AMD to Cut 5% of Workforce

[Tsavo]

I hate to jump on the bandwagon -- but Intel 1151 chips do seem pretty damn rare in my neck of the woods. Many of the unlocked sku's are constantly sold out on Newegg (which by my guess is among the highest volume online sellers in the USA).

I'm starting to wonder if the yields are not as rosy as Intel would have us believe -- because these chips remain pretty difficult to buy. Either that -- or demand is huge for a desktop product (which I'd say is somewhat unlikely given that PC's in general are still in decline).

PC's in general are in decline, but gaming PC's aren't. Besides, I wouldn't doubt for a second that most of Intel's current effort with Skylake is going into the production of mobile products.

 

[cytg111]

PC's in general are in decline, but gaming PC's aren't. Besides, I wouldn't doubt for a second that most of Intel's current effort with Skylake is going into the production of mobile products.

The market leading supplier of x86 cpu's, Intel, is producing one, ONE, core to span from low level tablets to full fledged workstations, going from 4 to 95 watts. Intels number one priorty has been getting to core-m levels, core-m levels where they meet surface3/4 kind of products. Throttling like hell, still, well done, clap, clap, well done.
But you have to wonder, if that stretch, 4W->95W is a tad but much for one uarch. What happens if Zen is a success, and ARM continues it's march on from the south. One could find ones lines stretched too thin. One core to rule them all?

 

[shady28]

This is what Intel is rightfully worried about. AMD just doesn't matter much on the desktop/laptop side.

And yes, this is a real shipping product. Windows Continuum. It makes me wish I had waited a couple more months on getting a new phone.

 

[ShintaiDK]

This is what Intel is rightfully worried about. AMD just doesn't matter much on the desktop/laptop side.

And yes, this is a real shipping product. Windows Continuum. It makes me wish I had waited a couple more months on getting a new phone.

Exactly. PC ending in the gadget segment.

Gadget segment essentially means all money get drained out of it.

 

[superstition]

Exactly. PC ending in the gadget segment.

Gadget segment essentially means all money get drained out of it.

A laptop looks more convenient than that.

 

[superstition]

I mean, I'm an economics nerd but I mean everyone should read economics in one less if you want to actually to understand what's going on in the world.
Otherwise, we're not having a remotely interesting conversation.

So many things posted here break basic rules of economics and it makes me want to cry.

I agree. I dealt with a Ph.D in Economics who abused people to extract extra money from the university.

It didn't matter how wrong he was because he wanted his money and that was that.

Sorta like Gramm-Leach-Bliley.

arrogance + greed = spectacle

 

[cytg111]

This is what Intel is rightfully worried about. AMD just doesn't matter much on the desktop/laptop side.

And yes, this is a real shipping product. Windows Continuum. It makes me wish I had waited a couple more months on getting a new phone.

I must admit, Microsoft is doing something differently these days, cause that surface lineup looks damn sweet, i want it, - and a windows10 phone to go with it that integrates seamlessly across devices. Thats value in my book. Make my desktop a 4c/8t, surface 2c/4t and phone 1c/2t skylake-like sku's and we're talking business.

 

[shady28]

I must admit, Microsoft is doing something differently these days, cause that surface lineup looks damn sweet, i want it, - and a windows10 phone to go with it that integrates seamlessly across devices. Thats value in my book. Make my desktop a 4c/8t, surface 2c/4t and phone 1c/2t skylake-like sku's and we're talking business.

Well, I'm thinking in about 5-10 years your phone will be your desktop. That's what Windows Continuum is about.

 

[Fjodor2001]

Yes, please give us less performance in smaller form factor. The tech companies should all vigorously strive to reduce performance and productivity. That should be the number one goal that they devote R&D to. Good times ahead.

 

[cytg111]

Yes, please give us less performance in smaller form factor. The tech companies should all vigorously strive to reduce performance and productivity. That should be the number one goal that they devote R&D to. Good times ahead.

If the extra performance is not warranted, why keep the desktop on lifesupport? Who cares?
Beyond that, im sure the upcoming VR/Augmented revolution will hit hard and keep the need for a household processing-nerve relevant. VR, specially augmented, IMO, will come to play a massive role over the next 10 years, and its gonna cost cpu cycles. lots of em.
edit: oh, and AI too..

 

[Fjodor2001]

If the extra performance is not warranted, why keep the desktop on lifesupport?

Chicken and egg. Why develop SW requiring more performance when that performance is not available anyway, since desktop performance improvement is so crappy? Nobody would use that SW anyway, since their PCs would be too slow for it.

 

[cytg111]

Chicken and egg. Why develop SW requiring more performance when that performance is not available anyway, since desktop performance improvement is so crappy? Nobody would use that SW anyway, since their PCs would be too slow for it.

Software devs always go big, then scale back cause crappy evolution wont give them enough compute cycles. The devs and their software is there, ill give you 99.
edit for context: ST performance

 

[ShintaiDK]

Its not a chicken and an egg issue. Its about cost, time and diminishing returns.

 

[Fjodor2001]

Software devs always go big, then scale back cause crappy evolution wont give them enough compute cycles. The devs and their software is there, ill give you 99.
edit for context: ST performance

Not sure what you mean. What SW are you talking about?

 

[Fjodor2001]

Its not a chicken and an egg issue. Its about cost, time and diminishing returns.

No. Why would it be? If the CPU perf was there to be used by SW, it could just be made use of. It's not about time, cost or diminishing returns. It's about crappy CPU perf increase in the mainstream desktop segment.

 

[Phynaz]

Yes, please give us less performance in smaller form factor. The tech companies should all vigorously strive to reduce performance and productivity. That should be the number one goal that they devote R&D to. Good times ahead.

Yes, we don't want form factors that enable new computing paradigms. PC master race and all that.

I assume you are too young to remember when the PC was lower performance in a smaller form factor.

 

[ShintaiDK]

No. Why would it be? If the CPU perf was there to be used by SW, it could just be made use of. It's not about time, cost or diminishing returns. It's about crappy CPU perf increase in the mainstream desktop segment.

No. And if you really cracked the code for multithreaded software in this scale you need to go work on it. Because there is countless billions of easy money waiting for you then. You would make Steve Jobs look like a poor homeless fool then.

Even servers with countless cores depends mainly on concurrent users rather than direct multithreading in many case. What can be multithreaded within the norm of reason have already been so.

 

[Fjodor2001]

I assume you are too young to remember when the PC was lower performance in a smaller form factor.

You're assuming that because...?

Also, what does it have to do with what I wrote anyway?

 

[Fjodor2001]

No. And if you really cracked the code for multithreaded software in this scale you need to go work on it. Because there is countless billions of easy money waiting for you then. You would make Steve Jobs look like a poor homeless fool then.

Even servers with countless cores depends mainly on concurrent users rather than direct multithreading in many case. What can be multithreaded within the norm of reason have already been so.

1. May I ask what educational and professional experience you have in the SW engineering field that makes you arrive at that conclusion? What experience do you have in designing SW?

2.The comment I made is not specific to MT performance. ST performance has not improved much either.

3. Did you read the recent AT article on mobile CPU core counts, concluding that SW was in general already heavily multi-threaded?

 

[ShintaiDK]

I have major enterprise class experience in the usage of the matter. But I question what experience you have, if any. Because what is seen and what you proclaim seems to be dimensions apart.

Why do you think ST performance haven't moved much? Even with billions thrown at it.

Using browsers and such as a way to show multithreading on mobile devices. You can do better. Why do you think Apple only uses 2 fast cores, even using a heavy 8MB cache with them. I mean they must be completely idiots according to your statement. Slap in a cheap 10 core MediaTek CPU and be done with it, right?

 

[Phynaz]

You're assuming that because...?

Also, what does it have to do with what I wrote anyway?

I assume that because your posts are nonsensical and show a distinct lack of experience.

As far as your second question, perhaps it would be better if you were to go back and read what you wrote.

 

[Fjodor2001]

I have major enterprise class experience in the usage of the matter. But I question what experience you have, if any. Because what is seen and what you proclaim seems to be dimensions apart.

Why do you think ST performance haven't moved much? Even with billions thrown at it.

Using browsers and such as a way to show multithreading on mobile devices. You can do better. Why do you think Apple only uses 2 fast cores, even using a heavy 8MB cache with them. I mean they must be completely idiots according to your statement. Slap in a cheap 10 core MediaTek CPU and be done with it, right?

"major enterprise class experience"? What does that mean? What SW engineering have you studied, and have you actually designed and written any SW?

As for ST performance improvement, I don't see why you're even asking about it. Just look at the ST perf increase we've seen the last few years on desktop and it's obvious how crappy it's been. Same goes for MT performance. Core count has been standing still for a long time at 4 too.

Regarding MT SW, the AT article did not only mention browsers. The conslusion was that SW in general was heavily multi threaded. Read the article again if you don't believe me.

 

[ShintaiDK]

"major enterprise class experience"? What does that mean? What SW engineering have you studied, and have you actually designed and written any SW?

As for ST performance improvement, I don't see why you're even asking about it. Just look at the ST perf increase we've seen the last few years on desktop and it's obvious how crappy it's been. Same goes for MT performance. Core count has been standing still for a long time at 4 too.

Regarding MT SW, the AT article did not only mention browsers. The conslusion was that SW in general was heavily multi threaded. Read the article again if you don't believe me.

So let me see. You know better than the entire industry and the only issue is...more cores missing. Because that would solve everything?

And for ST performance you simply dont even a tiny clue on the matter. And all you can come up with is "crappy".

I read the article. There isn't any easy solution, nor did they show anything new.

If you think you are so good at coding multithreaded software for what is inherit serial code. I think you should use some more time on that. Because multiple billions and fame is waiting for you, perhaps even a Nobel prize. Unless of course...its hot air and no show.

 

[Fjodor2001]

I assume that because your posts are nonsensical and show a distinct lack of experience.

Just because you don't understand them does not mean they are nonsensical. And FYI, I have a M.Sc. in C.S. and 10+ years of experience in that field. What relevant merits do you have yourself?

As far as your second question, perhaps it would be better if you were to go back and read what you wrote.

Your comment still makes no clear connection to what I wrote. As usual.

 

[Fjodor2001]

So let me see. You know better than the entire industry and the only issue is...more cores missing. Because that would solve everything?

Not what I wrote, read again. I specifically said my comment was not only for MT. How did you miss that!?

And for ST performance you simply dont even a tiny clue on the matter. And all you can come up with is "crappy".

So you consider the desktop ST performance improvements the last few years to be good/impressive? Then we'll simply have to disagree on that!

I read the article. There isn't any easy solution, nor did they show anything new.

If you think you are so good at coding multithreaded software for what is inherit serial code. I think you should use some more time on that. because multiple billions and fame is waiting for you. Unless of course...its hot air and no show.

It concluded that most of the SW was heavily multi threaded, and that more cores was the way to go (especially to reduce power consumption). So nothing new you say? Well it contradicts your claims anyway...