Intel To Kill Off The Desktop As We Know It?

And in my experience, most people that are given a tablet as a laptop replacement, quickly ask for the laptop back. They are a pain when it comes to real work still.

Yeah but for home use.

I have seen many of my customers move to tablets and the laptop sits gathering dust.

The small business folks are still using the PCs/laptops but the domestic market is dying off hardware wise.

Folks are in denial if they think otherwise.

Take me for instance. I was sat most evenings with one of my laptops.

Three months ago I bought a Playbook for pennies.

The laptops now all sit there with flat batteries. Sure I still do all my daily work from my desktop work PC but 99% of my domestic stuff is now done on tablet.
 
"Intel To Kill Off The Desktop As We Know It?"
i was actually concerned with the headline until i saw it was on cnet. They run bogus shit like that all the time, because they have been losing readers for years. Unless every business in the world decides to run tablets, this is not happening any time soon. However, at the consumer level things are diff. If it was not for me and my gaming habit, we would not have any desktops in our house.
 
The article is wrong. I assure you that mobile processors (as in laptops) are not soldered to the circuit board. I know this from actually working on laptops.

Yes, and no. Apple's more recent notebooks have all had soldered CPUs. And, even though the market for VIA chipsets is miniscule, some of VIA's more recent MOBO/CPU combos have been soldered.

Is nice for the size factor, but would make it so that the most valuable piece which could be switched out in a notebook PC before would be unchangeable now. Also: MOBO goes, you pretty much are buying a new laptop.

I don't like this.
 
I see the sad future as a shift to "web appliances" for a lot of people.
There will be a niche for power users (animators, video editors) etc, but this will be a small niche compared to the majority of people that are doing data entry and reporting of some sore.
That kind of computing is moving to web based "cloud" computing.
 
I see the sad future as a shift to "web appliances" for a lot of people.
There will be a niche for power users (animators, video editors) etc, but this will be a small niche compared to the majority of people that are doing data entry and reporting of some sore.
That kind of computing is moving to web based "cloud" computing.

Yeah I'm continually surprised that how many on this forum somehow get the impression they and others like them are some kind of valuable/sizable proportion of the whole worlds IT community. When in fact they are probably less than half of 1%.

Sure they make a lot of noise but in terms of most corps revenue, pretty negligible.
 
There certainly is likely to be a convergence between mobile and desktop CPUs as the mainstream desktop actually does "die" some time in the future. Whether that's coming in a year-ish time frame is pretty unlikely. I do think the lowering of mainstream desktop TDPs is paving the way for this. Certainly some manufacturers, particularly like Apple with very few SKUs, will choose BGA SoCs and solder them on the board, but other manufacturers have 1) many SKUs using dozens of CPU models and 2) require flexibility that sockets provide for just in time manufacturing. Sockets aren't going anywhere for years because of that.

Enthusiasts are a separate category. For Intel to not make variants which cater to that segment is unlikely to end. Whether that's the repurposing of server chips (a la LGA1366), or simply using the "convergence chips" at higher clock speed and more power (ex: 35W higher end "mobile" chip @ 75W and much higher clock speed, with a heatspreader).

Other than tin foil hat theories, which are just silly, I don't see a single compelling reason that Intel would ditch sockets soon. One of the silliest claims is charlie's which he writes is that Skylake would "bring back" sockets. :rolleyes:
 
qm.gif
 
Intel just did a sony timer.

Dead MB = change MB + CPU.

thats fucked up man.
 
Well if intel does kill it off that will make AMD the leader in the desktop market. Good god that would be scary.
 
The article is wrong. I assure you that mobile processors (as in laptops) are not soldered to the circuit board. I know this from actually working on laptops.

I don't think your talking "mobile' in the same sense. I believe it's in refence to smaller devices like smart phones.

I really cannot imagine this taking place in the 'full size board non-mobile' world. Mother board makers would have to create 1 of every possible configuration or wittle down thier offerings to a small and restricted catalog. I can see the sense in it but unless the desktop market really has collapsed I think it'll never happen.
 
I don't think your talking "mobile' in the same sense. I believe it's in refence to smaller devices like smart phones.

I really cannot imagine this taking place in the 'full size board non-mobile' world. Mother board makers would have to create 1 of every possible configuration or wittle down thier offerings to a small and restricted catalog. I can see the sense in it but unless the desktop market really has collapsed I think it'll never happen.

Thing is they could just sell one 2Ghz dual core machine and one 3Ghz quad core machine and it would cover 90% of folks out there.

Apple have shown folks are prepared to be happy with limited options.

All it has to do is look nice and desirable. Market it right and they will buy any old crap.

Would the general computing world really notice if Acer/Fujitsu and a few other manufacturers disappeared? Prices might go up but refer to my point on marketing.
 
There have been soldered cpus for decades, I don't see what the fuzz is about.

They've been doing this on the low-end for years.
I remember seeing 386sx chips soldered onto the motherboard, just like many of the netbooks with atom chips.
 
Yeah I'm continually surprised that how many on this forum somehow get the impression they and others like them are some kind of valuable/sizable proportion of the whole worlds IT community. When in fact they are probably less than half of 1%.
Sure they make a lot of noise but in terms of most corps revenue, pretty negligible.

Yeah, truth is web based apps are a win-win for the business bottom line.
It is enabling companies to move away from expensive per seat licence business software.
The primary software package my company used wasn't that good. It had some serious bugs that was always was to be fix in the next version (which you had to buy).
Amazing how software companies can get away with this kind of thing. The product don't work like promised and their response is "oh, buy the new version, that is fixed".
Since our company switch to web based apps, we went from 3 full time IT people to 1.
And he mainly supports the software and writes reports for it. PCs were replaced by pre-configured laptops. If one fails (or gets full of spyware) he hot swaps it with a freshly mirrored one. ;)
 
How does this benefit Intel? Intel is not producing its own motherboard. I believe Foxconn does it.
 
How does this benefit Intel? Intel is not producing its own motherboard. I believe Foxconn does it.

Yes but which fails more often? CPU or motherboard?

Think about it. I've never had a CPU fail, but had a few motherboards.

I've moved a CPU through several motherboards. If you BGA (whatever) it to the motherboard then its time to buy another whole unit.

Intel doesn't care as it would then still have a sale (its CPU) with every new computer purchased/replaced.

Business as usual for them. It's just another commodity item.
 
While this doesn't seem totally implausible to me, it looks like all the rumours are being extrapolated from one roadmap slide which may or may not be real, and if real may or may not show the full picture.

Time will tell.
 
One of the big issues with BGA is they are prone to failures if they get too hot. People will be heat-gunning their motherboards all the time. This also sounds a bit like Intel's failed Slot 1.. Remember when they said a PGA socket was not stable for speeds over 200MHz? I have not had time to read the article, so I'm unsure if they want to control the motherboard sales by integrating the CPU or if they will sell BGA parts to OEMs... If they sell the BGA parts, how difficult would it really be to engineer a socket that accepts a BGA CPU?

Slot 1 had nothing to do with processor speeds.

Originally, L2 cache was located on the motherboard. Slot 1 was used as a way to move the cache closer to the processor; the processor and cache were both located on a daughtercard and the processor had a dedicated link to that cache allowing for faster speeds. Once it become economical to integrate the L2 cache onto the processor die itself, there was no more need for slots.
 
Slot 1 had nothing to do with processor speeds.

Originally, L2 cache was located on the motherboard. Slot 1 was used as a way to move the cache closer to the processor; the processor and cache were both located on a daughtercard and the processor had a dedicated link to that cache allowing for faster speeds. Once it become economical to integrate the L2 cache onto the processor die itself, there was no more need for slots.

Exactly, I worked for Intel back then and it was simply a bridge solution until the other technologies they were working on enabled them to fully integrate the capability on die :cool:
 
Suuuurrrreeee... Lets just kill off 80% of our Revenue Generation. No problem.
 
Clearly nonsense, as desktop or no there is still a very obvious advantage to allowing manufacturers to offer a selection of different processors, video, and memory configurations at different prices in all but the tiniest of form-factors where its just not practical.
 
I'm not too worried.

As new ways of what "computers" can do for consumers come up, there will be more need for computing power in the home...

I think that we have not even began to scratch the surface of what computing can do for us.

As a hardware enthusiast, I think our future will be amazing, and that this soldered vs unsoldered stuff is completely irrelavent.

Personally I cant wait for the day the mobo, video cards, etc, go away...I want hardware that I can select and personalize, that allows me play games, compute, automate, enjoy media, wash my clothes, cook my food, get sexual satisfaction, educate me, vote, get rid of useless polititians and middle people, travel, time travel, get stuff I want, and work less, and bring peace to the Middle East.

Oh yeah, and spend more time in Genmay.

So fuck soldered vs unsoldered cpus.

/rant
 
Suuuurrrreeee... Lets just kill off 80% of our Revenue Generation. No problem.

How exactly? The computer/laptop whatever needs a CPU. If the CPU is soldered to the motherboard and the motherboard is more likely to fail and does fail then that's another CPU sale for Intel.

They win either way. I reckon it would increase sales as there are more points of failure that would increase the need for another CPU/board combo.

Capable PCs are pretty much as cheap if not cheaper than tablets. Plenty of folks will pay $500+ every 18 months for a new tablet.
 
This does not bother me at all :) When I upgrade every few years I allways buy a new motherboard for the new cpu anyways.

Except when you buy a motherboard and CPU, you get to chose what motherboard and what CPU you want to use.
 
How exactly? The computer/laptop whatever needs a CPU. If the CPU is soldered to the motherboard and the motherboard is more likely to fail and does fail then that's another CPU sale for Intel.

They win either way. I reckon it would increase sales as there are more points of failure that would increase the need for another CPU/board combo.

Capable PCs are pretty much as cheap if not cheaper than tablets. Plenty of folks will pay $500+ every 18 months for a new tablet.

Intel doesn't sell to consumers though and tablets aren't competing with each other beyond memory and size ... Intel's current customers (the OEMs like Dell, HP, Lenovo, ASUS, etc) are in intense and low margin competition with each other ... their primary competing points are in customer specific configurations for laptops and desktops (multiple processors, multiple speeds, different numbers of cores, etc) ... most of these configurations can use common hardware so they only need to replace the CPU ... with hard mounted CPU/MB combos this would not work ... I am pretty confident that this assumption is an error and not a strategy change from Intel since there is absolutely no way that Intel's customers (the OEMs) would want to go with this

The only way this model would work for Intel is if their customer base decided they only wanted 5-6 different CPU flavors ... and I find that EXTREMELY unlikely ;)
 
Let's not forget that Intel abandoned the memory market, but that was due to the Japanese was killing them on the market. The only competition that Intel has is AMD and ARM but both are hardly killing them. AMD processors are "inferior goods" (Not saying bad, just saying they are the cheaper alternative) and ARM produces designs that other companies manufacture.
Intel would be crazy to give up any market share.
 
The only way this model would work for Intel is if their customer base decided they only wanted 5-6 different CPU flavors ... and I find that EXTREMELY unlikely ;)

Yest people and corporations are not updating their PCs and laptops as often as they did before dual core CPUs came out.

Needing a huge range of CPU options for most businesses ended around 2006.

Sure some specialist operations need some juice but the rest of the world that gets by on email and Office just doesn't need it.

A $150 dollar commodity box will do them. You can strip out half the add ons we take for granted on a PC/motherboard. Basically a think client type box with a HDD in it.

As for average users they dont care either. Remember corps and average joes are the majority.

You and the rest of us here are a very very small minority.
 
It would make people less likely to upgrade. Instead of going from an i7 5600k to an i7 6600k for the cost of a CPU, you'd have to factor in the motherboard. If you got a higher end board, that would be a $1000 update, which would make you think much longer about it and make upgrades less tempting and makre more people sit on their purchases for longer. Breaking it up into parts makes everything seem much cheaper. :D

New features would be harder to add support for mid cycle too. As the adoption rates would be way way down as intel products would be way more expensive.
 
Essentially the world of computing as we know and love it, is changing.


And it ain't going our way.
 
After this has been put to work, we will look back at Extreme Edition Intel processors and giggle, wondering how we ever could have thought that was a rip-off, as Intel posts yet another record breaking quarter and AMD finally, finally dies for real.

Ma'Bell those criminals before it's too late.
 
Intel needs to be competitive in the phone and tablet market.
The phone and tablet market uses soldered soc chips since
it is more power efficient and going into the same device.
Samsung wants the chip on the phone's "motherboard".
THIS WILL NOT EFFECT HIGH END DESKTOP CPUs.
Asus will still make motherboards and Intel will still make socketed cpus.
The only debate is whether laptops will use soldered or socketed cpus.

P.S. Even if Intel was so stupid to make all cpu's not socketed,
a company would make a bridge to make the cpu socketed
and then sell motherboards that it plugs into.

Intel is not that stupid, they don't want to give all their high end cpu
market to AMD, they just wan't to be in cell phones and tablets.
 
Who's to say that going to a BGA format doesn't just mean Intel is making a return to a Slot A style processor?

There are advantages to such a layout. Like not bending pins. I'm sure there are many disadvantages too. I have no idea what, since that's not my trade. I'd assume something like harder to get power there or slower speed. *shrugs* Mounting a gigantic heatsink would suck too.
 
For all people's protestations to the contrary, Intel is not an Apple or a Microsoft. They don't usually try to force people or their customers into tight pigeon holes. They will make technology changes that benefit themselves AND their customers; and they will change form factors to suit their product and manufacturing requirements. However, these are usually done in close consultation with their OEM customers.

Generally consumers tend to be a little picky about what they buy. For cars they like different models (sports vs sedan), for TVs they like different sizes and resolutions, they are also very cost conscious (especially in today's economy). With tablets (because the form factor is relatively new, in its latest incarnations) they have been willing to treat the item as more of a value meal (upsizing screens and memory but leaving all other items the same). This has allowed that form factor to use SMT chips since their is only one configuration and they just mount different amounts of memory.

Intel's OEM customers (HP, Dell, Acer, Lenovo, etc) have survived with the car model. They produce a chasis that supports multiple configurations and they build to order rapidly so they customers still get quick delivery. They can only do this because the sockable technology allows them to build the core boards ahead of time and then insert the CPU, memory, and other internal components according to the customer's BTO instructions. The only way they could get rid of this model is if the customer was willing to pursue the value meal approach for computers (a highly unlikely scenario in my estimation). As long as the customers want to have different families of CPUs (i3, i5, i7, Xeon, etc) and different performance levels within those (clockspeed, cores, cache, etc) then it is too Intel's advantage to produce as many configurations as the market can support in reasonable volumes. There is little advantage for Intel to force an end to the socketable technology until their customers stop ordering in Chinese menu style (give me I5 with quad cores and enhanced speed; and an i7 with hexa core; to go please :) ) ... :D
 
As long as the customers want to have different families of CPUs (i3, i5, i7, Xeon, etc) and different performance levels within those (clockspeed, cores, cache, etc) then it is too Intel's advantage to produce as many configurations as the market can support in reasonable volumes.

Ask 100 people at random in the street what i3/i5 and i7 stand for and I bet 98 wont have a clue.

Most may have heard of Intel but they wont really care why.

I get this with most of my customers. You can ask them what type of CPU they want and if you try to explain they just glaze over and say "Look I really don't care! As long as it can do Ebay and Skype and doesn't cost me a fortune then I'll leave it up to you!"

Most customers don't care what's in the box already.
 
Ask 100 people at random in the street what i3/i5 and i7 stand for and I bet 98 wont have a clue.

Most may have heard of Intel but they wont really care why.

I get this with most of my customers. You can ask them what type of CPU they want and if you try to explain they just glaze over and say "Look I really don't care! As long as it can do Ebay and Skype and doesn't cost me a fortune then I'll leave it up to you!"

Most customers don't care what's in the box already.

Maybe there is a regional element at play here ... I think in the US and EU people are more likely to want the different options available (even if they don't fully understand what they do) ... the ones who don't are more likely to go the ARM or Atom route as they might be at the extreme low end of pricing which forces them into tablets or netbooks

Also, don't discount enterprise (one of the few places besides gamers that still buy desktops). Enterprise customers still have multiple needs AND they generally do have some idea what they want ... even if it is no more than they want i3's for the production desktops, i5's for their desktop replacement laptops, and i7's or Xeons in their workstations

Intel will switch to SMT chips when that is what their customers are demanding ... this report doesn't back up its assumption with any indications that Lenovo, HP, and Dell (three of the biggest computer manufacturers in the world) are desiring to pursue this type of strategy ;)
 
Back
Top