Lenovo released a Cannon Lake laptop

[jpiniero]

I don't really understand what is going on with this 10nm product. What is the issue with the lack of GPU? Why can't they get the GPU to work and why is the chip as a whole using so much power to produce such a low score?

As I said, functional yields are horrific.

 

[PeterScott]

I really have to wonder what the point of releasing something like this is. This is no revenue generator, it will end up as a low end, low priced, negligible margin product, and I imagine low volume as well. There is no money to be made.

All it does is serve as a big glowing neon sign that points out that Intel 10nm is desperately bad.

No monetary gain, and putting a spotlight on process weakness, just seems like a brain dead move.

Intel needs to start CEO search, when the man at the top approves products like this, or Skylake-X.

 

[jpiniero]

I really have to wonder what the point of releasing something like this is. This is no revenue generator, it will end up as a low end, low priced, negligible margin product, and I imagine low volume as well. There is no money to be made.

BK told investors early in 2017 they would ship 10 nm in 2017. Which they did...

Essentially they are selling test chips. Unless they are doing the contra-revenue there's really no harm since the alternative is the trash can as they continue to try to fix it.

 

[not_someone]

Contra-revenue is the only way to sell this, especially given the existence of https://ark.intel.com/products/137977/Intel-Core-i3-8130U-Processor-4M-Cache-up-to-3_40-GHz

Note that no Cannon Lake part is on ARK, they might feel shame about this even.

 

[FIVR]

What really perplexes me about this whole situation is that if Intel 10nm really is this bad then why aren't they designing a new architecture on 14nm as a stop gap until they get things fixed?


It is clear their 14nm process produces extremely high clock speeds and can still compete strongly against AMD and ARM products. They still have their OEM connections and if they produced a new chip on 14nm with an increase in IPC people would happily buy the chips. Just look at this forum, you have people who will gladly spend $700+ for an 8 core skylake-X even if it performs worse than TR and costs more.


Intel would be fine if it just made something new on 14nm instead of trying to shove this terrible 10nm stuff down OEMs throats.

 

[jpiniero]

What really perplexes me about this whole situation is that if Intel 10nm really is this bad then why aren't they designing a new architecture on 14nm as a stop gap until they get things fixed?

They thought they could fix the yield problems.

 

[not_someone]

What really perplexes me about this whole situation is that if Intel 10nm really is this bad then why aren't they designing a new architecture on 14nm as a stop gap until they get things fixed?


It is clear their 14nm process produces extremely high clock speeds and can still compete strongly against AMD and ARM products. They still have their OEM connections and if they produced a new chip on 14nm with an increase in IPC people would happily buy the chips. Just look at this forum, you have people who will gladly spend $700+ for an 8 core skylake-X even if it performs worse than TR and costs more.


Intel would be fine if it just made something new on 14nm instead of trying to shove this terrible 10nm stuff down OEMs throats.

They are, it's called Whiskey Lake (& Cascade Lake for server), but well, it's a blatant rehash.
Also, they did change stuff in the process, but it's too late to backport cores (could have been decided in 2016 and Piednoel said that he personally advocated and failed to do it).

 

[ZGR]

When 14nm first came out, clocks were low as heck. Hopefully 10nm may look pretty good in 3 years? :O

 

[Mockingbird]

I really have to wonder what the point of releasing something like this is. This is no revenue generator, it will end up as a low end, low priced, negligible margin product, and I imagine low volume as well. There is no money to be made.

All it does is serve as a big glowing neon sign that points out that Intel 10nm is desperately bad.

No monetary gain, and putting a spotlight on process weakness, just seems like a brain dead move.

Intel needs to start CEO search, when the man at the top approves products like this, or Skylake-X.

It's literally cheaper to give them away for free than to get rid of them (think: environmental fee).

 

[eek2121]

I really have to wonder what the point of releasing something like this is. This is no revenue generator, it will end up as a low end, low priced, negligible margin product, and I imagine low volume as well. There is no money to be made.

All it does is serve as a big glowing neon sign that points out that Intel 10nm is desperately bad.

No monetary gain, and putting a spotlight on process weakness, just seems like a brain dead move.

Intel needs to start CEO search, when the man at the top approves products like this, or Skylake-X.

So they can say that they shipped 10nm. It also allows them to fix yield issues without wasting silicon.

 

[Eug]

The i3-8121U gets 269 in Cinebench multi-core.
For reference, my Core m3-7Y32 fanless MacBook gets 265, although I believe it is with TDP up.

In Geekbench 4, the Core m3-7Y32 MacBook gets about 7050/3700.
The i7-7Y75 MacBook gets about 8500/4500.
The i3-8121U is in between at about 7800/4100, but as a 10 nm 15 W part with no iGPU.

 

[ksec]

I really have to wonder what the point of releasing something like this is. This is no revenue generator, it will end up as a low end, low priced, negligible margin product, and I imagine low volume as well. There is no money to be made.

All it does is serve as a big glowing neon sign that points out that Intel 10nm is desperately bad.

No monetary gain, and putting a spotlight on process weakness, just seems like a brain dead move.

Intel needs to start CEO search, when the man at the top approves products like this, or Skylake-X.

They need to keep producing the chip in order to fine tune the performance, yield and power usage. Sort of like Apple doing with iPad SoC, they need some small batch of chips as pipe cleaner to iron out lots of issues. And Intel will like to recoup some of the cost. So if it was 50% yield but the chip is only worth $50 in today's market, ( Performance and Power worst then 14nm, and no GPU support, ) they will need to sell it at the $50 price.

This is will likely hurt margin, and hence where the Contra Revenue comes in.

You know purely in terms of Revenue and Profits it is hard to fire BK. The two most important part of their Business, DC and Mobile, and third Desktop. Are all breaking records numbers and not hurt by AMD at all ( Which is rather surprising as if the whole market is expanding )

Now Qualcomm is out of Server business, they can continue to milk the market. After all the current main cost of Servers are now Memory and NAND. And the cycle doesn't seems to have an end. The market climate should, in theory allow Intel to sell their Optane with decent margin and competitive with DC SSD. Another bonus for BK. MobileEye is doing fine, and although very very late, they are finally moving into GPGPU business.

So unless BK mess up something big, I think he is pretty safe.

Note: Personally I do want BK out, but right now rationally speaking there just isn't an excuse. I love Pat Gelsinger.

 

[LTC8K6]

Hopefully it's just a way for Intel to use the early chips instead of having to trash them. Sell them on the China market.

 

[jpiniero]

They need to keep producing the chip in order to fine tune the performance, yield and power usage. Sort of like Apple doing with iPad SoC, they need some small batch of chips as pipe cleaner to iron out lots of issues. And Intel will like to recoup some of the cost. So if it was 50% yield but the chip is only worth $50 in today's market, ( Performance and Power worst then 14nm, and no GPU support, ) they will need to sell it at the $50 price.

BK did make it sound like transistor performance was within spec, which should still be better than the original 14. Variance might be high though, and does seem like yields are probably worse than 50%, even with 2/3 of the chip disabled.

 

[ZGR]

Cannonlake keychains coming soon? /s

 

[dark zero]

I

Serious question, what products outside of games do you use that might actually come close to utilizing 4 cores?

I think Intel intended Y to be much more popular by marketing fanless laptops. 15W you need a fan.

Products?
Excel, banking apps, virtual machines, JAVA, Firefox and pretty much more and more apps.

 

[Eug]

Serious question, what products outside of games do you use that might actually come close to utilizing 4 cores?

Multitasking.

 

[Gideon]

Serious question, what products outside of games do you use that might actually come close to utilizing 4 cores?

Programming. Many IDEs can easily use more than 4 cores, also some compilers. Overall running any non-trivial development environments for work. Databases, multiple micro-services, etc. I currently have a 13" Macbook Pro (2 cores 4 threads) as a work laptop, and for most everyday use it's totally adequate, but it could really use more cores for those tasks.

 

[krumme]

So someone can get their bonus

 

[LTC8K6]

So when would these chips have been shipped to Lenovo?

 

[Dayman1225]

So when would these chips have been shipped to Lenovo?

Last year till now probably

 

[french toast]

We some power numbers... honestly it doesn't look good. (Compared to 14++).

 

[Mockingbird]

This is a transcript of an actual conversation between Brian Krzanich and Yang Yuanqing, CEO of Lenovo.

Yang: Sup, Brian! What’s going on!

Brian: I need a big favor.

Yang: What’s going on?

Brian: I just got fined for destroying the environment

Yang: What did you do?

Brian: Err.. throwing the Cannon Lake chips into the landfill

Yang: It’s a good thing that, here in China, we don’t care about the environment

Brian: Precisely! Which is why I sent them to your office. You can do whatever you want with that: make incredible key chains, the possibilities are limitless!

Yang: You owe me big this time!

 

[dark zero]

This is a transcript of an actual conversation between Brian Krzanich and Yang Yuanqing, CEO of Lenovo.

Yang: Sup, Brian! What’s going on!

Brian: I need a big favor.

Yang: What’s going on?

Brian: I just got fined for destroying the environment

Yang: What did you do?

Brian: Err.. throwing the Cannon Lake chips into the landfill

Yang: It’s a good thing that, here in China, we don’t care about the environment

Brian: Precisely! Which is why I sent them to your office. You can do whatever you want with that: make incredible key chains, the possibilities are limitless!

Yang: You owe me big this time!

And the results: The first 10 nm laptops.

 

[LTC8K6]

And the results: The first 10 nm laptops.

Intel actually showed the first working 10nm laptop a very long time ago, January 2017 at CES.