BENCHMARK STAR 2.0

Discussion in 'SolidWorks' started by TOP, Apr 24, 2005.

  1. TOP

    TOP Guest

    Fellow Solidworks users. A new improved version of this benchmark is
    now available. It has been tuned to run as fast as possible on either
    2004 or 2005 and to be less dependent on your system or graphics card.

    My results for an Athlon XP2000+ with 1Gb RAM and 500XGL graphics card
    are:

    2004
    (3) 2.04 s
    (4) 8.09 s
    (5) 75.02 s

    2005
    (3) 2.32 s
    (4) 8.83 s
    (5) 79.23 s
     
    TOP, Apr 24, 2005
    #1
  2. TOP

    Alvin Leung Guest

    Hi Can you give me the link of download benchmark star. Thank you very much.
     
    Alvin Leung, Apr 25, 2005
    #2
  3. TOP

    TOP Guest

    Make sure you set TOOLS/REFERENCES

    If running on 2004 make sure SolidWorks2004TypeLib is checked.
    If running on 2005 make sure SolidWorks2005 TypeLib is checked.

    On an Athlon64 FX53, 2Gb, Nvidia FX3000
    SW2005

    (3) 0.71 s
    (4) 2.43 s
    (5) 39.0 s

    On a 650MHz PIII, ATI Rage graphics 392Mb RAM (Laptop)
    SW2005

    (3) 5.22 s
    (4) 22.0 s
    (5) 209.55
     
    TOP, Apr 25, 2005
    #3
  4. TOP

    TOP Guest

    Alvin,

    As before is is posted on www.solidworks.com in the Discussion Forum
    under the performance tab.
     
    TOP, Apr 25, 2005
    #4
  5. Paul,

    I'm still getting wildly varieing times. The average is

    3 .90 sec
    4 4.5 sec
    5 42 sec

    Dual Opteron 246
    MSI K8T master FAR
    2Gb Ram
    PNY 980XGL
    Win2000, SP4
    SW2005 SP3

    When you said "make sure" tools/references was set to whatever, did you mean
    turn off 2004 if your running 2005 and vise versa ??? Or just check to make
    sure they're there ??

    When you turn either of them off, the macro doesn't work.

    Regards

    Mark
     
    Mark Mossberg, Apr 25, 2005
    #5
  6. TOP

    TOP Guest

    I noticed the variation too. Possible explanations:

    1. The Timer() function is used which measures real time elapsed. If
    there are other processes that run while STAR is running that can
    effect the time. Don't touch the mouse rule applies.

    2. There really is a variation within SW that I can't explain. That
    seems less likely unless there is some kind of garbage collection going
    on in the background.

    You shouldn't need the 2004 type library with 2005. Run it from the VB
    editor after you make the change. It seems to need a kick ITB to get
    going sometimes.
     
    TOP, Apr 25, 2005
    #6
  7. TOP

    Adam Reif Guest

    My humble results - from a system that is
    slightly antique by today's standards:

    AMD XP1800+
    ASUS A7V133
    1Gb PC133 RAM
    Nvidia Quadro2 Pro
    Win2k SP3
    SWX2005 SP2.0

    Level Time:
    3) 2.85 s
    4) 13.35 s
    5) 104.35 s
     
    Adam Reif, Apr 25, 2005
    #7
  8. TOP

    Art Woodbury Guest


    My system isn't quite in the same league as yours, but not too far
    behind. Two sets of times:

    (3) 1.47 & 1.58
    (4) 5.47 & 5.14
    (5) 42.11 & 41.51

    System specs:

    Mobo: MSI K8N Neo 2 Platinum
    CPU: Athlon 64 3500+
    Video: Quadro FX-1100
    Ram: 1 Gb PC3200
    OS: Win 2K SP4
    HD: 2x WD Raptor 74Gb SATA, in striped array
    SWX: 2005 SP1.1

    Art W.
     
    Art Woodbury, Apr 25, 2005
    #8
  9. TOP

    Rich Guest

    Mark

    I have a similar machine spec to yours but it seems to be *much* slower
    :-( is there anything special I am missing?

    3 = 1.4
    4 = 8.6
    5 = 85

    Dual Opteron 246
    2gig ram
    ATI FireGL X2
    SW2005 SP2

    Rich
     
    Rich, Apr 25, 2005
    #9
  10. TOP

    TOP Guest

    Mark and Rich,

    First off, I misremembered my level 4 results for the FX53 as I posted
    from home.

    SW 2005
    (3) .69, .71 s
    (4) 3.87, 4.03, 4.43 s for three runs.
    (5) 37.09 s

    SW 2004
    (3) .64, .81 s
    (4) 3.43, 3.45 s
    (5) 33.99 s

    I think scatter is normal, especially for the shorter runs.
    Have a good look at what is running in the background. Virus checker,
    services, network traffic.

    I don't have a virus checker running for this test and I trimmed
    services per the blackviper list. I am on a network switch so I don't
    see other network traffic. Even email should be shut down. My task
    list only has 28 processes running including SW.

    I will say it again that the timer times real time, so if some other
    process grabs a few clock cycles during the test it will show up in the
    results.
     
    TOP, Apr 25, 2005
    #10
  11. TOP

    TOP Guest

    Only have the type library checked for the version of SW you are
    running.
     
    TOP, Apr 25, 2005
    #11
  12. My results for this version. I have never disabled anything to change my
    results, I run just as I run through the day - network attached, AV on, etc.

    2005
    (2) 0.61 s
    (3) 1.63 s
    (4) 6.75 s
    (5) 68.58 s

    WT
     
    Wayne Tiffany, Apr 25, 2005
    #12
  13. TOP

    TOP Guest

    Remind me again what you are running on. TX
     
    TOP, Apr 25, 2005
    #13
  14. SolidWorks 2005 sp3.0
    Windows XP sp2.0
    Dell Precision 360
    P4 2.99 GHz
    2.00 GB of RAM
    ATI FireGL X2 256 MB AGP 8x, Driver 6.14.10.6422


    My backup location:
    C:\DOCUME~1\wmtk\LOCALS~1\Temp­SWBackupDirectory


    Conf. corner off.

    WT
     
    Wayne Tiffany, Apr 25, 2005
    #14
  15. TOP

    MM Guest

    Rich,

    I don't really know why my home machine is so fast. My work machine is
    virtually identical, except for more memory, and it's slower as well.

    3. .90
    4. 5.22
    5. 44.9

    The only thing I can think of is that I have full control over the machine
    set up at home, and it doesn't have any unnecessary junk running on it. I
    also set up my machine at work, but I have alot more stuff running.

    That still doesn't explain why your system is running at half speed, which
    is what it seems like. Check the BIOS and make sure the chips are clocked
    right. The 246 should be clocked at 2ghz.

    What mother board are you using ?? There aren't many dual Opteron boards.
    Some, like the Tyan, won't let you mess with the CPU clock, but should
    report the frequency.


    Regards

    Mark
     
    MM, Apr 25, 2005
    #15
  16. TOP

    Rich Guest

    Mark
    using WCPUID (http://www.h-oda.com/) I get :
    AMD Opteron 246 (two off)
    Internal Clock 1994.60MHz
    System Clcok 199.46MHz
    System bus 199.46MHz
    L1 cache 64K
    L2 cache 1024K
    L2 Speed 1994.60MHz

    Chipset 1022-7451.01, 1022-7468-05
    VGA 1002:4E4B.80

    Host Bridge vendor 1022, device 7451, rev 01, sub 534D108E, AMD 8131
    Soutch Bridge vendor 1022, device 7468, rev 05, sub 534D108E, AMD 8111
    It's a Sun system, their own? not sure
    Not much running in background, but i did notice the disk keeps chugging
    away and very noisey. its a Seagate ST373307LW SCSI Ultra 320

    Any other ideas?

    Rich
     
    Rich, Apr 26, 2005
    #16
  17. TOP

    TOP Guest

    The other thing to consider is that STAR measures real time, not just
    the CPU time devoted to running the benchmark. This means it will
    account for any slowdowns due to running other processes concurrently.
    This is, I suppose a good thing. AFAIK there isn't an easy way to time
    just a process like in the 'NIXes.

    To prove this do a level 4 and continuously move the mouse while
    running. You should see a higher value than when running without
    touching anything.

    BlackViper has a list of services you can shut down to speed up your
    system. Unfortunately it appears to be down.
     
    TOP, Apr 26, 2005
    #17
  18. TOP

    TOP Guest

    WCPUID EXPLAINS WHY MINE IS A TAD BIT FASTER:

    AMD ATHLON 64 FX53
    INTERNAL CLOCK 2414.99 MHZ
    SYSTEM BUS 210 MHZ
    MULTIPLIER 11.5
    L2 1024K
    LW SPEED 2414.99 MHZ
    CHIPSET
    1106.3188.01
    1106.3227.00
    MEMORY
    2048 MB
     
    TOP, Apr 26, 2005
    #18
  19. TOP

    Rich Guest

    Tell ya what, I'll make some jpegs of all the screens from WCPUID and
    thanks mark, I'll take a look, be very nice to get to the bottom of this ;-)

    Rich
     
    Rich, Apr 26, 2005
    #19
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.