also, which i really hope.

what do you think?
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top