Is It Micron Technology Memory or? The Hunt for SPD Data?

Discussion in 'Asus' started by W. Watson, Dec 9, 2004.

  1. W. Watson

    W. Watson Guest

    Hi, I have an ASUS A7S333 motherboard, which takes DDR333 DIMM 185 pin chips. I'm
    trying to get the SPD values for the Micron memory I have installed in the board. The
    outside of the chip reads Micron Technology, PC2100, 256MB, 266Mhz. I talked to MT
    about this and they doubt that it really their chip. It should have a bar code and a
    logo. It has none, just the info I stated above, which is on a plain white tag.
    According to my BIOS, SPD values correspond to items labeled CAS, Delay and two
    others. I'm only able to recall that much from my head.

    In any case, I went to the site where I bought the memory chip in March 2003, and
    looked for what they sell along these lines. (They are trying to find my order at the
    moment that will give better detail on what I really bought.) I see the required chip
    for sale, and it is shown in two flavors: Micron and Micron or Micron or Major 256MB
    PC2100. I have a feeling I bought the Major, but can't be sure.

    All of this boils down to asking what are the four values I need? The values are
    shown as 2T, 3T, 3T, and 6T. There are a few choices for each. Mayb setting them to
    their highest values or lowest is the best strategy.

    Oh, the reason I'm interested in these values is that my ASUS board refuses to let me
    set the cpu speed above 1.1GHz. When I try the other choice, 1.6GHz, I get a Win boot
    up message about a corrupt file. I called ASUS and they suggested setting the SPD
    values to those required by the mfger.
    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet
    (Formerly Homo habilis, erectus, heidelbergensis and now sapiens)

    Sign in Programming Department: Mistakes made while you wait.

    Web Page: <home.earthlink.net/~mtnviews>
     
    W. Watson, Dec 9, 2004
    #1
    1. Advertisements

  2. W. Watson

    W. Watson Guest

    Forget this one. I now have the machine running at a higher speed without adjusting
    the SPD. I went manual. They key to getting it going seem to specify the frequency
    ratio as 1:1. If let at auto, it still screws up.

    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet
    (Formerly Homo habilis, erectus, heidelbergensis and now sapiens)

    Sign in Programming Department: Mistakes made while you wait.

    Web Page: <home.earthlink.net/~mtnviews>
     
    W. Watson, Dec 9, 2004
    #2
    1. Advertisements

  3. W. Watson

    Paul Guest

    SDRAM CAS Latency = tCAS likely 2 try 3
    SDRAM RAS to CAS Delay = tRCD likely 3 try 3
    SDRAM RAS Precharge time = tRP likely 3 try 3
    SDRAM RAS Active time = tRAS likely 6 try 8

    You should not be testing the speed settings of the computer, using a
    live OS. You can get a copy of memtest86 from memtest.org and use
    that for testing. The program will format a floppy diskette for you,
    and put a memory test program on it. The floppy needs no OS, and the
    floppy has its own boot loader, to load the memtest program. The floppy
    also has no file system, so you cannot explore the floppy after it
    has been formatted by the downloaded program from memtest.org .

    If the memtest86 program crashes, or if you get a ton of errors, then
    you will know you've pushed some hardware too hard. High performance
    memory can be helped a bit by boosting Vdimm, while a processor can
    be helped by increasing Vcore, but with some processors at least,
    this will cause a worrying rise in temperature.

    I've also used a Knoppix (read-only Linux) distribution CD, disconnected
    the hard drive in the computer, and then you can watch for segmentation
    faults or other visual oddities, while the system boots from the CD.
    Knoppix boots right into X-Windows, and is ready to go with no work
    from the user. Knoppix even includes a copy of memtest86, and it can
    be loaded from the boot prompt at boot time. You can also download a
    Linux version of Prime95 from mersenne.org and use Prime95 torture
    test while booted in Knoppix. If Prime95 can run for hours, without
    reporting any numeric errors, then you know memory and CPU are adjusted
    properly. (Since Knoppix is read-only, I keep the Linux version of Prime95
    on a floppy, as Knoppix can read the floppy after it has booted.)

    If you set the clock speed in the BIOS, watch that the PCI bus speed
    is not set to 37.5MHz or higher. 33MHz is the normal value and if
    the PCI clock is set too high, hard drive corruption will occur. If
    you want to play with settings over 37.5MHz, that is a time to be
    disconnecting an IDE hard drive, and using one of the test methods
    above. Back up the hard drive, if you want to play with a live OS.

    With your quoted CPU frequencies, I'm not sure what CPU you've got
    there. Since you are likely trying 100MHz or 133MHz clocks, there
    should be a 1.333 ratio between the CPU core frequencies you quoted.
    In any case, here is a list of processors, their bus clock speed,
    their internal multiplier, and the core frequency that results.
    For example, maybe you have a T-bred 1900+, which has a core
    frequency of 1600MHz, a bus clock of 133MHz, and a multiplier of 12x.
    Such a processor would be nicely matched by some PC2100 DDR266
    memory, as DDR is double data rate, double as in twice the bus clock
    of 133MHz. Selecting a CPU:memory ratio of 1:1 would make the memory
    run at DDR266 when the processor clock is 133MHz. Be careful when
    doing your manual selections, that the memory is not operated
    faster than its rating, as that will give rise to memory errors.

    (Damn Google and their beta program!)
    http://groups-beta.google.com/group...nboard.asus/msg/a38e766d4afd1a50?dmode=source

    Using the relaxed memory timings above, a 1:1 CPU:memory ratio,
    and selecting a CPU clock which is its normal rated clock, should
    be giving you a stable system. The relaxed memory timings can be
    tightened again, once you've got the rest of the settings squared
    away. Testing with Prime95 on an Athlon, is essential to determining
    whether you are really stable or not, while memtest86 is a first
    step in seeing whether the memory you bought is actually good
    stuff or not. If a system is not Prime95 stable, then forget
    using the system for serious work (crash city awaits).

    HTH,
    Paul
     
    Paul, Dec 9, 2004
    #3
  4. As far as whether it's Micron or not, it may have Micron chips on it but
    it's quite likely the module itself is not made by them - and the module
    has a big influence on what speed settings it will be able to run.
     
    Robert Hancock, Dec 11, 2004
    #4
  5. W. Watson

    W. Watson Guest

    Thanks. I'll look for the program. I tried using cpuz and Everest and neither gave me
    a clue on SPD; however, when I was talking to a fellow at the place I bought the mem
    chips from, he mentioned a problem he had with his computer that was similar to mine.
    He mentioned the Auto setting for frequency ratio was key for him. I thought I'd give
    it a try. I had tried to set the frequency and multipler once before with the same
    bad results. When I tried again with the ratio 1:1, it worked! I'm now running at 1.5GHz.
     
    W. Watson, Dec 12, 2004
    #5
    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.