Retail Q9650 Overclocking Thread

[jeanjean15]

Hello .

I bought a Q9650 to know if i was able to overclock it at 4300 mhz with only 1.4 volt .

But i must say what i was very disappointed . First of all , i needed 1.36 volt to be stable at 4050 mhz and it was impossible to exceed 480 mhz with the FSB and even less be stable at such high fsb !

So my question is : where do you find your fabulous q9650 ??

Do you buy 10 processors to find a good one ?

[TooFastForYa]

Here is my Q9650 Batch # L836B361 w/ VID of 1.2250
As you can see I just bought this as an upgrade. I had an E6400 before. System specs are as follows:

Asus P5B Deluxe WiFi
G.Skill 2x2GB Pi Black DDR2 800 F2-6400CL4D-4GBPI-B
Zalman CNPS9700
Antec P180
Corsair 520HX

This board has a horrible vdroop. vCore is set to 1.450 in the Bios, reads 1.400 at idle and 1.320 under full load using IntelBurnTest. Overall, I think this is a decent overclock considering its on a P965 board. What do you guys think?

[Jor3lBR]

Hello .

I bought a Q9650 to know if i was able to overclock it at 4300 mhz with only 1.4 volt .

But i must say what i was very disappointed . First of all , i needed 1.36 volt to be stable at 4050 mhz and it was impossible to exceed 480 mhz with the FSB and even less be stable at such high fsb !

So my question is : where do you find your fabulous q9650 ??

Do you buy 10 processors to find a good one ?

No you just need a better board.

I just finished a system for a friend and it was Prime95 stable at 4500Mhz with 1.39vcore. Mine cpu also Prime95 above 4600Mhz.

Some need less then others but that wall of 480fsb was probably the board not the cpu.

[jeanjean15]

No you just need a better board.

I just finished a system for a friend and it was Prime95 stable at 4500Mhz with 1.39vcore. Mine cpu also Prime95 above 4600Mhz.

Some need less then others but that wall of 480fsb was probably the board not the cpu.

Actually , i own a Asus P5E3 premium X48 and i don't think what it is a bad board but perhaps you are right and a P45 board is better with quadcore ?

[Jor3lBR]

Actually , i own a Asus P5E3 premium X48 and i don't think what it is a bad board but perhaps you are right and a P45 board is better with quadcore ?

Same processors on my DFI X48 would hit 472-486 wall. Only explanation for great stable OC's: Gigabyte EP45 platform

[jeanjean15]

Same processors on my DFI X48 would hit 472-486 wall. Only explanation for great stable OC's: Gigabyte EP45 platform

Ok , thank you for your advice and congratulations for your records !

[deachus]

I bought my Q9650 from NewEgg two weeks ago. I own both the P5E3 Premium (x48) and P5Q3 Deluxe (P45) and my chip does 4Ghz p95 blend stable at very low volts on both boards (1.29v-1.31v) and the highest I have hit is 495FSB so far. I did not buy 10, only 1.

[jeanjean15]

I bought my Q9650 from NewEgg two weeks ago. I own both the P5E3 Premium (x48) and P5Q3 Deluxe (P45) and my chip does 4Ghz p95 blend stable at very low volts on both boards (1.29v-1.31v) and the highest I have hit is 495FSB so far. I did not buy 10, only 1.

I bought mine in France this week and i hoped it would have been rockstable at 4.3-4.5 ghz with 1.45 volt max . But i have got no chance : only 4.05 ghz with 1.36 volt : not enough for me to change my qx9650 !

So , i returned it to my supplier .

[Osterman]

Mine


Its under testing!

[bito]

I bought mine in France this week and i hoped it would have been rockstable at 4.3-4.5 ghz with 1.45 volt max . But i have got no chance : only 4.05 ghz with 1.36 volt : not enough for me to change my qx9650 !

So , i returned it to my supplier .

Finished playing with mine and its up for grabs....VID is 1.225v

[neo_rtr]

Does any one having the following batch L825B425 DATE 9/17/08 899134 SLB8W
i just need to compare before i buy it. tks

[greg.m]

Does any one having the following batch L825B425 DATE 9/17/08 899134 SLB8W
i just need to compare before i buy it. tks

Here is 2 more batches plus neo_rtrs:

L825B425
L3844A356
L844B802

Guys, does someone of you have or know something about any of those 3 batches?
We are about to take a serious decision here

[deachus]

Here is 2 more batches plus neo_rtrs:

L825B425
L3844A356
L844B802

Guys, does someone of you have or know something about any of those 3 batches?
We are about to take a serious decision here

Here's mine:

SLB8W
L844B781
VID: 1.235v

My L844B781 runs 4Ghz+ low volts and clocks so far up to 495FSB, still haven't hit a wall yet. I would say go for the L844B781 which is pretty close to mine. I have read that A chips are not as good as the B ones. Just my 2 cents.

[greg.m]

Here's mine:

SLB8W
L844B781
VID: 1.235v

My L844B781 runs 4Ghz+ low volts and clocks so far up to 495FSB, still haven't hit a wall yet. I would say go for the L844B781 which is pretty close to mine. I have read that A chips are not as good as the B ones. Just my 2 cents.

I saw your temps under prime @ 4230mhz!!! Are they with your TRUE or with a wc system

[deachus]

I saw your temps under prime @ 4230mhz!!! Are they with your TRUE or with a wc system

There with my lapped true. Q9650 is not lapped yet. I have (1) 1200MM fan running half speed on that p95 run. I just installed a Delta last night running 5v and the temps are even better and perfectly quite.

[pascalbrown]

I've got a problem that I just can't solve and I've spent MANY (far too many) hours trying to fix.

When I got my Q9650 I got it fully prime stable at 470*9 without any difficulty. It passed over 12 hours of priming and I was happy with it. I also managed to get it to pass the 12 hours of OCCT at 475*9. A few weeks ago I had a problem whilst video editing and I was convinced it was due to an upgrade I'd just done, so I made a bunch of complaints. Boy did I feel like a fool when I fired up prime95 and one of my cores failed within a few minutes. Turns out the stability (or rather, the lack of) was causing the problem.

Since then i've been trying to get it stable again at 470*9. I'm using a p5q-e mobo and I've tried a whole variety of BIOS's, included Ket's modded ones. I've played with every settings in the bios and I can not for the life of me get this rig stable again at 470*9. I did manage to get it stable at 467*9 for an 8 hour prime95 run, but anything more and it konks out on me. This is where it gets more interesting. It always, always, always fails on core 2 (assuming you assign them core 0/1/2/3). I've tried every VTT value from 1.3v to 1.46v and tried a range of GTL ref's for each value. None of them seem to get me stable. SmallFFT will always fail within 10 minutes. I don't understand what is going on.

What should I try next? What have I missed? Could my chip be dying? Could I simply have hit the exact correct settings by fluke when I first got the chip? For reference I've been using CPU voltage of 1.425 with LLC enabled. This gives a load voltage of about 1.43v which some might see as high, but I personally think it's fine. I've tried NB values from 1.3v up to 1.5v but that doesn't seem to help.

Why does core 2 always fail? What should I tweak to get this core back in line and playing happily? The chip isn't a duffer as it's stable enough to load windows and be used at 500*9 without any more vcore. I just can't get this one core to play ball. I've also tried the CPU clock skew and the NB clock skew, going through the possible options up to 200 (ie. 0/0,100/0,200/0, 0/100,1/200, 100/100, 200/200 etc), but nothing is helping. I think this chip can give more but stability is killing me at this point. I'm literally spending every waking minute rebooting, changing one value, and then rebooting to run prime95. Please help me to find some stability both for my PC and for myself!

ps. I've even thought about buying another Q9650! That's how frustrated I've become!

[Hoss331]

That might be the max stable for that chip with the 9x multi. Mine acts similar to that but at the mid 480s with the 9x multi. It will go higher FSB with the 8.5x but I see no point in running these chips with the 8.5x. I have a P5Q dlx and the UD3P and it doesnt matter what board its on, how much voltage I give it, or what settings I adjust it will not pass small fft prime. It always fails within the first 4 runs on small fft prime at the mid 480s but will pass OCCT in to the mid 490s with the 9x. Im guessing OCCT is just not as stressful as prime. The chip doesnt need much voltage to run what it does so I was kinda pissed at first when I got to this wall. Ive pretty much given up for anything higher thats prime stable unless someone has some suggestions.

[vws]

Reseating heatsink and reapplying the thermal paste may help. I had a similar situation with a Q6600 back in the days and turned out there was a bubble amid the MX-2. No, you can't find it by simply checking core temps; you have to open it up and check physically. Due to the MX-2's watery nature I changed back to my good'ol thicky Ceramique. Since then I no longer suffered the clock degradation again.

[Martin.v.r]

4.410GHZ vcore 1.376v L844B809 VID: 1.250
http://valid.canardpc.com/show_oc.php?id=517196

ned help whit GLT to the cpu

[Grnfinger]

Late to the game here
Just pick'd up a Q9650 FPO L828A446

4.2GHz was a walk in the park, but my concern is the temps.
3 cores are nice and cool 29c-31c but the 1 core is a staggering 40c, this is water cooled and yes I have checked the mount and reseated twice.
Anyone else have this issue or am I the lucky winner today?
I'm in no hurry to lap the the chip for I might sell it if I cant do 500x9

[mreld3r]

Late to the game here
Just pick'd up a Q9650 FPO L828A446

4.2GHz was a walk in the park, but my concern is the temps.
3 cores are nice and cool 29c-31c but the 1 core is a staggering 40c, this is water cooled and yes I have checked the mount and reseated twice.
Anyone else have this issue or am I the lucky winner today?
I'm in no hurry to lap the the chip for I might sell it if I cant do 500x9

I have the same thing on my 9650. One of my sensors is always a LOT higher than the rest. I attribute it to sensor error. I have a hard time believing that one core on the same die could be 10c higher than the rest. I just ignore it.
These chips have a lotta messed up sensors, many reporting stuck sensors and stuph. As long as mine dont shut down on me due to heat; Im not really concerned too much. My sensors are great at idle, but as the temps go up one of them is about 8-10 degrees higher than the rest. When I run IntelBurnTest it gets kinda alarming (low quality air-cooling here). But; if i never even looked at my temps and I relied on the cpu-overtemp-protection I bet I could use this chip for years, and NEVER have a problem.
Just my 2-cents.

[Martin.v.r]

ASUS P5Q Deluxe Bios 1805 Q9650 L844B809 VID: 1.250
volt set to the fsb
watercooling
4.4GHZ 1.36v
4.5GHZ 1.52v xtream high or
4.6GHZ 1.56v

[Grnfinger]

I have the same thing on my 9650. One of my sensors is always a LOT higher than the rest. I attribute it to sensor error. I have a hard time believing that one core on the same die could be 10c higher than the rest. I just ignore it.
These chips have a lotta messed up sensors, many reporting stuck sensors and stuph. As long as mine dont shut down on me due to heat; Im not really concerned too much. My sensors are great at idle, but as the temps go up one of them is about 8-10 degrees higher than the rest. When I run IntelBurnTest it gets kinda alarming (low quality air-cooling here). But; if i never even looked at my temps and I relied on the cpu-overtemp-protection I bet I could use this chip for years, and NEVER have a problem.
Just my 2-cents.

Thanks for the reply,
I was a tad alarmed, my Q9550 is always cool and idle/load temps are very decent.
When I saw core1 on my Q9650 running 10c higher than the other 3 I was rather bummed but now not so much...

No tweaks, first run

[mreld3r]

As a matter of fact. my core 1 is also the high one. Other 3 are 'normal'.

[vws]

Folks, don't bother with any core temp readings below 50°C - they are simply not accurate. Quote from this recently released Intel datasheet regarding the thermal sensor issue:

The digital thermal sensor (DTS) accuracy is in the order of -5°C +10°C around 90°C ; it deteriorates to ±10°C at 50°C. The DTS temperature reading saturates at some temperature below 50°C. Any DTS reading below 50°C should be considered to indicate only a temperature below 50°C and not a specific temperature. External thermal sensor with “BJT” model is required to read thermal diode temperature.

From my own tests, I found it better be corrected to "60°C" to be closer to reality. Anything below 60°C in my experience doesn't make much sense at all.

Also here's an article at Tom's HW that's worth reading. It talks about how to set the Tj Max value accurately for your CPU.