V revision with Amber Lake?


#1

So Intel’s Core-Y is getting an update with over 10% clock increase. To put it short, the new i5 will be 10% faster than the old i7. Interestingly, however Dell still includes the 7th gen on their XPS lineup linked below, so perhaps the 8th gen will come a little bit later.

Any chance we could have this on the V?


#2

Very doubtful, as far as I’m aware all the chips have been paid for, they’re just waiting to be assembled when they get financing in place.

No question a speed bump would be nice but I wouldn’t expect one on the V.


#3

The components are in the warehouse… we only need unfrozen fund to process the assemble… bla bla bla

By the time they fulfilled the flash sales orders, how many customers are willing to join the fund freezing saga? It depends on the product satisfaction of the V owners, which I am not optismistic on.

Edit… I mean where should they prioritize themselves? Allocating resources in after sales services (such as the loose ports, hinges screws, ghost touches… etc) OR R&D in another device (heat distribution, energy, ) while not fixing the previous issues


#4

I remember Eve saying, that in regard to Intel’s roadmap the next Y-processor upgrade (would that be amber?) Is not worth the trouble of redesigning.
The one after that seems to be a different animal though. However I have no idea, where amber falls on the roadmap…


#5

Iirc it was Cannon Lake. This Amber Lake was just announced like a month ago, as an emergency release due to delay with Intels 10nm


#7

The fabrication process didn’t change. There is absolutely nothing to expect from Amber Lake. If you get 10% more clock increase, you will get almost exactly the same performance as a KabyLake/SkyLake at the same clock speed.
It means only one thing: a higher power consumption.


#8

Not necessarily. Take overclocking (without overvolting) for example, you can run the CPU at higher clock rate than its rated for, if youre lucky.

Or undervolting, you can run the CPU at the exact same clock, but at lower voltage (and therefore lower power consumption).

Case on point: The base clock, which is limited at 4.5 W TDP ceiling, is yet still increased. That indicates higher clock/watt than before.

Also keep in mind that this is sort of what Intel has been doing in the tock cycle, in their tick-tock tradition. The difference is for the last 4 generations (since Skylake), it has been tick-tock-tock-tock.


#9

Yes, necessarily, wether you like it or not.

Undervolting is an other story.
TDP is not power consumption.


#10

You do realize the clock v. voltage curve were made up by Intel, right? There is no written physics law that a certain clock has to have a certain voltage.

If they put the voltage to be 1 V yesterday, and 0.9 V today (hence lower power consumption), then its technically not an undervolt, but its the same effect as if you take a 1V CPU and set it to 0.9 V.


#11

Which they don’t do.


#12

There is no other way they could maintain a higher base clock, which is power-limited at 4.5 W.


#13

Of course yes. By having a more reliable fabrication process which leads to less etching (microfabrication) failures, itself leading to more operational transistors in the die.


#14

You’re right, it’s another point of improvement that I didn’t touch.

However, that also means you can indeed gain performance without increasing power consumption


#15

Sorry, but still no.
Having and using more operational transistors in a die means you can reach a higher and still stable frequency, but it will of course use more power.
The only way to have a die using less power is to reduce the size of the transistors via the fabrication process (14-nanometer process to 10nm for example).
(or changing drastically the architecture that leads to a better ipc).
That’s why Skylake=Kabylake=Amberlake, or Zen=Zen+ at equal frequency.

Therefor we go back to my first intervention: “The fabrication process didn’t change. There is nothing to expect from Amber Lake”.


#16

That still doesnt explain why the base clock can now be higher (Amber vs Kaby, both 14 nm), despite being limited at 4.5 W


#17

Because 4,5W is the TDP, not the power consumption.


#18

Arent all computers essentially space heaters? -> the vast majority of the electricity is turned into heat


#19

One can say that, indeed :joy: