Announcement

Collapse
No announcement yet.

New PC crashing - RAM related

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

  • New PC crashing - RAM related

    PC spec

    Windows 11

    Intel Core i7 12700F

    Gigabyte S1700 ATX Z690

    2x16gb DDR4 G.Skill 3200Mhz Aegis RAM (F4-3200C-16S-16GIS)

    Gigabyte RTX3080 10GB Eagle

    1TB Samsung 970 Evo plus M.2

    2TB Samsung 870 QVO SATA

    I tend to play Warzone the most. Every now and again the game will either freeze and kick me out of just the game, or BSOD. All the recent crashes have been very similar and i believe it is RAM related. However could be IMC in motherboard? However am unsure as analysis says could be software bug or misbehaving driver? Dump file data below, whenever i blue screen that last line 0x00007... is the final line.

    STACK_TEXT:
    ffffc181`d44771b8 fffff805`3de2d9a9 : 00000000`0000000a ffffc100`07d35be8 00000000`00000002 00000000`00000000 : nt!KeBugCheckEx
    ffffc181`d44771c0 fffff805`3de29b00 : ffffc601`e4e89180 ffffc601`e4e8c358 00000000`00000202 fffff805`3dcccbf5 : nt!KiBugCheckDispatch+0x69
    ffffc181`d4477300 fffff805`3dcaa88f : 00000000`00000001 00000000`00000000 00000000`00200200 ffffaed7`6ba05320 : nt!KiPageFault+0x440
    ffffc181`d4477490 fffff805`3dca9c86 : 00000000`00000000 00000299`1309c000 ffffc181`d44775f0 00000000`00001000 : nt!MiGetPageProtection+0xbf
    ffffc181`d44774f0 fffff805`3dc7f9fe : ffffd78d`73728740 00000000`00000000 00000000`00000000 fffff805`3dd02792 : nt!MiQueryAddressState+0x2e6
    ffffc181`d4477710 fffff805`3e12ae71 : ffffc181`d4477850 00000299`13078000 ffffd78d`7d310000 ffffd78d`7d3130c0 : nt!MiQueryAddressSpan+0x15e
    ffffc181`d44777d0 fffff805`3e12a735 : 00000000`00000000 00000000`00000000 00000000`0000002f 0000000b`c0bfee8a : nt!MmQueryVirtualMemory+0x721
    ffffc181`d4477960 fffff805`3de2d375 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!NtQueryVirtualMemory+0x25
    ffffc181`d44779b0 00007ffd`59bc51b4 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!KiSystemServiceCopyEnd+0x25
    000000e3`aeda2f08 00000000`00000000 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : 0x00007ffd`59bc51b4

    I have tried messing around with the RAM frequency, disabling XMP and upping frequency and voltage to XMP levels and leaving timings on auto, still crashing. Problem is the crash is only once a week.

    The most stable i had the machine was when i had XMP enabled and upped voltage to 1.43V (normally 1.35V with XMP) however am not sure if this is a good idea long term?

    The other solution i am tying to avoid is disabling XMP completely however then i lost a bunch of frames and performance. I have also not tested this long enough to see if a crash would occur however do not see it is a viable solution since i lose 40-50 frames. If XMP is not performing to advertised specs can this be cause for return? Or since it's technically an OC, won't be accepted.

    What voltage is safe to run at constantly for this RAM?

    memtest and windows diagnostic tool all run multiple times and no issue with RAM.


    Click image for larger version

Name:	20220922_202009.jpg
Views:	622
Size:	764.1 KB
ID:	168608​​ Click image for larger version

Name:	20220922_202027.jpg
Views:	525
Size:	706.2 KB
ID:	168609
    When i look in BIOS, the JEDEC timings, module organizations, Single sided vs Double sided, SPD rev 1.0 vs 1.1, 8 vs 16Bit are all different between the sticks. Could this be causing the problems? Both sticks are F4-3200C-16S-16GIS.

  • #2
    It is possible since they are not a matched set. Send them in for RMA to see if replacements can work better together.

    https://www.gskill.com/rma

    Comment


    • #3
      I went back to the shop where i got the PC built and they swapped out the RAM for another kit. Went home, turned XMP on, still crashed. Upped voltage to 1.4V, still crashed.

      I have read the XMP is not guaranteed stability and quite a common problem to cause crashes. I have turned XMP off for now (however comes with a performance loss).

      I am curious however what i can safely up the voltage to on these sticks? I have read 1.45V should be safe, even 1.5V. I had the best stability when i was running 1.43V however am not sure if this is ideal long term. Obviously i could just keep XMP off but trying to weigh up my options.

      Comment


      • #4
        Did you test the new kit alone or mix with other? Voltage can be increased and stay there as long as the system is stable. The memory kit at it's rated XMP specifications should not need much more than it's XMP Voltage.

        Comment


        • #5
          When i first got PC and it was crashing, i brought it back to shop and they told me they swapped out the RAM and CPU. Turns out this past weekend when i narrowed it down to RAM and they had a closer look, they only swapped one of the sticks. They discovered that even though it had the same part number, the original was "single sided/bank" and the one they swapped in was "double sided/banked", which is where you see the different timings and frequencies. So they swapped out the one that they swapped out the original for. I hope that made sense but basically i have one stick from the original build, and another new one now that is identical to the original stick.

          Tried to run extended memory diagnostic tool with XMP off, first time got stuck at 21%, then second time got stuck at 88%. Going to run memtest86 tonight.

          ***EDIT: memtest86 passed with no errors***

          maybe some games just do not like XMP?
          Last edited by Ninja Party; 09-27-2022, 01:03 PM.

          Comment


          • #6
            It may be best to send them in for replacements as it may be a mismatch between modules causing a problem. Hopefully two modules near in production can work better.

            Comment


            • #7
              so just to confirm, games should be able to run quite stable and not crash with XMP on? I have read before the XMP is not always guaranteed and has been known to cause crashes.

              Last night tried playing fifa 23 with XMP on and game just quit to desktop. Turned XMP off and game ran perfectly fine rest of the evening.

              i have run memtest86 again and it passed again with XMP on.

              Is there anything in the dump code above or here below that might indicate what the exact problem is?

              SYMBOL_NAME: nt!MiGetPageProtection+bf

              MODULE_NAME: nt

              IMAGE_VERSION: 10.0.22000.978

              STACK_COMMAND: .cxr; .ecxr ; kb

              IMAGE_NAME: ntkrnlmp.exe

              BUCKET_ID_FUNC_OFFSET: bf

              FAILURE_BUCKET_ID: AV_nt!MiGetPageProtection

              OS_VERSION: 10.0.22000.1

              BUILDLAB_STR: co_release

              OSPLATFORM_TYPE: x64

              OSNAME: Windows 10

              FAILURE_ID_HASH: {41372a35-237d-1281-82f6-8f5f9bf6044a}

              Followup: MachineOwner

              Comment


              • #8
                Correct, try manually lowering DRAM Frequency with XMP enabled to see if say DDR4-3000 or 2800 can be more stable. It is not XMP itself causing issues, it is typically settings or frequency too high for the system to natively support that causes instability. So if this is the case, a lower frequency should work better. If a lower frequency is still unstable, the modules may be severely unmatched that they do not work well together at all. This can happen when mixing separate modules or kits as the modules are not matched and tested from factory.

                Dump codes do not help much for this other than knowing there is an issue. The system should not produce error code at all.

                Comment


                • #9
                  so that's where it gets interesting. with xmp off, i have yet to experience crashing of any sort. i did also find it quite stable when i upped DRAM voltage to 1.43V. Is it safe to run this RAM at say 1.45V long term if system boots and is stable? what is the downside? how will i know if voltage is too high? won't boot/isn't stable? i know upping voltage would create more heat. does this shorten lifespan of RAM?

                  Comment


                  • #10
                    Higher DRAM Voltage is fine as long as the system is stable. Did you happen to try any lower value to see if it can be stable? Too high of Voltage can cause instability as well, so main thing to look out for is stability and temperature. Too high of temperature can also cause instability, so you will know, however the modules should handle 1.45V without a problem. Lifespan is difficult to say, however many run 1.45V+ and never have an issue. However in theory, cooler is always better.

                    Comment

                    Working...
                    X