I'll just leave this here. ;)

imo intel should release one celebratory "10K" cpu, which is a 7700k / 8700k / 9900k but with 10 cores 20 threads.
then for finally something ACTUALLY new start a new naming scheme.

ZoneDymoimo intel should release one celebratory "10K" cpu, which is a 7700k / 8700k / 9900k but with 10 cores 20 threads.
then for finally something ACTUALLY new start a new naming scheme.

no need for mainstream intel parts to go 10c.9900k is already overkill.They should stick to the ring bus design instead of using mesh or adding more dies.
Instead, 9900k should be an i7 and priced under $400 with 4.7GHz non-K variant available at ~$320.

Four clock states and the one these CPUs will actually run at most of the time, is 1.something. In other words slow as molasses.

But, at least you have a fancy number.

At least Intel is ensuring their naming scheme isn't confusing consumers, if it isn't & still using 4-digit numbers, it'll sound like Nvidia's Pascal GPU lineup.

Vayra86Four clock states and the one these CPUs will actually run at most of the time, is 1.something. In other words slow as molasses.

Flexibility of these -U CPUs has been one of Intel's biggest strengths. They're frugal in idle and very close to desktop counterparts under load. AMD wasn't able to match this yet.
Of course they're able to hold that boost for short periods of time, but in many use cases that's all you need.

Vayra86Four clock states and the one these CPUs will actually run at most of the time, is 1.something. In other words slow as molasses.

But, at least you have a fancy number.

These are all mobile parts, they normally have very low base clocks, so I don't really see the issue here.

Vayra86Four clock states and the one these CPUs will actually run at most of the time, is 1.something. In other words slow as molasses.

But, at least you have a fancy number.

When was the last time you used a skylake\cfl\kbl based CPU with these rated speeds?
In a proper environment, they usually clock very high.

But what am i saying, why would you miss an opportunity to jab another one into Intel

TheLostSwedeThese are all mobile parts, they normally have very low base clocks, so I don't really see the issue here.

My experience with that is that for anything more than light use these CPUs will not be very quick. I agree with you and @notb - its good for some use cases. But the impression you get with these CPUs is that they are 'lean', but fact of the matter is, for any prolonged, high load, they fall apart fast. Prolonged being anything more than a few minutes on end.

Good for business laptops and light office work / cloud based stuff. Not so good for, for example, gaming, any crunching workloads, and most other things that ask more than opening some pages and files. Anything you do in short bursts will probably work fine.

If I compare these specs to previous gen, such as the i5 8350U I have in front of me, the trend is clear: loss of base clock frequency to accomodate higher boost frequencies. I'm not so sure that is a positive trend. Already this CPU is erratic in its power and temperature behaviour. I have serious doubts if the overall performance is actually improving...

dj-electricWhen was the last time you used a skylake\cfl\kbl based CPU with these rated speeds?

Today and for the last couple of years. I also have a Skylake HQ-quadcore for a personal laptop and its a world of difference. More consistent, less erratic, much better performance. Of course it also guzzles more power.

dj-electricBut what am i saying, why would you miss an opportunity to jab another one into Intel

Yeah I'm totally anti-Intel indeed :laugh:

Intel been there done that.( 80286 80386 80486)
then they name changed as they could not copyright/patent numbers
So Intel gave us the Pentium/celeron and so on

So,
19980xe or 10980xe ?
Both sounds stupid.

dorsetknobIntel been there done that.( 80286 80386 80486)
then they name changed as they could not copyright/patent numbers
So Intel gave us the Pentium/celeron and so on

You forgot 8086 :)

They should simplify and focus what people know.

Core i9 X90
Core i9 X80
Core i9 K90
Core i7 K70
Core i5 K60
Core i5 S40

Next generation,

Core i9 X290
Core i7 K270
Etc..

I9, i7, i5 = class
X/K/S/U = features
First number = generation
Last two digits = specific SKU

kastriotYou forgot 8086 :)

It didn't sound stupid! It's commemorative CPU and 8086K makes sense.

Vayra86My experience with that is that for anything more than light use these CPUs will not be very quick. I agree with you and @notb - its good for some use cases. But the impression you get with these CPUs is that they are 'lean', but fact of the matter is, for any prolonged, high load, they fall apart fast. Prolonged being anything more than a few minutes on end.

Most people don't apply "prolonged, high load" to CPUs on regular basis.
This is the compromise we make for mobile parts.

Frugal operation and high, rapid boosts are perfect for mainstream PC loads: browsing www, office suites, programming, analytics.

Gaming aside, what long, heavy tasks are normal people running?
The only thing I could think of is watching video. But that's why Intel invests so much in hardware decoders. :)

Good for business laptops. Not so good for, for example, gaming.

Good for business laptops, good for personal laptops, good for NUCs/fast embedded. I.e. good for many things it is advertised for.
Not good for gaming, not good for long-running computing, not good for servers. I.e. not good for many things it is not advertised for.

MassmanI9, i7, i5 = class
X/K/S/U = features
First number = generation
Last two digits = specific SKU

Looks bad and way too few combinations.
Intel utilizes 3 digits for CPU versions and adds a letter for variant. All of that has to stay. Generation gets a second digit, that's it.

notbMost people don't apply "prolonged, high load" to CPUs on regular basis.
This is the compromise we make for mobile parts.

Frugal operation and high, rapid boosts are perfect for mainstream PC loads: browsing www, office suites, programming, analytics.

Gaming aside, what long, heavy tasks are normal people running?
The only thing I could think of is watching video. But that's why Intel invests so much in hardware decoders. :)

Good for business laptops, good for personal laptops, good for NUCs/fast embedded. I.e. good for many things it is advertised for.
Not good for gaming, not good for long-running computing, not good for servers. I.e. not good for many things it is not advertised for.

Looks bad and way too little combinations.
Intel utilizes 3 digits for CPU versions and adds a letter for variant. All of that has to stay. Generation gets a second digit, that's it.

I've already said I agree on that, didn't I? :nutkick:

The key part of my post you omitted. These CPUs, especially for someone less in the know, give us the impression they run at super high clocks. In reality, though, these CPUs have been steadily losing base clock frequency with every new iteration. I ask the legitimate question of whether that is positive. Do you really need 4.5 Ghz and up for those basic office tasks?! Of course not. But you do need a high sustained base clock for some other tasks, gaming being one of them.

Bottom line, the super high boosts actually reduce the versatility of this type of CPU. Who cares if it can do burst workload half a second faster, especially in the segment they're meant for?

This is Intel applying a strategy to its product portfolio that isn't necessarily in our best interests. Lower baseclock U-parts, enable a selling point for higher baseclock performance parts. Do we really need that distinction just so we can have a U-stack that is closing on 5 Ghz?

kastriotYou forgot 8086 :)

*8088 :p

kastriotYou forgot 8086 :)

:) not 5 digit :)
and not forgotten ( i once had a 8086 laptop).

notbFlexibility of these -U CPUs has been one of Intel's biggest strengths. They're frugal in idle and very close to desktop counterparts under load. AMD wasn't able to match this yet.

Of course they're able to hold that boost for short periods of time, but in many use cases that's all you need.

notbMost people don't apply "prolonged, high load" to CPUs on regular basis.

About 5 years ago, you could buy an "ultrabook" and it would be just fine with surfing the web. Try the same thing today, with a XPS 13, Macbook Air, etc., and it will struggle quite a bit, while getting hot and noisy. This is of course not the computer's fault, but bloated web pages full of JavaScript.
"Close to desktop computers" would be a stretch, unless you mean like a low-end one, and most laptops can only do it in short bursts, and usually throttle more as they get hotter. Many laptops are also not that useful when not plugged in, which kind of defeats the purpose of a laptop in the first place.

Too bad most small laptops these days are flimsy and super-thin. :(

Vayra86The key part of my post you omitted. These CPUs, especially for someone less in the know, give us the impression they run at super high clocks. In reality, though, these CPUs have been steadily losing base clock frequency with every new iteration. I ask the legitimate question of whether that is positive. Do you really need 4.5 Ghz and up for those basic office tasks?! Of course not. But you do need a high sustained base clock for some other tasks, gaming being one of them.

Bottom line, the super high boosts actually reduce the versatility of this type of CPU. Who cares if it can do burst workload half a second faster, especially in the segment they're meant for?

You raise some valid and very relevant concerns.
While CPUs (even desktop ones) do run at lower power states most of the time, the "base clock" represents the highest clock it should be able to sustain with all features enabled, and unfortunately, it seems like the increase in core count comes at a cost of base clock.

The way I view it is that too high variance between sustained performance and peak performance is bad, and it negatively impacts the user experience more than reflected in benchmarks.

almost as easy to understand as USB Gen 4.3 rev 2a subsection xy or w/e it is going to be called in the future

14nm and same architecture. Nothing to see here..

ICL isn't supposed to be 14nm though that naming scheme & number of SKU o_O

5 digit model numbers to go nicely with their 5 digit pricing?

sutyiI'll just leave this here. ;)

Unbelievable, either these models have no cache, or turn off all other cores and their cache to reach boost speeds over 4Ghz, perhaps believable if they were talking about 3nm production that's mature.

so how long can they sustain a 4+GHZ clock at 15W? .5 sec?

What does it pull when turbo is engaged on all cores?