Manufacturers push for USB 3.0 adoption, Intel holds it back

Status
Not open for further replies.

Jos

Posts: 3,073   +97
Staff

With Computex 2010 underway, several manufacturers have taken the occasion to launch new USB 3.0 peripherals in hopes of increasing sales as the new interface slowly gains traction. Among them is Iomega, which today revamped its eGo line of portable and desktop hard drives with a total of five new drive models, including a pair equipped with SuperSpeed USB 3.0. Kingston showed off an upcoming external SSD that hooks up to the next-gen interface, while SuperTalent continued to expand its USB 3.0 flash drive lineup with the new Express RAM Cache model.


But while each of them is doing their part in helping bring this technology to mainstream ground, it seems Intel is not very interested in giving them a hand. Initially expected to hold back adoption until 2011, unconfirmed reports now claim Intel will not be packaging USB 3.0 support into its chipsets until 2012. The SuperSpeed specification was introduced in November of last year and already there are plenty of vendors shipping USB 3.0 motherboards based on NEC's controller. AMD plans to integrate this chip in its reference designs and other cheaper alternatives are on the way too.

Given that vendors like Asus, Gigabyte and ASRock are already building USB 3.0 into their products, it's hard to see why Intel would have any problems supporting the standard. Perhaps it's not so much a technical matter but rather the chipmaker pursuing its own agenda -- that is, trying to make the market more receptive to Light Peak. The optical cable interface does have its advantages over USB 3.0 so we're interested to see how things play out.

Permalink to story.

 
I think it is pretty clear Intel doesn't want USB 3.0 to become what USB2.0 is today. Look what happen to firewire. Firewire was better in almost every way but so many people where already using USB 2.0 that firewire just confused the masses. I am pretty sure Intel wants Lightpeak to be on a level playing field as much as possible. With the resurgence of AMD this might force Intel's hand to move a little faster.
 
It's nice to hold confidence in a good idea. And Intel is definitely sticking to its guns on this one. But the rest of the world is pretty much moving on without them. Lightpeak may have its benefits, but if they're not going to get it into shape soon, then everyone will move to USB 3.0 just because it's here and Lightpeak isn't. Given that this seems to be the case, Intel might have yet another idea that becomes abandoned because they didn't put enough resources into it, or because the rest of the world moved ahead before they could deliver.

With all these other mainstream companies delivering USB 3.0 like crazy, if and when they do put Lightpeak out there, it'll have a bit of catching up to do.
 
Well, it's good to see Intel isn't swayed by the needs of the industry or the desire to stay compatible with the latest peripheral standards... Much better to keep shutting out USB 3.0 and turning a deaf ear to the outcry.

I find it a bit telling, that Intel seem so worried about the competition USB 3.0 might give to their LightPeak interface. If LightPeak is so much better (and by all accounts, it stomps USB 3.0 on almost every front), then maybe let the actual hardware performance and benefits sell the interface, rather than holding the industry hostage to try to FORCE your new technology on them. There has to be a driving motivation behind it... Is LightPeak going to be much more expensive or something along those lines, which would deter adoption and encourage the USB 3.0 route?
 
Ivan and Varmithrax each make some good points. I could possibly see it this way:

USB 3.0 will be the standard, while Light Peak would be the more expensive yet faster alternative...possibly to be used in businesses. And I would see devices being able to connect to both if Intel gets this project in gear so that manufacturers can actually make devices suitable for both USB 3.0 and Light Peak.

I could be wrong in all of that, but at this point we just don't know enough. Time will tell.
 
If USB 3.0 catches on with the masses, and it looks like it is at this point, I see Firewire going the way of the floppy disk because firewire is an Apple creation and it doesn't look like anything is in the wings to go beyong their latest offerings.

Between Firewire and USB 2.0 downloading and uploading was much better with firewire. We loved it for downloading the mini-dv content because the feed and quality is so much better, but try find the mini-dv in the latest camcorders. You can't and they don't even offer firewire ports anymore.

As for Intel, the USB 3.0 is here, now. LightPeak isn't. The latest motherboards offer 3.0 and camcorders and other peripherals aren't far behind (and few peripherals are here already). We all know very well that technology pushes forward at a fast rate in the tech world and it isn't going to wait for LightPeak.

Question: Who owns the rights to USB? Or is it a universal standard like an open source? If it is an open source I can see even less enthusiasm by manufactures to utilize Intel's LightPeak.

Bottom line, IMO, Intel needs to have this as a viable working alternative soon.
 
Question: Who owns the rights to USB? Or is it a universal standard like an open source? If it is an open source I can see even less enthusiasm by manufactures to utilize Intel's LightPeak.

Bottom line, IMO, Intel needs to have this as a viable working alternative soon.

USB is an industry standard specification. The developers are : NEC, ST Ericsson, Intel, Microsoft, HP, Philips, Alcatel-Lucent. The original spec also involved IBM, DEC, Nortel and Compaq.

Why does Intel need to implemement USB 3 at CPU level ?
Since the ICH 10R will be the last iteration of a seperate "southbridge" I/O hub (all future I/O's will be built into the CPU itself starting with P55) would you have Intel waste die space with what they hope will be a stepping stone to Light Peak ?
Is it plausible that Intel would build in a redundant feature into it's upcoming LGA1155/2011 processors ? Especially since the feature can be implemented just as well by adding a third party <$5 controller chip.

BTW: This latest/greatest chipset from AMD is likely to be hanging around for some time yet, but you and a segment of the tech readership seem to be ok with AMD following the same path of USB 3 implementation via onboard third party controllers.
 
matrix86 said:
Ivan and Varmithrax each make some good points. I could possibly see it this way:

USB 3.0 will be the standard, while Light Peak would be the more expensive yet faster alternative...possibly to be used in businesses. And I would see devices being able to connect to both if Intel gets this project in gear so that manufacturers can actually make devices suitable for both USB 3.0 and Light Peak.

I could be wrong in all of that, but at this point we just don't know enough. Time will tell.

Perhaps. But that depends on how well USB 3.0 works. If you got USB 3.0 standard on a whole bunch of machines, and the speeds are very good, why spend a bunch of extra money if 3.0 works just fine? That's the only quandary I can see. Even businesses will go with what works and skimp where they can. I doubt that a lot of people would adopt lightpeak if that's the case. If anything, I could see lightpeak pandering to the enthusiast market, but not very many other places.
 
Light Peak is going to be the next Firewire if Intel doesn't get on the ball. USB 3.0, however, is affordable and, most importantly, available now. I love the idea of light peak, though I don't think it will be nearly as successful as USB 3.0 in the consumer market.
 
Honestly, I think the niche that LightPeak could completely dominate is in media streaming... I've seen some demos with multiple streams of HD video running through a single LightPeak pipe, without a hitch. The throughput capability is truly amazing. Seems like with the ever increasing computer platform presence in the consumer media experience (nettop PCs, HTPCs, etc) and the dominance of HD, that blazing speed available in LightPeak would be put to good use on all the digital media goodness.
 
Intel actually has conceded to USB 3.0, at least to some degree. As far as I know, all their H57 chipset boards have it.

IMO external SSDs have to catch on at the same time as USB 3.0 to gain real utility. Even at that, USB 2.0 transfers a bunch of data really fast.

I was wondering who to blalme for "Firewire", Apple makes sense for this. Maybe they can get together with Iomega and turn out some Firewire Zip Drives. That would be cute, funny, and exponentially useless.
 
LightPeak is the future it is good and fast and future proof, but i'm sure that USB 3.0 will win for now and about 4-5 years we will be switching to lightpeak instead to USB 4.0 I think.
 
IvanAwfulitch said:
If you got USB 3.0 standard on a whole bunch of machines, and the speeds are very good, why spend a bunch of extra money if 3.0 works just fine?

Do you get out very much? I don't know about you, but I know people who will spend everything they have on the best. If Light Peak is indeed much faster, why not get it instead of USB 3? Take a look at processors. If an i5-750 works just as well as one of the i7's (can't remember which one, now) then why waste the extra money for that particular i7? Well...people do it every day.

Sure, the average consumer will say "meh, USB 3 is good enough." But a power user (gamer, big businesses, people who do a lot of file transfers) may want those extra juices since they are available. Lord knows once I graduate and start making good money, i'll be one of those who want the latest, greatest, fastest stuff that comes out. But only time will tell what will happen. If Light Peak can surpass USB 3 by great leaps and bounds, then there will be people who will want it.
 
The fact of the matter is that USB 3.0 will be fast enough for a long time.... and intel knows it. SSDs can barely use the bandwidth provided by this interface, and the fact of the matter is that no one in there right mind is going to use an SSD as an external hard drive (unless they drop below the price of a spinning platter... which I don't see happening this decade.) Besides external storage what else is going to use this bandwidth?

The real advantage that something like light peak would have is adaptability. I can see how having your video-out ports and data transfer ports being the same would be nice. For this reason it will probably get adopted (because intel will start shoving it on there hardware rather or not there really is a need), but not completely replace usb 3.0.

Although I have to admit, they are doing a good job of killing the market for it. How many usb 3.0 devices can one actually purchase these days?
 
Intel actually has conceded to USB 3.0, at least to some degree. As far as I know, all their H57 chipset boards have it.

IMO external SSDs have to catch on at the same time as USB 3.0 to gain real utility. Even at that, USB 2.0 transfers a bunch of data really fast.

I was wondering who to blalme for "Firewire", Apple makes sense for this. Maybe they can get together with Iomega and turn out some Firewire Zip Drives. That would be cute, funny, and exponentially useless.

Couldn't agree more captain. Right now there are a number of Intel board offerings with USB 3.0. Intel may be holding back but mobo manufactures like ASUS aren't waiting. I guess they too belong to that segment dividedbyzero accused me and a bunch of teach readers who are okay with using third party controllers. I didn't even know I belonged to a "segment." :rolleyes:

As for firewire there is the 1394A @ 400 Mbs and the newer 1394B @ 800Mps but there isn't a lot of demand especially for the latter except for professional musicians who record and people with mini-dv camcorders. Even some of the newer mobos don't offer 1394 ports.
 
Even with current platter HDs, USB 2.0 is often the bottleneck. You need to understand the difference between the theoretical bitrate (480 Mbps) and the real-world throughput factoring in bus traffic, CPU requirements, and other parts of the system. USB 2.0 can maybe sustain 30 MBps (MegaBYTES/sec) which is far lower than desktop drives can deliver.

Who cares about Inhell and their still-born Light Peak. It might make make an impact in 10 yrs, but between now and then, us consumers would like something faster than USB 2.0 and affordable too.
 
... Intel may be holding back but mobo manufactures like ASUS aren't waiting. I guess they too belong to that segment dividedbyzero accused me and a bunch of teach readers who are okay with using third party controllers. I didn't even know I belonged to a "segment." :rolleyes:.

The point is that unless you specifically want an Intel branded board, as opposed to an Intel chipset board then getting USB3 on a X58/P55/H55/H57 platform doesn't represent a problem. Motherboard manufacturers are only willing and able to add new bells and bows to their lineup if there is a buck to be made.

The issue I take with a "segment" of the readership is that the title of the thread is "Manufacturers push for USB 3.0 adoption, Intel holds back" so...
(these are the questions that naturally came to mind when I read the title and then the article)
QUESTION: How many mainstream consumer motherboard chipset manufacturers are there ?
ANSWER: Two. Intel and AMD

If, judging by the facts in the article, Intel is holding up adoption of USB3 because of Light Peak until 2012. I would ask myself
1. Why
2. Since there are only two mobo chipset makers then the article infers via omission that AMD has USB3 support-or at least will have much sooner than 2012, otherwise the article would read "...chipset makers (or Intel and AMD) holds it back"

So...
As for the 1.Why...
X58 currently uses the ICH10R I/O controller hub. To implement USB3 would require a re-working (ICH 11R ?). So between the present day and when the X58 is scheduled for replacement -probably a year more or less, Intel would need to build, test and qualify the I/O hub and motherboard manufacturers would then need to implement it (revise layout, rearrange traces, reconfigure PCI lanes for increased bandwidth) and then get it onto shelves.
P55/H55/H57/Q57...Same deal...except the timeframe is around six months since Sandy Bridge (LGA1155) is pretty much here.....and since Sandy Bridge is already done and dusted, you would need to scrap the CPU (including the ES already in the hands of motherboard manufacturers, third party controllers etc.) and retool the architecture since the USB controller is part of the CPU....Sandy Bridge is due to be replaced by Haswell in...you guessed it...2012 - the next architecture from Intel that isn't already made.

As for 2. AMD adoption of USB3...
This is AMD's latest architecture. You'll notice no USB3 support. Bulldozer is AMD's next desktop architecture to replace Phenom/Athlon II CPU's- it will also likely be a drop-in replacement on these AM3 boards. Moreover AMD has announced no USB3 support for any upcoming southbridge...

So basically the story should have read: " USB3 is made available via third party controller at the motherboard makers discretion, neither consumer motherboard chipset manufacturer will be offering on-die or southbridge support before the 2012 timeframe"....not very sensational as a story line is it?

So basically, when I read a tech story like this I look at what isn't said as well as what is. All I did was a little reseach to acquaint myself with the facts that were missing from my knowlege base- and indeed from the article before commenting.
It is the nature of such articles that they are designed to polarize opinion and generate comment (and of course pageviews)- some of it immediate and assumptive in nature, and some of it deliberately placed there by a few well placed trolls and shills. Unfortunately, it is my nature to call-it-as-I see-it if I see false, manipulative or incorrect information represented as fact- Hence the frequent links I use to help distinguish my opinion from those of others (or known fact as the case may be). I also don't believe in sugar-coating my comments if a short-sharp-shock is likely to generate more thought from the recipient.
 
Well, I think it's wonderful all the pent up feelings and frustrations this thread has produced. Now let's see, what have we learned.

You can get USB 3.0 in an Intel chipset board >> right this very minute....! It's going to ba a 3rd party controller though. Does that really matter? Probably not.

Intel is being rude and selfish to the enthusiast sector of the market because they don't put USB into their own branded boards. Guess what, the "enthusiast sector" doesn't buy their boards anyway. Why complain about something you'll never use? Now there's a conundrum. Let's go with "because we can", and we like to hear ourselves talk.

AMD isn't doing anything with USB 3.0 either, but we've learned that at least they have enough sense, and the good taste, to keep their mouths shut, to avoid running afoul some obscure member of Techspot, who may be thinking about buying a new board, God knows when.

"When I get enough money saved from my paper route, I swear, as God is my witness, I'll never go USB 3.0 hungry again" (Well said Ms O'Hara).

As for firewire there is the 1394A @ 400 Mbs and the newer 1394B @ 800Mps but there isn't a lot of demand especially for the latter except for professional musicians who record and people with mini-dv camcorders. Even some of the newer mobos don't offer 1394 ports.
In some instances this is becoming a legacy application of Firewire. I have a Nikon D-90 SLR (with HD movie) and it, as well as my D-80 offload via USB 2.0. (IDK how many of the very new DV actual movie cams are still using Firewire).
 
So...
As for the 1.Why...
X58 currently uses the ICH10R I/O controller hub. To implement USB3 would require a re-working (ICH 11R ?). So between the present day and when the X58 is scheduled for replacement -probably a year more or less, Intel would need to build, test and qualify the I/O hub and ,,, [blah blah]
Yeah god forbid Intel spend any resources developing a new chipset. I mean, who do they think they are, Intel?
It's called Planning. Intel could have easily incorporated USB3 by planning for it a year ago when the USB3 spec was released. But they didn't because they have ulterior motives, mainly Light Peak. It doesn't matter that much for desktops since a simple chip takes care of it, but many notebooks don't have the space for an extra chip & wiring, or can't afford the power premium.

I'm not sure why you're so adamant about your rant, but in the end it's a rant like the other guy's. Get off that high horse, mkay?
 
No offence but say a few new business's open up in your area and brought brand new servers and desktops (so they have USB3 in as its already here) are they really going to upgrade all their equipment to use LightPeak? I don't think so, but Business's don't want to upgrade to a technology which they may not use, actually after supporting soo many company's I can happliy say that unless they see a real benifit to themselves (believe it or not lots did actually upgrade to Windows Server 2008 R2 & Exchange 2007, Note not many have switched to 2010 Exchange yet but I guess its only time, lots of company's are virtualising servers at the moment as they can see the benifits but i'll shut up now) anyway i don't think people would upgrade to LightPeak if the price is higher than USB 3's.
 
Intel could have easily incorporated USB3 by planning for it a year ago when the USB3 spec was released.
IIRC The USB 3 spec was finalized in August 2008 - not last year.
The Sandy Bridge CPU architecture design started in 2006- (Sandy Bridges' original code name was "Gesher") you do the math.

But they didn't because they have ulterior motives, mainly Light Peak.
LIght Peak probably plays a part in the decision-so what ? And what's AMD's excuse ? And what's VIA's excuse ?
Obviously no chipset maker thinks integrated USB3 offers any benefit over an onboard chip - That includes Intel, AMD and VIA
It doesn't matter that much for desktops since a simple chip takes care of it, but many notebooks don't have the space for an extra chip & wiring.
http://www.everythingusb.com/usb-3.0-notebooks-18331.html
USB3.0 chip Physical dimensions: 10mm x 10mm x 0.65mm
or can't afford the power premium.
NEC µPD720200 USB3.0 Host Controller (Power requirement 1.0 watts)
I'm not sure why you're so adamant about your rant.
Who knows....maybe it's because it's based in fact
Get off that high horse, mkay?
Giddy up!
 
matrix86 said:
IvanAwfulitch said:
If you got USB 3.0 standard on a whole bunch of machines, and the speeds are very good, why spend a bunch of extra money if 3.0 works just fine?

Do you get out very much? I don't know about you, but I know people who will spend everything they have on the best. If Light Peak is indeed much faster, why not get it instead of USB 3? Take a look at processors. If an i5-750 works just as well as one of the i7's (can't remember which one, now) then why waste the extra money for that particular i7? Well...people do it every day.

Sure, the average consumer will say "meh, USB 3 is good enough." But a power user (gamer, big businesses, people who do a lot of file transfers) may want those extra juices since they are available. Lord knows once I graduate and start making good money, i'll be one of those who want the latest, greatest, fastest stuff that comes out. But only time will tell what will happen. If Light Peak can surpass USB 3 by great leaps and bounds, then there will be people who will want it.

I get out plenty enough. You just didn't read what I was saying correctly. I already said that "enthusiasts will probably adopt this". What I was trying to explain with my comments earlier was that from a BUSINESS standpoint, nobody would adopt it. Hence why it would be an enthusiast market. Those people that do spend the extra money for the latest and greatest are the ones I consider to be part of that market.
 
I suppose there will be USB 3.0 addon cards available both for older machines that do not have it on their mobo, and new intel stuff that refuses to adopt it. So whats the brew ha ha!
 
Guys forget about USB 3.0. Why?
Lightpeak has speed
eSATAp (eSATA/USB) has compatibility/speed/easy
Don't forget SATA is in all machine/HDD/optical drives
Implementing eSATAp simply requires a USD 12 Delock bracket to connect to onboard SATA.
ZERO drive, IDE, no need BIOS AHCI, 99% compatible with motherboard & USB.
Can HotSwap! (freeware).
Almost 50% of the notebooks from Sony, HP, Dell, Toshiba, Lenovo have eSATAp port.
Almost 80% of the NAS storage have eSATA (need port multiplier).

Fusion Cat
 
This was an interesting read, thanks DBZ for all the links, I've actually felt like I've learned something among all these posts. :D
 
Status
Not open for further replies.
Back