Announcement

Collapse
No announcement yet.

Need help with F3-16000CL9T-6GBTD kit!

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

  • Need help with F3-16000CL9T-6GBTD kit!

    I'm having some issues getting my F3-16000CL9T-6GBTD to stabilize completely. My system specs are:


    M/B: Asus P6TD DELUXE
    CPU: Intel core i7-920 @ 4.20GHz D0
    RAM: F3-16000CL9T-6GBTD
    PSU: 850W Corsair HX-850 ATX Modular
    HDD: Samsung "SpinPoint" 1TB HD103SJ
    GPU: Sapphire ATI HD 5870 PCI-E 2.0 1GB
    PhysX GPU: nvidia GeForce 9600GT PCI-E 2.0 1GB
    I'm running a single water loop just for CPU cooling using
    Radiator: MCR320-QP
    CPU Block: Apogee XT
    Pump: MCP-655

    The thing is that I can't get it to stabilize at 2000MHz and since I'm no RAM expert I could use some help. Here's a brief history of what I've done so far:

    I managed to boot it @ 2000MHz and validate

    but upon starting linx or even prime95 the system froze.

    I then had it running at 1600MHz and it's stable but no matter what linx couldn't get past the second pass. It always stoped due to an error around 5 to 9 mins in the test. CPU temps peeked at 70-71 degrees. My settings were:
    BLCK:200
    CPU Voltage: 1.325v
    DRAM: 1.66v
    QPI/DRAM Core: 1.36250v

    So I pushed the voltages a little and managed to get it stable for a short linx test at 1600MHz 9-9-9-24



    I then tried setting it back to 2000MHz 9-9-9-24 with the settings I found here. That resulted in slightly better stability but still not good enough, here's what I got.


    This was better than the first attampt @ 2000MHz but still not stable with
    BCLK frequency - 200
    PCIE frequency - 100
    DRAM frequency - 2005 MHz
    UCLK frequency - 4009 MHz
    DRAM CAS# Latency - 9
    DRAM RAS# to CAS# Delay - 9
    DRAM RAS# PRE Time - 9
    DRAM RAS# ACT Time - 24
    DRAM REF Cycle Time - 72 (there was no 74 option in my BIOS and then it was going up to 84)
    DRAM Timing Mode - 1N
    CPU voltage - 1.325V
    CPU PLL voltage - 1.88V
    QPI-DRAM voltage - 1.55V
    IOH voltage - 1.35V
    IOH PCIE voltage - 1.60V
    ICH voltage- 1.20V
    ICH PCIE voltage - 1.60V
    DRAM Bus voltage 1.66V

    Final attempt was at a 4.22GHz OC with:
    BCLK frequency - 201
    PCIE frequency - 100
    DRAM frequency - 1683 MHz
    UCLK frequency - 3368 MHz
    DRAM CAS# Latency - 7
    DRAM RAS# to CAS# Delay - 7
    DRAM RAS# PRE Time - 7
    DRAM RAS# ACT Time - 20
    DRAM Timing Mode - 1N
    CPU voltage - 1.325V
    CPU PLL voltage - 1.90V
    QPI-DRAM voltage - 1.40V
    DRAM Bus voltage 1.66V
    and all the rest in auto. results were similar, pi running ok, prime95 small for 7 hours but linx crashed on the 2nd-3rd pass.

    So I only managed to stabilize it completely at CPU 4.2GHz, DRAM 1600MHz 9-9-9-24 that is not acceptable timings for 1600MHz especially from this kind of chip. I wouldn't mind running at 1600-1683MHz instead of 2000MHz but I would expect tighter timings of at least 7-7-7. I think I saw somewhere that it is capable of 6-7-6-18 but I have no idea how to get there... As I said I don't know much about RAM settings so If anyone could help I would greatly appreciate it. Even some theory of how this whole RAM thing works so I can understand how I can adjust the voltages rather than randomly playing around with them and hoping to stabilize it. Thank you all in advance.

    Last edited by Malus; 01-11-2010, 05:00 PM.

  • #2
    Isn't there anyone that can help me with this? Can't stabilize it above 1600 8-8-8-24...

    Comment


    • #3
      Try DRAM Timing Mode 2N instead of 1N. CPU PLL 1.90V. Let me know how that goes.

      Thank you
      GSKILL SUPPORT

      Comment


      • #4
        Originally posted by GSKILL TECH View Post
        Try DRAM Timing Mode 2N instead of 1N. CPU PLL 1.90V. Let me know how that goes.

        Thank you
        GSKILL SUPPORT
        That is to run at factory 2000MHz 9-9-9-24? And if yes with what min DRAM voltage? What about tighter timings in 1600MHz to try and keep the voltages lower? Thanks for the reply btw!

        Comment


        • #5
          DRAM Voltage at 1.65V should be enough. At DDR3-1600 you can use 7-7-7-21 with 1.60V and everything should be fine. Although if you want maximum performance, you can try 6-7-6-18 as that is the lowest tested timings.

          Thank you
          GSKILL SUPPORT

          Comment


          • #6
            Originally posted by GSKILL TECH View Post
            DRAM Voltage at 1.65V should be enough. At DDR3-1600 you can use 7-7-7-21 with 1.60V and everything should be fine. Although if you want maximum performance, you can try 6-7-6-18 as that is the lowest tested timings.

            Thank you
            GSKILL SUPPORT
            Thanks for the reply G, will try 7-7-7-21 at 1600 and then maybe go for 6-7-6-18, what voltages do you think a 6-7-6 will require? Over 1.65? Don't want to max it out at 1.7...

            Comment


            • #7
              DDR3-1600 6-7-6-18 should only need 1.60V. You will never need over 1.65V unless over DDR3-2200+

              Thank you
              GSKILL SUPPORT

              Comment


              • #8
                Originally posted by GSKILL TECH View Post
                DDR3-1600 6-7-6-18 should only need 1.60V. You will never need over 1.65V unless over DDR3-2200+

                Thank you
                GSKILL SUPPORT
                I tried it at 1600 6-7-6-18 with the settings you can see in the screenshot, still no good...



                goes unstable way too early... maybe some other voltages I need to alter asides the QPI and the PLL? Maybe the CPU needs more to cope with this memory settings? damn it's getting frustrating

                Comment


                • #9
                  At 4.2GHz, you may need more VCore. Have you tried 7-7-7-21? I'd go with that first before attempting to use lowest timings. From here on out, it could be too low of timings (memory can't do it), or the CPU has too much stress it is unable to do DDR3-1600 at 4.2GHz. Typically what you should do is find the limits first, then work on up. So for example, use default CPU speed to test memory capabilities then combine them.

                  Thank you
                  GSKILL SUPPORT

                  Comment


                  • #10
                    Originally posted by GSKILL TECH View Post
                    At 4.2GHz, you may need more VCore. Have you tried 7-7-7-21? I'd go with that first before attempting to use lowest timings. From here on out, it could be too low of timings (memory can't do it), or the CPU has too much stress it is unable to do DDR3-1600 at 4.2GHz. Typically what you should do is find the limits first, then work on up. So for example, use default CPU speed to test memory capabilities then combine them.

                    Thank you
                    GSKILL SUPPORT
                    Tried it at 7-7-7-21 with exactly the same voltages and still nothing so I have settled down to 8-8-8-24 for now that runs great at 1.4 QPI and 1.325v CPU(did 50 full passes in Linx with no error). And that's what makes it more confusing, 8-8-8-24 at 1.4v works great but 7-7-7-21 at 1.6-1.65v still unstable even with CPU at 1.4v. Same goes for the factory 2000MHz 9-9-9-24 (that is what I was originally aiming for) at up to 1.6v QPI and 1.4 CPU. Still can't go over 8-9mins in Linx. As for the stress of the CPU I don't think that's the case cause CPU is stable even at 4.4GHz but I'm just running it @ 4.2 for some extra peace of mind and better temps!

                    Comment


                    • #11
                      Originally posted by GSKILL TECH View Post
                      DDR3-1600 6-7-6-18 should only need 1.60V. You will never need over 1.65V unless over DDR3-2200+

                      Thank you
                      GSKILL SUPPORT
                      one more thing, when you were saying 1.6v were you refering to DRAM core voltage or QPI/DRAM? Cause the issue seems to be the QPI/DRAM voltage here...

                      Comment


                      • #12
                        hi i'm in the same situation..i have the ripjaws ddr3-2000 3x2gb...i can right now the only stable setting with i7-920 P6T dlux v2 is at 4.0 ghz 1600 8-8-8-24 @ 1.6 bclk 200 multi 20 cpu 1.3 cpu pll 1.8 vtt 1.4

                        here's a list i tried and failed:
                        1600 7-7-7-21@ 1.6 bclk 200 multi 20 cpu 1.3 cpu pll 1.8 vtt 1.4 linx failed after a few passes
                        1600 6-7-6-18@ 1.6-1.66 bclk 200 multi 20 cpu 1.3 cpu pll 1.8 vtt 1.4 linx failed after a few passes
                        2000 9-9-9-24 or 27 1T or 2T@ 1.66 bclk 200 multi 20 cpu 1.3 cpu pll 1.88-1.9 vtt 1.55 can't even post

                        I've also tried the settings in the sticky and also a few other people's settings @ 2000 i can't even post with any of them. However i haven't raised my dram volt past 1.66 or VTT over 1.55.

                        Can you let me know how i can run 2000 if not i would like to run 1600 at a tighter timing, preferable 6-7-6-18. Right now at 1600 8-8-8-24 its not worth the premium i paid and i'd return them

                        thanks!

                        Comment


                        • #13
                          Originally posted by jagger333 View Post
                          Can you let me know how i can run 2000 if not i would like to run 1600 at a tighter timing, preferable 6-7-6-18. Right now at 1600 8-8-8-24 its not worth the premium i paid and i'd return them

                          thanks!
                          couldn't agree more, ain't worth the extra cash to run it at 8-8-8-24 1513MHz like I am atm as well. Hope we can get some detailed answers soon or I'll have to RMA them and get another brand... I would like to run it at 1600 with tighter timings as well so help us out here. and asides the DRAM voltage help with QPI, ICH, IOH voltages etc... I had to downclock my cpu too @ 4.0 down from 4.3 too even use it at 1513 8-8-8-24... It works at 7-7-7-24 too but unstable like hell, I even get "overclocking failed" at startup sometimes with CPU 1.3125 DRAM 1.66 QPI 1.475...

                          Comment


                          • #14
                            Malus,

                            When running a high overclock on the CPU, it dramatically increases the amount of stress, thus limiting the memory controller abilities as well. This is why a high QPI Voltage is needed. Consequently, not all i7 920 CPUs can run 4.0GHz+ with DDR3-2000. To test this theory, I would suggest reducing the CPU overclock, get the memory to operate at your rated specs, then see how high you can go on the CPU overclock with the proper voltages.

                            In post #7, I was speaking of the VDIMM DRAM Voltage. Although, for DDR3-2000, most motherboards will require just about that much QPI/DRAM Voltage (1.60V).

                            Jagger333,

                            DDR3-2000 will require a slightly higher QPI/DRAM.

                            Tell you guys what, I'll pull out this motherboard tommorrow and retest it with the latest BIOS then report here with the settings I used. Meanwhile, make sure you have the latest BIOS as well.

                            Do you guys have C1E, TM, load line, spread spectrum, and SpeedStep disabled? That may be causing the instability.

                            Thank you
                            GSKILL SUPPORT
                            Last edited by GSKILL TECH; 02-09-2010, 04:20 PM.

                            Comment


                            • #15
                              thanks! sounds good

                              I've tried QPI/DRAM up to 1.55 and i'm scared to go over that so i guess i'm more interested in getting a better timing for 1600 6-7-6-18 if possible without any errors in linx, memtest, prime

                              currently i have
                              c1e disable, TM enabled, LLC enabled, spread spectrum auto, speedstep disabled.

                              I had disabled them all before in testing but it made no difference

                              Comment

                              Working...
                              X