Lenovo released a Cannon Lake laptop

[jpiniero]

Most software does. Unless you're a content creator almost everyone is exclusively using a browser. Chrome is the #1 browser and it saturates 8 threads easy. Clock an 8 core to it's lowest and check CPU utilization.

Chrome's display renderer is multithreaded but that's only a part of the process. Javascript is single threaded. So you will get some parallelism but nothing of the sort where 8 threads (or even 4 really) would be fully loaded.

Now if Windows 10's spyware hogs up a core, well that could be a problem.

And it looks like CNL Y STILL exists
https://twitter.com/TUM_APISAK/status/996706570552524800?s=19

Ooh, interesting, although with no GPU that would be kind of silly.

 

[Dayman1225]

Ooh, interesting, although with no GPU that would be kind of silly

Is it possible that CNL Y is the 2+2 part that was mentioned in the Intel documents?(assuning this one is active) otherwise it wouldn't make sense to push out

 

[Eug]

And it looks like CNL Y STILL exists

https://twitter.com/TUM_APISAK/status/996706570552524800?s=19

Ooh, interesting, although with no GPU that would be kind of silly.

Is it possible that CNL Y is the 2+2 part that was mentioned in the Intel documents?(assuning this one is active) otherwise it wouldn't make sense to push out

Where in that post is GPU mentioned?

Interesting clock speed and interesting naming. Too bad the boost speed isn't mentioned, but a 1.5 GHz base clock seems like a big jump. TDP up?

It would seem again to be a low volume part, but if so, who would use it?

 

[Dayman1225]

Where in that post is GPU mentioned?

I was referring to Official Intel documents for Meltdown and Spectre mitigations mentioned a 2+2 CNL (which this could be because you don't put a dGPU onto a fanless laptop most of the time).

Interesting clock speed and interesting naming. It would seem again to be a low volume part the

I have no idea who would use it. Maybe one or two OEMs.

 

[Eug]

I was referring to Official Intel documents for Meltdown and Spectre mitigations mentioned a 2+2 CNL (which this could be because you don't put a dGPU onto a fanless laptop most of the time).

Yes I realize that, but I was wondering why jpiniero was saying it had no iGPU.

It makes me think that this m3-8114Y part is the main part, and the i3-8121U is the result of binning with the iGPU deactivated and at higher power.

But for most of us, it's probably still mostly irrelevant since it seems to be a low volume part.

 

[jpiniero]

Yes I realize that, but I was wondering why jpiniero was saying it had no iGPU.

Maybe I was a bit hasty, since the 8121U doesn't have the GPU and yields are obvs terrible. I guess the question would be what Intel would do with the very small % of dies that actually get through with the GPU usable, even if partially disabled. I kind of assume that even now that % is close to zero though.

A 1.5 Ghz base is still better than what Apple is offering with the Retina Macbook for instance, although the turbo would have to be there as well.

 

[Eug]

A 1.5 Ghz base is still better than what Apple is offering with the Retina Macbook for instance, although the turbo would have to be there as well.

That’s why I was wondering if that might be TDP up.

It’s clearly listed as an m3 too, not an i7.

The optimist could claim it was evidence of a significant clock speed boost but I’m not optimistic enough to think that, esp. for a real volume part.

 

[FIVR]

That’s why I was wondering if that might be TDP up.

It’s clearly listed as an m3 too, not an i7.

The optimist could claim it was evidence of a significant clock speed boost but I’m not optimistic enough to think that, esp. for a real volume part.

There is no way apple would accept this junk. You can basically write off any new MacBooks in 2018 unless Apple switches to AMD (unlikely).

 

[not_someone]

Maybe I was a bit hasty, since the 8121U doesn't have the GPU and yields are obvs terrible. I guess the question would be what Intel would do with the very small % of dies that actually get through with the GPU usable, even if partially disabled. I kind of assume that even now that % is close to zero though.

A 1.5 Ghz base is still better than what Apple is offering with the Retina Macbook for instance, although the turbo would have to be there as well.

The turbo is 100MHz higher than base or so...
EDIT: turbo doesn't seem to get read properly by that toolset

 

[Eug]

There is no way apple would accept this junk. You can basically write off any new MacBooks in 2018 unless Apple switches to AMD (unlikely).

They’d be more likely to add A12. It benches way faster than Intel Y series and i3-8121U.

I personally probably wouldn’t buy one though in the near term, for compatibility reasons.

 

[Hitman928]

From the picture posted on twitter, it looks like it has a boost clock of 2.2 GHz, but who knows if that is the actual max boost or not.

 

[jpiniero]

They’d be more likely to add A12. It benches way faster than Intel Y series and i3-8131U.

Too early for that. They would have to announce that they were moving to ARM first, give developers time to get their apps ported. But in the long term, yes, that is where they are going.

 

[Eug]

Too early for that. They would have to announce that they were moving to ARM first, give developers time to get their apps ported. But in the long term, yes, that is where they are going.

That’s why I said “add”, not “switch”.

They could add it with no pre-announcement. It would happen at WWDC along with 10.14.

 

[Eug]

From the picture posted on twitter, it looks like it has a boost clock of 2.2 GHz, but who knows if that is the actual max boost or not.

No the boost clock listed is 1510 MHz, which is 10 MHz higher than base. So effectively the same and clearly wrong.

 

[Hitman928]

No the boost clock listed is 1510 MHz, which is 10 MHz higher than base. So effectively the same and clearly wrong.

Ok, yeah, I see it now. I read it wrong and didn't realize there were two CPUs side by side.

 

[LTC8K6]

KL-Y M3 chips turbo-ed up to 3.0, so would we expect better than that from the CL-Y M3 chips?

Or have they possibly gone with a higher base clock and no turbo?

 

[Eug]

KL-Y M3 chips turbo-ed up to 3.0, so would we expect better than that from the CL-Y M3 chips?

i3-8121U turbos up to 3.2 GHz, which is the same as the KBL-Y i5-7Y54. So it would seem that something like 3.0-3.4 GHz would be a reasonable target for a CNL-Y m3.

Or have they possibly gone with a higher base clock and no turbo?

Nah, that wouldn't make sense, and it would suck.

It's just the software not reading things correctly. That twitter post says the turbo clock for i3-8121U is 2208 GHz, but as mentioned, it's 3.2 GHz.

 

[LTC8K6]

I gotta' remember that CL could be confused as Coffee Lake, and use CNL for Cannon Lake.

 

[LTC8K6]

https://ark.intel.com/compare/137977,136863

With the 8121u vs 8130U, we get a lower boost clock, more PCIE lanes, more memory bandwidth, LPDDR4/X support, and no igp.
Package size is slightly larger, and the temp is 5C higher.

 

[dullard]

https://ark.intel.com/compare/137977,136863

With the 8121u vs 8130U, we get a lower boost clock, more PCIE lanes, more memory bandwidth, LPDDR4/X support, and no igp.
Package size is slightly larger, and the temp is 5C higher.

That seems like 10 nm (not 10 nm+) since we know from Intel's own marketing documents that Intel's use of 10 nm was lower performance than 14 nm++. Hence the lower turbo and higher temperatures.
https://images.anandtech.com/doci/11722/kaizad-mistry-2017-manufacturing(1)_29.png

 

[LTC8K6]

Why is the max memory bandwidth higher for 8121U? Both support DDR4-2400.

EDIT: Just an ARK error.

 

[jpiniero]

That seems like 10 nm (not 10 nm+) since we know from Intel's own marketing documents that Intel's use of 10 nm was lower performance than 14 nm++. Hence the lower turbo and higher temperatures.
https://images.anandtech.com/doci/11722/kaizad-mistry-2017-manufacturing(1)_29.png

Intel's definition of transistor performance is a bit murky. After all, according to the graph regular 10 is still higher than Kaby's 14+. It's obviously not going to hit the max clock speeds that 14+ or 14++ can but if yields were not so horrific they could certainly offer products which would sell.

The higher temps and bigger package size is probably due to the FIVR.

 

[Bouowmx]

So, on Wikipedia, a user added that the i3-8121U has 3.1 GHz all-core frequency. Is there an independent source to verify?

 

[LTC8K6]

So, on Wikipedia, a user added that the i3-8121U has 3.1 GHz all-core frequency. Is there an independent source to verify?

Intel doesn't publish that info anymore, so you'd have to test the chip, I guess. Published info is 3.2 single core.
It's only two cores, so how much does it matter?

 

[Dayman1225]

Intel Confirms that the 8121U supports AVX512, updated the ARK Page

https://ark.intel.com/products/136863/Intel-Core-i3-8121U-Processor-4M-Cache-up-to-3_20-GHz