USB 3.0 on Hold Until 2011?

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
EETimes, citing an anonymous source at a “top tier PC maker,” says that Intel is essentially putting off USB 3.0 until 2011. The article goes on to say that an Intel spokesperson hadn't heard of any delays but declined to comment further.

Without chip set support from Intel for USB 3.0 aka SuperSpeed USB, adoption in 2010 will be limited to "a few high-end graphics workstations and consumer systems," said the source. That's because system makers will be forced to buy discrete host controllers for their motherboards, a relatively high cost. "It's hard to commit to an emerging technology like this when the key silicon enablers are not making it a priority," said the source, referring to Intel.
 
ever since the core chip made intel into a near monopoly again, it seems they went right back into "fuck the customers" mode.
 
USB3 is dead, anyway. Short of replacing all USB2 plugs, I doubt USB3 will get used much.

Why?

eSATA.

No need for USB3 attached hard drives if you can get a riser slot and turn an internal SATA plug into an eSATA one... for those without a built in eSATA jack.

For everything else, USB2 is 'fast enough' and Intel delaying it will essentially make sure that people end up missing USB3 completely during the Windows7 new hardware upgrade cycle.

They fucked up.
 
Is intel planning any chipset refreshes in 2011 at all?

My speculation from the thread about the upcoming asus USB3/SATA6 mobos is that DMI is too large a bottleneck to implement either of them on the southbridge (2 sata6 or 3 USB3 devices could saturate it); and that Intel won't be providing them until they provide a replacement technology. Putting PCIe 2.0 lanes on the SB suffers the same problem (DMI is only 4 1.0 lanes wide). I haven't seen any informed speculation about what Intel intends to do about this problem.

One option would be to just widen or overclock the DMI bus to increase its bandwidth. Unless there are technical limitations that prevent a higher clock rate (doubtful since it's widely believed to be a bastardized PCIe interface) I'd say clocking it faster is probably the most likely option. This would keep costs down (more pins/traces is more expensive) and allow for back compatability with current hardware.

A second option for LGA1366 would be to combine the NB and SB into a single chip and eliminate DMI from the equation entirely. Dropping down a process node would let them keep the silicon size in control; the number of pins on the chip could become an issue though. Probably not though since IIRC nVidia chipsets have been monolithic for a few generations.

Making LGA1156 a short lived architecture and replacing it with a model which uses QPI instead of DMI to connect to the chip would be a somewhat longer term option. The problem here would be the sort of consumer backlash that AMD took over the rapid cycle from S754/S939/AM2; so while I suspect something like this is probably in intels longer term gameplan I doubt they'll do it any sooner than Sandybridge. Sandy bridge is expected in early 2011; which would fit with rumors of not doing anything major in the chipset front until then. Even if tigerpoint ups the specs somewhat for nettops or atom based HTPC's I suspect it'll stay significantly below what they offer on LGA1156.

DMI is likely to linger longest on the atom platform (or its successor); but the amount of data being shoved through it it much lower than on conventional platforms. The USW15 (Poulsbo) chipset for the zXXX atoms (and the only dedicated atom chipset intel has made) only supports a pair of 1.0 lanes.
 
USB3 is dead, anyway. Short of replacing all USB2 plugs, I doubt USB3 will get used much.

Why?

eSATA.

No need for USB3 attached hard drives if you can get a riser slot and turn an internal SATA plug into an eSATA one... for those without a built in eSATA jack.

For everything else, USB2 is 'fast enough' and Intel delaying it will essentially make sure that people end up missing USB3 completely during the Windows7 new hardware upgrade cycle.

They fucked up.

eSata has built-in power?
 
Unless (or until) eSATA becomes more common it means nothing to me.

My external hard drive and usb thumb drives would beg to differ.

^hence, eSATA for your external drive. Anybody could argue their point. As techrat mentioned USB2.0 is plenty of speed for things like printers, keyboars, mice, etc. However, if people aren't jumping the eSATA bandwagon(and why not?) then USB3.0 would be nice for items like external HD, flash, etc.
 
USB is already the popular standard for external devices and you don't see eSATA nearly as often. eSATA doesn't allow cable lengths to be as long.
 
Anyone thinking eSATA will be more popular or is actually well known has to be kidding me.

eSATA is nice for those who know about it, but USB just has too much awareness/popularity and ease of use for the normal consumer (plug it in, that's about it for most device except hard drives or other devices requiring separate PSA).

If I was old I would probably say "E say what?" when asked about eSATA.
 
as long as the plugs are universal it don't matter much.

Problem is they aren't. They had to add pins for USB 3.0. You can plug USB 1/2 devices into a USB 3.0 port, but not the way around.

USB 3.0 flash drives won't take off until USB 3.0 becomes ubiquitous like current USB. Well, unless the come with an adapter cable; but that would just be too much a pain in the ass for most people.

External HDDs will fair better, they will just have to come with 2 cords, one for each standard. I would hope the USB 3.0 controllers will be smart enough to use both.
 
but isnt the lack of supplied power in eSATA, and what was also mentioned- cable length, really the main reasons you cant compare sata to usb?
 
USB3 is dead, anyway. Short of replacing all USB2 plugs, I doubt USB3 will get used much.

Why?

eSATA.

No need for USB3 attached hard drives if you can get a riser slot and turn an internal SATA plug into an eSATA one... for those without a built in eSATA jack.

For everything else, USB2 is 'fast enough' and Intel delaying it will essentially make sure that people end up missing USB3 completely during the Windows7 new hardware upgrade cycle.

They fucked up.

Wrong.

Ask anyone on the street what eSATA is and they'll look at you funny. eSATA thumb drives are super rare. Only external hard drives use the eSATA connection.

USB is well recognized and has been firmly established as an all-in-one connection. USB 2.0 and 1.1 devices being backwards compatible with 3.0 make this upgrade in speed even more advantageous.
 
Didn't Asus just announce an upcoming motherboard using USB 3.0 connections? Weird.
 
USB2 will be dead b/c USB 1.1 is fast enough! I think I've heard that argument before too.
 
They did. They're using 2 3rd party chips to implement it. One multiplexes 2 PCI 1.0 lanes into a 2.0 lane, the second has 2 USB3.0 ports. (the multiplexer chip also connects a SATA6GB chip)
 
Intel "innovation" is a deceiptful lie. Intel only makes what is easy to do. The hard things always escape them for years after everyone else has managed a given technical feat.

AMD did memory controllers years before Intel managed to figure it out in mainstream chips. Now the first USB upgrade in 8 years and Intel can't figure it out. Put their mobile VP on unemployment and I'd bet the rest of the company would get a fire lit under them.
 
eSata has built-in power?

Every external drive I've got that was USB2 required a double USB-A plug cable or external power wart. Whether or not eSATA has built in power is rather negated by USB2 not having enough.

Unless (or until) eSATA becomes more common it means nothing to me. My external hard drive and usb thumb drives would beg to differ.

eSATA is potentially as common as SATA is. You need nothing more than a riser slot adapter that goes from SATA on the motherboard to eSATA on the back of the case. Even Acer's little Nettop had an eSATA plug on the front. Cost of SATA to eSATA plug conversion? About 5 bucks.

http://www.newegg.com/Product/Product.aspx?Item=N82E16812816069

Compare that with the potential cost for an USB3 addon card for every system through the next year until they come standard. Say, $69.99 or thereabouts when they're common.

Your thumbdrives, by the way, are bottlenecked by the flash chips inside them. Not the USB2 bus. For as long as they remain a commodity and budget item, people are not going to be buying higher priced thumbdrives for top speed.

but isnt the lack of supplied power in eSATA, and what was also mentioned- cable length, really the main reasons you cant compare sata to usb?

My three external hard drives and my external DVD recorder all have power bricks. They're all USB, USB+eSATA or USB+FW. Laptop hard drives in external enclosures still need a Mini USB to Double USB-A Y cable to get enough power to be used. Soooo... really not a significant argument for now.

USB3 may help with the power issue, but how long before that reaches critical mass? Not for a while. People are gonna be stuck with USB2 for external drives or end up using... yep, eSATA.

Funny thing about cable length. Not sure what the maximum is for eSATA, but the eSATA cable I got with my enclosure is longer than all but one of the USB2 cables I have for other devices. It's almost as long as my desk is wide. How much longer do you really need it to be?

Backwards compatibility?...

It's like you didn't even read my post. And I quote: "Short of replacing all USB2 plugs, I doubt USB3 will get used much."

That's backwards compatibility right there, genius.

Wrong.
Ask anyone on the street what eSATA is and they'll look at you funny. eSATA thumb drives are super rare. Only external hard drives use the eSATA connection.
USB is well recognized and has been firmly established as an all-in-one connection. USB 2.0 and 1.1 devices being backwards compatible with 3.0 make this upgrade in speed even more advantageous.

The guy I quoted before you didn't read my post. Looks like you didn't, either. I never said anything of getting rid of USB entirely. I never said anything of using eSATA to replace all USB devices. Did I say "USB is dead"? No. I said "USB3 is dead." Stillborn. It'll exist, but its adoption window of opportunity was completely missed. That won't happen again for a while. It won't be anywhere near as ubiquitous as USB2 has become for -specific- devices. USB3 specific devices (that will actually consume the bandwidth it allows) will probably be less common than eSATA ones... eSATA plugs are on nearly every single new brand name desktop or laptop I've seen in the past year. Hell, even Nettops have them.

Mice, keyboards, card readers, webcams, yeah. They're going to still be USB2 for a long, long time. Hell, most all of those are still running at USB1 speeds.

USB3 offers little over an eSATA attached drive. You may still need a special Y cable or a brick (for a larger one) for an external hard drive, dvdrom, bluray drive, etc. due to power draw.

Hence my specific example: Hard drives. "No need for USB3 attached hard drives if you can get a riser slot and turn an internal SATA plug into an eSATA one... for those without a built in eSATA jack."

For everything else, it'll probably be advertised as "USB3 Compatible" but really running like many USB2 devices run now. A USB2 compatible device running at one of the lower USB1 speeds.

USB2 will be dead b/c USB 1.1 is fast enough! I think I've heard that argument before too.

Far from what I've claimed.

Show me a USB keyboard that needs a USB3 connection. Even USB2 keyboards are still running at "low speed" or "full speed" mode. Hint. That's USB1. USB2 specifically added "high speed."

Or a mouse. Yeah. Do you really think that USB2 mouse is running at 480Mbps "high speed" when attached?

Webcams. Built in or external, still aren't saturating USB2. People aren't going to be webcamming in 1080p.

The one example I DID give in my argument, however, was hard drives... Catch the drift?

Like I said...

Intel missed the boat here. They fucked up. People who were looking to replace their systems now that Win7 is out... or anytime for the next year... and everyone who has a computer now, whatever it is... Will be in the vast, vast majority of those without USB3.

They'll have USB2 and SATA.

USB2 drive is too slow. What can we do now? Can't buy a USB3 card or enclosure right now. Didn't come with the computer. Won't be available until 2011 rolls around. Then you gotta wait for manufacturing and support. And for the devices to show up. But SATA plugs are already here and offer more bandwidth with less overhead than USB3.

eSATA will always have the benefit of SATA's backwards compatibility. SATA 6.0? Yeah. You'll be able to do that over eSATA, too.

For hard drives, we'll be seeing more eSATA than we may see in USB3.

I never claimed that eSATA was going to be replacing USB2 or eliminating USB3's use. I said "short of replacing USB2 plugs, I doubt USB3 will get used much...", meaning it'll be there, it just won't have much of a distinction beyond existing as a plug that is USB3 and being used primarily for USB2 devices.

Reading Comprehension. It's what's for dinner.
 
Let's not go back to having a different connector for each peripheral...

I hope eSATA goes away in lieu of a universal connector fast enough to replace it. Be it USB3 or Lightpeak I don't really care as long as it works and can be used easily for any device, easily split with hubs for more ports, and include power. Also eSATA has terrible connectors, requires very short cables and can be problematic on the configuration side for some users (requires AHCI mode, conflicts with onboard RAID on some boards etc.). It's a kludge because USB2 is too slow. It's a fine kludge for now, but to suggest that USB3 is pointless is silly.

USB2 video dongles? HD capture devices? Network cards (wireless or wired)? Niche applications you haven't heard of? Laptop hard drives? There are lots of things that could use better performance. To ignore those *and* write off every future development is just foolish. We need USB3 or some equivalent, and we need for it to start appearing now so that adoption is high enough when it's really needed. It's already late, given all your arguments about eSATA - eSATA should never have existed.

Your thumbdrives, by the way, are bottlenecked by the flash chips inside them. Not the USB2 bus. For as long as they remain a commodity and budget item, people are not going to be buying higher priced thumbdrives for top speed.
They're pretty close, if not there already, and will shortly surpass USB2.
 
USB3...for what...all those USB3 HD's out on the market?

lulz

Anyone whining about their USB HD and wanting USB3...guess what:
you USB2 HD ain't going to run any faster on USB3.

USB is the worst standard for storage, hell FireWire or eSATA beats the crap out of it in performance.

Use the right tools for the right jobs.
 
USB3 is dead, anyway. Short of replacing all USB2 plugs, I doubt USB3 will get used much.

Why?

eSATA.

No need for USB3 attached hard drives if you can get a riser slot and turn an internal SATA plug into an eSATA one... for those without a built in eSATA jack.

For everything else, USB2 is 'fast enough' and Intel delaying it will essentially make sure that people end up missing USB3 completely during the Windows7 new hardware upgrade cycle.

They fucked up.

USB is more versatile than eSATA. Gadgets, printers, thumbdrives, smartphones, PMP, webcam, and a zillion other things.

USB isn't going to replace eSATA anytime soon.
 
USB is the worst standard for storage, hell FireWire or eSATA beats the crap out of it in performance.

Use the right tools for the right jobs.

Just to put it into perspective.

USB2 is worse than eSATA piped through a PCMCIA slot.

Did my own tests using a Seagate 7200.10 Barracuda SATA2 drive, in an enclosure that does both eSATA and USB2. Used a Rosewill budget eSATA Cardbus card bought off NewEgg for less than $20. It has a Silicon Image 3512 SATALink Controller inside of it. The case used was an Antec MX-1 USB2/eSATA external enclosure.

And no, the hard drive was NOT connected through a USB hub. No additional USB devices were plugged in during the tests.

USB2, 29MB/sec burst, 27.1MB/sec average. CPU utilization ranged from 12 to 17%.
eSATA through PCMCIA, 70MB/sec burst, 61MB/sec average. CPU utilization ranged from 0 to 2%.

You'll probably end up needing a dedicated core on that Core2Quad just for USB3 utilization.

USB has its uses for simple devices, but for disc and hard drives? eSATA or Firewire all the way.
 
Too bad Firewire is not as mainstream as USB. It was faster and less CPU intensive than USB. And no less complicated for the average layman to use.
 

Probably based on a drafted spec. Makes me wonder if there's any logic involved in the USB3 plug traces... and if there is, how they're going to resolve that if anything ends up getting changed by the time Intel finalizes the spec. Probably might be as simple as a firmware update or they'll code around any issues in a driver.

In any case, uncertainty of compatibility is something that makes me want to stay far, far away from a product.
 
Here's my problem w/ esata. If you do not have and external esata port, you have to run a cable from one of the internal channels out the back, and then use a conversion cable to get to esata. USB2 run a cable from the back of the computer to the external hard drive. The copy speed isn't that much slower, and USB2 is a hell of a lot more reliable. I think esata was more of a flop than usb 3 will ever be. Motherboard manufacturers are going to start putting the USB3 slots on their boards before devices are released. By the time the devices are released a lot of people are going to be ready for a hardware refresh, thus USB3 will take off.
 
Let's not go back to having a different connector for each peripheral...

I hope eSATA goes away in lieu of a universal connector fast enough to replace it. Be it USB3 or Lightpeak I don't really care as long as it works and can be used easily for any device, easily split with hubs for more ports, and include power. Also eSATA has terrible connectors, requires very short cables and can be problematic on the configuration side for some users (requires AHCI mode, conflicts with onboard RAID on some boards etc.). It's a kludge because USB2 is too slow. It's a fine kludge for now, but to suggest that USB3 is pointless is silly.

USB2 video dongles? HD capture devices? Network cards (wireless or wired)? Niche applications you haven't heard of? Laptop hard drives? There are lots of things that could use better performance. To ignore those *and* write off every future development is just foolish. We need USB3 or some equivalent, and we need for it to start appearing now so that adoption is high enough when it's really needed. It's already late, given all your arguments about eSATA - eSATA should never have existed.


They're pretty close, if not there already, and will shortly surpass USB2.

I agree. eSATA is great in theory, I have it myself, but the physical connector sucks and the connection doesn't deliver power. For portable storage it is horrible, Firewire 800 is still the best all around when taking notebooks into account. Let's hope that Lightpeak or USB 3 wins out and that eSATA ends up being a fad.
 
Back
Top