Intel planning for thousands of job cuts, internal sources say

[Adored]

Its based entirely on future hopes. You know, how stocks work

Also if I was you, I would see if it sticks or drops again.

Well this is funny because most people say that AMD is always looking to the future and missing out on the present.

Money now is what they need. If the Intel licensing agreement comes to fruition then that's even more free money. AMD has a ton of IP that they just haven't been monetizing because of their belief that they can make more in future.

Semi-custom taught them the value of money today. They should have licensed Mantle to Valve as well, makes me wonder what they are really getting out of it by giving it away to Khronos instead.

 

[ShintaiDK]

Well this is funny because most people say that AMD is always looking to the future and missing out on the present.

Money now is what they need. If the Intel licensing agreement comes to fruition then that's even more free money. AMD has a ton of IP that they just haven't been monetizing because of their belief that they can make more in future.

Semi-custom taught them the value of money today. They should have licensed Mantle to Valve as well, makes me wonder what they are really getting out of it by giving it away to Khronos instead.

They should be doing more licensing deals. Because their own products doesn't sell well to put it bluntly. They have missed huge opportunities in the graphics part for example, due to missing out of licensing. And today most of those are gone.

The future is questionable. Because growth is gone in all but servers. And that's not just for AMD. Its for Intel, Qualcomm, MediaTek, Apple etc as well.

 

[frozentundra123456]

Well, I *did* say it was an over-reaction. I have no AMD stock at all, but if I had bought it at 2.00 or less, I would be jumping with joy and dumping it ASAP at 3.00.

It is just that your statement of "worst quarter" definitely did not tell the whole story. Margin was up, loss was less than expected, and projection for next quarter was way above expectations.

 

[Adored]

They should be doing more licensing deals. Because their own products doesn't sell well to put it bluntly. They have missed huge opportunities in the graphics part for example, due to missing out of licensing. And today most of those are gone.

The future is questionable. Because growth is gone in all but servers.

I think they will do more in future.

The sad thing about all this is, the companies doing the actual hard work are being left behind to the marketing companies and the software companies. There just isn't enough money in chips compared to the money built on them.

 

[ShintaiDK]

I think they will do more in future.

The sad thing about all this is, the companies doing the actual hard work are being left behind to the marketing companies and the software companies. There just isn't enough money in chips compared to the money built on them.

There never was. Hardware is the ungrateful task. :\

But consolidations and buyouts will pave the short term future. Specially in a time where money cost nothing. (Zerorates).

 

[ShintaiDK]

Well, I *did* say it was an over-reaction. I have no AMD stock at all, but if I had bought it at 2.00 or less, I would be jumping with joy and dumping it ASAP at 3.00.

It is just that your statement of "worst quarter" definitely did not tell the whole story. Margin was up, loss was less than expected, and projection for next quarter was way above expectations.

A funny note in terms of investment. If I borrow money in my home, its at a -0.44% interest rate. (Yes I get money for borrowing.).

Can you say housing and stock bubble?

Its really dangerous when investors got free money this way, because risk will increase.

 

[Dresdenboy]

Its based entirely on future hopes. You know, how stocks work

Also if I was you, I would see if it sticks or drops again.



This is the driver. Plus the 320M$ from ATMP joint venture completion.

There were some fast money profit takers already. Expect some fluctuations on the open (this statement is most often true and brings no new information ^^).

The ATMP deal is no surprise anymore.

But if you say, that short term goals (related to the current quarter) are the driver, you contradict your own statement, that future hopes are the driver.

I don't know, should I buy a tray full of BDW-E's now or keep the shares longer? This is my last long term trade. ^^

 

[ShintaiDK]

Q2 is still the future both financially and practically. Not to mention beyond that.

In terms of the stock you should ask yourself what you believe in. Because the world isn't going to recover. So they need to capture revenue faster than its lost for the entire segment.

Even smartphones, "the great PC killer", is in decline now.

Personally I wouldn't invest in any production company currently, unless its base essentials for survival. But there isn't any money in that.

AMD was also 3$ around new year. Then down to 1.8$, now up around 3.15$.

 

[Dresdenboy]

Q2 is still the future both financially and practically. Not to mention beyond that.

In this strict sense you're right, but I see a difference in "future hopes" (something might happen) and same or next quarter business results (it's very likely going to happen), not talking about single percent fluctuations here.

But to understand these discrepancies, I think, it is necessary to leave the seemingly established factual discussion and look more for interesting things beyond this level, like the use of natural language and especially the individual way of thinking and perceiving the world. But this is entirely another topic. ^^

In terms of the stock you should ask yourself what you believe in. Because the world isn't going to recover. So they need to capture revenue faster than its lost for the entire segment.

Even smartphones, "the great PC killer", is in decline now.

Personally I wouldn't invest in any production company currently, unless its base essentials for survival. But there isn't any money in that.

Nvidia stock jumped from 25$ to 35$. But on what?

AMD was also 3$ around new year.

I don't believe, I try to get the picture with as much plain (unmodulated) information as possible. And I also take care of what I don't know. But this is more just for fun and my last long term trade, so no more market risks in the long run. Looking for price patterns with some more advanced metrics is my thing, as they are there, and tell you, what the market participants are actually doing. And these methods are different than most charting books and courses try to teach. That's how the world works. But there is often the pattern that money flows from the less informed or "positioned" ones to the better informed/"positioned" ones.

But money will continue to flow into the markets thank to retirement funds and other money seeking for a home based on hopes (more general one like "avg. 5% p.a.").


That's quality analysis ^^:

Source: @asymco

 

[Systems analyst]

IBM is a goner long ago. Qualcomm is struggling for survival now in the MPU section. I have a feeling their days of custom cores is soon, if not already over. Google is buying Intel and Nvidia, just like Baidu, Facebook, Microsoft, Amazon etc.

Google have said that they will replace current servers with ARM or OpenPower servers, if it saves money; they develop and test their software for all three solutions. Furthermore, they can and will change to one supplier for one generation, and then change to another supplier for the next generation. Currently they are evaluating Qualcomm servers; ARM have issued V8.1 and V8.2 architecture for server use. Qualcomm may be limiting their custom micro-architecture to server SoCs, and possibly Premium smartphones.
For ARM at least, they are increasing their profits in mobile-phone SoCs, due to the increasing penetration of ARM V8, Mali, physical IP etc.
All the top hyper-scale companies are evaluating the options. Software is, presumably, not a problem to them, due to their scale.

 

[ShintaiDK]

Google have said that they will replace current servers with ARM or OpenPower servers, if it saves money; they develop and test their software for all three solutions. Furthermore, they can and will change to one supplier for one generation, and then change to another supplier for the next generation. Currently they are evaluating Qualcomm servers; ARM have issued V8.1 and V8.2 architecture for server use. Qualcomm may be limiting their custom micro-architecture to server SoCs, and possibly Premium smartphones.
For ARM at least, they are increasing their profits in mobile-phone SoCs, due to the increasing penetration of ARM V8, Mali, physical IP etc.
All the top hyper-scale companies are evaluating the options. Software is, presumably, not a problem to them, due to their scale.

We hear that story over and over, yet it never materialize. Its all about TCO, something some people forget all the time. The CPU cost itself is close to irrelevant in that matter.

 

[imported_ats]

Mellanox provide the TILE-Gx72 Processor, a cache-coherent SoC, based on 72 ARM A53 cores I believe:
http://www.mellanox.com/page/products_dyn?product_family=238

This features ARM cores:
https://careers-mellanox.icims.com/jobs/3075/software-architect---multicore-processors/job

Sandia Labs have tested multi-processor ARM based SoCs:
http://www.mellanox.com/related-docs/solutions/SB_M400.pdf
They say:
"... we were able to fully exploit all the cores on the processor to achieve linear scaling."
James Ang, Sandia National Labs.

LOL, you do realize that an m400 can't even beat out a ~3 year old ATOM C2750 in performance right? The performance of the core is literally so bad that the rest of the system isn't a bottleneck in anyway. Yes it is possible to get linear scaling sometimes: when your performance is so bad that its pathetic. Its like in the bad old days when Sun was touting linear performance of their servers, neglecting to mention their linear scaling servers had worse performance than their competitor servers that were like 1/4 the size and 1/4 the price.

And even then its highly likely that they were running embarrassingly parallel kernels with no communication.

The SoC fabric provides the cache coherency across the SoC.

Mellanox merged with EZchip, who had bought Tilera.

Maybe X86D: does not scale linearly, but ARM V8 does.

I sure you could get a bunch of x86 quarks to scale linearly to some point, I'm also pretty sure that like your example it would still be significantly slower than a ~3 year old Atom. And again, only on embarrassingly parallel toy programs that don't do anything useful.

ARM V8 doesn't provide ANYTHING that assists MP scaling. And it actually lacks significant features compared to x86 for MP scaling.

Oh, and you might not want to get all gung ho about tilera and then post a quote about a different product...

 

[imported_ats]

We're not talking about re-runs of Knight Rider, dude.

Prior to the Telsa, there was never a car that could autopilot/drive itself available for purchase to a consumer. So what ever you are smoking -- I'd like some....

But thanks for the laughs.

Autopilot vehicles have been for sale for over a decade. The only thing that's changed is that they've come down in size and cost (you might want to pay attention in what's been going on in heavy machinery and farming for quite some time now... )

And hell, the Tesla really doesn't autopilot/drive itself. Oh, it can backup out of a garage isn't driving itself.

 

[imported_ats]

Google have said that they will replace current servers with ARM or OpenPower servers, if it saves money; they develop and test their software for all three solutions. Furthermore, they can and will change to one supplier for one generation, and then change to another supplier for the next generation. Currently they are evaluating Qualcomm servers; ARM have issued V8.1 and V8.2 architecture for server use. Qualcomm may be limiting their custom micro-architecture to server SoCs, and possibly Premium smartphones.
For ARM at least, they are increasing their profits in mobile-phone SoCs, due to the increasing penetration of ARM V8, Mali, physical IP etc.
All the top hyper-scale companies are evaluating the options. Software is, presumably, not a problem to them, due to their scale.

Hi, this is 7 years ago calling, why are you repeating the same story?

Cloud providers always have their software running on "alternative".
Cloud providers always say they are evaluation other "options".
Cloud providers always end up buying Intel.

I think you are confusing negotiation tactics with actual purchases.

 

[Dresdenboy]

Autopilot vehicles have been for sale for over a decade. The only thing that's changed is that they've come down in size and cost (you might want to pay attention in what's been going on in heavy machinery and farming for quite some time now... )

The necessary tech fits into one or two breadboxes. The differences are also in situational complexity, safety, acceptance, vehicle dynamics, and so on.

 

[MiddleOfTheRoad]

Autopilot vehicles have been for sale for over a decade. The only thing that's changed is that they've come down in size and cost (you might want to pay attention in what's been going on in heavy machinery and farming for quite some time now... )

No they haven't. There is a big difference between navigation-based autopilot and adaptive cruise control. You are clearly mistaken on the differences.
Nobody has built a system as sophisticated as the Tesla's prior for a consumer product. Their implementation was a huge step forward.

The next gen will likely have additional sensors for a completely autonomous drive.

 

[positivedoppler]

Meanwhile,
http://techcrunch.com/2016/04/22/in...ved-threats-for-companys-stance-on-diversity/

 

[Systems analyst]

We hear that story over and over, yet it never materialize. Its all about TCO, something some people forget all the time. The CPU cost itself is close to irrelevant in that matter.

"PayPal releasing information demonstrating the benefits of using ARM based servers with a 50% reduction in capex, an 85% reduction in running costs and a 10-fold increase in server density compared to traditional data center equipment"

ARM's 2015Q2 results refers:
http://media.corporate-ir.net/media_files/priv/ccbn/event_help/smalldownload/pdf.gif

 

[Systems analyst]

LOL, you do realize that an m400 can't even beat out a ~3 year old ATOM C2750 in performance right? The performance of the core is literally so bad that the rest of the system isn't a bottleneck in anyway. Yes it is possible to get linear scaling sometimes: when your performance is so bad that its pathetic. Its like in the bad old days when Sun was touting linear performance of their servers, neglecting to mention their linear scaling servers had worse performance than their competitor servers that were like 1/4 the size and 1/4 the price.

And even then its highly likely that they were running embarrassingly parallel kernels with no communication.




I sure you could get a bunch of x86 quarks to scale linearly to some point, I'm also pretty sure that like your example it would still be significantly slower than a ~3 year old Atom. And again, only on embarrassingly parallel toy programs that don't do anything useful.

ARM V8 doesn't provide ANYTHING that assists MP scaling. And it actually lacks significant features compared to x86 for MP scaling.

Oh, and you might not want to get all gung ho about tilera and then post a quote about a different product...

My understanding is that many different SoCs will be designed and developed, for many different purposes; it is not a competition to which is the most powerful, in some sense, but to provide varied SoCs that are fit-for-purpose in different applications.

At present, ARM based SoCs are being used in the field, where there are suitable use-cases, to try them out. More powerful SoCs, for more demanding cases are being developed.

But you know a lot about Intel; can they develop a SoC with, say, 72 Atom cores of some description? - or would it be prone to 'dark-silicon', due to heat dissipation?

ARM have announced that they are developing IP for SoCs of up to 256 cores, presumably for a case like web-serving, perhaps, or network processing. This is enabled by SoC-level design, not by the ISA so much, although that does play a role, so I understand, in the efficiency of some SoCs.

Are Intel developing IP for up to 256 cores on a SoC?

 

[jpiniero]

But you know a lot about Intel; can they develop a SoC with, say, 72 Atom cores of some description? - or would it be prone to 'dark-silicon', due to heat dissipation?

That's pretty much what Knights Landing is. Well, it also has the AVX-512 unit.

At this point I don't know serious the cloud companies are with ARM servers. It could be seen as just simply flirting to scare Intel enough to get better deals.

 

[Exophase]

Meanwhile,
http://techcrunch.com/2016/04/22/in...ved-threats-for-companys-stance-on-diversity/

There's much interest in the demographics of the people they're hiring. I wonder what the demographics are like for the people they're laying off?

 

[pitz]

There's much interest in the demographics of the people they're hiring. I wonder what the demographics are like for the people they're laying off?

Caucasian US citizen males, of course. By far the most expensive part of their workforce.

The really good stuff at Intel hasn't even been designed in the US for a long time now. Most of the engineering in Intel's modern CPU and chipset products happens at their Haifa complex.

Having said that, companies of Intel's size tend to get really bloated as people who should have been laid off, aren't. For instance, those Intel dancing bunnies that were seen on the TV ads, it wouldn't surprise me if they were still working at Intel, doing not much. So a purge, of course, takes them out.

 

[Thala]

ARM V8 doesn't provide ANYTHING that assists MP scaling. And it actually lacks significant features compared to x86 for MP scaling.

What features in particular?

 

[videogames101]

That's pretty much what Knights Landing is. Well, it also has the AVX-512 unit.

At this point I don't know serious the cloud companies are with ARM servers. It could be seen as just simply flirting to scare Intel enough to get better deals.

If you asked most small to medium sized web companies if they would re-compile and switch to ARM in order to cut their AWS bill in half, 100% of companies would "yes, how about before 5pm tonight?"

The ARM dream: ARM-based AWS servers that cost half as much (iso-performance) as Intel servers to AWS end-customers.

The question is whether or not ARM, its partners, and the software ecosystem can pull together to actually deliver a product such that Amazon could offer that option and still make money.

My understanding is that there are still many parts of the stack missing to offer something like that. So if it happens, it's still a few years out.

 

[custombuild]

Sensationalist headline: Intel running scared of AMD's Zen CPU, massive layoffs ahead of release!