Intel Overclocking. | Forum profile

Hunter, Math can only tell you what GTL would be in theory, and not being able to measure it off the board to compare to its theoretical value means there is no certain answer anyway. Really, there are a limited number of gtl setting combinations, and you just test them out to find the "best". With a low vtt GTLs may not be needed or may not make a difference, but +.1 or +.15 probably would affect gtl by 450fsb. Also what...

No I don't mean that, I mean that some of the benches I ran at say 6-6-6 were slower than the same benches at 6-7-6, but as I say, it may just have been the IMC on my 860 at it's limit.

One thing you could try before af13's, is to set CPU Clock Skew Control to 200ps and see if that gives you any more stability.

What are the upward limits of the 'safe zone' for an i5 then? 1.5?

Well....I dropped the vPLL from 1.7v to 1.6v and i got OCCT test failed after ~2 minutes...All setting still the same, vcore , vnb , vfsb, vdimm , charge overvoltage etc...

Tnx for the response well yea its ok at 442mhz atm, but it gets more difficult above 450mhz.. If i want to OC to 3.66ghz that's 458mhz i need to pump allot more volts, but its not stable 100% so im thinking about those GTLRs and whats the right one, when i force fsb overvoltage (and cpu overvoltage) since those 2 (3) work with one and another. I know its a simple math thing but i'm not such a math genius here is what i used for...

Most Q9450s werent great clockers tbh, not over 3.6/3.8Ghz anyway. Same with most 45nm C2Qs. The later stepped Q9650/Q9550s would often do 4Ghz but almost all of the others didnt do so well. Try running your RAM on a lower strap, that should rule it out as the cause, but my money is on the CPU just not liking higher clocks.

i would try setting trd to 8

Honestly I wouldn't run a PLL of 1.7 volts on a daily basis.

GtlRef also works less well the hotter the chip is running, if running >60C ie an average air cooled oc on load , the gtlref has little effect.