The latest news about RK3288

[no_spam_for_me]

... I'm pretty much accepting the boxes I have are Cortex-A12 and I'm not all that displeased, they work well and that's all that matters in the end ...

As I wrote

I can live with an A12 but I don't like to be fu**ed...

.

...The A80 development board's new price: http://www.aliexpress.com/store/prod...015735541.html
Shipping is really expensive I think. ...

US $87.71 to Germany via Fedex IE

[Javimetal]

While it's slightly disappointing to hear that the (1) Rk3288 is only an a12, its really the only game in town right now. I was interested in the a80 chip, but the only board taking orders (Optimus board) (2) was selling for $345! That's insane.

I'm waiting a Rockchip reply yet...


The A80 development board's new price: http://www.aliexpress.com/store/prod...015735541.html

Shipping is really expensive I think. I don't know what price I can get here in China... if have time we will ask Merrii.

(Edit): Just asked Merrii about the A80 board, price is same in China.

[Shomari]

If you e.g. mean cpu-z v1.08, I think it doesn't work correctly, because it also says it's an rk3066... and 40-nm is ARM Cortex-A5 (<=1GHz)
(45-nm is ARM Cortex-A9 (0.8 -2GHz))

But it says 'ARM 0xc0d' which seams to be an A12 ("The CPU architecture is 0xc0d, which stands for ARM Cortex A12, whereas Cortex A17 should be 0xc0e" http://www.cnx-software.com/2014/08/...ndroid-tv-box/) "seams to be" because the blobdiff isn't longer available at the kernel.git but you can also find it e.g. at http://gcc.gnu.org/ml/gcc-patches/2013-12/msg01390.html

Here you will find a blobdiff including the define for the CORTEX_A17 https://git.linaro.org/arm/ds5/gator...driver/gator.h

Have you other useful tools?

Really? Mine reads RK3288 in CPU-z. I saw the same in other CPU info apps, in fact all of them. Same with the part codes and drivers. I'm pretty much accepting the boxes I have are Cortex-A12 and I'm not all that displeased, they work well and that's all that matters in the end ... tried like heck to get one of the manufacturers to revise their product listing but they say if Rockchip lists it as A17 they can't list it as A12 ... hilarious.

[no_spam_for_me]

Well, from all the (admittedly questionable) CPU info apps I've tried.

Even as I posted I was thinking it doesn't seem possible that it can be built using 40nm. No way the processors would run at the speeds they're running at current voltage levels.

edit: ...also, why would they go from 28nm to 40nm after releasing the RK3188 on 28nm. Doesn't make sense but who knows.

If you e.g. mean cpu-z v1.08, I think it doesn't work correctly, because it also says it's an rk3066... and 40-nm is ARM Cortex-A5 (<=1GHz)
(45-nm is ARM Cortex-A9 (0.8 -2GHz))

But it says 'ARM 0xc0d' which seams to be an A12 ("The CPU architecture is 0xc0d, which stands for ARM Cortex A12, whereas Cortex A17 should be 0xc0e" http://www.cnx-software.com/2014/08/...ndroid-tv-box/) "seams to be" because the blobdiff isn't longer available at the kernel.git but you can also find it e.g. at http://gcc.gnu.org/ml/gcc-patches/2013-12/msg01390.html

Here you will find a blobdiff including the define for the CORTEX_A17 https://git.linaro.org/arm/ds5/gator...driver/gator.h

Have you other useful tools?

[no_spam_for_me]

From Arm's A17 webpage performance tab:

	Example Topologies: 1.5GHz – 2GHz Dual-core Cortex-A17 processor and Dual-core Cortex-A7 big.LITTLE 1.0 – 2.0 GHz Quad, Dual and Single-core Cortex-A17 solutions

...

THX, I missed this...

[Shomari]

Re: The latest news about RK3288

I would tend to concur.

When we see mass shipping of 4288 devices and the media techs get hold of them for full reviews, it will be interesting to see how many ask the same questions.

...probably none. So far maybe only one reviewer has even broached the topic.

[bluesmanuk]

I think it's more likely to be #4, loosely applying the term 'fast one' though that is what it seems to be after all is considered.

I would tend to concur.

When we see mass shipping of 3288 devices and the media techs get hold of them for full reviews, it will be interesting to see how many ask the same questions.

[Shomari]

Re: The latest news about RK3288

The fact that Arm licence their own technology to companies like Rockchip and they must be well aware of how things are being advertised by RK and all the vendors promoting the 3288, I can only see four possibilities.

1. Both companies are colluding to mislead, which I can't really believe would be true.

2. The naming conventions are not strict enough to allow every combination to have some kind of unique identifier, which would be down to Arm to resolve for the benefit and clarity of both vendors and buyers.

3. The 3288 has potentially had further development since it's original marketing that gives it credibility.

4. Rockchip are pulling a fast one and hoping that Arm aren't that interested enough to make them change the marketing. After what they did with the secret release of the 3188T, I would not put it past them.


With Intel partnering up with Rockchip, the relationship between them and Arm should become interesting.

I think it's more likely to be #4, loosely applying the term 'fast one' though that is what it seems to be after all is considered.

[bluesmanuk]

The fact that Arm licence their own technology to companies like Rockchip and they must be well aware of how things are being advertised by RK and all the vendors promoting the 3288, I can only see four possibilities.

1. Both companies are colluding to mislead, which I can't really believe would be true.

2. The naming conventions are not strict enough to allow every combination to have some kind of unique identifier, which would be down to Arm to resolve for the benefit and clarity of both vendors and buyers.

3. The 3288 has potentially had further development since it's original marketing that gives it credibility.

4. Rockchip are pulling a fast one and hoping that Arm aren't that interested enough to make them change the marketing. After what they did with the secret release of the 3188T, I would not put it past them.


With Intel partnering up with Rockchip, the relationship between them and Arm should become interesting.

[Haba]

He said it is A17, just wants to ask ARM about why that "ARMv7 Cortex-A12" is there.

I did not pick up on this right away that we also have a Rockchip rep state that the RK3288 is an A17 core along with all the documentation. Cpu detection apps may be out of date and the bits they are reading off the chips may have not been updated properly since Rockchip got into trouble promoting the A17 before they were legally allowed. The drivers may be nearly the same since the cores are nearly the same. I would think Rockchip would get into legal issues if they are making false claims. Why would they risk legal issues? Arm/Rockchip should explain themselves for everyone!

[Shomari]

From Arm's A17 webpage performance tab:

	Example Topologies: 1.5GHz – 2GHz Dual-core Cortex-A17 processor and Dual-core Cortex-A7 big.LITTLE 1.0 – 2.0 GHz Quad, Dual and Single-core Cortex-A17 solutions

Everywhere it shows RK3288 is a 28nm chip. Shomari where did you get the info that the RK3288 is built on 40nm process?

Well, from all the (admittedly questionable) CPU info apps I've tried.

Even as I posted I was thinking it doesn't seem possible that it can be built using 40nm. No way the processors would run at the speeds they're running at current voltage levels.

edit: ...also, why would they go from 28nm to 40nm after releasing the RK3188 on 28nm. Doesn't make sense but who knows.

[Haba]

OK, but if you read ARM-page it looks like a MUST to have at the A17...

From Arm's A17 webpage performance tab:

	Example Topologies: 1.5GHz – 2GHz Dual-core Cortex-A17 processor and Dual-core Cortex-A7 big.LITTLE 1.0 – 2.0 GHz Quad, Dual and Single-core Cortex-A17 solutions

Everywhere it shows RK3288 is a 28nm chip. Shomari where did you get the info that the RK3288 is built on 40nm process?

[Shomari]

According to somebody over at the Arm community site, it seems that on the Cortex-A17 it is quite possible to use it on its own or in a big.LITTLE system, so it may possibly a marketing thing that has no real differentiation.

...yup.

[bluesmanuk]

According to somebody over at the Arm community site, it seems that on the Cortex-A17 it is quite possible to use it on its own or in a big.LITTLE system, so it may possibly a marketing thing that has no real differentiation.

[no_spam_for_me]

bigLITTLE is an option for A12-17 and is not required. ...

OK, but if you read ARM-page it looks like a MUST to have at the A17...