Intel CES 2014...what the?

[norseamd]

these new applications could be a good direction for them in my opinion. but they are the cyberpunk evil mega corporation before the time so buy them if you want but do not be surprised if nano robots suddenly infect you (not yet at least).

Intel doesn't care about desktop users

been know for a long time. does not seem like intel even cares for themselves.

maybe the workstation and gamer markets will merge. then you will still have the powerful hardware.

 

[BrightCandle]

This presentation really shows how much Intel is struggling with its CPU. Its one thing for all of us to realise this miniscule improvements in IPC aren't adding up but it's a whole different problem when Intel's picture of it's own future is not really in high performance CPUs. This whole approach to performance stopped working 10 years ago which is why they went many core. Then just two steps into that they stopped again because that didn't work and they have done very little to correct the problem ever since.

 

[norseamd]

This presentation really shows how much Intel is struggling with its CPU. Its one thing for all of us to realise this miniscule improvements in IPC aren't adding up but it's a whole different problem when Intel's picture of it's own future is not really in high performance CPUs. This whole approach to performance stopped working 10 years ago which is why they went many core. Then just two steps into that they stopped again because that didn't work and they have done very little to correct the problem ever since.

multi core has worked i think. they could put more cores on but there is not very much software support. that is getting better though. mores cores would definitly be usefull. hsa might offer better improvements though.

 

[Homeles]

This presentation really shows how much Intel is struggling with its CPU. Its one thing for all of us to realise this miniscule improvements in IPC aren't adding up but it's a whole different problem when Intel's picture of it's own future is not really in high performance CPUs. This whole approach to performance stopped working 10 years ago which is why they went many core. Then just two steps into that they stopped again because that didn't work and they have done very little to correct the problem ever since.

This is one of the most ridiculous statements I've seen in a while. How many times must I say this: Intel's much quieter under the leadership of their new CEO. This has absolutely nothing to do with how Intel's processors are doing.

I'm extremely confused how you came to this conclusion. It couldn't be more contrived.

 

[WaitingForNehalem]

This is one of the most ridiculous statements I've seen in a while. How many times must I say this: Intel's much quieter under the leadership of their new CEO. This has absolutely nothing to do with how Intel's processors are doing.

I'm extremely confused how you came to this conclusion. It couldn't be more contrived.

You know, if you are going to keep making arrogant comments and act as if your opinions are greater than everyone else's, at least lay out your credentials.

 

[Homeles]

You know, if you are going to keep making arrogant comments and act as if your opinions are greater than everyone else's, at least lay out your credentials.

Same to you.

Isn't that what this thread is all about for you? Intel's pleasing the masses at my expense, and they're stupid for it?

 

[Lonyo]

Intel has always been about pleasing the masses.
The masses just used to be pleased by mainframes, then desktops, then laptops.

Now they are pleased by phones and tablets, and in the future they (it is assumed) will be pleased by wearables and computers being everywhere.

How do you remain relevant when you are Intel? You get x86 everywhere (hey, who's said that before?) by making everything you can with an x86 processor. Which is what Intel spent most of the last few decades doing except where markets were small or irrelevant.

Now those markets are no longer irrelevant, so Intel is making a push everywhere.

And like others have said, Intel are top dog because they spend a lot to stay top dog.
How do you spend a lot? Earn a lot first. How do you earn a lot? Ship a lot. How do you ship a lot? Be in large or growing markets. The PC market is large, but not really growing. So they expand to growth markets alongside their core market.

Qualifications: I work in finance.

 

[tential]

This presentation really shows how much Intel is struggling with its CPU. Its one thing for all of us to realise this miniscule improvements in IPC aren't adding up but it's a whole different problem when Intel's picture of it's own future is not really in high performance CPUs. This whole approach to performance stopped working 10 years ago which is why they went many core. Then just two steps into that they stopped again because that didn't work and they have done very little to correct the problem ever since.

I think you're just completely out of touch with the needs of the masses and are focusing only on yourself.

I don't care about CPU performance anymore. If I'm the AVERAGE consumer, Conroe is fast enough for what I need. If they make a processor TWICE as fast as Haswell only enthusiasts will be excited that's it. The REST of people, instead of buying that 200 dollar Haswell processor, will buy the bottom of the line Haswell since ALL THEY NEED is something to do web browsing/word processing. This idea that intel is "struggling" to make faster processors is crazy. There isn't a DEMAND for faster processing.

Trust me, I've told so many people (class mates, parents, etc.) about intels new processors including Haswell, and they couldn't care less about upgrading, even if they have a PC that's 5 years old (or laptop). It just doesn't matter to them. Their current hardware is enough for them that they can't justify spending even 500-600 dollars to replace what they have when what tehy have works.

Intel has always been about pleasing the masses.
The masses just used to be pleased by mainframes, then desktops, then laptops.

Now they are pleased by phones and tablets, and in the future they (it is assumed) will be pleased by wearables and computers being everywhere.

How do you remain relevant when you are Intel? You get x86 everywhere (hey, who's said that before?) by making everything you can with an x86 processor. Which is what Intel spent most of the last few decades doing except where markets were small or irrelevant.

Now those markets are no longer irrelevant, so Intel is making a push everywhere.

And like others have said, Intel are top dog because they spend a lot to stay top dog.
How do you spend a lot? Earn a lot first. How do you earn a lot? Ship a lot. How do you ship a lot? Be in large or growing markets. The PC market is large, but not really growing. So they expand to growth markets alongside their core market.

Qualifications: I work in finance.

Exactly. Most people ITT post from an engineering/enthusiast perspective. Intel doesn't care about "being the best" or making "huge gains every year". They care about ONE thing, making profits. That's why you start a business. I don't doubt they can make a processor twice as fast as Haswell(not literally but you know what I mean). It just is that the demand ISN'T there for such a thing.

 

[witeken]

Intel has always been about pleasing the masses.
The masses just used to be pleased by mainframes, then desktops, then laptops.

Now they are pleased by phones and tablets, and in the future they (it is assumed) will be pleased by wearables and computers being everywhere.

How do you remain relevant when you are Intel? You get x86 everywhere (hey, who's said that before?) by making everything you can with an x86 processor. Which is what Intel spent most of the last few decades doing except where markets were small or irrelevant.

Now those markets are no longer irrelevant, so Intel is making a push everywhere.

And like others have said, Intel are top dog because they spend a lot to stay top dog.
How do you spend a lot? Earn a lot first. How do you earn a lot? Ship a lot. How do you ship a lot? Be in large or growing markets. The PC market is large, but not really growing. So they expand to growth markets alongside their core market.

Qualifications: I work in finance.

Exactly. Intel did a bad thing by not putting Intel inside the iPhone. They (BK) don't want make that mistake again, so at CES they're focusing at the future. It makes sense to me, but they get a lot of criticism for that. They have a decent roadmap for mobile which they showed at IDF so they wanted to show other things at CES. This presentation doesn't change any roadmap (mobile or desktop).

 

[Qwertilot]

Medium term its even a bit worse than core being good enough - at some point not that terribly far away stuff like atoms/arm based mobile SOCs will also be good enough for mainstream lap/desktop use.

That'll throw core out of quite a few machines and inevitably rather dent Intel's margins in some of their markets.

Nothing they can do about that now they're committed to making Atom good. So they're hunting for lots of other revenue streams/models before it gets too bad. Plenty of time/resources and seemingly the will.
(They do look to be coping with all of this much better than microsoft.).

 

[sandorski]

I'm biased against them, so use salt, but Intels record on coming up with ideas is a mixed bag. They've had so many flops, so many Products/Technologies that were going to change everything, but didn't or in some cases sorta did, just not in the way intended.

Flopish:

AGP-an interface to give Graphics cards needed Memory Bandwidth from System Memory as demoed by the I740 video card. AGP itself was a success, just not in the way intended, although that was largely due to an unexpected fall in Memory Costs.

Itanium-was going to change everything, hardly changed anything.

RDRAM-was the future of Memory, obsolete by the time it was used

Pentium4-was going to reach 10ghz! Barely achieved 33% of that

64bit-was totally unnecessary

A lesser company would have failed with these kinds of errors, but Intel has been so successful in other areas and so dominant in their Marketshare/Revenues that they can just throw ideas onto the Market hoping for success.

 

[scannall]

I don't see a problem. CES is a good place to introduce new products/categories. The OEM's really don't care what they introduce there, and it won't change their buying decisions. While speeds and feeds are the fun stuff for us techies, we're a pretty small market.

 

[norseamd]

64 bit has been a success in my opinion. we would be stuck at 4 gb without it. most programs coming out today now require that or more. that 16 gb of ram would not be possible without 64 bit.

 

[sandorski]

64 bit has been a success in my opinion. we would be stuck at 4 gb without it. most programs coming out today now require that or more. that 16 gb of ram would not be possible without 64 bit.

Yes it has been, but it was AMD who pushed it. My point probably wasn't clear about it, but Intel is on record at the time as saying it was unnecessary.

 

[norseamd]

well i do not really listin to intel. they think of marketing and profits. i think we could use more core and more memory easily but the programmers will not make software that does. they say it is complex but programming already is. we will have much better programs when they decide to stop being lazy.

 

[tential]

Yes it has been, but it was AMD who pushed it. My point probably wasn't clear about it, but Intel is on record at the time as saying it was unnecessary.

I have no ill will against intel, really don't care who makes what I use as long as it's not terrorists.

But they have been behind consistently, and I think they FINALLY realized what consumer trends are. The issue is people here want CPU updates and things of that nature. Sorry, "Intel debuts 14 nm process at CES!!!!!" Doesn't get you a headline on a majority of websites. NO ONE outside of people on tech websites know what that means

"Intel debuts Jarvis, your own personal assistant!" Is something that gets people interested.

Meanwhile, us in the tech world, we already know 99% of what intel is planning, and if we want to know more, we don't need CES, or any huge convention to do it. Intel can simply leak the slides, and we have all the information we need until we get the product on hand. A show like CES though it doesn't benefit intel much to talk about processors when there are tons of press at these shows that may not understand most of the terminology talked about. It does better for them to talk about their other products.

 

[DigDog]

a management screwup doesn't mean the engineers aren't working.

 

[ShintaiDK]

Yes it has been, but it was AMD who pushed it. My point probably wasn't clear about it, but Intel is on record at the time as saying it was unnecessary.

Intel actually wanted us to leave x86 for IA64 at that time. Its a shame we didnt go the IA64 route on the long term. But thats a bill we have to pay later on.

You should really review your list again. If thats all you could find, its not looking bad at all.

AGP was a success and leaped graphics forward. Nothing more to say about that.

RDRAM got killed by the corrupt cartel. Execs jailed later on. Consumer yet again stuck with the worse solution. And now its getting really painful with DDR4 and one DIMM per channel only. And all the future hopes are put on stacked memory.

Pentium 4 and 10Ghz was before speedracer designs ran into physical trouble. And remember the ALUs actually ran up to 8Ghz on stock chips. Something AMD funny enough repeated with BD that was supposed to run 1-2Ghz higher. However both companies should have seen this from their process engineers warnings and I agree it was a silly PR stunt/management screwup for both. Only real way to (drasticly) increase performance today (with x86/ARM) is by new instructions. Frequency is dead so to say. And "moar cores" is simply a bandage. ARM is in the same trouble as x86 already. TDP increase doesnt work anymore either as we already seen with AMDs 220W chips and the failed big.LITTLE chips.

64bit, again, read about IA64. There is a reason why IA64 today is 11 issue wide. And Core is 4 issue wide.

 

[NTMBK]

RDRAM got killed by the corrupt cartel. Execs jailed later on. Consumer yet again stuck with the worse solution. And now its getting really painful with DDR4 and one DIMM per channel only. And all the future hopes are put on stacked memory.

Was RDRAM really superior, though? From what I've read it did have higher bandwidth, but it also suffered from high latency. (Wasn't as up on computers 15 years ago, so I don't know this stuff too well.)

 

[pmv]

Flopish:

AGP-an interface to give Graphics cards needed Memory Bandwidth from System Memory as demoed by the I740 video card. AGP itself was a success, just not in the way intended, although that was largely due to an unexpected fall in Memory Costs.

Itanium-was going to change everything, hardly changed anything.

RDRAM-was the future of Memory, obsolete by the time it was used

Pentium4-was going to reach 10ghz! Barely achieved 33% of that

64bit-was totally unnecessary

A lesser company would have failed with these kinds of errors, but Intel has been so successful in other areas and so dominant in their Marketshare/Revenues that they can just throw ideas onto the Market hoping for success.

The Pentium 4 clearly didn't work out the way Intel planned, but I'm not sure if it was entirely a 'flop' (if anyone knows better I'm happy to be corrected). Don't know enough to say about some of the others.

The only one of those that I would definitely agree about was RDRAM - I seem to remember many people (e.g. Tom of Tom's Hardware, who went on about it at length with Germanic thoroughness!) thinking that was a bad idea right from the start, and it seems like events bore them out (I may be biased by having had one of those infamous faulty RDRAM memory controllers in my work PC at the time! Constant BSODs).

Seems to be true though that if you are big enough you can get away with mistakes that kill smaller entities and just rumble on - the principle seems to apply to the careers of some Hollywood stars also!

 

[pmv]

Intel has always been about pleasing the masses.
The masses just used to be pleased by mainframes, then desktops, then laptops.

I mostly agree, but I'm not sure mainframes were ever a product for the 'masses'!

Is it not more that the tech business has become generally more about 'the mass market' over the decades, and Intel move with it like everyone else?

 

[DaZeeMan]

As I remember, in the early RDRAM days, the biggest problem with RDRAM was the expense. SDRAM was significantly cheaper by comparison, so a lot of people kept going the SDRAM route.

If RDRAM had been implemented more cheaply, it might have been another story. Plus the other reasons mentioned above.

I get that Intel was pandering to the audience this week, and trying to get people a little excited about the possibilities. If I had to guess as to what their 'goal' was, I'd say that they were looking to lure the up and coming 'out of box thinkers' to the Intel brand.

Which ends up not being as much fun for us number guys...


I linked a presentation in another thread where some guy was talking about the transition to 14nm, which he noted as being more effective for power savings than for performance. Intel's latest CPUs seem to bear that thinking out, and indeed with the 'high performance user' market not being what it once was, you have to respect the strategy.

The thing that Intel has done in the last couple of years which has been impressive to me is drastically increasing the performance of their iGPUs. They essentially went from lackluster performance (most computer enthusiasts did not take them seriously) to quite respectable in just a couple of years. With the amount of intellectual muscle that Intel has at their disposal, I can see the next generation(s) of Intel iGPUs giving NVidia and AMD some serious headaches.

It does look like Q3 2014 should be very interesting for you Intel gamer types:
http://www.guru3d.com/news_story/intel_haswell_e_roadmap_confirms_launch_in_q3_2014.html


For me, the bigger news this week was Alan Mulally deciding to stay put at Ford. Kinda makes you wonder where the next Microsoft leadership is going to come from...

 

[sandorski]

The Pentium 4 clearly didn't work out the way Intel planned, but I'm not sure if it was entirely a 'flop' (if anyone knows better I'm happy to be corrected). Don't know enough to say about some of the others.

The only one of those that I would definitely agree about was RDRAM - I seem to remember many people (e.g. Tom of Tom's Hardware, who went on about it at length with Germanic thoroughness!) thinking that was a bad idea right from the start, and it seems like events bore them out (I may be biased by having had one of those infamous faulty RDRAM memory controllers in my work PC at the time! Constant BSODs).

Seems to be true though that if you are big enough you can get away with mistakes that kill smaller entities and just rumble on - the principle seems to apply to the careers of some Hollywood stars also!

P4 was definitely a Marketing/Sales success, just that it wasn't what Intel had hoped it would be. That's why I used the word "Flopish".

 

[witeken]

Can someone explain what happens at ~41:00?

http://www.intel.com/content/www/us/en/events/intel-ces-keynote.html

 

[WaitingForNehalem]

Can someone explain what happens at ~41:00?

http://www.intel.com/content/www/us/en/events/intel-ces-keynote.html

What do you mean? Are you asking about the DreamWorks CEO?