Intel Delays Skylake-X and Kaby Lake-X release to August

Maybe if AMD becomes competitive by then Intel wont charge our souls for this chips.

 

Non of the companies are like they used to be before, say, 2010. Overproduction combined with general social (not just "economical") crisis is taking it's toll.

 

This article has some extra salty editorializing: "Intel has not been the company it has been for years lately" (because there is a rumor some new cpu/chipset will release in August instead of June?) and "processors once again will need a new lga 2066-socket, and yes that means that once again you have to purchase a new motherboard" (nice trudging sluggish use of once again and once again).

I am not sure what the first quote means. If it suggests Intel is not the same because they make more money and have a higher share value, then ok. In 2010 Intel stock sold for US$20 a share. Friday they closed at US$37. And as for a new socket, get over it. I get the feeling some think Intel should make a socket and stick with it for a decade. Socket 2011 was not released last year, it was released back in 2011. They have changed chipsets in that socket but you can only go so far. Would we prefer the same socket used for 10 years with tiny processor clock increases with who cares feature upgrades? That model didn't work for another company I know which was selling drywall and chairs for capital loans.

I for one can't wait for the RyZen release. Perhaps then we can stop with the trolling of Intel and pumping of AMD and move on to pumping the competition between both.

 

Cpu market become so boring for the enthusiast...

 

I didn't even know they aren't caring about press in Europe anymore.

 

Only he has a point, and he probably has seen more new hardware the last decade than all of us in the thread combined. And Intel definitely is not the same company as it was, and not only in the profit margin area. They exited mobile after throwing billions at it, and they seem to have failed to produce a new desktop CPU design, although they were basically left alone for more than a decade. Their main investment has been in GPUs, despite lacking the actual design and software talent for maintaining said division. They ended up having to purchase AMD patents to keep the ship rolling. Their modems are sh*t compared to Qualcomm's, and their only "success" is the "diversivication" of their workforce, which means tens of thousands of people getting fired and people getting hired sometimes more based on color and gender rather than ability. This led them to even have bad relationships with traditional OEM partners due to the fall of quality in personnel.

The socket changes are pure marketing. Even an Intel engineer half-admits in his Reddit AMA that socket changes are responses to markets, not technical needs.

 

Well setting a socket type for a particular motherboard does ensure compatibility I suppose while retaining most of the feature set that the CPU supports although from a tech point I suppose those ~2011 or so pins they had back for the x79 would have been enough but required some more backwards compatibility work.
(The v3 just switches things around a bit far as I remember.)

At least they kept the socket type between the x170 and x270 motherboards although the x270 doesn't seem to have that many new features either.
(But if I remember a earlier post about these EX models they actually do use a different socket configuration requiring that x290 board or what it was to be called.)

 

If you see the AMA link, that's a big excuse. Socket 1156 and socket 2011 do have the pin counts to support everything needed, it was clearly something the "markets" decided, not the engineering team.

 

Interesting, will have to give that a read then though yeah marketing wise why sell just the CPU when you could sell it with a new motherboard too.

Likely also pleasing the various third party motherboard manufacturers they have contracts with now that Intel isn't manufacturing their own reference motherboards which for the x270 boards release it looks like most companies had some three or four different designs lined up in time for release covering everything from entry to enthusiast level. (Which in itself isn't a bad thing I suppose, no need to stick everything on the board or how to say.)

As far as these actual CPU's go (And motherboards should they indeed use a new socket.) I guess this is perhaps in response to AMD Ryzen allowing Intel to monitor that a bit and then price these appropriately although likely still more towards enthusiast level pricing or how to call it.

Altering pricing on the 6000 or 7000 CPU series or adding some new models might also happen, kinda difficult to say until Ryzen is actually released and we know how well it holds up and at what pricing point.

That and I guess if these won't bring any major performance gains (Even if they are "x" instead of "z") I guess waiting a bit with their release probably makes sense also allowing for more sales of the 6000 and 7000 series of CPU's.

Also forgot about core count, guessing most of the 7000 series (Kaby lake.) will still be quads so having hexa core or what it was (hexa and octa I believe, one was six and the other eight.) that could also make them a bit more attractive now that software and even games are regularly using more cores.
(Effectiveness of said usage varies a bit still though but it's improving. )


EDIT: Also come to think of it from reading the above then review wise if Intel is ("sorta") giving the finger to most of Europe in regards to PR and what not how do you even get hold of review samples of these CPU's and motherboards in time for whenever the NDA expires or when the product launches, can't imagine having them shipped from the US or bought directly from store.

Direct connection with some of the hardware manufacturers perhaps, that or co-operating with other tech sites I suppose which come to think of it I think was mentioned as how it was done with the Kaby Lake CPU's.
(Although limiting the amount of time available to use said CPU's.)

 

C´mon this is Intel we are talking about... Intel will only cut their CPUs prices when their sales start to decrease until then...

As for Intel´s problems, they only exist because of greedy shareholders that want more and more even when they already reached the top of the mountain...
Luckily for Intel, they have so much money, they can afford some bad years in the future...

 

I don't believe that most of you realize how much money Intel has literally thrown into the bin the last few years. Not that they have a cashflow issue, but aside from desktop and server x86 CPUs, they really have no other source of income. Ryzen (if is as good as reported), now threatens the market that keeps them alive basically.

 

I mean, they can just do what they did the last time and just give vendors exclusive rebates for Intel products. They got slapped with a 1.4B fine that they never paid (still disputing a decade later). But even still, $1.4B for ~90% marketshare seems like a good deal. That's like half what they pay for a new architecture.

 

That´s true. but the reason why they wasted so much money it´s because they have much more!... It´s like MS, Google, Facebook and other giant companies, they have so much money in the bank that they can´t resist blowing in the most stupid manner possible...

And as long as Intel can keep their monopoly in the server market, they will be fine even when the desktop market disappears, sooner or late... But if they lose the server market, then they are ****ed!!!

 

Not really.

x86 CPUs are likely less than 20% of their income.

They have a huge market in chipsets, SOCs, SSDs, Networking/IO, software, boards, etc.

Ryzen won't really touch the server market, only the consumer market(if it's good). PC microprocessors sales are basically inconsequential.

 

That's quite the sad truth. On the other hand, if Ryzen and especially Ryzen+Vega APUs are as good as they hint to be, I don't believe that any OEM will bother with colluding with Intel.

I don't say this from the naïve side, just from the side that today things are much more open information-wise, and even a rumor like that would hurt the flailing mainstream PC sales, and the underwhelming performance in a market where even a hint of a better alternative in an interview, can sink your product.

I honestly don't believe that Intel is holding the market as much as it used to do.

I don't believe that anybody has the full list, but here are some examples:

Larrabee.

IPAD (yeah, seriously).

Intel Play brand as a collaboration for toys with Mattel.

The mobile market. After spending more than 11 billion in acquiring various mobile-related companies, Intel sold its unit to Marvell for 600 million in 2006.

There is a whole list in Wikipedia on Intel's acquisitions for the 2000-2016 period. With the exception of Altera which is an FPGA company that will help them to add auxiliary CPUs in their servers, something that AMD already does, all the rest are a bunch of failed crap. Even that purchase wasn't really well received, considering Intel's history with 3rd party acquisitions.

The greatest failure there being McAffee, there were other horrible ones there too. Infineon, for example. And there haven't been any concrete numbers on the development costs on their Atom Z line, which is now defunct. There are some in there that are almost comical, like the 5 billion fab in Arizona that has never worked.

I won't mention the costs for the personnel "diversification" and the loss of quality OEM-Intel people over the years because they were too male and too white (and I happen to know a few personally actually).

All that failed crap is on the back of the CPU division outdoing AMD on the server and desktop market. Intel gave billions for useless software companies, useless fabs, useless gpu and mobile patents, yet it hasn't refreshed their main life blood line since 2011.

All of their profit stands on the profit margins of selling <1bn transistor CPUs for $300+. This can't and won't go on forever.

 

You are completely wrong.

Furthermore, if you don't believe that upsets in the server market can happen, you have failed to notice NVIDIA. If AMD offers a 32-core CPU with better interconnects than Intel and similar per-core IPC in the same TDP, they will be just on time for the first big data center refreshes that will happen in 2018-2020. Most companies are on a ten-year data center refresh cycle these days, and if AMD offers more cores per socket on the same TDP, that's what they'll get.

From the 55.36 billion of income they had, 48.2 billion was the desktop and server markets from the PC client and Data center groups. All the rest combined are barely 11% of their income.

Whoops.

As you can see, all the crap they have wasted their money on, has failed to be anything more than a dent in their actual income coming from the products they seem to neglect the most.

 

PC client group includes 'PCs, servers, tablets, phones'

Which that includes a multitude of things, a few of which i mentioned already.

So you won't get an exact number for CPUs in the consumer market. As i said, likely under 20%. (im not including server side)

Intel has ~99% odd market share in the server market.

Ryzen would have to be significantly better than intel's offerings, which is not likely.

Even then, chipset features are a big part of the market too. Consumer wise, they don't have anything that will really compete with intels top end C610(X99) and upcoming skylake-x chipsets.

Server side is unknown at this point. But X99 is similar to server chipsets for Intel at least.

Plus the fact that they will have to compete with ARM and NV in the coming years as you said.

Potential to do well in the market? Hopefully.
But if you think that that will end Intel's rein on having most of the market share in the next few years, that is wishful thinking.

What you are saying will take a decade or longer.

 

When Intel united all of their "client side" into the PC Clients Group last year, the mobile part of it had a 4 billion net loss while the "just desktop" part was profitable at 14 billion. Repeating a myth doesn't make it more real.

A lot of analysts also speculated that Intel basically "hid" the massively failed mobile division into the PC client one, whose profits could cover the net loss from the failed mobile ventures. They killed the Atom because it wasn't working, and they practically have nothing in the mobile space.

All of the 'multitudes of things' are either x86 CPUs, or supporting chipsets of various kinds (Wi-Fi etc). The mobile division was losing a billion per quarter before the merger.

I don't know why you ignore Intel's own reports that I constantly link. From the 55 billion Intel made last year, 32 billion was made by selling x86 CPUs and the supporting chipsets. The mobile division (now part of the PC clients), was losing a billion per quarter. If anything the profits would be higher if it wasn't for the mobile division.

This is true for the traditional server market. They are losing badly to NVIDIA for the newer "smarter" data center with AI operations. That market share is also a "traditional x86 CPU" market share. Again, the sector that Intel seems to have ignored the most, despite relying on it.

A 32-core Naples CPU at the same TDP as a 22-core Xeon, means that the Xeon will lose badly.

You're right, for that side of the things, they don't. On the other hand, what percentage of people will be willing to go from a ~$350-$400 AMD 8-core CPU with practically the same TDP, to a $1,000+ Intel CPU for an extra couple of cores? I'm sure that there will be situations that that's called for, just a lot less than in the past.

I don't believe that Intel's grasp of the market will end any time soon. I just believe that their total and absolute monopoly in some sectors is done, and that if they don't reorganize themselves to do something, they are possibly in trouble in the coming five years.