USB 3.0 Preview @ [H]

I'm looking forward to widespread USB 3.0 adoption more than any other technology in 2010. I really hope they will release cheap PCI or PCI-E USB 3.0 expansion cards for people who don't intend to upgrade their whole system just for USB 3.0.

For me USB 3.0 will mean incredibly faster system backups on usb thumb drives, music library transfer/creation on MP3 player and high res photo transfer from my camera (or does this depend on the write speed of the media i'm transferring from?) This means alot more to me than any general performance gain I'd see from upgrading to LGA 1156/1366, 5000/300 series graphics cards or 6gb ddr3. In fact the only upgrade I'd rate of a similar importance would be SSD/RAM drive, neither of which I'd currently upgrade to. What does USB 3.0 mean for the rest of you?
 
@C-rizzle
I think kencheeto should have used the 3.5"HD external enclosure as an example, as we all know the 3.5" cases all need a wallwart while the laptop sized 2.5" HD cases only need one USB plug and in worst case scenario 2-USB plugs.

So, that said, I think we'd still like to know, is the 3xPower feature going to get rid of the wallwarts for 3.5" HD enclosures and external CD/DVD cases?
 
I wonder if this new standard means external video cards may become more commonplace in the future, especially for laptops?

cool... external Video Cards! you can easily add Crossfire without opening your case :D
 
I think you should have also used an SATA II drive to see if there is any advantage in that scenario. Theoretical bandwidth of SATA II is 300MBs. SATA drives will do that much bandwidth, but we never see it--ever. I was under the impression the bottle neck was the MB BUS.

(1) If true, and the new MB you used to test USB 3.0 opens this bottle neck, then using USB 3.0 and a new USB 3.0 MB, shouldn't we should see huge improvements in SATA II drives also?

(2) Wouldn't the same be true of any USB 2.0 device that is capable of 480Mb/s when plugged into a USB 3.0 port?

(3) If the bottleneck is lifted on MB's when using USB 3.0, but not the SATA BUS, then also, wouldn't it be better to use all USB HD's and bypass the SATA 3 BUS altogether? I understand that the controller on the USB device is also sometimes the limiter. But if there is a good controller on a USB II device that would allow 480Mb/s, and the device itself is capable of at least that much bandwidth, then the limiting factor would have to be the BUS on the MB. So again, if you have a quality USB controller, a device that is capable of saturating the USB2.0 pipe, then USB 3.0 should uncork the MB bottleneck.
 
Do SATA II drives do 300MBs across the SATA II bus? No......

1. Not true.
2. No.
3. The bottleneck in the scenario shown, and I thought thoroughly discussed, is the drive itself. Sorry I was not clear on that.
 
What's funny is that IF Intel had actually managed USB 3.0 notebooks by now I'd already have a 920 mobile one. This is one of those "once in a decade changes" that literally is worth waiting to buy a PC for. 10x the speed 10 years in the making. Failure to include it when the final spec was ratified, on a "new" chipset, was a pure goof. Oh well guess this means I'll get 32nm cpu's as well.
 
Accepting the pre-Intel P55 implementation options available off the shelf now with the pcie lane sharing, is there any reason to avoid one of the bridge motherboards with the intent to add pcie USB3.0/SATA 6.0 adapter cards instead? I wouldn't expect there would be any bandwidth advantage with add-on adapter cards but have I missed reading something between the lines?
 
Can BIOS updates implement USB3.0 into current USB2.0 motherboards? Or is this hardware related as well?
 
ok, soz, didn't even know USB 3.0 was released, bit behind current tech news unfortunately.
 
Maybe USB 3.0 will finally give way to viable external GPU's?
 
I'd be as interested in power capability of USB 3.0 just as much as bandwidth. Having two plugs for an external dvd-rw is a PITA.

+1 I'm actually MORE excited about the power capabilities than the speed.
 
Which is still a fraction of the amount available in USB, not tomention hard drives are hardly the only USB devices I've got.
 
Which is still a fraction of the amount available in USB, not tomention hard drives are hardly the only USB devices I've got.

I was just responding to the complaint about needing two USB plugs for an external drive.

*shrugs*
 
I understand... I'm very excited to see more USB powred devices. Scanners for one, the ones currently available are painfully slow. Force feedback steering wheels might be another possibility now.
 
Remember also that it matters how the USB 3 chip is hooked to the chipset. x1 pci-e isn't enough bandwidth, and many "bolt on" chips use just a simple single lane.

It's nearly idiotic for Intel to have not put this into it's chipsets so far. But that's the nature of things with no competition.

Actually USB 3.0 operates at a theoretical maximum of 4.8Gbps or 480MB per second. PCI-Express 2.0 x1 slots are capable of reaching speeds of 500MB per second. They are sufficient for one port. If the chip uses more than one PCI-Express lane for connectivity it might be enough bandwidth depending on how many ports they support. Even using a single PCI-Express lane USB 3.0 will still be FAR faster than USB 2.0 is.
 
No, that is the limits of a rotating hard disk. The transfer speed on the outside of the disc is faster than the inside. If you look at harddrive reviews for internal drives, you see the same thing.

That's what everyone thinks, but it doesn't mean it's faster transferring data. Realize that the inner part of the disk is also smaller in area. Just like runners running on a circle track, the runner on the outside may be faster, but may be past by a slower runner on the inside track.
 
I'd like to know what bottlenecks current SATA drives when they are transferring data to and from each other then, if not the MB? And I'm not talking about your scenario--only.

I take it you didn't test a current SATA II drive because there aren't any SATA II drives that would utilize a USB 3 controller, or a USB external HD reader? In other words, the SATA II drive would default back to USB 2.0.

So I guess what I was saying is that if you could run an SATA II drive on the USB 3.0 MB and have it use the USB 3.0 archetecture, the SATA II drive would see an increase in speed.

The drive itself is the bottleneck. It can't physically move fast enough to access the data any more quickly. The data coming out of the drive and into the SATA cable simply can't go any faster than ~150MB/s on consumer drives. No matter what fancy SATA interface or motherboard you hook it to, it simply won't be any faster, due to the laws of physics. SSDs are faster than HDDs, but they do also have a maximum speed that their components are able to reach, and they simply can't go any faster than that.

The actual drive inside that enclosure is most likely a regular SATA II 1TB drive, just like you would buy from Newegg to put inside your PC. It's been a long time since I've seen an enclosure that was anything more than a standard internal drive with an adapter hooked to it. That's why Kyle mentioned trying to take it apart to swap in the Intel SSD. If he could've gotten the enclosure opened, he most likely could have just unplugged the 1TB drive and plugged in the SSD.

So to answer your question, this almost certainly is a SATA II drive plugged into USB3. SATA II drives wouldn't "default" to USB2, because SATA and USB have nothing to do with each other. This enclosure is just a USB3-to-SATA adapter, plugged into a regular SATA drive. The adapter is upgrading from USB2 to USB3, but the drive isn't changing. Any drive you plugged into it would "become" a USB3 drive.

It's possible they're using a SATA III (6Gb) drive, but 150MB/s is SATA I speed, so II vs. III doesn't really matter.
 
That's what everyone thinks, but it doesn't mean it's faster transferring data. Realize that the inner part of the disk is also smaller in area. Just like runners running on a circle track, the runner on the outside may be faster, but may be past by a slower runner on the inside track.

Actually it does. If you compare the inside track to the outside track, the inside one is much shorter, as you point out. The longer outside track allows it to hold more sectors. In one rotation of the platter, the read/write head may cover 50 sectors on the inside track, while it may cover 90 sectors in the outside track. Because the outside track is longer, it has room for more sectors, meaning more data is handled in the same amount of time.

The sectors are no longer like pieces of a pie, where there are the same number of pieces at the inside and outside, but they're wider at the outside. Current drives actually divide the outer tracks into more pieces of similar size to the pieces on the inside track.

With runners on a circle track, a lap on the inside is a shorter distance than a lap on the outside (hence the staggered starting locations to give each runner the same total distance at the end). Likewise, one time around the inside of a drive platter is less data than one time around the outside of the drive platter, but the rotation takes the same amount of time regardless of platter position (because it's a solid single piece).
 
The drive itself is the bottleneck. It can't physically move fast enough to access the data any more quickly. The data coming out of the drive and into the SATA cable simply can't go any faster than ~150MB/s on consumer drives. No matter what fancy SATA interface or motherboard you hook it to, it simply won't be any faster, due to the laws of physics. SSDs are faster than HDDs, but they do also have a maximum speed that their components are able to reach, and they simply can't go any faster than that.

The actual drive inside that enclosure is most likely a regular SATA II 1TB drive, just like you would buy from Newegg to put inside your PC. It's been a long time since I've seen an enclosure that was anything more than a standard internal drive with an adapter hooked to it. That's why Kyle mentioned trying to take it apart to swap in the Intel SSD. If he could've gotten the enclosure opened, he most likely could have just unplugged the 1TB drive and plugged in the SSD.

So to answer your question, this almost certainly is a SATA II drive plugged into USB3. SATA II drives wouldn't "default" to USB2, because SATA and USB have nothing to do with each other. This enclosure is just a USB3-to-SATA adapter, plugged into a regular SATA drive. The adapter is upgrading from USB2 to USB3, but the drive isn't changing. Any drive you plugged into it would "become" a USB3 drive.

It's possible they're using a SATA III (6Gb) drive, but 150MB/s is SATA I speed, so II vs. III doesn't really matter.

Yes, excellent explanation indeed. 150MB/s end of the line, no matter what other speeds are. I had no idea 150MB/s was the physical end of consumer HDs. So if your SATA connection is giving you enough bandwidth, the limit is the drive itself. This means that SATA I drives suing SATA II, or III are the bottleneck themselves, like you say. So are there any drives on the consumer market that will actually hit their theoretical limit, or close to it?

And now I under stand what Kyle was talking about. Thank a lot for that explanation. I really appreciate it. And so using any HD with a USB 3 connection should give good boosts in transfer speeds, since USB 2 was a bottle neck for high speed hard drives.
 
Actually it does. If you compare the inside track to the outside track, the inside one is much shorter, as you point out. The longer outside track allows it to hold more sectors. In one rotation of the platter, the read/write head may cover 50 sectors on the inside track, while it may cover 90 sectors in the outside track. Because the outside track is longer, it has room for more sectors, meaning more data is handled in the same amount of time.

The sectors are no longer like pieces of a pie, where there are the same number of pieces at the inside and outside, but they're wider at the outside. Current drives actually divide the outer tracks into more pieces of similar size to the pieces on the inside track.

With runners on a circle track, a lap on the inside is a shorter distance than a lap on the outside (hence the staggered starting locations to give each runner the same total distance at the end). Likewise, one time around the inside of a drive platter is less data than one time around the outside of the drive platter, but the rotation takes the same amount of time regardless of platter position (because it's a solid single piece).


Yep. I agree and another clear explanation. I read somewhere recently, and admittedly not too closely, that this was not the case. Can you imagine what was being said that I must have misinterpreted?
 
USB 3.0 just sounds nice haha, and with more speed now.
 
Remember something very important as you create (Kyle) and the rest of us "read" the reviews.

The connection hosting the USB adapter has an impact on the USB 3 speeds. Quite literally an expresscard can only deliver 2.5Gb/s and thus the full 5Gb/s USB 3 potential is not possible. That's why it's pretty important that laptop vendors integrate them onto a bus that can handle the full speed. Further, this is very hard to test because after you get a full speed pci-e slot (must be gen 2 or x4 to have 5Gb/s possible) you then run into what the drive itself can transfer. You really need a high speed SSD to really test for max speed. Now of course, you'll never "hit" the 5Gb/s max so there is also overhead to consider as well.

Kyle is to be commended for leaping in with this test. Just be careful as you read other reviews that are not so thorough. There are a lot of links in the chain that can limit speeds and they may not all be "real" as in what you'll see at the end of the day. What I think is next is testing each laptop and motherboard to see how it holds up because internally "included" USB 3 ports may not all be hooked up to a high enough host port to give full speed.

This makes Intel dragging it's feet on getting USB 3 into chipsets a real pain to the end customer. An expresscard is only going to give a fraction of the speed.

$.02
 
I am not sure, but if Light Peak really does become a reality and Apple is really going to try to push the technology, USB 3.0 might be short lived.
 
I am not sure, but if Light Peak really does become a reality and Apple is really going to try to push the technology, USB 3.0 might be short lived.

Even so, the gap is there. I would rather have USB 3.0 challenging the market to increase speeds instead of maintaining it's stagnant position because Intel hasn't figured out how to release it's latest cash cow. USB 3.0 isn't going to slow Intel/Apple's effort to bring about LightPeak, if anything it should motivate it.

And I doubt USB 3.0 will be short lived, it brings too many upgrades to USB 2.0 (duplexing, power, streaming 1080p). In order for LightPeak to be an industry change, you'll first have to see an article that shows LightPeak cheaper than USB. I think you won't find this happening any time soon.

Given the nature of LightPeak, this may be something great for the enthusiast, but I don't see it usurping USB's dominance in the market ANYTIME soon.

To put it plainly, I'd sure as hell buy a motherboard with USB 3.0 support on it for a few bucks more. I'd rather not have to deal with a lightpeak pci-e card AND a USB 3.0 pci-e card. You may argue you won't need USB 3.0 if you have lightpeak, but it's one thing to implement a LightPeak connection between two powered devices, it's quite another to see a LP Flash drive arrive on the market.
 
To add to the above, I'd love to use FW 800 but it's hard to find the externals I want with FW, or for that matter, and internal USB card reader, to be specific. It's unfortunate that the better technology loses to an inferior technology, but that's the way it's been since the US decided to go with 120v instead of 220v in households (so people selling copper would make more money).
 
Well, for an external HDD/optical drive I'd go with eSATA. I really don't see the selling point of USB 3 there.
 
Earmarks and policy makers thank you :)

To add to the above, I'd love to use FW 800 but it's hard to find the externals I want with FW, or for that matter, and internal USB card reader, to be specific. It's unfortunate that the better technology loses to an inferior technology, but that's the way it's been since the US decided to go with 120v instead of 220v in households (so people selling copper would make more money).
 
Simply that USB 3 is powered and eSata requires separate power interface such as a power adapter or a USB 2/3 for power.

Less is more

Well, for an external HDD/optical drive I'd go with eSATA. I really don't see the selling point of USB 3 there.
 
Simply that USB 3 is powered and eSata requires separate power interface such as a power adapter or a USB 2/3 for power.

Less is more

ESATA is faster, though, and it's not like even USB 3 has an easy time powering an external 3.5" HDD or Blu-Ray burner.
 
Just wondering, how much faster is a 2.5" 500GB notebook drive in a USB3.0 enclosure compared to an off the shelf USB2.0 Seagate/WD external drive? It's not quite clear in the article what drive is inside that Buffalo case.
 
Most modern drives should be faster on usb 3 just because data can be sent full duplex instead of the half as it was in usb 2. Performance should basically be the same as if it was a sata connection.
 
Back
Top