Announcement

Collapse
No announcement yet.

F3-12800CL9D-8GBRL problems (random BSoD & weird timings)

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • F3-12800CL9D-8GBRL problems (random BSoD & weird timings)

    System Specs or click here
    Memory: G.Skill Ripjaws F3-12800CL9D-8GBRL
    Mother board: ASRock P67 Pro3 (B3) - UEFI P1.90 (5/13/2011)
    CPU: Intel Core i5 2500K

    System tests passed
    Memtest 86+ 20+ passes with both modules in the banks at base clock.
    IntelBurnTest 30+ passes with the CPU clocked at 4.3GHz.
    Prime95 6+ hours blend passed with the CPU clocked at 4.3GHz.

    Problem 1
    The first problem I have ran into are random 124-BSoDs when I clock my 2500k to 4.3GHz (x43) with the memory at specification timings (9-9-9-24-2N), voltage (1.5v) and speed (1600MHz). Any activity that keeps the multiplier at max seems to be enough to allow my system to run for many hours, but if I let it rest for a while, it gets cranky. Multiple hours of idling will almost always turn out into a bsod while I've had it occur during minor usage (such as folder opening or simple web browsing) from time to time too. To summarize, it occurs completely random when the computer isn't doing much.

    I have tried most if not every sensable thing I have come accross on Google and your forums but still haven't found a solution. One thing left to try is to find out if either of the memory modules is malfunctioning, so I decided to give it a try with just one of the two modules in the banks (as I suspect the dual channel might require more than the default voltage) - this brings us to the second problem which is the one I require assitance with before I can confirm this is a memory (controller) issue.

    Here are the settings used for my OC:
    1 2 3 4 5

    Problem 2
    When I insert either one out of the two modules into the memory banks, no matter what slot it is in, the timings end up being 6-6-6-20-2T as shown below (occured during OC and during factory clock).



    Below is a quick memtest I ran to confirm these timings (oddly enough, it passed).


    These odd timings make it impossible for me to get an accurate test, as they're way off from the specifications. The odd thing here is that the settings in the UEFI are set as per specs (I've also tried XMP):



    To make it clear, this only occurs with 1 module in the mother board, but it doesn't matter which one it is. With 2 modules, memtest reports back normally;


    So how can I fix the timings for these modules so I can run a proper test? Or better, any suggestions that might resolve problem 1?

  • #2
    Do you have the latest BIOS? Think I would suggest raising your vCore, and check your power settings


    Pls offer comments on support I provide, HERE, in order to help me do a better job here:

    Tman

    Comment


    • #3
      I do have the latest BIOS (UEFI) version available at the ASRock website and have previously ran at 1.4v (1.26v passes all stress tests I ran) which is the max I feel comfortable running at. It strikes me as odd that I would need more voltage when the multiplier is lowered during idle while it is sufficient during full load (from IBT, Prime95 or a game like Crysis 2).
      Last edited by Zerotorescue; 06-10-2011, 12:02 AM.

      Comment


      • #4
        It is bit weird to me also, but have found it to work w/ quirky BIOSs, can also disable some of the energy saving settings in BIOS, some of them can cause this and similar problems


        Pls offer comments on support I provide, HERE, in order to help me do a better job here:

        Tman

        Comment


        • #5
          Check the SPD tab to see the memory profiles.

          Manually set everything to see if that works.

          Thank you
          GSKILL TECH

          Comment


          • #6
            Is there any point in also trying the settings in the profiles provided as JEDEC? They are much different from the XMP-profile.

            SPD tab:


            However if I set the 9-9-9-24-2N provided at the product page with a single module online, the ratings end up different in CPU-Z and Memtest86+. I don't think a tRas of 1 higher will make much of a difference when the modules aren't OCed, but I will certainly give it a try.

            Thanks for your help so far, it is much appreciated. Especially Tradesman who is willing to help even though he feels the memory isn't to blame.
            Last edited by Zerotorescue; 06-11-2011, 02:58 AM.

            Comment


            • #7
              Might try CPU-Z 1.57.2 and see if their updates provide better info


              Pls offer comments on support I provide, HERE, in order to help me do a better job here:

              Tman

              Comment


              • #8
                Originally posted by GSKILL TECH View Post
                Manually set everything to see if that works.
                I set the timings to the 9-9-9-25-2N specified in the XMP-1600 profile without any luck. The question still remains if trying any of the JEDEC profiles is advisable?

                Originally posted by Tradesman View Post
                It is bit weird to me also, but have found it to work w/ quirky BIOSs, can also disable some of the energy saving settings in BIOS, some of them can cause this and similar problems
                Back to the raising the vCore suggestion, I had to set a fixed voltage so the vCore of my idle CPU is already deci-volts higher than stock.

                As per energy saving settings, I have disabled all C-states as you can see below, if this helps, I will toggle them one by one until I find the problem causer, if it doesn't work, then I'll also try disabling SpeedStep / EIST (I guess that might just work as the CPU wouldn't be downclocked then, but it is not an optimal solution).

                ("Package C State support" provides the options Auto, C2, C3 and Disabled, the other options are just Enabled or Disabled.)

                Originally posted by Tradesman View Post
                Might try CPU-Z 1.57.2 and see if their updates provide better info
                I can only find 1.57.1 and that does not adjust any info displayed.

                Comment


                • #9
                  Apparently 1.57.2 is specifically for the ASUS ROG (sorry bout that)...will try to see if I can find anything later this AM when I get back, should be in before noon.


                  Pls offer comments on support I provide, HERE, in order to help me do a better job here:

                  Tman

                  Comment


                  • #10
                    No, JEDEC settings are not near what the memory is actually capable of.

                    Fail proof settings are DDR3-1333 8-8-8-24 1.50V, so you can try that.

                    Thank you
                    GSKILL TECH

                    Comment


                    • #11
                      Originally posted by GSKILL TECH View Post
                      No, JEDEC settings are not near what the memory is actually capable of.

                      Fail proof settings are DDR3-1333 8-8-8-24 1.50V, so you can try that.

                      Thank you
                      GSKILL TECH
                      Thanks, these settings seem to have improved stability; I have been able to leave my computer running fine for the past two nights where it would previously crash at the first night. I will continue to keep you posted.

                      Comment


                      • #12
                        Sounds good, we'll keep an eye out


                        Pls offer comments on support I provide, HERE, in order to help me do a better job here:

                        Tman

                        Comment

                        Working...
                        X