No announcement yet.

G.SKILL + HYNIX compatible? 800+800=666

  • Filter
  • Time
  • Show
Clear All
new posts

  • G.SKILL + HYNIX compatible? 800+800=666

    System arrived with three 1GB HYNIX, which were running at 800.
    Had a G.SKILL 4gb pack (2x2GB), so removed one HYNIX and installed the two G.SKILL.
    Why are they together running at 666? Is there a fix to get them to run at 800?

    Oh, by the way, removed the two HYNIX and the two GSKILL were running at 800.
    click here for pic1
    Any suggestion(s) to get the 6GB to run at 800, instead of 666.
    What is the real difference?
    click here for pic2
    I cannot make adjustments in BIOS, but read on another forum re: SPDTool_063.

    Will the use of SPDTool void memory warranty?

    Thank you in advance for any help, tips or tricks for this fix.

    Dell Studio Desktop 540
    Intel Core2 Quad Q8300 @ 2.5 GHz
    Intel G45 Express Chipset (

    G.SKILL F2-6400CL5S-4GBPQ
    (2) 2GB
    CAS_Latency 5-5-5-15
    Speed DDR2-800 (PC2 6400)
    Test Voltage 1.8 ~ 1.9 Volts
    PCB 6 Layers
    Error Checking Non-ECC
    Type 240-pin DIMM

    memory depth 128mb
    DDR2-800 (PC2 6400) 6-6-6
    1.8v 240 pin unbuffered dim
    Up to 8GB 800 MHz Dual Channel DDR2 SDRAM3
    Four (4) user-accessible DIMM connectors, installed in pairs

  • #2
    Not ignoring you but for specs, you'll have to wait for the actual GSkill folks, they should be in Tues AM


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



    • #3
      Thanks for the reply Tman.

      Playing around I see WHY (running at 667Mhz) this is occurring.

      I guess this is why THE BOOK says to not to mix memory.

      Maybe G.SKILL (who knows more about memory than me) may offer a solution.


      • #4
        Haven't looked real close but think because the HINIX are CL^, to run 800 you may have to manually set the speed, and the timings to 6-6-6-18 to run all 4 at 800, got a message from Dumaine on this and will look a little closer in a bit.

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



        • #5
          Tman, thanks.
          I read your Car analogy (RPMs) was trying to discern real difference.

          I read and saved via SPDTool the info from G.SKILL and HYNIX memory.
          link to GSKILL SPDTool file
          ink to HYNIX SPDTool file

          Never done anything like this before, so any help will be very appreciated.

          I opened SPDTool and read GSKILL, then opened SPDTool and read HYNIX...
          side by side I changed the GSKILL to match the HYNIX, but must have done something wrong, as the system still 667MHz.

          The GSKILL CAS was 400MHz 6-6-6-18, 333MHz 5-5-5-15.

          I suppose I can trial and error, but do not want to kill anything.

          G.SKILL appears to be better memory, I am familiar with Micron Tech. ( and they show CL6 on their web page.

          I purchased this G.SKILL from newegg when they were on sale months ago $10/GB, simply on price alone.


          • #6
            ...maybe this will help you save time and offer suggestions...
            I will check back tomorrow, I know you are busy ,so as time permits...


            • #7
              Yes, the G.Skill memory will be able to run on looser timings. It may be more beneficial to run the lower speed with lower latency. People overclock the memory for many reasons. Mostly to attain the highest speed possible, but also getting "more than what you paid for". A lot of the computer hardware are under spec'd, so they have room for overclocking. If you can increase your overall computer speed by 20%, it would be a big deal to many people.

              For example, instead of spending $1000 on a i7 975 CPU, simply purchase the 920 for $300 and overclock it to 3.4GHz+ to achieve the same or even better results than the stock i7 975.

              Thank you
              GSKILL SUPPORT
              Last edited by GSKILL TECH; 02-04-2010, 01:39 PM.


              • #8
                Finally was able to change GSKILL CAS to match HYNIX using SPDTool,

                but the system recognition did not change.

                Handicapped by the bios limitation.

                "It may be more beneficial to run the lower speed with lower latency."-GSKILL TECH

                Was not quite sure of your meaning...but if I plugged in the numbers correctly, then someone's explanation makes sense.

                "The Core 2 Quad has an 8 Byte (64 bit) FSB that operates at 1333 MHz (effective). So...

                8 B x 1333 MHz = 10664 MB/s, or about 10.67 GB/s bandwidth.

                Now lets compare the bandwidths of two different RAM speeds, a dual channel pair of DDR2 667 RAM, and a pair of DDR2 800 RAM. If you have a dual channel 16 Byte (128 bit) DDR2 667 MHz (effective) setup...

                16 B x 667 MHz = 10672 MB/s, or about 10.67 GB/s bandwidth.

                You can see it is equal to the FSB, so your FSB to DRAM ratio is 1:1. This is good, since the one component that has to be fast enough to keep up with the RAM, CPU, and rest of the system is the FSB. But if you have a dual channel 16 Byte (128 bit) pair of DDR2 800 MHz (effective) RAM...

                16 B x 800 MHz = 12800 MB/s, or 12.8 GB/s bandwidth.

                With the DDR2 800 RAM, the bandwidth of the RAM exceeds the bandwidth of the FSB, and the extra speed of the RAM is wasted."

                Someone's thoughts...maybe true.
                "Lower CAS values are faster and give better performance, most notable in games and where multimedia or computational work is concerned."

                I should have used an exclamation mark, instead of a question mark (re: overclocking 667 to 800 & negligible difference).
                Understand your point, just trying to get the most with what I have on hand.
                Could justify spending $34.99 (20090113) ,won't spend $84.99 (20100203).
                Usually try to install max. ram, but will wait for price to come back.

                Originally, missed your Feb 1 email response.
                Thank you for reading and responding.
                Did not mean to be troublesome.