[allthingsd.com] AMD getting ready for another round of Layoffs

[anongineer]

The thing I don't get is how AMD could have put the development pipeline on "hold"?! Does this mean that while AMD kept producing designs, it didn't push them forward to tape out?.

For individual designs that go into a chip and where work is already ongoing, it may be difficult to yank funding. But apparently it's not that hard to kill a chip at the very end, i.e. Wichita and friends.

Provided a chip is killed before litho masks are produced, a decent amount of money can be saved. Plus, all those individual designs (blocks, IP, whatever you want to call it) may be useful to other chips, so at least there's some consolation.

The company could also choose to understaff development teams. Save on salary, take a hit on schedule.

And of course, there is always the possibility that some dumb corporate initiative caused delays across the board.

 

[WhoBeDaPlaya]

I just hope in 3 - 5 years time a mediocre performing Intel x86 CPU does not cost 500+ dollars. I lived through that once and it was not fun (Pentium 233MMX).

That's nothing. I paid nearly $1,000 for a P2-400 back in the day

 

[CTho9305]

Are you talking about Rory or Charlie?

Rory's memo to the downtrodden was vacuous to the point of embarrassment. The guy truly does not help himself in avoiding garnering the perception that he is all talk and no thought.



with leadership like that oozing from above down onto the heads of the plebeians, how can the company not be kicking Intel's butt left and right?

Are executive memos not always like that?

I can speak from a perspective of personal experience since I was present in an HP campus when I watched HP layoff thousands about 8 years ago (ish), and, I had to personally lay someone off once - who ironically is over at AMD where he's been pretty happy, I think (I hope).

The people I know who got laid off from AMD generally ended up in better positions. The ones who were gone early didn't have to sit through multiple subsequent rounds of layoffs, 10% salary cuts, etc...

As far as the layoffs at AMD, I wish them the best possible. About the only nice thing is that electrical engineers can usually find another job fairly quickly... if they are willing to move.

The Austin market seems pretty hot right now (ignoring some Intel rumors), with Samsung, Oracle, Qualcomm, Apple all hiring. At least some people should be able to get new jobs pretty quickly.

 

[CHADBOGA]

Why do people in the CPU forum think AMD is about to die and in the video forum they think AMD is on top of the world?

Because all the bright people hang out in the CPU forum.

 

[Dribble]

Why do people in the CPU forum think AMD is about to die and in the video forum they think AMD is on top of the world?

The AMD cpu fanboys have mostly given up hope, the ATI gpu fanboys still believe.

 

[NTMBK]

You lived through an era of continually falling CPU prices in the 1990's and at one point the Pentium 233MMX was a top end CPU in its day.

Prices were continually falling due to good competition in the x86 market with rivals like Cyrix, Centaur, AMD, and Texas Instruments, with real pressure from RISC competitors in the workstation market. If AMD fall behind as badly as VIA do, that leaves Intel with no real competitors outside of phones and tablets. What do you really think will happen to their prices? Just look at what's happened with this round of Xeon pricing.

 

[CHADBOGA]

Prices were continually falling due to good competition in the x86 market with rivals like Cyrix, Centaur, AMD, and Texas Instruments, with real pressure from RISC competitors in the workstation market.

All computer components were continually falling in price.

And you seem to miss the point, that by the late 1990's, only AMD was left as a viable competitor, the other names you mentioned had fallen away, yet still prices fell.

The time of the Pentium 233MMX was one where there were more competitors, yet that is the time that Madpacket claims was a diabolical time for CPU prices.

If AMD fall behind as badly as VIA do, that leaves Intel with no real competitors outside of phones and tablets. What do you really think will happen to their prices?

They will go up somewhat.

 

[Keysplayr]

Why do people in the CPU forum think AMD is about to die and in the video forum they think AMD is on top of the world?

The 5 Stages of grief is in effect in the VC&G sub-forum.
Kubler-Ross model
http://en.wikipedia.org/wiki/Kübler-Ross_model

 

[pelov]

The Austin market seems pretty hot right now (ignoring some Intel rumors), with Samsung, Oracle, Qualcomm, Apple all hiring. At least some people should be able to get new jobs pretty quickly.

So it has begun...

Regardless of whether Intel has competition in the x86 space or not, they're going to have to drastically change the way they conduct business in the next 1-2 years. Haswell looks great, but unless they price it accordingly and are able to funnel worthy products into consumers hands, they're going to be the next big chip maker to be heading toward a downward slide.

 

[Keysplayr]

Well, truth be told, setting aside all the pokes and jabs that are really meant in jest, I actually thought (and continue to think) that if anyone can save AMD and morph it into a company that creates products that serve a market in a viable fashion (not this for-loss mode that AMD has been in almost always) then it would be Rory that could do it and it would take a Rory (experience-wise) to be able to have the vision of what it would take to do it.

His memos may be vacuous and full of buzzwords but that may be an unfortunate side-effect of what HR requires him to not say in a time of considerable internal headcount flux.

I doubt Rory is as one-dimensional as the snippets of his caricature seem to be from a handful of online videos and a few leaked internal memos.

But it does make for good sport to lampoon the guy for the strawman that he can be spun to be on the basis of those video snippets and memos :whiste:

I don't. I think Rory Read is the exit CEO. Have thought that since he took the job. But, there is plenty of room to prove me wrong. He could be the next JHH, or Steve Jobs perhaps? Time will tell, but there isn't a lot of that left. People continue to say, and truly believe, that, "AMD isn't going anywhere." . I can say I'm amazed they held on for this long, but then realized they were infused with billions from Abu Dhabi and Intel. If that hadn't happened, AMD would be dust by now or in another form with it's IP sold off. Now, the money is running out, the CPU line is all but irrelevant and the graphics division can't do enough to overcome that.
Prove me wrong Rory Read. God Speed.

 

[WhoBeDaPlaya]

The Austin market seems pretty hot right now (ignoring some Intel rumors), with Samsung, Oracle, Qualcomm, Apple all hiring. At least some people should be able to get new jobs pretty quickly.

Indeed. I wish the Dallas market was a vibrant as it once was

 

[Idontcare]

That's nothing. I paid nearly $1,000 for a P2-400 back in the day

QX6700, $1500

At least the chip worked and met expectations, that crappy striker mobo though...

 

[Lonbjerg]

QX6700, $1500

At least the chip worked and met expectations, that crappy striker mobo though...

Well...going none-Intel with the chipset on an Intel CPU has always been bad in book ^_^

 

[NTMBK]

Well...going none-Intel with the chipset on an Intel CPU has always been bad in book ^_^

Well, except for Ion, which was the only way of getting useful graphics on Atom

 

[pelov]

Well, except for Ion, which was the only way of getting useful graphics on Atom

At the cost of typical nVidia power consumption.

Does anyone remember just how damn hot those nForce chipsets ran? I remember buying a nForce chipset and promptly going right back to VIA

 

[AdamK47]

QX6700, $1500

At least the chip worked and met expectations, that crappy striker mobo though...

And people complain that I don't need a $1000 i7 3960X. That CPU was $500 more.

 

[sm625]

AMD will get the same fate as VIA.

LOL Via is part of Formosa which is a $20 billion company. They arent doing too badly and they arent going anywhere.

 

[Olikan]

LOL Via is part of Formosa which is a $20 billion company. They arent doing too badly and they arent going anywhere.

well, if amd stay competitive atlest with nvidia, is all good
i mean, the cpus haven't been competitive for years anyway

 

[Idontcare]

They arent doing too badly and they arent going anywhere.

That would be the point, they aren't going anywhere.

Is that what we've come to as business success now? No way forward = good news for the consumer?

 

[pelov]

That would be the point, they aren't going anywhere.

Is that what we've come to as business success now? No way forward = good news for the consumer?

Hasn't that been the motto for the past 4-5 years for both Intel and AMD anyway? Incremental increases for nobody in particular.

At least netbooks and Ultrabooks had a semblance of thought and innovation as far as product lines go, but the prior failed to keep momentum and the latter hasn't even gotten off the ground.

 

[Idontcare]

And people complain that I don't need a $1000 i7 3960X. That CPU was $500 more.

Keyword being "need" versus "want".

In my case the QX6700 was purchased for business reasons, an investment that was made to generate profits.

I suppose one could take up the philosophical argument and say I did not need to make money to pay my bills, that I merely wanted to do that, and as such I did not need the QX6700.

But from the business perspective, a $1500 quad-core was worth the price premium in terms of the returns that would come from that investment versus the returns expected from a cheaper dual-core at the time.

 

[Keysplayr]

At the cost of typical nVidia power consumption.

Does anyone remember just how damn hot those nForce chipsets ran? I remember buying a nForce chipset and promptly going right back to VIA

I still have a 650i and a 790 Ultra still running every day.

 

[AdamK47]

Keyword being "need" versus "want".

In my case the QX6700 was purchased for business reasons, an investment that was made to generate profits.

I suppose one could take up the philosophical argument and say I did not need to make money to pay my bills, that I merely wanted to do that, and as such I did not need the QX6700.

But from the business perspective, a $1500 quad-core was worth the price premium in terms of the returns that would come from that investment versus the returns expected from a cheaper dual-core at the time.

I suppose needs have changed then or else you would be running an Intel hex core.

 

[Phynaz]

Rory's memo to the downtrodden was vacuous to the point of embarrassment. The guy truly does not help himself in avoiding garnering the perception that he is all talk and no thought.

with leadership like that oozing from above down onto the heads of the plebeians, how can the company not be kicking Intel's butt left and right?

"ambidextrous architectures"

WTF??!

Please AMD, get your stuff together. Please.

 

[Idontcare]

I suppose needs have changed then or else you would be running an Intel hex core.

I have long since shifted the programming model to use more cores that don't have to be in the same box now. (distributed computing if you will)

Paying a price premium for more cores in the socket only makes sense if it reduces the costs over that of the shared hardware in multiple boxes, but it rarely does in the consumer space (when trying to get the same performance out of that hex-core versus two quads in two boxes).

For example you might think you stand to save money buying just one PSU instead of two, but if you are scaling your PSU's for actual system loading then two good quality low-wattage PSU's can be bought for nearly the same price as a single good quality high-wattage PSU.

Likewise there is no savings in terms of ram/core when I consolidate boxes because I still have to buy the same amount of ram in terms of total GB.

Mobo's cost scale with complexity as well, slower quads are less demanding than high performance hex-cores, etc.

Of course it isn't one-for-one. There are some cost savings to be had in consolidating, but the end result tends to favor distributed computing if your programming model can support it.

The one thing that makes distributed computing expensive on a per-box footprint is the operating system expense. Microsoft carves out that $100+ per box, but if you consolidate you can get quite a few cores onto a single Win7 x64 Ult license.