Announcement

Collapse
No announcement yet.

Can't figure out timings with Shuttle SX58H7 and 12800CL8TU

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

  • Can't figure out timings with Shuttle SX58H7 and 12800CL8TU

    I just got the SX58H7 with a I7 920 and can't get the F3-12800CL8TU-6GBPI to stay stable even at default settings. At default, it detects 6GB and runns at 1066MHz. With XMP, it detects only 4GB but runs at 1600MHz. When it does boot up, it's a crap shoot whether it's gonna stay on. Sometimes it crashs upon login, when opening picture files or running games(Farcry2/WoW) it will BSOD a nvlddmkm.dll error. I tried to adjust the timings and voltages manually but it just boots in loops and I have to clear the CMOS to get back into BIOS. I RMA'd the memory but the new chips are still giving me problems. I figure I'm screwing something up somewhere. I've got the newest BIOS, running Windows 7 and have a nVidia GTX260 with the latest Win7 drivers.

    you can download the motherboard manual here: http://image.shuttle.com/ResourceCen...?file_id=12733


    Any help is greatly appreciated.

  • #2
    I don't think it's the memory in general.

    You can run the 'Auto' and do an Memtest86+ (http://www.memtest.org/#downiso), If there' no errors it's not the memory. I just had conversation on this same issue on another board. Seems the issue might be in latencies and slots memory is used at also might be that the board voltages are too low. When you push it up in XMP/EPP profile the board sure does hit the first 4-5 values from the memory profile, the rest of the values are the same what board decided to use for 1066Mhz which is probably the main problem here.

    Try this setups:
    #1:
    - Make sure memory in in slots 2-3-4 (the red ones only)
    go to 'Frequency/voltage functions':
    - Change 'DRAM voltage' to 1.65v
    - Use Manual Setup of memory latencies:
    * tCL: 8 tRCD: 8 tRP: 8 tRAS: 21
    * tRRD: 7(this value might be wrong, if you can see board default please use +1t from it)
    * tRFC: 84

    and try it with this setup..

    If it still doesn't work couple you please take 'Everest' from http://www.lavalys.com/products.php?...ang=en&page=48 and go to motherboard > chipset page copy the page in here [alt-PrtScr drop it to paint ctrl-v > upload it to somethinglike tinypic.com and add as image here] with any setup you can run. So, we can see what the automatic memory setup looks like.
    Last edited by genetix; 07-08-2009, 03:52 PM.
    "Sex is like freeware, shareware on weekends. When do we get to open source?" -TwL

    Thanks AMD/ATI for banning legit customers who asks questions of your screw-ups:
    http://i45.tinypic.com/30j0daq.png

    Comment


    • #3
      With those settigns it would crash after logging in. Here's the Everest chipset report.

      Comment


      • #4
        how about clicking the IMC memory controller of that screen.

        The crash you see is most like an Vista error in this case also. I know for sure there's an bug in the system when sync on memory to core and the crash was probably caused by can't read driver file from memory.

        Anyway, I am out of ideas, if that indeed crashed. Perhaps you could try the MemTest or 8-9-8-24 timings. See, if that also crashes since on that point I'd suggest running the memory test as there's probably a bad stick.
        "Sex is like freeware, shareware on weekends. When do we get to open source?" -TwL

        Thanks AMD/ATI for banning legit customers who asks questions of your screw-ups:
        http://i45.tinypic.com/30j0daq.png

        Comment


        • #5
          IMC memory controller?

          Comment


          • #6
            Originally posted by Hagarat View Post
            IMC memory controller?
            your screenshot shows no information about memory controller which we are working with. 'Northbridge: Intel Nehalem IMC' probably would show timings of memory & general.
            "Sex is like freeware, shareware on weekends. When do we get to open source?" -TwL

            Thanks AMD/ATI for banning legit customers who asks questions of your screw-ups:
            http://i45.tinypic.com/30j0daq.png

            Comment


            • #7
              oh, ok. here ya go.

              Comment


              • #8
                Wonder why the board prints 1066Mhz as last supported. Could add this picture to board manufacturers pages also because the problem there is that, if windows will see this Vista/7 has an Overclocking block in some % scale (the system won't boot even while CPU or memory would be stable). This means if windows sees RAM over clocked from 1066Mhz it would be 50% OCs. Windows sure wouldn't accept this in any case.

                (I don't now know is that 8-9-8-24 spinning on 1600Mhz, but if it is. Run MemTest on it to make sure none of sticks are bad.)

                This could allow you to run this memory one spec:
                About the timings and correcting the 1600Mhz. You might wanna raise the tRFC to 70-85T on your BIOS 59T is too low so is the tRRD should be minus 1-2T from tRCD at minimum. tREF 509T is literally something damn weird, but I think with those changes you can push it to CL8-8-8-21.

                Anyway, even on current 8-9-8-24 on 1600Mhz board should be pretty near as fast as the spec. You can test everest cache/memory benchmark does that speed satisfy. +1T on minor values ain't exactly that huge lose to speed actually just adds a bit more scalability.
                Last edited by genetix; 07-09-2009, 01:29 PM.
                "Sex is like freeware, shareware on weekends. When do we get to open source?" -TwL

                Thanks AMD/ATI for banning legit customers who asks questions of your screw-ups:
                http://i45.tinypic.com/30j0daq.png

                Comment


                • #9
                  I can't get it stable at any decent setting. Right now I'm at 1066MHz and the CPU is 1.8GHz. Prime95 runs ok at this setting but totally ridiculous to have to turn it down this much. Also, I noticed the memory doesn't like 1.65v or 1.625v at all. It never boots up at those voltages.

                  Comment


                  • #10
                    hmm, that memory should hold clean on 1.7v even. Could you check me what does the everest 'Computer->sensor'page say the vDIMM voltages would be, atm, volts.

                    Also to stabilize the RAM. This is in Windows or in MemTest86+? (no offense to either side board manufacturer, G.Skill or Microsoft, but I still think that the board BIOS is faulty & Windows seeing too low maximum of board is the cause of all of this. You could prove my theory false or correct by running memtest86+ on it, if memtest doesn't error on 8-8-8-21 this means it's either on faulty BIOS or Windows would restrict where we should search the problem.)

                    Yeah, I understand it's totally ridiculous situation to run low speeds on specially on Core i7 which can take it up. As far I looked on the the timings on your Everest it should be stable as those timings, but dunno why it isn't need G.Skill help on this one. Assuming you have pumped CL9 and see what's the last stable under 1600Mhz or is there any.
                    Last edited by genetix; 07-09-2009, 03:37 PM.
                    "Sex is like freeware, shareware on weekends. When do we get to open source?" -TwL

                    Thanks AMD/ATI for banning legit customers who asks questions of your screw-ups:
                    http://i45.tinypic.com/30j0daq.png

                    Comment


                    • #11
                      Just got off the phone with GSkill. we tried a number of settings and nothing but the default 1066MHz would boot. He seems to think it's the board.

                      I don't see vDIMM voltage on the sensor page

                      Comment


                      • #12
                        Even with loose timings, it doesn't want to post at any speed but 1066. You may want to shoot an email to Shuttle and see if they have anything up their sleeve. Meanwhile, try testing each stick with the XMP to make sure a stick is not defective. It is odd that XMP will show up as 4GB not 6GB. I will see what more I can find out tommorrow.

                        GSKILL SUPPORT

                        Comment


                        • #13
                          Originally posted by GSKILL TECH View Post
                          Even with loose timings, it doesn't want to post at any speed but 1066. You may want to shoot an email to Shuttle and see if they have anything up their sleeve. Meanwhile, try testing each stick with the XMP to make sure a stick is not defective. It is odd that XMP will show up as 4GB not 6GB. I will see what more I can find out tommorrow.

                          GSKILL SUPPORT
                          Got to agreed on there. This just has to be BIOS issue, but... (I'd still do the test I asked for on my last message).

                          This same happens on 'Asus P6T Deluxe V2' and as same as here even, if we tweaked on board it wouldn't recognize clearly the 6GB on every POST even on loosed timings or higher voltages only lower speeds. Neither in there was anything wrong on the sticks. (actually there was nothing wrong in DOS bootable MemTest86+, but in windows where the problem began).

                          Somehow, call me paranoya I still think this issue has something to do with Vista overclocking limit when it only sees 1066Mhz DDR max at 1600Mhz run.
                          "Sex is like freeware, shareware on weekends. When do we get to open source?" -TwL

                          Thanks AMD/ATI for banning legit customers who asks questions of your screw-ups:
                          http://i45.tinypic.com/30j0daq.png

                          Comment


                          • #14
                            Vista overclocking limit??? Does win7 have the same issue??

                            Comment


                            • #15
                              Yes, Windows 7 has 'lowered' limit. As microsoft obviously wants to 'punktuate' Windows 7 now. So, it has smaller limit, but it's still there. It's like different component based 'last % value you can go' and then it will start random Blue Screening for no reason or simply failing on boot. Only solution for this is to install the system with the higher overclocks funny thing is people seems to think installing windows in 'stock' is a better option, lol, while it's the exact opposite. Perfect example would probably be DDR2 at 1200Mhz CL6 I recently tested under Windows Vista x64 SP2 latest system installed to clean disc with up-to-date drivers at 800Mhz on RAM and switching to 1200Mhz every single test in this world will be rock solid, but windows won't while this same OS disc installed with the 1200Mhz clocks CL6, driver updates, system and applications are rock solid.

                              This is something no one wants to talk about. Since no one can verify this to be true.


                              Don't really even care what people thinks of this 'theory' or deny it (who knows perhaps it has something to do with HAL detection while installing system), but I am saying it's there. you can listen or not for my paranoya mind.
                              Last edited by genetix; 07-10-2009, 02:48 PM.
                              "Sex is like freeware, shareware on weekends. When do we get to open source?" -TwL

                              Thanks AMD/ATI for banning legit customers who asks questions of your screw-ups:
                              http://i45.tinypic.com/30j0daq.png

                              Comment

                              Working...
                              X