Announcement

Collapse
No announcement yet.

Mushkin 998679 Breaks 2700Mhz Triple Channel

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

  • Mushkin 998679 Breaks 2700Mhz Triple Channel

    While these are not the greatest of timings, I did stop myself from going to CAS 10, which I think now maybe I should have to see how high that would take me?


    This is Mushkin 998679 7-8-7-20 1600Mhz kit
    998679 - 6GB (3x2GB) DDR3 PC3-12800 7-8-7-20 Blackline (Triple Kit)








    I had a terrible time this session and chose the wrong OS and disk setup for much of what I planned to accomplish, so that may have also hindered my memory clocking progress too...

    Not sure, I can only hope for a better session next time around.

    I also did not have a second sensor wire, so I could not monitor the memory temps vs voltage, so I wasn't sure what temps the memory was reaching, maybe once I get another sensor I can do better. Not even sure if the memory was actually getting all that cold, maybe not really so these could probably go a bit higher once I can do things properly



    Max 9-11-x-x @ 2691.6Mhz




    CPU-Z Validator 3.1


    Max 9-12-x-x @ 2713Mhz




    CPU-Z Validator 3.1


    Air testing results here (2500-2600Mhz)
    Last edited by Lsdmeasap; 06-27-2010, 03:04 PM. Reason: Re-Arrange Result Images

  • #2
    Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

    LN2 ftw! Yea those guys did it with Windows Server 2003 Enterprise Edition but going with 'casual' OS(es) like Vista or Win 7 does shed a light to casual others, in term of how far such common thing can go under extreme conditions. Champion among commoners they may say but let see what others can do about it.

    All the best!


    "If Andes is just a grain then take my words with it"

    Comment


    • #3
      Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

      It's Dry ice, I wish I could afford a Ln2 dewar!

      Did what with server 03? I should have used XP on a IDE drive, always have great luck that way and not sure why I didn't this time around

      Comment


      • #4


        "If Andes is just a grain then take my words with it"

        Comment


        • #5
          Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

          Ohh ok, thanks I wasn't sure what you meant.

          I am sure if I had the right setup I could do it in XP. I can only do dry ice though, no Ln2 here, so I am kinda temp limited, plus I don't have tons of spare hardware like many who make those records do.

          When I do this, the hardware is what I use daily, so I can't really break it and I keep that in mind when doing such things, otherwise I might be able to push 1.9 Vcore and 2.0V dram if I had tons of spare hardware laying around.

          Comment


          • #6
            Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

            impressive stuff Ed
            congrats mate
            GIGABYTE Australia

            Comment


            • #7
              Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

              Thanks Dino!

              I didn't get too much further cooling the memory, I think I am at the edge of what this CPU's uncore can do on the cooling I have.

              I could get 2500-2600 on air cooled CPU and memory, so I need Ln2 to get further I think.

              Comment


              • #8
                Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

                Absolutely awesome LSD, keep it up. Also, enjoy the summer weather!

                Since your help, my machine has been completely stable, not one hickup, unless it's from a failed driver uninstall/install!
                Antec 300, E8400/CO-6 4.05GHz CPUID 1.344/BIOSvcore1.3750 VTT1.320 CPURefv0.835/0.76 PLL1.57 Cooled by~
                DD-CPX-Pro
                12V Pump, 3/8"ID-1/2"OD tube, Dual bay reservoir,
                86 Camaro Heatercore, x2 140mm push/pull, MPC-CPU Quad core WB. Hot Day 35c/58c IBT
                KHX8500D2K2/4G 2x2GB HyperX 1080MHz/CAS5-5-5-5-15-tRD9-tRFC54-tWTR13-tRRD3-tRD_WD9-tRTP7-tREF4213-CR2T DRAMv2.2V DRAMRefv1.1 -Getting tighter!
                GA-EP45-UD3R v1.1 F12 MCH1.3 MCHRefv0.835 I/O ICHcore - eVGA GTX275 @ 691/1512/1224
                Games, Media, Programs - WD5001AALS, System - WD500AAKS, Corsair TX650

                Comment


                • #9
                  Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

                  Thanks man!

                  Glad to hear your rig is running smoothly, aside from those driver issues which sounds like you fixed up already.

                  Comment


                  • #10
                    Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

                    have you tried clocking this CPU high to see how it goes mate?
                    GIGABYTE Australia

                    Comment


                    • #11
                      Re: Mushkin 998679 Breaks 2700Mhz Triple Channel

                      I have ran 5.25Ghz, I only use dry ice here and I have not pushed the voltages very hard on it yet because I only have one 980X, so really not sure how high it will go.

                      Colin did tell me it wasn't a very good CPU for Ln2 though, so that may mean it wont clock super high or just don't like Ln2 temps - not sure?

                      I can boot 246 Bclk on water using low voltages, 1.34 or so, so I think it may be a decent CPU overall. Sadly though I can't get the BIOS team to change/fix the QPI Slow mode for me, so I am stuck at that exact Bclk even at Dice temps.

                      Current QPI Slow mode uses 0.5 multi, when it should be x24 multi (GBT Terms), I have talked this over with Colin for weeks and he has not had any luck getting them to change it. They say it is not worth it, or will not help me achieve anything, yet I see many other boards hitting 160+ with the use of QPI Slow mode. They did supposedly fix it with UD7 F7q BIOS, but it was a flash fail from the get go, and F7r still uses 0.5 so it will fail to load as that is WAY to low. Hopefully when Colin gets back my last email will convince him to get them to change this to the proper value of 12 (24 GBT BIOS Terms)

                      I know Bclk isn't everything to many people, but that is something I am always trying to reach for so it just bothers me about this setting, sorry for the rant :)

                      Maybe you could get on them about it too :) Load it up on any X58 board and you will see, some will boot to OS, some will not, but no matter what once you hit the OS it still uses x36 anyway despite setting slow mode - or fails to boot because 0.5 is too low (Some will allow it to stick in slow mode with the 0.5 multi, but it wont boot this way).
                      Last edited by Lsdmeasap; 07-01-2010, 02:34 AM.

                      Comment

                      Working...
                      X