Intel Skylake / Kaby Lake

[Ajay]

http://wccftech.com/intel-kaby-lake-x-core-i7-7740k-gigabyte-x299-gaming-3/

a Cpu with 4 threads on HEDT ? Is it True?

I hope not, talk about a waste.

 

[formulav8]

a Cpu with 4 threads on HEDT ? Is it True?

Ouch, what a fail Intel, pathetic if true. Another site mentioned similar as well it seems for Kaby Lake-X (Kaby Lake X only 4 core?). Only dual channel available on the 4core that is supposed to be high end? Not sure why Intel would do this aside from a knee jerk to Ryzen catching them completely off guard. I have to be reading things wrong or the news isn't completely true or something.

Intel X299 platform launching on May 30th

 

[jpiniero]

http://wccftech.com/intel-kaby-lake-x-core-i7-7740k-gigabyte-x299-gaming-3/

a Cpu with 4 threads on HEDT ? Is it True?

Maybe if you had a legit source and not WTFTech....

 

[Shivansps]

I still dont get it, i dont see any avantage of going X299+KBL-X vs mainstream. Maybe they are targeting OEMs? It will be easier to offer HEDT options if there is a cheap one.

 

[.vodka]

Maybe if you had a legit source and not WTFTech....

CPC had this up a while ago (February 8, 2017) about early KBL-X samples. It's a nice read with an excellent explanation on how these are just mainstream Kabylake dies slapped on LGA2066 packaging.

https://www.cpchardware.com/core-i7-7740k-et-i5-7640k-kaby-lake-x-lga2066/

Google Translate'd

Quoted conclusion:

However, other information has been received, which complements those published by our colleagues.

• Core i7 7740K (QMM9 ES): Its basic frequency is at 4.30 GHz (100 MHz more than the 7700K) and its Turbo frequency at 4.50 GHz (identical to the 7700K). Hyper-Threading is enabled and has 8 MB of L3 cache. There is, however, a small subtlety. The Core i7 7700K is specified at 4.4 GHz in Turbo mode with all active cores and 4.5 GHz with a single active core. The Core i7 7740K is designed to operate at 4.5 GHz in Turbo mode, regardless of the number of active cores. This makes it, in practice, a CPU at 4.50 GHz.
• Core i5 7640K (QMMA ES): Much more disappointing than expected, the Core i5 7640K is well clocked at 4.00 GHz base against 3.80 GHz for the Core i5 7600K on Socket LGA1151. A priori it is a gain, except that here is its mode Turbo is 4.2 GHz for a single active core, 4.1 GHz with two active cores and ... 4.0 GHz - its basic frequency - with 4 active cores ! Values strictly identical to the Core i5 7600K and which come back to the removal of Turbo mode in Quad Core mode. This is enough to annihilate the interest of the increase in the base frequency. Regarding the presence of Hyper-Threading, Intel's internal documents indicate that it is disabled, but we still have no confirmation of a source that actually has a functional CPU. As soon as our sample has found a compatible motherboard (which will not be long), we will confirm this point. Nevertheless, given the marketing abuse that seems to constitute this range of processors, we now have little hope.

It remains with these CPUs that Intel intends to compete with Ryzen. The acceleration of the timing - they are now in the final validation stage - demonstrates this. Without even pronouncing on their performance, these Kaby Lake-X resemble for the moment a joke of bad taste. Intel is obviously seeking (yet) to create novelty by all means to mask the lack of evolution in its architectures. These practices make us think of the detestable waltz of labels in the world of GPU. Should it be interpreted as a sign of excitement due to unexpected competition? Clearly !

Take a 7700k, set manual turbo to 4.5GHz on all cores and you've got a 7740k... unless things changed at the die level in a matter of two months... which is impossible. So, KBL-X is.. pointless. Yeah. I don't know what was the motivation to do this apart from unifying sockets like AMD has done with AM4... I reckon that big LGA2066 socket and the PCB to support that won't be cheap..

Kabylake-X is more or less a known quantity at this point, Skylake-X is what remains to be seen and is actually the exciting part for the LGA2066 platform. If I'm not mistaken, LGA1151 gets 6C12T Coffeelake as the new flagship and it should actually be a great CPU considering 14nm++, probably an easy 5GHz on 6C12T if you're able to keep that cool enough. I suspect delidding+CLU will do the trick as it does on the 7700k. High speed DDR4 should be able to feed the beast on dual channel, seeing how 7700k benefits from that.

I still dont get it, i dont see any avantage of going X299+KBL-X vs mainstream. Maybe they are targeting OEMs? It will be easier to offer HEDT options if there is a cheap one.

There is none. It's a quick money grab at average joe's expense who sees bigger numbers in 7740k and 2066 over 7700k and 1151. That's it as I see it. At least Intel should have the decency to solder these parts as they do on HEDT CPUs... at least that.


Watered down mainstream Skylake/Kabylake has been a pretty decent step for Intel when paired with high speed DDR4, I can't wait to see what full fat true Skylake can actually do, that 1MB L2 should do wonders for some stuff out there.

 

[IEC]

Yeah, KL-X is pointless for me. Skylake-X on the other hand could be interesting depending on the performance.

 

[formulav8]

Yeah, KL-X is pointless for me

If true, its somewhat pointless for most everyone, in itself, on its own standing. Just IMO of course.

 

[beginner99]

I still dont get it, i dont see any avantage of going X299+KBL-X vs mainstream. Maybe they are targeting OEMs? It will be easier to offer HEDT options if there is a cheap one.

OEMs make sense. However we should not forget that Coffelake was planed for Q12018. In that case Kaby-X made a lot more sense. Lack money? buy the kaby-X first upgrade to a 8+ core later when you have the funds without needing to replace the whole platform. Last possible reason is that these are soldered and not TIM. So you save on the deliding work and risk to get it ti 5 ghz+. But yeah all in all pointless and a tiny niche market.

 

[nvgpu]

http://nucblog.net/2017/04/kaby-lake-i7-nuc-review/

 

[IndyColtsFan]

Yeah, KL-X is pointless for me. Skylake-X on the other hand could be interesting depending on the performance.

And price as well. I'm expecting SKL-X to perform very well, but if Intel keeps up the pricing shenanigans it started with Broadwell-E, SKL-X is a non-starter for me and I'll look to Coffee Lake. IMO, Intel should drop their current KBL-X plans, drop SKL-X's 6800k equivalent into that slot, eliminate the 6850k model, and then move the new models into the hierarchy with price reductions. A $1000 octocore isn't going to do it here - they need to drop that down in the $600 range and the 10 and 12 core models need to come down in pricing as well.

 

[jwapk]

Is there any indication whether or not Intel could be officially supporting higher memory speeds with the new chips? This would basically give them a free performance-plus unless they want to continue to artificially limit their platform…

 

[csbin]

Intel to brief press on Sandy-E/X on May 2

http://semiaccurate.com/2017/04/24/intel-brief-press-sandy-ex-may-2/


Intel has another top-secret gathering soon for briefings on Sandy-E/X, Kaby-X and X-299. If you weren’t already invited, you might want to ask Intel why they stiffed you, tell them SemiAccurate says hi.

In just over a week, Intel will be flying hordes of press to San Fransisco for a briefing on the upcoming 6-12 core Basin Falls platform. You might know this as Sandy Bridge-E or the 1-socket Purley minus a lot of features. They are also likely to try and convince the gathered masses that the Kaby Lake version that fits into the same socket, minus half of the memory channels, many PCIe lanes, but with a massive price increase is not a bad joke. Feel free to differ with their opinions here but recall these are the same clowns who claimed you needed a 10-core Broadwell-E for video encoding while gaming.

So the short story is that if you haven’t been invited yet, Intel doesn’t think you are worth bothering with any more. If this story is like the last one, expect a hastily noticed conference call email next week with excuses about it being planned all along. Please pretend to believe them again this time. And smile and nod when they show off the usual hilarious scenarios about why 10+ cores are useful on the desktop, if you are outwardly critical you will be cut off, they don’t tolerate dissent any more, true or not. Worse yet your ad revenue will be in jeopardy if you so much as snicker.

 

[LTC8K6]

Ouch, what a fail Intel, pathetic if true. Another site mentioned similar as well it seems for Kaby Lake-X (Kaby Lake X only 4 core?). Only dual channel available on the 4core that is supposed to be high end? Not sure why Intel would do this aside from a knee jerk to Ryzen catching them completely off guard. I have to be reading things wrong or the news isn't completely true or something.

Intel X299 platform launching on May 30th

When we discussed KL-X here a couple months ago, the consensus seemed to be that it was probably intended as an entry level HEDT chip.
It was not likely a reaction to RyZen, given that it must have already been in the pipeline for the X299 platform.

 

[IndyColtsFan]

Intel to brief press on Sandy-E/X on May 2

http://semiaccurate.com/2017/04/24/intel-brief-press-sandy-ex-may-2/


Intel has another top-secret gathering soon for briefings on Sandy-E/X, Kaby-X and X-299. If you weren’t already invited, you might want to ask Intel why they stiffed you, tell them SemiAccurate says hi.

In just over a week, Intel will be flying hordes of press to San Fransisco for a briefing on the upcoming 6-12 core Basin Falls platform. You might know this as Sandy Bridge-E or the 1-socket Purley minus a lot of features. They are also likely to try and convince the gathered masses that the Kaby Lake version that fits into the same socket, minus half of the memory channels, many PCIe lanes, but with a massive price increase is not a bad joke. Feel free to differ with their opinions here but recall these are the same clowns who claimed you needed a 10-core Broadwell-E for video encoding while gaming.

So the short story is that if you haven’t been invited yet, Intel doesn’t think you are worth bothering with any more. If this story is like the last one, expect a hastily noticed conference call email next week with excuses about it being planned all along. Please pretend to believe them again this time. And smile and nod when they show off the usual hilarious scenarios about why 10+ cores are useful on the desktop, if you are outwardly critical you will be cut off, they don’t tolerate dissent any more, true or not. Worse yet your ad revenue will be in jeopardy if you so much as snicker.

Uhhhh....Sandy Bridge E? I checked and the article wasn't dated 4/1/17.....

 

[Shivansps]

I do wonder if 1151 cpu are compatible with x299... I still remember the 478 to 775 adapter i had one for a while to play with, it made a 478 celeron d to work with every 775 i tryied, somehow.

 

[LTC8K6]

Uhhhh....Sandy Bridge E? I checked and the article wasn't dated 4/1/17.....

Strange piece. It's written as if they haven't already heard of KL-X, when we all know it was in the news at least two months ago, and as if they are unaware of the name of the new Intel HEDT chips.
It also seems to be an attack on Intel.
The bit about needing a 10 core chip almost reads as an attack on RyZen as well.

 

[crashtech]

Sandy Bridge? It's a brain-fart typo, right? Although it appears twice three times in the article. Strange. I think it's supposed to be satirical, if so it went right over my head.

 

[w3rd]

HEDT sku for 2018:
$799 12core
$599 10 core
$399 8 core
$299 6 core


Based^ on the rumors of a Ryzen (16c) & Intel (24C) monster chips.

 

[IEC]

That's... optimistic, to say the least.

I fully expect Intel HEDT SKUs to cost more than that.

 

[pantsaregood]

I'm not holding my breath, but I'm really hoping that there's an 8-core Skylake-X in the sub-$700 range since core counts are being shifted around. I've been waiting for HEDT Skylake and now I'll feel a little silly if I get a 6-core HEDT CPU when they're going to be available on LGA 1151.

 

[Phynaz]

Uhhhh....Sandy Bridge E? I checked and the article wasn't dated 4/1/17.....

It's Charlie, enough said. Same bucket as WCCF, the trash bucket.

 

[IndyColtsFan]

That's... optimistic, to say the least.

I fully expect Intel HEDT SKUs to cost more than that.

I expect Intel will drop prices on the HEDT chips but they're still going to cost more than the equivalent Ryzen. I think at best, we'll see the HEDT chips slide down a rung in pricing (10 core will occupy the current 8 core pricing, etc), but the pricing a few posts above is way too optimistic (and believe me, I'd LOVE to be the one wrong in this case). I mean, I guess I could possibly see Intel dropping 6 core HEDT and releasing a crippled 8 core at the $399 price point, but why would they? The HEDT market is a niche for them and they are still faster than Ryzen so they can command a premium.

 

[Ajay]

It's Charlie, enough said. Same bucket as WCCF, the trash bucket.

Even a broken clock is right twice a day

 

[CakeMonster]

If I'm not mistaken, LGA1151 gets 6C12T Coffeelake as the new flagship and it should actually be a great CPU considering 14nm++, probably an easy 5GHz on 6C12T if you're able to keep that cool enough. I suspect delidding+CLU will do the trick as it does on the 7700k

How do you justify claiming it can do 5GHz? So far the HEDT 6c and 8c have had a distinctly lower OC potential than mainstream 4c, even on the same generation (mainstream on average being at least one generation ahead).

 

[dahorns]

How do you justify claiming it can do 5GHz? So far the HEDT 6c and 8c have had a distinctly lower OC potential than mainstream 4c, even on the same generation (mainstream on average being at least one generation ahead).

It is a pretty reasonable guess in this instance. (1) current 6 cores can push to 4.3 or 4.4, and that is with design and process combination that was notably poor at overclocking even on the mainstream; (2) in contrast to Broadwell's poor showing, a good percentage of 7700ks can push to 5; and (3) Coffeelake will actually be on a superior process to the 7700k as it will be built on 14nm++.