Quote Originally Posted by eva2000 View Post
closest would be my info at http://i4memory.com/showthread.php?t=4743

Just think of clock fine delay like GTL tuning only difference, is there is no set table to go by like GTL. each memory module, different mem divider, mem clock speed and timings will alter the clock fine delay values. The info at http://i4memory.com/showthread.php?t=4743 just sheds some light on what to do when you hit a memory clock speed/instability ceiling.

Use http://i4memory.com/showthread.php?t=4743 with memtest86+ v2.01 to figure out initially what clock fine delay values will help lessen or eliminate memtest errors.

Weird part is, now i tried QX9650 first time on DFI LT X38, i find it helps stablise FSB too 9x445FSB 32M pi stable so far at CPU VTT 1.28v - no need for crazy CPU VTT volts >1.5-1.6v
yeah thanks i did read threw ur guide , its good info . but the problem is this bios wont allow me to set it the previous stable settings. sometimes the settings could be more that 50 in difference. i tried setting them as close as possible but it just became even more unstable.
other thing is it was stable at 40. 200 etc very low settings. i increased fsb and used similiar settings. then it was unstable. now i started again and got to the same fsb and every setting was identical. but fine tune delay was stable at oposite settings ie 1246 311 etc. its weird that it can be stable at a low 40 or so then another time be stable at a high 1246? with the same fsb and volts and timings.
anyway im at 475/1140 again on occt but only for 2hours after that i get a 12v error which i heard is related to nb . so ill up my nb to 1.6 and see if that helps. i have my vtt at 1.553 i just couldnt get stability any lower.