503+ BIOS settings and random lockups

Discussion in 'FIC' started by Den, Jan 31, 2004.

  1. Den

    Den Guest

    I have been getting very random lockups that the only
    choice is reset or switch off. This happens within a few minutes
    to a few days, usually in W98SE but not in Linux. (dual-boot)

    In the Chipset Features Setup;

    SDRAM Cycle Length
    it was set at 2 but have now set to 3
    SDRAM Bank Interleave this I found set at 4 bank.
    Changed this to 2 bank would this be correct ? as I do not
    recall when this setting was changed.
    this was at default as far as I can recall, but until I upgraded the
    memory with a Crucial recommended 256 & an IBM 128 which
    have been OK with cycle length at 2

    Den
     
    Den, Jan 31, 2004
    #1
    1. Advertisements

  2. Try testing your RAM with Memtest86+, which you can get from
    http://www.memtest.org . Just load it to a floppy, boot to the floppy,
    and let it go. If you do not get any errors, then it is is probably
    something else.

    By the way, I believe 4-way memory interleaving should be OK, but I am
    not 100% sure.

    --Alex
     
    Alex Zorrilla, Feb 2, 2004
    #2
    1. Advertisements

  3. Den

    Den Guest

    OK thanks will test with disabled, 2 & 4 and see what happens, I see that
    this program has just been updated, the first since mid 2002.

    Den
     
    Den, Feb 3, 2004
    #3
  4. Den

    Robert Akins Guest

    If you can still find the utility ctspd it will tell you all the specs from
    the spd on the sticks of memory including interleave.
    Robert
     
    Robert Akins, Feb 3, 2004
    #4
  5. Den

    Den Guest

    Just tried tests but found the default test would after a short time give
    a (I assume) pattened crash until I set 384MB setting, I assume that
    the extra memory count on boot of 393216K is with possibly some of the
    video card memory which is a 32MB one, as L1 Cache is 64K and
    L2 cache is Unknown according to Memtest Only a guess on my part.


    So as when it reaches 384MB this is when it usually fails regardless
    of default, 2 or 4 set.
    Setting memory to 384 within program the default, 2 or 4 and also if I set
    SDRAM Cycle Length=2 or 3 is all OK.

    Latest test of 60 minutes at these settings show no errors, so I assume
    it could be possibly mains borne interference, despite filtered power strip.
    SDRAM Cycle Length=2
    SDRAM Bank Interleave set at 4 bank
    PS
    Talking of power surges;
    http://news.bbc.co.uk/1/hi/england/staffordshire/3457965.stm
    Thank goodness for the good old fashioned manual loos.

    Den
     
    Den, Feb 4, 2004
    #5
  6. Den

    Den Guest

    Thanks for that, according to that program the Crucial 256MB;
    (http://www.crucial.com/uk/store/PartSpecs.asp?imodule=CT32M64S4D7E&cat=RAM

    Is made by Micron, as is the IBM 128MB stick and both show it
    to be 133MHz and CAS 2/3 in fact the only thing to change between
    the two is the Micron DIMM type number and size.

    Though the 503+ has been on for the last 48 hours without any lockups
    and Memtest86+ had not shown any errors when run.
    Check in other post (BBC) link.

    Den
     
    Den, Feb 4, 2004
    #6
  7. Actually, 384 MB = 393126 kB. This is because 1 MB = 1024 kB, at least if you're talking about RAM. I think that by
    manually setting 384 in the program, you are actually bypassing the bad part of memory.

    Try testing the memory one stick at a time, allowing the program to set its own memory. I wager that one of the sticks will
    test OK, but the other one will give errors near the end of its size.

    By the way, for a K6-2, the L2 cache shows up as "Unknown" because it is located on the motherboard, not on the CPU itself.

    --Alex
     
    Alex Zorrilla, Feb 4, 2004
    #7
  8. Den

    Den Guest

    As the program goes straight into testing it is showing every single test
    as an error then screen changes to the unreadable display, this is with
    both sticks and each one on its own as it reaches around its size.

    So I wonder if I can find an early version of Memtest86 or a similar
    memory test, by the way the 128 stick was a Crucial not an IBM
    as the IBM stick is in this PA-2013 along with the original 128 which
    refused to work in the 503+ after 3 or 4 years.

    Yes I agree with the L2 cache on the MB, I recall back in 1999 not
    being able to get the larger memory option.

    Den
     
    Den, Feb 5, 2004
    #8
  9. Den

    Roger Hunt Guest

    I have versions 2.7(07/01), 2.8(10/01) and 3.0(05/02)
    email if you would like a copy.

    Cheers
     
    Roger Hunt, Feb 5, 2004
    #9
  10. Den

    Den Guest

    "Roger Hunt"
    Thanks Roger but I had found and tried version 2.9 and that crashed before
    completing the first screen.

    I will test vers 2.9 on this PC which has the original 128 that refused to
    work on the 503+ after 3 or 4 years.

    Well this PA-2013 passed with flying colours.

    While retrying on the 503+ with version 3 but setting option 4 to probe
    it appears to be testing OK ? I'll be honest these memory tests are beyond
    my understanding, it had three options
    1 BIOS-standard
    2 BIOS-extended
    3 probe (this is the one I picked) (clutching at straws really)
    It got as far as pass 67% before any errors started to show.

    Den
     
    Den, Feb 6, 2004
    #10
  11. Den

    PC Ninjas Guest

    I never did find memtest to be user friendly, and I am fairly
    experienced system builder. there has to be a better program ~
     
    PC Ninjas, Jun 6, 2004
    #11
  12. Den

    Kylesb Guest

    If you want a "dumbed down" test, docmem is also free, but much less
    thorough, memtest86 has found errors not found by docmem.

    I don't see the problem with using memtest86, you make the test
    floppy, you setup machine to boot from floppy, you insert floppy and
    reboot machine.

    --
    Best regards,
    Kyle
    | I never did find memtest to be user friendly, and I am fairly
    | experienced system builder. there has to be a better program ~
    |
    |
    |
    | Den wrote:
    | > "Roger Hunt"
    | >
    | >>I have versions 2.7(07/01), 2.8(10/01) and 3.0(05/02)
    | >>email if you would like a copy.
    | >>
    | >
    | > Thanks Roger but I had found and tried version 2.9 and that
    crashed before
    | > completing the first screen.
    | >
    | > I will test vers 2.9 on this PC which has the original 128 that
    refused to
    | > work on the 503+ after 3 or 4 years.
    | >
    | > Well this PA-2013 passed with flying colours.
    | >
    | > While retrying on the 503+ with version 3 but setting option 4 to
    probe
    | > it appears to be testing OK ? I'll be honest these memory tests
    are beyond
    | > my understanding, it had three options
    | > 1 BIOS-standard
    | > 2 BIOS-extended
    | > 3 probe (this is the one I picked) (clutching at straws
    really)
    | > It got as far as pass 67% before any errors started to show.
    | >
    | > Den
    | >
    | >
     
    Kylesb, Jun 7, 2004
    #12
  13. Den

    Kylesb Guest

    A small followup, I will agree that invoking specific tests with
    memtest86 is not straightforward, though I have succeeded in running
    single tests after fiddling with the input commands.

    --
    Best regards,
    Kyle
    | If you want a "dumbed down" test, docmem is also free, but much less
    | thorough, memtest86 has found errors not found by docmem.
    |
    | I don't see the problem with using memtest86, you make the test
    | floppy, you setup machine to boot from floppy, you insert floppy and
    | reboot machine.
    |
    | --
    | Best regards,
    | Kyle
    | | | I never did find memtest to be user friendly, and I am fairly
    | | experienced system builder. there has to be a better program ~
    | |
    | |
    | |
    | | Den wrote:
    | | > "Roger Hunt"
    | | >
    | | >>I have versions 2.7(07/01), 2.8(10/01) and 3.0(05/02)
    | | >>email if you would like a copy.
    | | >>
    | | >
    | | > Thanks Roger but I had found and tried version 2.9 and that
    | crashed before
    | | > completing the first screen.
    | | >
    | | > I will test vers 2.9 on this PC which has the original 128 that
    | refused to
    | | > work on the 503+ after 3 or 4 years.
    | | >
    | | > Well this PA-2013 passed with flying colours.
    | | >
    | | > While retrying on the 503+ with version 3 but setting option 4
    to
    | probe
    | | > it appears to be testing OK ? I'll be honest these memory tests
    | are beyond
    | | > my understanding, it had three options
    | | > 1 BIOS-standard
    | | > 2 BIOS-extended
    | | > 3 probe (this is the one I picked) (clutching at straws
    | really)
    | | > It got as far as pass 67% before any errors started to show.
    | | >
    | | > Den
    | | >
    | | >
    |
     
    Kylesb, Jun 7, 2004
    #13
  14. Den

    farmuse Guest

    yes, no problem running the test but the interpretation is a bit
    screwy as I recall although it has been quite a while. maybe I can try a
    third time and see how it goes. just that the interface is not so great,
    and it takes forever to run ~
     
    farmuse, Jun 8, 2004
    #14
    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.