VPC/WMPlayer comparison?

Discussion in 'Virtual PC' started by Vincent Fatica, May 13, 2009.

  1. good point, I guess if the goal of the malware is to prevent you from
    using VT or your system though.

    With VPC2007 and VMW running both with VT, you'll get a host
    blue-screen.
     
    Steve Jain [MVP], May 15, 2009
    #21
    1. Advertisements

  2. Vincent Fatica

    Dave Warren Guest

    In message <> "Steve Jain
    I think you've misunderstood. My assumption isn't that the goal of said
    malware is to prevent you from using anything, most likely it's using VT
    to hide it's goal is to stay hidden, doing whatever it wants in the
    background without annoying the user enough to go looking.

    This fits the model of most modern malware, which is to either collect
    data about the user, act as a spam zombie or something similar.

    Such malware which attempts to use VT to hide can be trivially
    discovered by attempting to start any other VT enabled application,
    which will cause a bluescreen on a machine which formerly did not
    bluescreen in this circumstance, thereby defeating the point of hiding
    using CPU based VT capabilities.

    Having VT enabled when not in use might be a possible security issue,
    but if you're regularly using VT, malware isn't.
     
    Dave Warren, May 16, 2009
    #22
    1. Advertisements

  3. Vincent Fatica

    VanguardLH Guest

    What happens when you run an actual benchmark program?

    http://www.futuremark.com/products/pcmark05/
    (there is a free version)

    http://www.passmark.com/products/pt.htm
    (not free but the 30-day trial version is fully functional)
     
    VanguardLH, May 16, 2009
    #23
  4. Vincent Fatica

    Dave Warren Guest

    You get unpredictable and incorrect results.
     
    Dave Warren, May 17, 2009
    #24
  5. Vincent Fatica

    VanguardLH Guest

    Oh, and a for-loop is more predictable? Prove your claim. So far, I've
    only folks making such a claim but not actually pointing to any proof.
    Have you ran the benchmarks? After running the 1st one and discarding
    it and then running, say, 5 more do those last 5 have wildly different
    results? Are you making sure that nothing is running on the host to
    make sure they don't impact VPC? Did you make sure the VM kept focus
    for its window so its priority didn't drop?
     
    VanguardLH, May 17, 2009
    #25
  6. |>>What happens when you run an actual benchmark program?
    |>
    |> You get unpredictable and incorrect results.

    I get very consistent results timing a very simple routine. They're nearly as
    consistent on the guest as on the host. They also show that, ignoring GUI and
    network activity, my virtual Vista is pretty fast. It wouldn't surprise me if
    *real* Vista were about the same.

    #include <windows.h>
    #include <stdio.h>
    INT wmain ( INT argc, WCHAR **argv )
    {
    srand(GetTickCount());
    INT sum = 0;
    for ( INT i=0; i<1000000000; i++ )
    {
    sum += ( rand()%2 ) ? 1 : -1;
    }
    wprintf(L"%d\n", sum);
    return 0;
    }

    Most recently I tested thus (using 4NT):

    for /l %i in (1,1,20) (timer > NUL & test.exe > NUL & timer)

    Guest Host
    0:00:20.27 0:00:18.06
    0:00:20.36 0:00:18.08
    0:00:20.29 0:00:18.06
    0:00:20.31 0:00:18.06
    0:00:20.32 0:00:18.08
    0:00:20.28 0:00:18.08
    0:00:20.30 0:00:18.06
    0:00:20.27 0:00:18.07
    0:00:20.38 0:00:18.06
    0:00:20.31 0:00:18.08
    0:00:20.29 0:00:18.06
    0:00:20.27 0:00:18.06
    0:00:20.32 0:00:18.06
    0:00:20.35 0:00:18.08
    0:00:20.26 0:00:18.06
    0:00:20.35 0:00:18.07
    0:00:20.28 0:00:18.06
    0:00:20.28 0:00:18.06
    0:00:20.26 0:00:18.06
    0:00:20.27 0:00:18.07

    I don't see how one could adequately test anything but raw computation speed.
    Video and networking (with NAT) are very crippled inside VPC.
     
    Vincent Fatica, May 17, 2009
    #26
  7. More than I cared to.
    IIRC, I did a set of 3 per machine/VM/VM app
    Yep, I was/am quite familar with how VM tech works ;-)
    Yep, or set the correct parameters in the VM to always run the VM at
    full speed.

    I ran many benchmarks back at Connectix, the results depended on the
    benchmark and what it was actually measuring. Some memorable results:
    VMWare 4.x VM had 2x oe 3x the "disk access" speed of the host machine
    SiSoft Sandra's memory access benchmarks aren't.
    Benchmarks that make use the the internal clock go crazy when the VM
    additions or VM tools adjust the time to match the host...more
    noticable on older and slower CPUs (i.e. P3, P4, Athlons)

    I don't think there's any "office" benchmarks anymore like the old
    PCMag one that try and emulate what a user does.
     
    Steve Jain [MVP], May 18, 2009
    #27
    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.