Driver download services

Discussion in 'DIY Computers' started by Philip Herlihy, Nov 2, 2009.

  1. I've used a couple of services which identify the hardware on a computer
    and provide updated drivers where they are available. These include
    driveragent.com, and driver genius. Both are just a little too
    expensive for me to want to use them on every machine that comes my way.
    Any recommendations for alternatives? A utility which listed all the
    hardware on a machine would be almost as good.

    Phil, London
     
    Philip Herlihy, Nov 2, 2009
    #1
    1. Advertisements

  2. Philip Herlihy

    Andy Guest

    I think Sisoft's Sandra does this. Ther used to be a free version I
    think, but it doesn't look to be around now.

    However there seems to be a version that runs from a memory stick.
    It's not free but it might well meet your needs as after the initial
    payment you could use it on various machines.

    I've not tried it, but there seems to be a trial version so it might
    be worth a look.

    Here's the link:

    http://www.sisoftware.net/


    Andy
     
    Andy, Nov 2, 2009
    #2
    1. Advertisements

  3. Philip Herlihy

    Rob Morley Guest

    If you boot a recent Linux live CD and run "lspci -v" you get a pretty
    comprehensive list of the hardware configuration.
     
    Rob Morley, Nov 2, 2009
    #3
  4. Thanks, I'll try that!

    Phil
     
    Philip Herlihy, Nov 2, 2009
    #4

  5. Looks promising, but it's very expensive (almost £350). Wouldn't use it
    quite often enough to justify this (and the need to keep it up to date.
    Thanks anyway - an interesting option.

    Phil
     
    Philip Herlihy, Nov 2, 2009
    #5
  6. Even more with lspci -vv.
     
    Mike Tomlinson, Nov 2, 2009
    #6
  7. The message <_KxHm.17910$2>
    Try googling for "UnknownDeviceIdentifier.exe". This is a useful tool
    for identifying and allowing you to search the internet for sources of
    drivers for whatever you choose from its list of discovered hardware.

    Admittedly, it produces an overly detailed search term for its
    submission to google but you can easily enough trim this to produce more
    hits if need be. Most of such searches pick up hits on the likes of
    driverguide.com et al but you're free to try the manufacturer's sites if
    you're not happy with the likes of driverguide.

    It doesn't run directly. It needs to be installed but it only takes
    moments to do so (and the same to uninstall when you've finished with
    it). I don't use it often, only when I'm dealing with really obscure
    hardware or else a problem arising with the automatic verification that
    the system has the appropriate hardware to which the driver software
    installer relates _is_ actually fitted (typically, Nvidia's driver
    software installer).

    The last time I used UnknownDeviceIdentifier.exe was to verify that the
    hardware (a TNT2 adapter) was detectable in spite of the Nvidia
    installer claiming otherwise and refusing to run as a result. In the end
    I successfully forced an install via windows driver install, pointing it
    to the unpacked driver files created by the Nvidia installer prior to
    it's refusal to accept there actually was a TNT2 card in the system.

    In this case, I think the Nvidia driver 'leftovers' for a later, but
    incompatable, card[1] that I'd had to pull were the source of confusion
    to the Nvidia installer which usually "just gets it right".

    [1] Well, I'm hoping it was simply an incompatability issue between the
    MX440 SE AGP adapter and the SiS chipset used by the Elite K7S5A skt A
    MoBo I had used to replace the burnt out Packard Bell 'special' skt 478
    (also a SiS chipset) MoBo.

    Initially, after a successful repair install of winXP HE on the
    repaired system, I was seeing rather peculiar behaviour on display
    updating when opening folder windows, which after updating the SiS AGP
    port driver software and applying the latest Nvidia update, upgraded
    itself to random system crashes. Replacing the MX440 with the TNT2 card
    totally resolved the problem.

    Since neither ATI or Nvidia actually make graphics adapters (just the
    chipsets) and leave it to third parties to create product from the chips
    and reference designs, you sometimes see 'broken implementations'
    (usually by MoBo makers who "Don't Quite Get It").

    I'm thinking of an MX440 based card by,iirc, a MoBo maker calling
    itself 'Mercury' which produced jumpy motion in OpenGL games. Obviously
    an artifact because exactly the same MX440 graphics made by Jetway
    (amongst others) did not suffer from this defect.

    Looking at the initial display when booting a VIA chipset MoBo from a
    Knoppix Live CD to test the card, suggests it might be product from some
    nameless and clueless manufacturer. I've yet to install a win2k test
    setup to fully test this card to decide whether or not it's one for the
    bin.
     
    Johnny B Good, Nov 2, 2009
    #7

  8. Found it - http://www.zhangduo.com/udi.html

    There appear to be versions of this around with trojan passengers.

    Thanks - I'll study it carefully!

    Phil
     
    Philip Herlihy, Nov 2, 2009
    #8
  9. The message <LFDHm.39191$%%2>
    You're right to be careful. FYI, the version I've been using seems to
    be just over 5 years old (it was nearly a year old when I downloaded it
    on 22Jul2005). File properties doesn't reveal a version number on this
    download whereas it does for the version you found.

    The version I have is 836,464 bytes long and has an MD5 checksum value
    of: 4d1f9458cd5b5d41fda864cb2cc3e12f. Submitting it to a clamwin scan
    suggests it is free of malware ( and post install runs of Spybot S&D on
    systems I've used it on backs up this finding).

    I've just downloaded the current version you found and scanned it with
    clamwin which gave it a clean bill of health, as did submitting it to
    the virustotal website. However, with recent files (say less than 6
    months old), the fact that all 41 AV engines give a negative result is
    no guarantee that it isn't infected with a cunning zero day threat :(

    The real test only comes _after_ you've installed it and you do another
    SpBot scan and, (for the realy paranoid) re-run the ComboFix tool to
    verify the lack (or otherwise) of a zero day threat this may have
    introduced.

    Although this new version looks to be virus free, I'll not be entirely
    confident until I've had a chance to install it on a test system and run
    the SpyBot scans and the ComboFix tool. Since I need to create such a
    test system to test that suspect MX440 graphics adapter anyway, I might
    be able to confirm this in a day or two's time.
     
    Johnny B Good, Nov 2, 2009
    #9
  10. Much appreciated!

    Phil
     
    Philip Herlihy, Nov 2, 2009
    #10
  11. Thanks - driverscollection.com looks promising. I've used driverguide
    before, but it was very hit and miss, and I didn't feel it was worth it
    when they started charging. Driverscollection seems to be funded by
    advertising.

    I don't mind paying for a utility - it's just a question of perceived value!

    Phil
     
    Philip Herlihy, Nov 2, 2009
    #11
  12. Philip Herlihy, Nov 2, 2009
    #12
  13. The message <ojUHm.30615$2>
    from "Mike the Brewer" <> contains these words:

    ====snip====
    Or use UnknownDeviceIdentifier.exe instead. ;-)

    I used to have a free membership account with driverguide.com many
    years ago. Actually, I've had two (they invalidated all existing
    accounts when they made some changes, offering free membership signup on
    their revamped service). The second account _might_ still be valid but I
    can't recall what password I'd used and I now can't be bothered to sign
    up yet again.

    As I recall, the most problematical piece of hardware I had to track
    down drivers for were dial up modems. The driverguide site always
    produced the goods in this regard except for one particular modem. The
    one exception being a host controllerless PCI modem card I'd sold to a
    friend about a year or two before.

    He kindly donated it back to me when his need had become redundent.
    Unfortunately, he'd never thought to hang onto the all important driver
    floppy disk so all I got off him was the bare card. "No problem!" I
    thought, "Driverguide to the rescue!".

    Little did I know that finding a driver for this particular modem was a
    quest doomed to utter failure (and, believe me, I tried off and on over
    the next 12 to 18 months before I finally gave it up as a bad job).

    I'm not decrying driverguide.com (well, not as it was when I _did_ have
    a need to use its services - it might be a different case now with all
    the rampant commercialism that has poisened most internet resources
    today).

    TBH. I'm reluctant to sign up to any such "Free Service" since I feel
    as though I'm selling my soul to the Devil whenever I've done so in the
    past. Knowing that the internet is awash with scumware merchants and
    scamware operations, you need to be very circumspect when considering
    such "Free Services".
     
    Johnny B Good, Nov 3, 2009
    #13
  14. Thanks, Mike - there's a free version (non-commercial use) and a
    commercial version which isn't too expensive. I'll try it out next time
    I need to audit a machine.

    Phil
     
    Philip Herlihy, Nov 3, 2009
    #14
  15. Yup. I took up binning them and replacing them instead - they were
    under £15, so there's no sense spending more than 20 minutes on the
    job.

    Cheers - Jaimie
     
    Jaimie Vandenbergh, Nov 3, 2009
    #15
  16. The message <>
    I totally agree, however, I wasn't referring to what you call a
    winmodem (which, properly speaking. is actually a softmodem) but a host
    controllerless modem (a completely different animal to the much despised
    softmodem).

    A softmodem is essentially a glorified soundcard with a phoneline
    interface and absofuckinglutely no DSP capabilities whatsoever which
    leaves not only the (trivial) overhead of supervisory control to the
    main CPU but also the strain of DSP to be handled in software.

    The host controllerless modem _does_ have DSP built in, it merely
    leaves the higher level supervisory functions to software. The
    difference is that whilst you needed a minimum of a Pentium MX 166 to
    support a soft modem, you could fit a host controllerless one into a
    486DX66 powered system and not spot any performance issues compared to
    using a traditional external rs232 connected modem.

    What annoyed me the most about softmodems was the marketing bullshit
    regarding the 'benefit' of the softmodem being that you'd be able to
    upgrade the transmission speed in software to meet any future speed
    hikes that, in the past, had required you to buy yet another model to
    take advantage of such. Yeah, right! That _might_ have been useful if it
    had been introduced ten years sooner rather than when the max possible
    line speeds had already been reached.
     
    Johnny B Good, Nov 4, 2009
    #16
  17. I had a similar situation with Driveragent.com. Paid for a lifetime
    subscription (about £25 ISTR) which was fine until they introduced a
    limit to 10 machines, and then only for a number of years. Didn't like
    them changing the terms unilaterally and haven't used them since.

    Phil
     
    Philip Herlihy, Nov 4, 2009
    #17
  18. The message <>
    No need to apologise, you're quite correct. :)
    I don't see how you come to that conclusion. I'd estimate the
    controllerless bit in the soft modem represents no more than 10% the
    total software overhead involved in making the main CPU "Do the Voices
    as well". The host controllerless at least deals with the donkey work of
    "Doing the voices" with its built in DSP.

    Since these cards were only available in PCI form, we'll never be able
    to demonstrate the superiority of a real modem card over a host
    controllerles modem in something like an 8MHz clocked 80286 powered AT.

    By the time PCI slots appeared, the oldest technology MoBos that were
    so blessed would have been 80486 powered which could comfortably afford
    the processor overhead required by the host controllerles type.

    The softmodems required a minimum of the much later 166MHz clocked MMX
    pentium in order to function and leave just enough cpu resources free to
    service the needs of a web browser or whatever comms app that was in
    need of a dial up connection.

    The big hint that you had bought a softmodem, if the packaging hadn't
    made that fact immediately obvious, was the "Recommended minimum system
    requirements" (usually a P166mmx or higher cpu). If no such minimum
    requirements were stated, you'd either bought a 'real modem' or (more
    probably) its braindead cousin, the host controllerles type.

    Quite simply, the manufacturers knew there was no need to spell out a
    minimum cpu requirement since this would most likely have been "a
    386DX33" or lower cpu and, to my knowledge, no PCI slotted MoBos for
    such a cpu were ever manufactured. In short the PCI interface
    Wrong again. There was very little difference in price between the two
    and, quite often, no price difference or even a negative difference in
    favour of the host controllerles over the soft modem.

    Remember, the major cost of such adapters comes down to the cost of the
    PCB the chips are mounted on. In the case of discrete memory chips a
    decade or so before, it wasn't the complexity of the silicon chip, but
    the number of pins on the DiL package that decided production costs.

    HTH
     
    Johnny B Good, Nov 5, 2009
    #18
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.