Announcement

Collapse
No announcement yet.

Quest to stabilize 2133 w/ASRock Z68 Extreme 4 Gen 3

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

  • Quest to stabilize 2133 w/ASRock Z68 Extreme 4 Gen 3

    Hi Tradesman / GSkill Tech
    This rig 2500K@4.4ghz (rock solid even at 4.8ghz) / HSF H100 max load temps 60-64c / MB ASRock Z68 Ext4 Gen3 bios1.10 (newest non-beta) / GSkill F3-17000CL11D-8GBXL.

    Been a steady GSkill promoter since 2005 and the 939/Opteron OC craze and receiving my 1st hand picked/binned DDR500 (New 283 4m passed) sticks from LA/GSkill. I'm no expert, but I've been around the block a few times OC'ing ram. If you remember Oscar Woo and the DFI Lanparty bios options for ram tweaking then you know what I mean! ... In all that time I've never had to RMA any ram from GSkill! I thought I may have finally run into a weaker than spec'd kit, AND initially could not get the sticks to boot with the XMP profile. After hours of testing manually setting timings/volts combinations(4-6hrs), I finally gave up @2133 and settled on stability @ 1866, not the speck'd 2133, I ran accross a fairly simple solution while reading up at GSkills Forum, something we use to do in the "old days", but was contrary to ASRocks recommendation for initial set/boot-up ...

    XMP finally booted when I switched DIMM slots from A1/B1 to A2/B2 - manually set ram voltages 1.65v / VCCSA - 1.107v VTT 1.5v? Passed all my prelim tests/benches (HyperPi 1m/8m LinX short runs/3DMark/Heaven/Crysis benches) ... WTF problem solved right? ...

    BUT NOOO P95 still failed within minutes It was so stable in the other tests, I thought I had it solved ... I'm maxed at DRAM volts (1.65v) and really don't like it that high, guess i could raise VTT more? VCCSA more? BUT what exactly does this do? Could go back to manual timings greater than spec'd (12-12-12 :O), but that's not what you pay a premium for? Oh well here goes my 1st RMA with GSkill ...

    Why I did this next, I still don't know, ready to give up while making some final notes on the bios settings, I noticed XMP had slightly raised/lowered? my VTT voltage to 1.142v when set to "AUTO" and against any overclockers inner instincts, I decided to set the DRAM and VCSSA back to "Auto" just for the heck of it and see if GSkill and ASRock would play together.
    Let me just say right now I really like this ASR board as it made "manually" OC'ing my other kits of ram and the 2500K a breeze (see notes below) ...

    BAM wouldn't you know it ASRock "auto" sets the DRAM @ 1.530v and the VCCSA back to default 0.925v AND it is NOW Prime stable including the dreaded 1344/1792 FFT's ... Lowering VOLTAGE? it has worked for me before, BUT RARELY!!!

    Moral of this story is don't give up, sleep on it, do some more reading no matter how smart you think you are :P ... and even though these new platforms take a lot of skill out of "old school" manual OC'ing with "modern presets" it still pays to individually clock your ram seperate from your CPU ...

    Sorry this post is sooo long, it's like my girlfriends "blog" :P ... but I hope someone finds the information useful!

    Notes I was putting together for RMA ...
    Problem ... Will NOT boot with XMP settings, [Bios XMP 1.2 profile 1 : 11-11-11-30 1.50v DDR3-2133][CPU-Z XMP-2134 profile=1067MHz 11-11-11-30-50-1T-1.500v] CPU@ stock or OC'd 4.4GHz to 4.8Ghz / DRam 1.54v - 1.65v / VCCSA 0.925 (auto) - 1.107 (manual) ...

    Will boot with manual timings 11-11-11-33 (1 or 2T) But immediately errors in P95 Torture and 1344/1792FFT's / passes HyperPi 1m but will sometimes error @8m / somtimes passes short run LinX.

    I've tried various other timing combo's with the only ram settings allowed on this board ... as "loose" as ...
    tWR-16 / tRFC-170 / tRRD-7 / tWTR-8 / tRTP-8 / tFAW-50
    Tried all kinds of voltage "boosts" / combo's ...

    Note: this board runs XMP profiles just fine with 2 other kits I have ...
    1) GSkill RipJaw-X F3-12800CL7D-8GBXH (nice ram by the way! stable@1866 9-10-9-28 1N 1.605v VCSSA "Auto" (0.925v) tWR-12 / tRFC-130 / tRRD-5 / tWTR-6 / tRTP-6 / tFAW-26 (haven't had time to go for 2133 ... yet?)

    2) GSkill RipJaw-X F3-12800CL9D-8GBXL ... also OC's to 1866 11-11-11-28 1N 1.605v VCSSA "Auto" (0.925v) tWR-12 / tRFC-150 / tRRD-6 / tWTR-7 / tRTP-7 / tFAW-28
    failed 9-10-9-28
    Last edited by Tomcat; 01-21-2012, 07:48 AM.

  • #2
    Can never read enough and thanks for sharing this with everybody, I too am a fan of ASRock (mobos anyway, not to impressed with their tech support)...Sometimes it takes them a bit to get the BIOS set up to take in the higher performance sets of sticks and their XMP settings, but I normally allways set up manually anyway.....Thanx again and always feel free to jump in with Words of Wisdom


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

    Tman

    Comment


    • #3
      Originally posted by Tradesman View Post
      I too am a fan of ASRock (mobos anyway, not to impressed with their tech support)...Sometimes it takes them a bit to get the BIOS set up to take in the higher performance sets of sticks and their XMP settings, but I normally allways set up manually anyway.....Thanx again and always feel free to jump in with Words of Wisdom
      NP Ditto ... I normally always set-up manually also, and that's what through me off here. No stability until I switched Dimms, I also forgot to mention, when I went to A2/B2 dimms one of the sticks wouldn't work at all, maybe because of the overvolting? Who knows but when I switched the sticks within A2/B2 they worked! This was truly an exercise in patience and involved almost every troubleshooting trick in the book. The new trick I learned here (GSkill Forum) that was part of the solution was to try and boot the sticks while "overclocked" :O ... that goes against all the old school philosophies of when putting in new hardware always set your OC's back to stock! :P

      I'm still not convinced these sticks are going to be good to go over the long haul, who would after jumping through hoops like that, but for now I thank this forum for saving me from my 1st RMA w/GSkill ... I like perfect/clean records!

      Comment


      • #4
        Well, keep us informed, and keep an eye on ASRock for BIOS updates....with all the newer Intel and AMD sockets, it often seems more of a challenge than the old socket 775 days, but then the additional power seems worth the extra effort when required.


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

        Tman

        Comment


        • #5
          Nice read. Can I ask 1 thing thou. When you switched rams or between trying new settings, did you cut power to mobo ? (unplugging power cord or using the psu switch)

          I had a little adventure myself today and I'm having a hard time believing the instability cured itself by just changing ram sticks around and having them working at the starting positions after all troubleshooting I did.
          This cat scratches free.
          Intel i5 2500k @ 4ghz Offset - 0.015v | Asus P8P67 EVO | 12gb GSkill RipjawsX F3-14900CL8D-4GBXM 2x2gb & F3-14900CL9D-8GBXL 2x4gb 1866 9-10-9-28@1.6v | AMD HD6850 1GB | Enermax Infiniti 650w (28a,28a,30a) | 2xWD500 Caviar Black & Blue | CM Haf-912 Advanced | Noctua NH-U12P SE + Akasa 12cm Viper | Windows 7 Ultimate x64 SP1 | AMD Catalyst 13.10b

          Comment


          • #6
            Originally posted by Catscratch View Post
            Nice read. Can I ask 1 thing thou. When you switched rams or between trying new settings, did you cut power to mobo ? (unplugging power cord or using the psu switch)

            I had a little adventure myself today and I'm having a hard time believing the instability cured itself by just changing ram sticks around and having them working at the starting positions after all troubleshooting I did.
            ALWAYS! I don't just use the PSU power switch (residual "juice" left in the capacitors) I religiously fully "UNPLUG" when switching ram/components but I rarely fully power down when CMOS reseting(hangs/black screens) or trying different timings. That's the beauty of these newer, higher end boards, mine even has a CMOS reset right on the exterior I/O panel!) ... glad you had the patience to solve your problem, kind of the whole point of this thread

            EDIT: Your ? about powering down between timing changes/boot hangs(black screens)? ... Without the convient features of some of these newer boards, ie CMOS reset buttons (instead of jumper pins) and multiple "Bios User Profile" save features, I may not have had the patience to stick with it :O, hope you had a board with at least 1 CMOS reset button (not pins)
            Last edited by Tomcat; 01-22-2012, 06:47 AM.

            Comment


            • #7
              Might possibly have been a bit of lint, even a bit of hair in the socket or on the contacts of the sticks, doesn't take much, not a bad idea to keep some rubbing alcohol around and some foam applicators to clean contacts and such....while many advocate Q-tips, have to often been called out to a problem to find a fine cotton strand (from the Q-Tip of course) on a stick of DRAM, even on a graphics card and the bottom of a CPU at times, to be the problem causer I was called for


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

              Tman

              Comment

              Working...
              X