Bookmark and Share


Growing importance of ultra-thin notebooks that require chips in ball-grid array (BGA) form-factor as well as shift towards low-power micro-architectures will reduce importance of microprocessors in LGA and µPGA packaging for Intel Corp.

Nowadays, the vast majority of central processing units manufactured by Intel Corp. come in either land-grid array (LGA) or micro pin-grid array (µPGA) packages and are tailored for maximum performance on the micro-architecture level. When Intel introduces its code-named Broadwell micro-architecture in 2014, the company will shift focus to lowest possible power consumption and the amount of processors with 15W (ultra low voltage, ULV) and ~10W (extreme ULV) thermal design power will increase considerably.

Thanks to increased amount of chips with very low power consumption, the share of ultrabooks, ultrathin laptops and various ultrathin convertibles will get considerably higher than it is today and the demand towards chips in µPGA form-factor will therefore decrease. Considering the ongoing shift to notebooks, the demand towards LGA processors will also get lower by 2014 than it is today.

In a bid to lower the amount of similar chip models in various packages, Intel will either completely cease to make mainstream client chips in µPGA and LGA packages that allow interchangeability of microprocessors, or will at least dramatically reduce their amount, reports PC Watch web-site, sometimes in 2014, when Broadwell-generation chips hit the mainstream market.

The web-site cites sources with knowledge of Intel’s plans as saying that for the company it may simply become economically inefficient to test and package LGA and µPGA products for clients. Eventually, this could lead to extinction of upgradeable PC platforms in general. On the other hand, it is hard to believe into economic inefficiency of certain Intel products, given its dominating position on the market of CPUs and keeping in mind the fact that its smaller rival AMD manages to be almost profitable selling chips in various form-factors.


Tags: Intel, Broadwell, Haswell, Core, 22nm, 14nm


Comments currently: 15
Discussion started: 12/26/12 01:23:13 AM
Latest comment: 12/27/12 01:30:55 AM
Expand all threads | Collapse all threads


There are big costs involved in setting up infrastructure and boards to test LGA, uPGA, PGA etc type socketed builds. There are also test program changes between socketed and non socketed models. This means completely different bring up teams for both types. This means more manpower expenses, time co-ordination. If they see LGA,uPGA as a very small market in the coming years, abandoning it is an obvious business choice.

Sales are not going to drop if the enthusiast species ceases to exist. That's just sensationalism. The overclocker/extreme benchers will still buy highest end hardware even if its non socketed to go about their benching. Realistically how often do people upgrade? 2-3 years? Intel changes or forces the buyer to change mobos anyway so this is really not going to change anything on that front.

But this does force you to make a purchase choice on a combo rather than independent CPU and motherboard choices at time of purchase. Intel knows the market well, tech savvy people who read websites like this are not on their minds or their major revenue generators. You have to wake up to the market reality that we are just not that important.

They have understood how the vast majority of the people will think for months to buy a car but would buy a very costly smartphone in a heartbeat as if its some daily use cosmetic. Its now fashionable to buy and change phones like its some sort of use and throw item. Trade in programs by many telcos are a reflection of that. Use that logic and they can force the buyer to make a quick choice now Combo 1, Combo 2 or Combo 3 etc no more mix-matches now. It maybe hard for people like us to imagine it but this means lesser complex choices to make, its a simple matter of combos. We on the other hand like lots of permutations in these choices, that's just us not the average joe.

This move, should it happen, will make buyer choice simpler and serve Intel's refined engineering options moving forwards. Many companies will go down this route.

The market will dictate it not the egos of a few enthusiasts.
7 3 [Posted by: vanakkuty  | Date: 12/26/12 01:23:13 AM]
- collapse thread

@Vanakkuty - I agree that computers are becoming more and more like a white good because it is now a necessity in a modern society, having become "main stream" sometime around Facebook/ Apple (2006/7). But do you also not agree that Intel is hitting a couple of birds with the one stone? Meaning, Intel sees a way to concentrate their profits through controlling the motherboard market. I believe it is a trend with the big American tech companies, progressively cutting out Asian OEM's that have been important for over the past 2 decades. What's happening to the internet (software walled gardens) is also starting to be reflected in the hardware world. It's frightening to think a handful of companies could end up controlling access and also the means of access to the internet.
6 1 [Posted by: linuxlowdown  | Date: 12/26/12 04:05:41 AM]
linuxlowdown . you are right this really is going to suck big time with people who like to build their own computers , no more interchangeable parts or upgrades.Well I just hope AMD will see the light of day and take advantage of Intel monopoly arrogance but that is doubtful.
5 1 [Posted by: Urhu  | Date: 12/26/12 06:19:12 AM]
AMD will say they are for sockets for the foreseeable future, but that "foreseeable" is a nice word they can play around with later when they do go back on this policy. Time frames are neither fixed nor mentioned.
5 1 [Posted by: vanakkuty  | Date: 12/26/12 08:22:06 AM]
I can understand that enthusiasts are being dwarfed by the mountains of computer illiterate end users who view computers as a means to an end in the most simplistic way - ie a device that lets them keep in touch with their friends and shop for shoes. The problem I have is the decreasing competition that will play out over the next 5 to 10 years in hardware. By 2022, I can see perhaps 6 big companies whereby consumers will be practically forced to buy their branded hardware to access the internet in their walled gardens. At this point the leaders are Apple, Google, Microsoft, Amazon, Sony. Will Intel even go the vertically integrated model and do software? Its partner Microsoft are not helping them at the moment. Will they keep the faith? I don't think it is an absurd idea if the ARM based ecosystems start completely dominating the consumer market.
4 1 [Posted by: linuxlowdown  | Date: 12/26/12 08:31:59 AM]
Depends if Intel wants to stop motherboard partners from getting rights to make new boards, they can simply leave it to them to market their chips on partner boards, that way the motherboard market enters into a new "features based" warfare. This can be a reference board like design for the partners and Intel leaving the motherboard market out completely, focusing the now freed engineering resources on for other projects aimed at emerging markets.

High-end boards getting K series unlocked chips for overclocking? A possible idea. While mainstream boards get regular vanilla chips?. This could also mean lesser steps in processor frequencies reducing over all complexity of SKU lineups for consumer market.

I can find out more from my friends, to stress more on this issue its still not clear if Intel wants to shut out motherboard partners, ASUS's tone suggests that is not the case.

Lets think of it this way, if PC is such a shrinking market why bother sticking your fingers into a small pie thats going to be small anyway when there are tablets and smartphone markets to fight in?
6 2 [Posted by: vanakkuty  | Date: 12/26/12 08:19:23 AM]
Enthusiasts might be a small market for Intel but you forgot another factor - when my computer illiterate friends go out and get a new system/upgrade, they call me. What enthusiasts do is generate positive publicity for the product and spread it via word of mouth. You can't buy that alone with marketing. The minute Intel forces us to choose Combo 1, 2 or 3, I'll go with a more flexible option of picking and choosing a CPU and Mobo to fit specific needs of people who ask me for advice. Chances are once this happens, I'll only recommend AMD CPUs. This is not like losing a one time sale, forever. I won't support a company that turns its back on enthusiasts. This round NV locked voltage control on their GPUs, too bad, lost my sale. It's a matter of principle and giving consumers choices. If Intel cares only about their profit margins, and they show me that they don't care about the enthusiast market by making me buy bundles, I'll stop buying all their products and won't recommend them to hundreds of my friends. And because my illiterate computer friends don't understand computer tech, they can easily be swayed because they trust opinions of PC enthusiasts. I can see many other PC enthusiasts doing the same to preserve our hobby and the importance of choices/flexibility. Intel can do all it wants for dual cores, OEMs and budget systems. Leave the enthusiast DIY market alone.
0 0 [Posted by: BestJinjo  | Date: 12/26/12 10:49:06 PM]
All true, but you don't have say 400 million friends to help build computers do you ? So Intel or any company wouldn't care for such sentiments. People like us can only influence a few people in this way. The marketing will get into the heads of millions of others and thats more than enough for any company to sell and make money.
0 0 [Posted by: vanakkuty  | Date: 12/27/12 01:06:05 AM]

Do you want your PC to become a complete throwaway item? Do you want to have to upgrade your motherboard to upgrade your CPU? That is what Intel wants. Intel wants you to have a piece of crap Xbox or Playstation. And you are going to like it too.
4 1 [Posted by: Urhu  | Date: 12/26/12 06:58:21 AM]
- collapse thread

Exactly, not only that, I buy Core i7 and $140-170 mobos because I don't care for features on $250-400 boards I'll never use. Mobo+CPU combos will be a nightmare to manage. Intel is committed to LGA1150 for at least 2 more years. Because Intel is sticking with quads once again for Haswell, and Steamroller aims to increase IPC by 20-30%, I can easily switch to the competitor if I see them catching up and offering me an LGA option. Once enthusiasts are abandoned by a company, the brand equity erodes. I don't think Intel wants to lose its core enthusiast base to AMD.

Many enthusiasts share my sentiment on other forums. Here is a good summary of many such thoughts:

"I have to have some help understanding here. I'm working under the assumptions that: 1) Overclocking is for enthusiasts only. 2) The mainstream wants lighter/smaller devices. 3) The mainstream doesn't care about raw performance. 4) Intel wants to make money, so they largely cater to the mainstream. Utilizing these assumptions, Intel moving to BGA with their mainstream offerings makes sense. A person willing to spend $600 on a computer requires more flexibility (LGA) than someone looking to spend $400 (BGA). As long as Intel focuses BGA packaging on the lower end, no harm no foul. I will raise hell if the 3570s descendants wind-up as BGA. That will get me to switch to AMD over night (performance be damned). Intel might do some stupid things, but they aren't stupid enough to kill all of the motherboard manufacturers (read: we still aren't the reason they keep LGA, there's too much money at stake for Intel to cut out the other manufacturers). For Pete's sake, get some perspective here people. Intel is aiming itself at ARM and tablets. The traditional PC is going to take a back seat for a while. Intel already confirmed this with socket 2011. The back seat sucks, but what sucks worse is if PCs were to have the gloomy outlook of the home console."
0 0 [Posted by: BestJinjo  | Date: 12/26/12 10:57:36 PM]
Like I said enthusiasts are nothing infront of the volume sales to customers on a world wide scale. How many enthusiasts are posting on these sites? 500? 1000? 10,000? Doubt anything more than that on the very top sites at best. Rest have much smaller numbers.

Running an engineering program to satisfy a 50k-250k Or even a million member market is nothing financially useful compared to the world at large. Many enthusiasts who post on these forums don't even overclock they are just there to talk tech. So that makes enthusiasts rather a small group no more significant than the occasional protesters in a rally on the streets. I was once an avid overclocker when I was younger, but it wears off when you work in industry and see the direction where the whole market is going now.

If they think they can raise hell they are mistaken. Its a bit funny to hear such grande empty statements. They are in no real position to do anything. They could though start a new chip company if they have what it takes
0 0 [Posted by: vanakkuty  | Date: 12/27/12 01:19:30 AM]
Once enthusiasts are abandoned by a company, the brand equity erodes

This is way far from reality BestJinjo.

When I was dealing with people during the heights of AMD's K7/K8 people were still buying Intel P4s. USA is not the only country on the planet.

Everywhere its the same thing, "Yeah I hear this AMD thing is good and all but I dont know, Intel is the bigger brand I don't know AMD...."

And then the Intel Rep comes in with the classic FUD "I once talked to a Japanese Intel engineer, I asked him how can we compete with AMD's products? He says there is no need for that at the end of the day you know there is one reliable and dependable company that is Intel, AMD sells crap"

Sure lot of BS there, but it worked it will work again and there is nothing anyone can do about it, AMD has to work on its products and work hard for several years to come to change that image.

"AMD is not good in laptops they explode" this line still lives on today and its working.

You should check the revenues on Intel's P4 during that period, virtually no real hit, some enthusiasts raved about AMD, made AMD stocks go high upto 40s was it? Yeah and after that came the big fall and no one who didnt care much about the tech like we do really bothered with any of it. For them it was Intel number 1 as always, AMD's small time at the top is not even believed by many who will tell me that I am making things up. This is not a small minority of people sharing this sentiment mind you, its the vast majority world over.

Intel know this all too well.
0 0 [Posted by: vanakkuty  | Date: 12/27/12 01:30:55 AM]

Another thing to keep in mind is that CPUs have reached a plateau. They are not progressing at any where near the rate they did for the last 2 decades. At 3 GHz, today's average CPU is plenty powerful enough for 99%. So, think forward, not backward.

If CPUs are not advancing rapidly, and they are good enough for the majority of buyers for the life of the device, then what is the use of a socketed CPU? You don't upgrade the CPU on your DVD player, right? If the DVD player stops working, you throw it out and get another one. Same will be true of future computers. They'll be the size of a Mac Mini, cost $150 and all your documents will be in the cloud. Stops working? Who cares, throw it out and get another one.

The only interesting question is what happens to the 1%. In addition to the enthusiasts, that 1% also contains software developers, architects, content creators, etc., that need above-commodity hardware to do their jobs, which are essential. So how do vendors justify the costs of meeting the needs of that tiny population? We need someone to do it, but there won't be much money in it.
0 1 [Posted by: guyr  | Date: 12/26/12 08:38:03 PM]
- collapse thread

Your argument is somewhat contradicting itself. If today's average CPUs are plenty powerful for 99% of users, then there needs to be no discussion at all regarding LGA or BGA. That means 99% of people have no reason to buy anything after Haswell in 2013. So essentially everyone can just upgrade one last time next year and never buy a single CPU again. You know this isn't true as software, artificial intelligence and so on are bound to become more advanced.

CPUs don't appear to be advancing as fast because Intel has little competition and instead wastes valuable transistor space on worthless APUs that we don't care about. Get rid of the GT2 in Haswell and give me a 6-core Haswell for $325. Intel won't do it, because they'll charge $500 for such a CPU on successor to socket LGA2011. Intel is just taking full advantage of no competition on the desktop above $200 level.

The die sizes go down but Intel keeps prices the same. They are ripping us off. The only reason Haswell is growing in size is the GPU is getting way more advanced than in IVB. Up to this point Intel is selling us less CPU die to prop up their margins.

If PCs become like appliances, I'll just quit the hobby entirely and do something where I can appreciate choices, flexibility and similar aspects of PC building, like high end headphones and amps, modifying a sports car, etc. The PC could become an appliance and be fine for 99% of users but once that happens, I'll never build another PC.
0 0 [Posted by: BestJinjo  | Date: 12/26/12 11:11:55 PM]

Broadwell will still have LGA1150 variants:

After that, no one really knows besides Intel. Maybe they'll still allow for upgradable chips but relegate that to $225 i5s and above.
0 0 [Posted by: BestJinjo  | Date: 12/26/12 10:40:49 PM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture