Become a fan of Slashdot on Facebook

 


Forgot your password?
Close
typodupeerror
×
Businesses Hardware

USB 2 Devices Not Necessarily High-Speed 268

Posted by timothy from the words-by-any-other-names dept.
mgcsinc writes "Yahoo is running a story on how some manufacturers of "USB 2.0" devices are making hardware compatible with the USB 2.0 standard, but not necessarily its high-speed component." Sounds like the complaint raised earlier this year.
This discussion has been archived. No new comments can be posted.

USB 2 Devices Not Necessarily High-Speed More

USB 2 Devices Not Necessarily High-Speed

Comments Filter:

  • But do they NEED it? (Score:5, Insightful)

    by pfraser ( 651313 ) on Saturday October 04, 2003 @09:10PM (#7135117)
    Do the devices need that high-speed component? Does a USB mouse need to be able to transmit data in excess of 400mbit/sec? No?
    • Short answer: Yes

      Long answer:

      Have you never heard of a USB storage device? There are tons of uses for USB 2.0. Also, it's not just about speed, it's about false advertising. When I buy a USB 2 device, I expect a USB 2 device, not a USB 1 device.

      • Re:But do they NEED it? (Score:3, Insightful)

        by Kenja ( 541830 )
        Do you get this pissed off when your 100Base-T network dosn't give you 100 megabit throughput?
        • Do you get this pissed off when your 100Base-T network dosn't give you 100 megabit throughput?

          If someone labeled a 10Mbit card as a 100Mbit card and tried selling it, they would get sued for false advertising. And yes, I do get angry if I can't get 100Mb/sec over my ethernet connection.

          • Re:But do they NEED it? (Score:5, Funny)

            by Wumpus ( 9548 ) <[IAmWumpus] [at] [gmail.com]> on Saturday October 04, 2003 @09:32PM (#7135228)
            Then you must be real angry, real often.
          • If you're getting 100Mbps over 100base-T, you're doing somthing you should tell the rest of us about.

            'jfb
            • Interface Received(KBps) Transmit(KBps) Total(KBps)
              eth0 10663.285 54.074 10717.359
              max: 11355.107 58.747 11413.854

              A peak of 11355.107 KBytes/sec transfer over my 100mbit card and an average rate of over 10000 during the entire transfer, maybe not quite the full 100mbit speed, but its pretty close, and taking into account overheads etc.. This benchmark was taken from copying a file over NFS from a 250mhz IRIX machine to my linux workstat
        • No because I know that it's impossible to get 100 Mb (small b for bit).

        • Re:But do they NEED it? (Score:4, Informative)

          by Jucius Maximus ( 229128 ) on Saturday October 04, 2003 @10:39PM (#7135476) Journal
          "Do you get this pissed off when your 100Base-T network dosn't give you 100 megabit throughput?"

          No, but if someone took a 11 MBit device and labelled it as USB 2.0 then I would not like that at all.

          (Now with the new asinine naming convention, USB 2.0 "Full Speed" is actually 11 MBit, so it might not be false advertising.)

          You have to look for the "USB 2.0 High Speed" marking to be sure that it's 480 MBit/s.

    • Dude you have no idea how fast some of us point/scroll/click.

      Roommate1: Whoa, what's wrong with Tod? It looks like he's having a seizure or something?
      Roommate2: Naw, he's fine, he's just surfing the net after 2 quad-lattes & a couple of red bulls.

    • Does a USB mouse need to be able to transmit data in excess of 400mbit/sec?

      Not necessarily. But it does need to not force every other device on the same hub down to the lower speed. Supposedly, when you put a USB 1.0 device on a USB 2 hub, it can limit all the devices on that hub to USB 1.0 speeds. I haven't verified this personally, and don't know whether USB 2 devices that don't run at the high speed limit the speed of other devices on the hub or not - the article doesn't even cover this.

      • This is not true. If it were, then if you plugged ANY 1.1 device into your computer it would all slow down to 1.1 speeds. I have a 1.1 hub with all of my periphials (mouse, keyboard, PDA, 1.1MS reader, USB audio) and then I have a 2.0 CF reader attached to the other port on my computer (actually I have 5 more ports, but that's not the point...). When the CF reader is attached to the 1.1 hub I get 800KiB/s off of it; when it's not on the hub I get 4MiB/s. So if plugging in a 1.1 device slows stuff down i

        • Re:Unanswered questions (Score:2, Informative)

          by Obyron ( 615547 )
          Your assertion that plugging ANY 1.1 device into your computer's USB 2.0 ports should slow the whole thing down is slightly erroneous.

          The reason plugging a 1.1 device into a USB Hub that then plugs into a -SINGLE- USB 2.0 port slows everything else on the Hub down is that the Hub is plugged into a single port, which has its own bus. You slow that individual USB bus down to 1.1 speeds. The other USB 2.0 ports on your computer have a separate communications bus that continue to operate at the expected 2.0

        • Re:Unanswered questions (Score:2, Informative)

          by Anonymous Coward
          Uh. You're confused.

          When USB was new each USB port on your system was attached to the same USB backplane. The bandwidth was therefore shared between all ports on your system, exactly like a hub.

          Today it is commonplace to have a seperate backplane for each port. So plugging in a USB 1.1 device to one USB 2.0 port will not affect other USB 2.0 ports. That is because they're on different backplanes, not because the USB 1.1 device isn't slowing down the USB 2.0 bus.

          Now, if you were to plug in a USB 2.0 hub i
        • IIRC, the way USB shares bandwidth is by time slicing. That means that if you have a low-speed peripheral, and that peripheral is maxed out moving data, then it is using up time slices for inefficient transfers. If the device doesn't send much data anyway (like a keyboard), then it will have minimal impact on the bus because it hardly ever needs a time slice.

          I'm not familiar with the details of how USB arbitration works, but a simplistic example would say that if one slow device and one fast device were c

        • I said:

          Supposedly, when you put a USB 1.0 device on a USB 2 hub, it can limit all the devices

          on that hub to USB 1.0 speeds

          and you said:

          This is not true.... I have a 1.1 hub with all of my periphials (mouse, keyboard, PDA, 1.1MS reader, USB audio) and then I have a 2.0 CF reader attached to the other port on my computer

          Do you have a 2.0 hub to try plugging one of your USB 1.0 devices into, and then measure the speed of the flash reader? Has anyone else tried this to either confirm or deny its trut

    • Though it can be agreed that USB mice and keyboards don't really need that much bandwidth, you're forgetting that there are plenty of USB devices out there that DO.

      For example, there exist external hard disks and MP3 players that connect to the computer through USB. These need all the bandwidth that they can get. The more you can transfer, faster it can get done. Moreover, I actually have a USB mouse that doubles as a memory stick reader. Does it need transfer speeds of up to 400mbps? Well, it sure would

    • Does a USB mouse need to be able to transmit data in excess of 400mbit/sec?

      Not at first glance, but while a mouse by itself does not need this kind of bandwidth sustained other devices on the same chain just might. So if the mouse gets what little info it has to send out of the way 40 times faster then that just clears up the bus more quickly and lets the higher speed devices back in the action faster.

    • Comment removed (Score:4, Insightful)

      by account_deleted ( 4530225 ) on Saturday October 04, 2003 @10:32PM (#7135449)
      Comment removed based on user account deletion
      • The article mentions that many devices aren't high-speed capable. I have a ieee1394 CF card reader that in theory can transfer at 400 Mbps, but the CF card seems to be the limiter. I get 42 Mbps with this arrangement, which is satisfactory.

        To your point, though, if I were limited to 12 Mbps I'd definitely notice that. I might even be damned pissed. :)

        My HP 7550 printer has a CF card reader, and this is the slowest way I've found to transfer images to a PC. It must implement the low-speed 1.5 Mbps conn
    • Does a USB mouse need to be able to transmit data in excess of 400mbit/sec?
      Speedy Gonzales does.
    • Do the devices need that high-speed component? Does a USB mouse need to be able to transmit data in excess of 400mbit/sec? No?

      Well, actually, probably yes.

      If your mouse works at 480 mbit/sec then it wont slow down anything else on the hub. If it runs in a legacy mode such as 1.5 mbit/sec, everything else timeslices with it and runs much slower.

      Michael
    • What about a digital camera or a portable harddrive or a CD rewriter? If it says USB 2.0, is it high speed or just a relabelled low speed USB 1.1 device? Is it going to have slug like performance or perform the way expected from the implemented version USB? If I plug this device into my existing hub is it going to fuck up the performance for other devices?

      The fact is that allowing manufacturers to relabel previously USB 1.1 devices as USB 2.0 is confusing as hell. More honestly they would tag devices with

    • > Does a USB mouse need to be able to transmit data in excess of 400mbit/sec?

      Well, if the USB mouse sends slow packets, it blocks the bus for a long time. A lot of "fast" packets fit into one "slow" packet slot. Fast devices will suffer an unproportionally high bandwidth penalty, when you wiggle your el-cheapo USB2 mouse. After all, it's a shared resource.
    • This is not the right question to ask.
      obviously no, some devices do not and they should be manufactured and labeled as USB 1.0 or 1.1. More appropriately, why call a device USB2.0 if does not incorporate the 2.0 speed? IMHO the device should be proven capable of transmitting at a decided rate to meet requirements for the USB2.0 standard.

  • Is this disingenuous? (Score:4, Insightful)

    by BizidyDizidy ( 689383 ) on Saturday October 04, 2003 @09:12PM (#7135123)
    There's no real false advertising here; just an assumption on the part of consumer.

    Should it be necessary that they inform you of the lack of full speed utilization? What if it's faster than USB 1.0 but not FULL speed.

    IMO, the only clearcut measure is whether the standard is met, and it seems to be.
    • Generally it's implied when an external hard drive posts the "Hi-Speed" USB logo that the hard drive will perform to something over 11 mbps. Granted, with a mouse it's not too critical, but for hard drives and high resolution cameras it can be a pretty large issue.
      • Generally it's implied when an external hard drive posts the "Hi-Speed" USB logo that the hard drive will perform to something over 11 mbps. Granted, with a mouse it's not too critical, but for hard drives and high resolution cameras it can be a pretty large issue.

        But is that the fault of the manufacturer, or the standard setter? This logo is the generic USB 2.0 logo. They've chosen to use that logo, despite the fact that high-speed is only a part of the standard.

        A good review of USB 2.0 can be

    • Re:Is this disingenuous? (Score:2, Insightful)

      by pla ( 258480 )
      Should it be necessary that they inform you of the lack of full speed utilization?

      If Dell sells you a PC with a "Pentium 4 3.2GHz", would you feel a tad bit peeved to discover that, while it actually does have a 3.2GHz P4 in it, they chose a noname chipset that only clocks it at 800MHz?

      Because, that would satisfy your condition - It has the advertized part in it, but only clocks it at 25% of its rated maximum.

      Yes, people expect (and should expect) a product to make full use of the standards it suppos
      • First of all, your post reminded of the premise from this movie [imdb.com] where a con man sells old people "numbered copper engravings of Abraham Lincoln" for something like 49.99, then mails them a penny.

        USB 2.0 is a standard, not a component. When I buy a computer that advertises certain components, I better damn well get them.

        When I buy a product that advertises a certain standard, I better damn well get it. And I am.

        We both know there is a problem. I argue that it's not at all the problem of the person who

        • A better example is a P4 3.2ghz chip that clocks itself down when it gets too hot. They sold you a P4 3.2ghz chip. It's the fault of Intel and their design that makes it slower. You're absolutely right: the standard is the problem, since any of 3 different speeds can satisfy the standard.

          For a non-computer analogy, try black jelly beans. Black jelly beans can be either licorice or grape (depending upon the brand). If I buy a bag of black jellybeans, I assume they will be licorice. If they're not, and
    • The problem originated when USB 2 first made its rounds. For the first year or so, the only people using or selling USB 2 devices were manufacturers that were selling high-speed devices that utilized the extended bandwidth.

      I didn't even know until the first article that USB 2 -could- be dropped into lower bandwith. I am sure I'm not alone.

      Now, we start seeing companies without fully developed USB2 high speed implementations dumping USB2 compatible crap onto the market making many people pissed off. I was
    • Certian Notebook and chipset manufacturers petitioned the board to make this change because CONSUMERS were passing up USB 1.1 spec'd computers for USB 2.0 spec'd computers. There was no confusion at all here. For a matter of fact, they are sowing confusion, not fixing it.

      Because the USB 2.0 "mark" being used by many manufacturers was never offically sanctioned by the "committee" they decided to change the tradename rather than simply say "tough luck" So Now pretty much everything is USB 2 [note: not USB

    • There's no real false advertising here; just an assumption on the part of consumer.

      BS. This is the same attitude I was given when I built my first PC from parts. ("Gee, you didn't ask for a full-speed drive controller just a full speed drive!") It was dishonest then -- or at a minimum arrogant snobbery -- and this is no different.

      IMO, the only clearcut measure is whether the standard is met, and it seems to be.

      Yet, that's not really what was promoted, was it? USB 2 at USB 1 plus a little isn't wh

    • The biggest concern is that people might not notice the "full" speed and "high" speed being two vastly different things. It's not as if they are close, 12Mbps and 400Mbps are vastly different animals. It is completely concieveable that a maker of USB 1.1 12Mbps external drives could just touch up the firmware, relabel their drive as a USB 2.0 "high" speed when it is practically crippled compared to a drive that has a "full" speed.

      I look at the packaging of USB products and not all of them declare what co

  • Whaaa? (Score:2, Funny)

    by Anonymous Coward
    You mean my "USB 2.0" mouse is not high-speed?

  • So? (Score:4, Funny)

    by bconway ( 63464 ) on Saturday October 04, 2003 @09:13PM (#7135134) Homepage
    This isn't much of a revelation, it just means that the USB connection isn't the bottleneck. ATA133 drives won't run at 133 MB/s, either, I wonder if someone's going to start complaining about that now. ;-)

  • cost of backwards compatibility (Score:5, Insightful)

    by Creepy Crawler ( 680178 ) on Saturday October 04, 2003 @09:19PM (#7135165)
    Well, many devices do not need the 480Mb/s speed as said in the article however, the devices are backwards compatible.

    So why is everyone whining? Just have the anufacturers put max speed transfer on the boxes as spec sheet. Just dont buy anything from those manufacturers who dont.

    And as a sideline about the jab about printers not meing "full-speed", who cares? Paralell printers, in epp/ecp mode, could only transfer max 11Mbps. And since consumer printers dont print very fast, what's the big deal? And it it was made to be fast, it'd have a network jack for 100Mbps connectivity.

  • GIGABYTE GA-7VAXP Motherboard (Score:3, Informative)

    by Anonymous Coward on Saturday October 04, 2003 @09:30PM (#7135218)
    This explains why I receive a warning message saying "Hi-Speed USB 2.0 device plugged into a non Hi-Speed USB 2.0 port." whenever I plug in my fancy new Nomad JukeBox Zen NX. I tried everything: uninstalling and reinstalling the latest drivers, third-party drivers, microsoft's drivers, different ports on the motherboard...nothing would work. Instead, I had to wait many many hours to transfer my music over to my mp3 player. On the website [giga-byte.com] it doesn't make any distinction between Hi-Speed and Full Speed. Maybe I can return it for false advertising?
    • I have the same mother board, at least under Linux it shows me a 6-port USB 2.0 device (line from usbview: Speed: 480Mb/s (high)) with the EHCI driver loaded, and 3 2-port USB 1.0/1 devices as well (when the UHCI module is loaded). Could it be that what ever OS you are running (ie Win2K SP3 or earlier I believe is one that doesn't do USB 2.0 HiSpeed speeds ) doesn't support the USB 2.0 ports, or at least the EHCI interface? (And don't forget the BIOS options to turn on the USB 1.1 controller and the USB 2.0

    • This explains why I receive a warning message saying "Hi-Speed USB 2.0 device plugged into a non Hi-Speed USB 2.0 port." whenever I plug in my fancy new Nomad JukeBox Zen NX. I tried everything: uninstalling and reinstalling the latest drivers, third-party drivers, microsoft's drivers, different ports on the motherboard...nothing would work. Instead, I had to wait many many hours to transfer my music over to my mp3 player. On the website it doesn't make any distinction between Hi-Speed and Full Speed. May

  • Surprised? (Score:5, Insightful)

    by Deaper ( 659229 ) <deaper@[ ]il.com ['gma' in gap]> on Saturday October 04, 2003 @09:36PM (#7135246)
    Does this actually surprise anyone? I mean these are the same people who have been decieving consumers [slashdot.org] for years with hard drive sizes. Do you actually think they're going to tell you the device doesn't work the way it should by labeling it as such? The record industry doesn't label crippled crap [slashdot.org]. Why should the computer industry? Why do consumers actually trust producers that constantly try to implement new technology that assume that the consumer is the one that can't be trusted? Does it make me angry that the producers keep doing this kind of crap? Yes. Does it surprise me? Not in the least.

  • bad packaging (Score:3, Insightful)

    by chickenwing ( 28429 ) on Saturday October 04, 2003 @09:38PM (#7135253) Homepage
    The problem is that the boxes of most boxes are covered with pictures of people oohing and ahhing but don't have much information about what is actually inside the box. Even manuals (if you are lucky enough to get one) are very light on specs.

    I see this as one of the more unfortunate side-effects of the mass adoption of computers. Most people will never realize their hardware is crappier than they thought, and the rest of us are told to shut up and be good little consumers. I get my revenge by buying multiple models off the shelf, and returning all but the one that makes the cut.

  • Stick with Firewire (Score:3, Informative)

    by Otto ( 17870 ) on Saturday October 04, 2003 @09:45PM (#7135280) Homepage Journal
    If you need to attach a device that can use the high speed connection, like a hard drive or an MP3 player or something, then stick with firewire. It's easier to deal with. Of course, USB is fine for your mouse or keyboard or what have you, but trying to sort out the differences there is just too much of a pain. Firewire has various speeds too, but I've yet to see a firewire device that really needs a high speed work at a lower one.
    • FireWire doesn't divide the bandwidth among each device regardless of what they're using; I believe it's packet-based and each device uses only as much bandwidth as it needs.

  • When are they going to learn... (Score:5, Insightful)

    by A Commentor ( 459578 ) on Saturday October 04, 2003 @09:46PM (#7135288) Homepage
    So the industry keeps thinking it can pull one over on the consumers and make more money, but it's just the lawyers that will win on this one...
    • Just like:
    • Screen sizes include areas you can't actually see. 15" really means 13.8"...
    • Harddrive sizes are not what they seem. 1M == 1,000,000 bytes not 1024*1024.

    The computer industry will keep making disceptive ads and lawyers doing class-action suits against them will keep getting rich. My guess is that the computer industry still make more money than they lose to the lawyers, so everyone is better off but the consumer...
    • Except that when I read screen sizes, it almost always lists the viewable size of the screen nearby. And when I see hard drive sizes, I almost always see 1M = 1,000,000 bytes or something like this. I have never seen specs indicate high- or full-speed, though. Has anybody actually seen this on spec sheets, either in manuals or on ads?
    • "So the industry keeps thinking it can pull one over on the consumers and make more money, but it's just the lawyers that will win on this one...

      Just like:

      • Screen sizes include areas you can't actually see. 15" really means 13.8"...
      • Harddrive sizes are not what they seem. 1M == 1,000,000 bytes not 1024*1024.

      The computer industry will keep making disceptive ads and lawyers doing class-action suits against them will keep getting rich. My guess is that the computer industry still make more money tha

  • Choose Firewire. Problem solved. (Score:5, Informative)

    by The Ancients ( 626689 ) on Saturday October 04, 2003 @09:47PM (#7135290) Homepage
    IEEE1394 appears to be faster [techtv.com] than either USB 2.0 full speed or high speed.

    Disclaimer: This being /. the above is more with regard to cameras external HDDs, and other hardware which would benefit from the higher speed.

    • Not only is FireWire 400 faster than USB 2, but FireWire 800 (IEEE1394b) is even faster than that. Built into new PowerMacs and PowerBooks (except the 12"PB), and available here [orangemicro.com] here [ezq.com] here [lacie.com] and here [fwdepot.com] (quick Google results).

  • Article is rather pointless/common sense (Score:4, Insightful)

    by reiggin ( 646111 ) on Saturday October 04, 2003 @09:48PM (#7135297)
    What's so surprising here? There is nothing of real merit in this article. It's all stuff any educated consumer should realize. And as for the uneducated consumers.... well, they probably aren't reading Slashdot. My point is, why is this frontpage Slashdot news?

    All it's saying is that mostly only hard drives and burners are using "hi-speed" USB. We all knew that, right? And if we thought that maybe printer and flash memory readers were really using "hi-speed" we at least knew that they couldn't take advantage of the full bandwidth, right? And we all have come to expect manufacturers to lie to us on the packaging, right?

    Yes, it is disappointing that companies are using the USB 2.0 hype to sell lower speed products. But what's the big deal?

    The only thing that would really piss me off is if the hard drive and cd/dvd-r burners WEREN'T using "hi-speed" USB 2.0. But the article says they are! Or at least it doesn't say they aren't. And that's my second point: The article doesn't really say much of anything. It only puts the question of authenticity in the reader's mind. I think it is a poorly constructed article and not very worthy of Slashdot attention.

    I'm not trying to flame or troll. I am just really missing the significance of this article.

    • I'm a technical type and I got burned by this. I bought a new ink-jet printer which said "USB 2.0 Full Speed" on the outside of the box. I started looking for a cheap USB 2.0 PCI card so that I could dump bits to the printer at 400 Mbps. I didn't realize that "Full Speed" meant 12 Mbps.

      Do you know why I was fooled? It was because I had been reading Intel propaganda on USB 2.0 for many months. Every paper on USB 2.0 touted its blazing speed, allegedly making it the equal of Firewire. They didn't say anythi

  • This is old, USB already defined it (Score:5, Informative)

    by mlrtime ( 520968 ) on Saturday October 04, 2003 @09:50PM (#7135307)


    The USB2 does not signify high-speed of course. If you want USB High Speed you need to look for the high speed logo.

    as in this image here: http://www.usb.org/images/headermain/2logos.gif

    The one on the left is the high speed, one on the right is regular speed. Simple eh?

  • I'm confused (Score:4, Funny)

    by platypus ( 18156 ) on Saturday October 04, 2003 @09:59PM (#7135345) Homepage
    Weren't USB 2.0 "highspeed" devices actually the slow ones? So, if you have a slow device, it is highspeed, isn't it?
    Or was it Big Speed?

    Wait.

    USB2.0 Huge Speed. No, that wasn't it

    I'm seriously confused.

  • The standards: Defined... (Score:5, Informative)

    by Anonymous Freak ( 16973 ) <anonymousfreak@nOspam.icloud.com> on Saturday October 04, 2003 @10:33PM (#7135453) Journal
    Okay, there are two major 'versions' of USB, and three speeds.

    USB 1.1 is the 'old' standard. USB 1.1 defined the 2Mb/s and 12Mb/s speeds ('Low speed' and 'Full speed' respectively.) USB 1.1 devices are fully compatible with USB 2.0 devices, but, of course, can only run at 12Mb/s maximum. Note that hubs that are only USB 1.1 compatible will only allow 12Mb/s maximum through them.

    USB 2.0 is the current standard. It is fully compatible with Low and Full speeds, plus adds 480Mb/s 'Hi-Speed'. Any USB 2.0 compatible controllers (computers,) can run any device that supports any of the three speeds. USB 2.0 devices that are 'Hi-Speed' are also supposed to support Full speed for compatibility (For example, that 52x CD-RW drive should support Full speed, but will drop to 4x speed, when connected to a USB 1.1 controller.)

    The official 'branding' of devices is that they should *NOT* specify USB 1.1 vs. USB 2.0. They should only say the speed they operate at. So Low or Full speed devices (mice, keyboards, printers, etc,) should have a 'USB' logo, with no version numbers, just the USB logo. 'Hi-Speed' devices (hard drives, CD-ROM drives, camcorders,) are supposed to use the 'USB Hi-Speed' logo, which, again, does not say 'USB 2.0', only adds 'Hi-Speed' to the normal USB logo. Companies that use "USB 2.0" branding to advertise any device are not complying with the USB group's marketing standards.

    But, yes, a USB 2.0-compatible device can very well operate at 2Mb/s, or 'Low' speed. A good example is keyboards with built-in hubs. My old keyboard is only USB 1.1, so I can plug in any device I want, but it will run at 'Full' speed (12Mb/s) maximum. Newer keyboards have USB 2.0-compatible hubs, so even though the keyboard itself is 'Low' speed, you can plug in your external HD, and the hard drive will happily run at 480Mb/s to your host computer. (Obviously, you also need a USB 2.0-compliant host controller in your computer.)

  • If they did the same with processors a Pentium would be advertised as a Pentium IV because the Pentium IV is backward compatible with the Pentium but can also go faster.

    So a Pentium is now a low-speed Pentium IV; a Pentium II a full-speed Pentium IV and a Pentium IV a high-speed Pentium IV**.

    They may qualify under the technical standard as USB2.0 but it clearly is labelled as such to deceive the customers into expecting a faster device.

    *the fine print says it's a low-speed one.

    **I know it is not a perfe

  • A lot of people fail to notice the real issue. (Score:5, Interesting)

    by stvangel ( 638594 ) on Saturday October 04, 2003 @11:58PM (#7135743)
    USB as well as FireWire are still a single set of wires. Unless you use some sort of USB router or multiple controllers, all the devices must share a finite amount of bandwidth based on the ratio of their speeds.

    For example: A device running at 2mb speed that sends 500kb in a second uses a full 1/4 of the entire USB bandwidth. This automatically chops the 12mb down to 9mb, and the 480mb down to 360mb. A 12mb device that sends 6mb cuts it in half.

    By the time you have a keyboard, mouse, joystick, mp3 player, external drive, and who knows what else sharing the USB connection, you have a lot of things competing for limited bandwidth with the slower devices taking an inordinate share of the pie. This is one of the reasons I like sticking to the old PS/2 style Mouse and keyboard connectors. Keep these usually slow devices from flooding the connection. Particularily the high-res mice.

    And then when you consider the 2mb/12mb/480mb numbers are the absolute maximum theoretical numbers without overhead, you realize that you get nowhere near this kind of throughput in the first place. Things can get bogged down pretty quickly.

    Personally, I run two separate USB adapters. The built-in USB on the motherboard and a separate PCI USB controller. I leave all the slow things like keyboard and mouse and joysticks on one controller. I put the things that need speed like a dvd burner or mp3 player on the other one and make sure I don't use them at the same time.
    • You for get that most modern boards have several separate USB controllers. Who would ever want to connect keyboard, mouse, joystick, mp3 player, external drive off a single controller using a HUB???

      The minumom number of controllers I have seen on today's machines is three (two UHCI and one EHCI).

  • Host controllers are slower than stated too (Score:3, Informative)

    by tim.kerby ( 206359 ) * on Sunday October 05, 2003 @05:16AM (#7136062)
    I develop a number of video products for USB 2.0 and previously USB 1.1. I have some experience of driver development but I am technically a hardware engineer with a leaning towards microcontroller development.

    Recently I started developing with USB 2.0 assuming that I could get maybe 50MB/sec data through (480Mbps - overhead) the high speed mode of USB 2.0. Note that full speed is lower than high speed in the spec ?!?!?!?

    What I found was that on PCI / cardbus plugin cards, this was actually reduced to about 20MB/sec. This is less than half what you seen on product boxes.

    The issue is that host controllers are at fault. USB 2.0 contains a number of slots in each frame on the bus that can be filled with data. If I remember correctly, there are 13 available slots for bulk transfers that can take 512 bytes each. Technically, 12 of these shoulb be the theoratical maximum limit to fill. In practice, many controllers only fill 3 giving the poor bandwidth as they cant keep up with the data rate.

    The other issue is with the PCI bus. On many computers this is not fast enough to deal with a single device needing high speed bandwidth although in most cases it does not appear to be the bottleneck.

    Most add-on USB 2.0 host controller cards contain a chip from one manufacturer (who I choose not to mention). These suffer the worst performance of 18-20 MB/sec. They comply with the Intel EHCI 0.95 spec for host controllers although the manufacturer has offered a new 1.0 compliant chip offering some increases in speed.

    The best performance is when USB 2.0 is tapped from Intel North Bridges on the motherboard. 11 of the slots are filled with data and 35MB/sec can be achieved. Its still not the maximum performance though

    If you are buying a PC, make sure you insist on built in USB 2.0 or all your devices may run slow. Also make sure you only use the Microsoft drivers on Windows as they offer significant improvements over others. Win XP or Win2K SP4 contain these.

    Note that the USB 2.0 and EHCI 1.0 specs do not contain any specification as to the bandwidth a host controller must provide. Some chips may be better or worse than those mentioned above as there is no control on what a manufacturer should provide

  • If you are cursed with VIA USB 1.1 controlers (UHCI) having slow 2.0 device that will talk to the EHCI par of the controler makes LOTS of sense. The EHCI/OHCI are infinitely smarter and have a much lesser impact on the PCI than UHCI!

    So I'd be happy to find a USB 2.0 mouse, if I had a VIA chipset!
  • Remember when they renamed USB 1.1 to USB 2.0, and what had been USB 2.0 (the high speed) was renamed to USB 2.0 HighSpeed?

    The bottom line is that to claim USB 2.0 compliance, it doesn't actually have to be USB 2.0 compliant. If that makes any sense. My laptop was made USB 2.0 complaint overnight by that change. I didn't even have to install anything! Wheee! Feel the burn?
  • I bought a sandisk little keychain usb drive, because it costed 60$ with 256mb of ram, and it is supposed to be usb 2.0 hi-speed compatible, as advertised on the box.

    I only get 1.5 mb/second from it (12 mbps a.k.a 1.1)

    This is fraud, scam, ripoff
  • Let's see what they say about it.

    My box of my cruzer mini drive says "hi-speed usb 2.0" and displays the logo "hi-speed certified usb 2.0", so i expect 480mbps from it instead of the lousy 12mbps i am getting.

    Yes, i have usb 2.0 ports with their drivers installed, and yes, i have used true 2.0 devices at full speed before on them.

Slashdot Top Deals

The hardest part of climbing the ladder of success is getting through the crowd at the bottom.

Close