Name of Icelake successor leaks out

[NTMBK]

Nice scoop from Ashraf: http://www.fool.com/investing/gener...e-name-of-intels-third-10-nanometer-chip.aspx

 

[ShintaiDK]

Seems everything is named after lakes these days

Nice scoop.

 

[Dahak]

Seems everything is named after lakes these days

Nice scoop.

Not sure if its a coincidence or not but the "theme" for a lack of a better word seem to be related to the node that it should/would have been on

haswell, bradwell, Broadwell, all seem to be on 14-nm
Kaby Lake. Cannonlake, tigerlake all seem to be on 10nm

At least that is the feeling I get

 

[NTMBK]

Not sure if its a coincidence or not but the "theme" for a lack of a better word seem to be related to the node that it should/would have been on

haswell, bradwell, Broadwell, all seem to be on 14-nm
Kaby Lake. Cannonlake, tigerlake all seem to be on 10nm

At least that is the feeling I get

Skylake and Kabylake are 14nm, Haswell was 22nm.

 

[Arachnotronic]

Intel's CPU naming is wholly un-creative...enough with the lakes!

The next gen Atoms are also lakes...Apollo Lake, Mercury Lake, and Gemini Lake. So. Many. Lakes.

 

[Dahak]

Skylake and Kabylake are 14nm, Haswell was 22nm.

Hence the should/would have as the 14nm got delayed. i guess that theory went out the window for haswell tho

 

[The Stilt]

Intel's CPU naming is wholly un-creative...enough with the lakes!

The next gen Atoms are also lakes...Apollo Lake, Mercury Lake, and Gemini Lake. So. Many. Lakes.

IIRC Intel changed their naming policies to point places (usually in Oregon) after they got sued by Warner Bros (Batman / Batman´s Revenge) :biggrin:

 

[Techhog]

Hence the should/would have as the 14nm got delayed. i guess that theory went out the window for haswell tho

I believe that before it was based on architecture (Sandy/Ivy Bridge, Has/Broadwell, Sky/Kaby/Cannonlake), but Icelake threw that out of the window.

 

[CHADBOGA]

Sure I would rather know the name of this planned 3rd generation 10nm product, than not know it(coz why not?), but what difference does knowing this make?

Knowing the name won't give us any indication about its architectural features.

I'm surprised how these product codenames seem to have become a story that media outlets are keen to publish.

 

[BigDaveX]

When you consider that Intel had three 22nm desktop products (Ivy Bridge, Haswell and Devil's Canyon), and are apparently going to have three 14nm products (Broadwell, Skylake and Kabylake) and three 10nm products (Cannonlake, Icelake and Tigerlake), I think we can safely say that the classic "Tick-Tock" strategy from the Core 2 days is now a thing of the past.

 

[Arachnotronic]

Sure I would rather know the name of this planned 3rd generation 10nm product, than not know it(coz why not?), but what difference does knowing this make?

Knowing the name won't give us any indication about its architectural features.

I'm surprised how these product codenames seem to have become a story that media outlets are keen to publish.

The key is that it's a 3rd 10-nanometer product

 

[Dresdenboy]

Sure I would rather know the name of this planned 3rd generation 10nm product, than not know it(coz why not?), but what difference does knowing this make?

Knowing the name won't give us any indication about its architectural features.

I'm surprised how these product codenames seem to have become a story that media outlets are keen to publish.

Such codenames might help finding more information. For example try this in Google: myrtle site:ranker.sisoftware.net

 

[railven]

Sure I would rather know the name of this planned 3rd generation 10nm product, than not know it(coz why not?), but what difference does knowing this make?

Knowing the name won't give us any indication about its architectural features.

I'm surprised how these product codenames seem to have become a story that media outlets are keen to publish.

Going by the IceLake name, I can be assured it will not have TIM issues

 

[Fjodor2001]

I wish they could just compress all the *lake cores to a single one. No use upgrading from one *lake to another anyway. It's even borderline worth upgrading from *bridge to *lake, skipping *well.

 

[Arachnotronic]

I wish they could just compress all the *lake cores to a single one. No use upgrading from one *lake to another anyway. It's even borderline worth upgrading from *bridge to *lake, skipping *well.

So you want Intel to release one new processor every 3 years?

 

[ELopes580]

IIRC Intel changed their naming policies to point places (usually in Oregon) after they got sued by Warner Bros (Batman / Batman´s Revenge) :biggrin:

Actually, Intel has always (or mostly) named it after places in Oregon. Or Israel for those projects from there.

If my memory serves me, you are confusing it with nVidia. Had no idea they got sued about it.

 

[witeken]

The key is that it's a 3rd 10-nanometer product

Relax, it's just a back-up plan.

Just like they can do 7nm without EUV.

 

[Fjodor2001]

So you want Intel to release one new processor every 3 years?

Yes. Same as they used to do some decades ago.

The difference is that the jump from one generation to the next would still be greater back then compared to what it would be today with a release every 3 years. Back then we used to see more than doubling of performance in 3 years. Now we're not even close, perhaps ~10-15%.

But even better, I'd like to see some more progress, justifying a new release every year, or even every third year. Does not look like that is on the horizon though...

 

[ShintaiDK]

All that old performance came from frequency at the expense of power.

We would have 500W CPUs today then, lol.

165W on 14nm gives you a quad at 5.1Ghz.

 

[Fjodor2001]

All that old performance came from frequency.

No, IPC and frequency.

And so what by the way? Let them double the TDP today and see if they could achieve the same frequency jump from that as they did before. Take the Skylake 6700 (3.4/4 GHz @ 65 W TDP) as base. You think even a doubling of frequency to 6.8/8 Ghz Skylake at 130 W would be possible? Good luck with that. They won't even get close. Not even a suicide run on liquid nitrogen could save them.

 

[ShintaiDK]

I never claimed it would today.

Last big IPC jump was PPro.

Today everything is about performance/watt. Including GPUs. The one winning that wins the game.

 

[Fjodor2001]

Today everything is about performance/watt.

How much perf/watt have we seen on desktop? Regressing from 77W to 95W for the top end SKUs from IB to SKL. That while only seeing minuscule performance improvement.

 

[Arachnotronic]

Relax, it's just a back-up plan.

Just like they can do 7nm without EUV.

Hell of an expensive back-up plan.

 

[ShintaiDK]

How much perf/watt have we seen on desktop? Regressing from 77W to 95W for the top end SKUs from IB to SKL. That while only seeing minuscule performance improvement.

You ignore performance.

At 91W the 6700K is more than twice as fast as the 3770K at 77W.

In non AVX2/FMA3 loads the 6700K barely passes 45W. While the 3770K will sit at close to its TDP.

 

[NTMBK]

So you want Intel to release one new processor every 3 years?

It's worked out so well for AMD :awe: