Q9450 on P5K-Dlx

[sbinh]

Ran to Microcenter near by awhile back when they first carried Q9450 .......

popped it into P5K-Dlx board, reset CMOS .... booted it right into windows. .....Pushed it up to 3.4GHz

Anyway, up to 3.4GHz now .. kinda hot .. Idle: 43*C ... Might need to reseat heatsink



Pushed it a little more to 3.6GHz (vcore 1.25v)



Amazingly, it ran fine and passed OCCT test ... but the loaded temp 55*C (max) ... 2*C lower than it used to be at 3.4GHz.


Pushing it a little more after let it runs for awhile @3.6GHz ...

Now, able to push it to 3680MHz with vcore = 1.275v (in BIOS) ~ 1.25v in windows .... Max loaded temp 55*C.....
(Ealier, was not able to push FSB over 454MHz ) .. Now, it's stable at FSB=460MHz .. OCCT failed when FSB set to 463 (or above) within 10 minutes (with vcore set to 1.30v in BIOS)







It appears to me that this CPU need to be "burned" at new speed for awhile before it would be able to run stable at next 5 MHz FSB increasement.

After increasing vMCH from 1.55v to 1.7v :o ), I was able to raise FSB to 465MHz.......







Will try to push it up more ..

[Methious]

Why is Tj Max on core temp in the first two screenies at 105C and in the last 3 shots it's 95C on the same CPU?

[sbinh]

Why is Tj Max on core temp in the first two screenies at 105C and in the last 3 shots it's 95C on the same CPU?

According to uncleweb - who developed new tool to measure core's temp : RealTemp, the TJMax for 45nm chip is 95*C not 105*C as CoreTemp report.

On first two pictures, I was using old version of CoreTemp - that reported tjmax = 105*C
On the other hands, on last 2 pics, I used both RealTemp and newest version on CoreTemp (which allows user to change tjmax to 95*C)

[ibleet]

You don't burn-in, stress test your system at stock before overclocking?

[sbinh]

You don't burn-in, stress test your system at stock before overclocking?

Seriously, I have never done so ...