Q9650 Overclocking!!

P

prasvt

Gawd
Joined
Dec 8, 2005
Messages
557
I figure since there was a price drop on these suckers, there might be a few of you picking up Q9650s to OC. Since this is a slightly different animal from the Q6xxx series voltage-wise, I thought I'd start a thread on OC'ing them. Plus I can always use the suggestions.

About to start a Intel Burn Test 20 pass run with these settings:

467 x 9 = 4.2 ghz
Ram @ 935 (until I'm stable and then I'll see whether it's possible to run 8gb @ 1124)

Vcore - 1.35V
NB - 1.40V
FSB Vtt - 1.35V
VDimm - 2.075
CPU Clock skews -- all at normal
DRAM timings automatic

All the other bells / whistles at auto including GTL for the moment.
 
Looks fine. Overclocking the chip is not a problem, tweaking ram, voltages to keep it up and squeeze some more performance is.

So are you just sharing your overclocks or needing help? Cause 4.2ghz is not impressive for Q9650 ;)
 
LOL, no I'm not looking for speed records. I've seen people hitting 4.5ghz at 1.4V+, but I don't plan on pushing more than 1.3625V, which looks like may be 4.2 GHz. As much as I enjoyed abusing my Q6700, I knew it could withstand voltage more than these 45nm chips.

So yes, I could use some help tweaking some of the advanced settings (assuming they'd help) - like GTL, NB.

467 x 9 @ 1.35V failed prime on Core 0 after 2 minutes. Aside from the Vcore all the other settings are what I put in the first post.

What should do I do now..bump the vcore or mess w/ the GTL settings?
 
UtopiA said:
It isn't?
Click to expand...

Nope, well to me.

Bump the voltage a few notches, Your ram is rated to run @ 1066 so im sure ram is not cauing stability problems, Run Ibt instead of prime, more accurate.

You can always just bump the voltage to 1.4V, if its ibt (10 runs at full stress) stable, keep notching down the voltage til its unstable.
 
What's the max safe temp on these ..70C? I know intel burn test temps are higher than prime temps..

nm...60C all cores thus far. BTW, does IBT stop if it finds errors?
 
Nope, but you can stop it, best way to me for doing so is just leaving error detection off, so if i see a round comes uneven, i can stop it and tweak it and retest.

As for max safe temps, If your getting 70C+ temps, it might be time to tune it down, its far from intels max temp, but its still indeed high
 
Okay..well my temps are like 60 60 56 57 after 20 minutes.

What do you mean if you see a round comes uneven? Does the test show more info when you have error detection off? I think I turned it on...so that means I won't know until all 10 passes finish?
 
If you turn off error detection, instead of getting the percentage and showing how many passes it passed. With it off, it gives out Residual <norm> numbers, which should be all the same, if one even has a diff number than normal, Its unstable.

Its very nice to turn it off, since like i said, you can stop it and tweak it and retest, while with it on, you have to wait til all the passes are done.
 
Ok got it..I'll restart it with the detection off. I guess in the mean time, I'll go plug-in the ol' Q6700 into the 'family' machine. Time to retire the e2200.
 
UtopiA said:
It isn't?
Click to expand...

I think any OC is impressive. And I think 4.2GHz is very very good. Yeah, it might be far from a world record, but it's a damn fine OC man.

And if you wanted to know all 45nm Q9xxx chips have a TJMax of 100C.
 
Thanks BlueFireIce ..honestly, the only reason I even want it to go this fast is b/c I do alot of HD video encoding...so it would just speed things up a bit. Thanks for the info on the temperatures - unfortunately, CoreTemp is being very inconsistent. I just ran the burn test for 20 minutes and the temps were in the high 50s / low 60s. I stopped the test to restart with error detection off, and this time Coretemp is showing 70s/80s. I don't get it..
 
The TJMax shows 100C in CoreTemp 0.99.3. RealTemp is showing the same temps, which seem kind of high...48C idle on cores 0 and 1, 38C idle on cores 2 and 3 - on water? Asus Probe shows 34C.

1.356V Vcore right now.
 
prasvt said:
The TJMax shows 100C in CoreTemp 0.99.3. RealTemp is showing the same temps, which seem kind of high...48C idle on cores 0 and 1, 38C idle on cores 2 and 3 - on water? Asus Probe shows 34C.

1.356V Vcore right now.
Click to expand...

That does sound weird to me. I would assume a 45nm chip would run a little cooler than a 65, and my q6600 w/ 1.432v through it show's idle's like 10deg. lower..........on air. :confused:

Maybe your boards giving you innacurate readings? I dunno man, but that definately seems strange to me.
 
Very strange. I haven't had a bsod or anything thus far, but the temp reporting is seriously sketchy. I dropped down to 4.1ghz @ 1.3375V ... right now in OCCT the temps are 70 70 61 61. Ehh I dunno. I may just shut the computer down for an hour and let the TIM cure a little. I'll pick this up tomorrow..thanks for the help today guys.
 
You must log in or register to reply here.
Back
Top