Question How to overclock the Intel Q9550 to 3.6 GHz and higher ?

[LuKaWin10]

Hi, I need help on how to overclock my Q9550 to 3.6 GHz and Higher. My specs:

Motherboard: P35-DS3 rev 2.1 (Version F14 with injected SLIC 2.1 and Xeon microcodes)

CPU: Intel Core2 Quad Q9550 @3.55 (stable, but after 3.6 it freezes)

CPU cooler: AIGO ICE200PRO 95W TDP cooler

RAM: G.SKILL 4 GB 800 MHz CL5, 5-5-5-15, 1.8~1.9V (Speed is at 848) (and I will be getting 4 more GB soon)

GPU: Asus GT 730 2 GB GDDR5 Kepler

I have no Case fans, so yeah. This is my current config that I have now.

https://imgur.com/a/ARofM2Z

View: https://imgur.com/a/ARofM2Z


I hope someone can help me! Please! Thanks.

 

[SkyNetRising]

I need help on how to overclock my Q9550 to 3.6 GHz and Higher.

It's 16 years old system.
Why do you want to overclock it to permanent failure/death?

 

[LuKaWin10]

It's 16 years old system.
Why do you want to overclock it to permanent failure/death?

I just wanna test some stuff out. I am pretty curious about overclocking since I have nothing else to OC

 

[SkyNetRising]

Do you have a backup system, if you kill this one?

 

[LuKaWin10]

Do you have a backup system, if you kill this one?

Nope. Only an i5 laptop

 

[Misgar]

If you're not worried about electromigration and have a capable cooling solution, increase CPU Vcore gradually. I suggest steps of 10 to 20mV, then bump up the CPU clock rate by whatever method suits (combinations of multiplier and FSB) and run a stress test. If it crashes and the CPU isn't getting too hot during stress tests, increase the voltage. You might benefit from a slight NorthBridge voltage boost.

From the manual:-

CPU Voltage Control
Allows you to set the CPU voltage. Normal sets the CPU voltage as required. The adjustable
range is dependent on the CPU being installed. (Default: Normal)
Note: Increasing CPU voltage may result in damage to your CPU or reduce the useful life of the
CPU.

Someone on this forum got 3.6GHz at 1.325V. I generally limit my overclocks to 1.250V for similar processors. I wouldn't risk anything over 1.300V, but you might not kill the CPU until you hit it with 1.40V+. Don't blame me if it goes up in smoke. There are loads of guides available.
https://www.overclockers.com/forums/threads/core-2-quad-q9550-overclocking.695297/
https://www.scan.co.uk/images/shops/intel/Intel_Q9550_Core_2_Quad_basic_overclocking_guide.pdf

Remember, more volts = more heat = reduced life.

I'm not going to risk messing around with my Q9550 because it's installed in one of my old TrueNAS systems. It probably has a very mild overclock, but nothing fancy.

 

[LuKaWin10]

If you're not worried about electromigration and have a capable cooling solution, increase CPU Vcore gradually. I suggest steps of 10 to 20mV, then bump up the CPU clock rate by whatever method suits (combinations of multiplier and FSB) and run a stress test. If it crashes and the CPU isn't getting too hot during stress tests, increase the voltage. You might benefit from a slight NorthBridge voltage boost.

From the manual:-

CPU Voltage Control
Allows you to set the CPU voltage. Normal sets the CPU voltage as required. The adjustable
range is dependent on the CPU being installed. (Default: Normal)
Note: Increasing CPU voltage may result in damage to your CPU or reduce the useful life of the
CPU.

Someone on this forum got 3.6GHz at 1.325V. I generally limit my overclocks to 1.250V for similar processors. I wouldn't risk anything over 1.300V, but you might not kill the CPU until you hit it with 1.40V+. Don't blame me if it goes up in smoke. There are loads of guides available.
https://www.overclockers.com/forums/threads/core-2-quad-q9550-overclocking.695297/
https://www.scan.co.uk/images/shops/intel/Intel_Q9550_Core_2_Quad_basic_overclocking_guide.pdf

Remember, more volts = more heat = reduced life.

I'm not going to risk messing around with my Q9550 because it's installed in one of my old TrueNAS systems. It probably has a very mild overclock, but nothing fancy.

Okay, thanks! I overclocked my CPU to 3.6 GHz and the Vcore is around 1.39125V I think but it is at 1.39 smth like that!

 

[Misgar]

the Vcore is around 1.39125V

Has the magic smoke started to come out yet? You'll probably shorten the 9550's life if you keep it near 1.4V. Good luck.

 

[LuKaWin10]

Has the magic smoke started to come out yet? You'll probably shorten the 9550's life if you keep it near 1.4V. Good luck.

Nothing yet lol. No magic smoke. I may have just gotten a bad chip. Since the normal vcore of my q9550 is 1.2875V, an e0 stepping it is. I now went back to my E6750 to try overclock it for fun!

 

[LuKaWin10]

Nothing yet lol. No magic smoke. I may have just gotten a bad chip. Since the normal vcore of my q9550 is 1.2875V, an e0 stepping it is. I now went back to my E6750 to try overclock it for fun!

Update: BSOD🙏. Going back to the Q9550, keeping it 3.6 GHz and calling it a day. When I get a Xeon E5450 OCing will be easier (hopefully)

 

[BFG-9000]

my E6750 the Q9550, When I get a Xeon E5450

What's the fascination with low-multiplier Core 2? The quad-cores have severe FSB limits because they are multiple dual-core chips on the same bus, so it only makes sense to get the highest multipliers available, such as X5460 or X5470 which would give you 3.8GHz or 4GHz, respectively at the highest stock FSB of 400.

The dual-cores can work at very high FSB, but you have intentionally restricted yourself from being able to run those at all by only getting DDR2-800. You don't even know right now if that 424MHz FSB is a limit of overclocking the RAM or of your Q9550 chip. Or even your 95w TDP cooler, as that Q9550 uses approximately 125w at 3.6GHz.

The maximum safe voltage for 45nm Core 2 is 1.45v, and for 65nm it's 1.5625v. With enough voltage, plenty of people can get way over 500MHz FSB and 4GHz on the 65nm E6750 but that's more wattage on a dual-core than your quad currently uses. You just don't have enough cooler, or the correct RAM to run those kinds of speeds.

You were warned before that desktop Intel chipsets for Core 2 have no multiplier below 2x for the RAM to FSB. If you had wanted to run high FSB, then DDR2-1066 would have allowed you to run the RAM at its rated speed clear up to 533FSB.

 

[Syntaximus]

I would not push for much of an OC on that board.
It was not intended for it - only 4-pin CPU power, no VRM cooling etc.

If you want to see how far that Q9550 can go for fun try and find an old P45 board.
The Gigabyte EP45-UD3P (and UD3R) were among the best for pushing high FSB.

 

[LuKaWin10]

I would not push for much of an OC on that board.
It was not intended for it - only 4-pin CPU power, no VRM cooling etc.

If you want to see how far that Q9550 can go for fun try and find an old P45 board.
The Gigabyte EP45-UD3P (and UD3R) were among the best for pushing high FSB.

Good, thanks!

 

[LuKaWin10]

What's the fascination with low-multiplier Core 2? The quad-cores have severe FSB limits because they are multiple dual-core chips on the same bus, so it only makes sense to get the highest multipliers available, such as X5460 or X5470 which would give you 3.8GHz or 4GHz, respectively at the highest stock FSB of 400.

The dual-cores can work at very high FSB, but you have intentionally restricted yourself from being able to run those at all by only getting DDR2-800. You don't even know right now if that 424MHz FSB is a limit of overclocking the RAM or of your Q9550 chip. Or even your 95w TDP cooler, as that Q9550 uses approximately 125w at 3.6GHz.

The maximum safe voltage for 45nm Core 2 is 1.45v, and for 65nm it's 1.5625v. With enough voltage, plenty of people can get way over 500MHz FSB and 4GHz on the 65nm E6750 but that's more wattage on a dual-core than your quad currently uses. You just don't have enough cooler, or the correct RAM to run those kinds of speeds.

You were warned before that desktop Intel chipsets for Core 2 have no multiplier below 2x for the RAM to FSB. If you had wanted to run high FSB, then DDR2-1066 would have allowed you to run the RAM at its rated speed clear up to 533FSB.

You're completely right. I did give myself big restrictions unfortunately but it is okay, I will keep it at 3.6 GHz and call it a day, it is enough for me anyways! And thanks for telling me this, very useful in the future

 

[LuKaWin10]

If you're not worried about electromigration and have a capable cooling solution, increase CPU Vcore gradually. I suggest steps of 10 to 20mV, then bump up the CPU clock rate by whatever method suits (combinations of multiplier and FSB) and run a stress test. If it crashes and the CPU isn't getting too hot during stress tests, increase the voltage. You might benefit from a slight NorthBridge voltage boost.

From the manual:-

CPU Voltage Control
Allows you to set the CPU voltage. Normal sets the CPU voltage as required. The adjustable
range is dependent on the CPU being installed. (Default: Normal)
Note: Increasing CPU voltage may result in damage to your CPU or reduce the useful life of the
CPU.

Someone on this forum got 3.6GHz at 1.325V. I generally limit my overclocks to 1.250V for similar processors. I wouldn't risk anything over 1.300V, but you might not kill the CPU until you hit it with 1.40V+. Don't blame me if it goes up in smoke. There are loads of guides available.
https://www.overclockers.com/forums/threads/core-2-quad-q9550-overclocking.695297/
https://www.scan.co.uk/images/shops/intel/Intel_Q9550_Core_2_Quad_basic_overclocking_guide.pdf

Remember, more volts = more heat = reduced life.

I'm not going to risk messing around with my Q9550 because it's installed in one of my old TrueNAS systems. It probably has a very mild overclock, but nothing fancy.

I actually overclocked it. I undervolted it a lot. 1.325V at 3.65 GHz. I am gonna try 3.8. My cooler is the LC-Power LC-CC-120(180W TDP, so good headroom), got it new, and I added case fans, i even modded my case a bit. I will make cutouts for more fans maybe in the future. Max degrees approx for CPU is around 66°C. I did this formula:

Core1+Core2+Core3+Core4/4

Core1 referring to the temps. Just for avg

 

[USAFRet]

I actually overclocked it. I undervolted it a lot. 1.325V at 3.65 GHz. I am gonna try 3.8. My cooler is the LC-Power LC-CC-120(180W TDP, so good headroom), got it new, and I added case fans, i even modded my case a bit. I will make cutouts for more fans maybe in the future. Max degrees approx for CPU is around 66°C. I did this formula:

Core1+Core2+Core3+Core4/4

Core1 referring to the temps. Just for avg

OK, you OC'd it.

Now, what actual performance benefits did you see?
Apart from simple benchmark numbers...

 

[LuKaWin10]

OK, you OC'd it.

Now, what actual performance benefits did you see?
Apart from simple benchmark numbers...

Just a few more fps higher. The whole point of this is I just wanted to see how far I can push a chip that is almost 17 years old 😎. And this PC is super old and I don't really need it much anymore, it's already made from S**T and STICKS. Also I lapped the Q9550, temps aren't bad tbh I could've done a better job but whoopsies. Next goal is Xeon E5450. Microcodes are ready, cooler too, gonna get mini heatsinks which are cheap AF then try 4 GHz on air, i might sound crazy, but 99% of gamblers quit before they win, and I am taking that vow to the next level and I am gambling with old hardware

 

[kgt1182]

I would say that even 1.5V as a daily voltage for the Q9550 to reach 4.1-4.2 GHz is fine as long as you keep it under 75C. At higher voltages the safe temperature decreases since electromigration speeds up with temeprature which may start to noticably shorten chip lifespan - the chip requires more voltage to maintain stability.

I would say that 1.55V is the maximum for 65nm Core chips long term use with any cooling. If yours is the newer 45nm model then the limits should be scaled down by 0.1V or 1.4-1.45V maximum. Smaller nodes take less voltage because they are designed with lower default voltage, have thinner insulation.

Hence older 130nm chips run safely at 1.6V-1.7V while newer FinFETs are risky to run over 1.3V; FX SOI 32nm is an exception at 1.6-1.65V.

TL;DR - You will need to overvolt your CPU and need good cooling, and likely end up at 1.5V 4.1-4.2GHz.

 

[LuKaWin10]

I would say that even 1.5V as a daily voltage for the Q9550 to reach 4.1-4.2 GHz is fine as long as you keep it under 75C. At higher voltages the safe temperature decreases since electromigration speeds up with temeprature which may start to noticably shorten chip lifespan - the chip requires more voltage to maintain stability.

I would say that 1.55V is the maximum for 65nm Core chips long term use with any cooling. If yours is the newer 45nm model then the limits should be scaled down by 0.1V or 1.4-1.45V maximum. Smaller nodes take less voltage because they are designed with lower default voltage, have thinner insulation.

Hence older 130nm chips run safely at 1.6V-1.7V while newer FinFETs are risky to run over 1.3V; FX SOI 32nm is an exception at 1.6-1.65V.

TL;DR - You will need to overvolt your CPU and need good cooling, and likely end up at 1.5V 4.1-4.2GHz.

I think that my chip is just not good enough. 3.8 GHz and higher is just impossible. I tried everything. Literally every combination. I am gonna try getting a Xeon E5450, keep the side panel off then get results. Lapping won't rly help small differences only. And thanks for the clarification too.

And my revision is the E0 one!

 

[LuKaWin10]

Update: I'm just gonna buy an X5460 Xeon, try getting it to 4 GHz and call it a day