K7N2 Delta-L Poor memory results with dual channel, please help !

Discussion in 'MSI' started by Jean, Sep 10, 2003.

  1. Jean

    Jean Guest

    CPU Barton 2500+

    Memory 2x256meg

    memory setting 7-3-3-2.5 manual (spd is using timing of 8-3-3-3)

    Program Used Sandra Sisoft Pro 2003.7.9.73



    1-Should I use synch or asynchronous setting for the memory FSB, 166mhz or
    200mhz ?

    2-The result I got are 2500mb/s and 2400mb/s, very poor compare to Intel =
    around 5000mb, even my other computer with P4 2.8fsb533 rambus 1066 is
    having arounf 3300...

    BUT the problem is that memory results are the same using SINGLE or DUAL
    CHANNEL configuration...

    I show DUAL CHANNEL when it boot but the results are still the same...

    THROW ME a bone here ?? The Nforce 2 chipset SUCK or what ? On the MB box
    they wtote theorical memory bandwith up to 6.1... Sandra told me 2600 mb/s
    and it cannot identify the memory banks neither...

    PLEASE HELP. is it normal that the results suck so much and to see no
    difference between single and dual channel configuration, it is only a
    marketing ?

    thanks !
     
    Jean, Sep 10, 2003
    #1
    1. Advertisements

  2. Jean

    Robert Akins Guest

    One of the reviews that I read on Tom's Hardware of nForce2 boards said that
    CL2 DDR was a must for dual channel operation. Just what they said.
    Robert
     
    Robert Akins, Sep 10, 2003
    #2
    1. Advertisements

  3. Jean

    Inglo Guest

    Did you happen to look at the reference chipsets in Sandra?
    The nForce2 w/ PC2700 dual channel gets 2566/2456. So that's around
    what your results are. Almost all the tests I've seen show negligible
    differences between dual and single channel on nForce2 boards.
    The best Intel based boards with the best RAM should get about double
    the nForce2 memory bandwidth.
     
    Inglo, Sep 10, 2003
    #3
  4. Jean

    Jean Guest

    well thanks... precisely what I thought, dual channel using nforce2 and
    AMD is a bad joke... at least this is not my computer...

    and who design the retail heatsink ? it's loud and annoying, poor
    design... Silent boost from thermaltake will replace it soon...
     
    Jean, Sep 10, 2003
    #4
  5. 2-4% in those types of memory tests with non-integrated graphics, is about
    right. Your improvment is 4%. Check out this article:
    http://www.digit-life.com/articles2/nforce2-1vs2channels/


    Too_Much_Coffee ®
     
    Too_Much_Coffee ®, Sep 10, 2003
    #5
  6. Jean

    Inglo Guest

    Its not really that bad its just not as good as the top Intel based
    chipsets. And the dual channel memory can be significant in using
    onboard GPUs.
     
    Inglo, Sep 10, 2003
    #6
  7. Jean

    Jean Guest

    but should I used synch or asynch FSB for the memory ? 333 or 400 mhz ?

    I readed that if I run in synch mode it should be better but bring
    instability ?

    thanks
     
    Jean, Sep 11, 2003
    #7
  8. Jean

    Silvermane Guest

    Don't be so quick to dog that mobo. Just because a P4 system gets twice the
    mem bandwidth doesn't mean it is faster. The P4's need all that bandwidth
    just to remain competitve with the Athlons. Clock for clock a P4 gets it ass
    handed to it by AMD.
     
    Silvermane, Sep 11, 2003
    #8
  9. Jean

    LeeB18509 Guest

    What are you trying to say? Benchmarks don't mean EVERYTHING? What a
    concept.
     
    LeeB18509, Sep 11, 2003
    #9
    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.