Xtreme Low Q6600 G0 Temp on Air !!! (Lots of Pics)

[treatmentx]

XtremeTiramisu - did you lap your G0? Did you lap anything at all? What were your final voltages in BIOS? My G0 comes tomorrow, prepping up prior to.

[unclewebb]

graysky: Intel has never documented absolute temperatures for any of the desktop processors.

Anyone interested in an absolute temperature can use CoreTemp 0.95 and will get a realistic looking load temperature number but it is not documented and may or may not be 100% accurate.

What is fully documented by Intel is the DTS. It provides accurate temperature information that is relative to the thermal throttling point for any Core based processor. If you use CoreTemp 0.95 and set it to "Show Delta to Tjunction", you will be able to read the DTS in real time. If you do this, it doesn't matter what the real TjMax is and it doesn't matter what changes Intel made to the latest G0 steppings. The information returned by the DTS did not change.

Post a few links to any questionable G0 temperature readings and I'll have a look at them and try to come up with a reasonable explanation.

[XtremeTiramisu]

XtremeTiramisu - did you lap your G0? Did you lap anything at all? What were your final voltages in BIOS? My G0 comes tomorrow, prepping up prior to.

If you've seen the first post I made in the first page (All links have being fixed), you can see that my Thermalright Ultra120 Extreme is nicely lapped. The Q6600 G0's IHS isn't not lapped and it's not necessary to lap a G0 proc. The stock IHS is already preety flat and the Q6600 G0 runs very cool even on air cooled.
I used AC MX-1 thermal paste and I followed "AS AS5" official application method to apply the TIM for quad cores.

Thermal Compound Break-in Priod: it will take a up to 200 hours and several thermal cycles to achieve maximum particle to particle thermal conduction and for the heatsink to CPU interface to reach maximum conductivity. (This period will be longer in a system without a fan on the heatsink or with a low speed fan on the heatsink.) On systems measuring actual internal core temperatures via the CPU's internal diode, the measured temperature will often drop 2C to 5C over this "break-in" period. This break-in will occur during the normal use of the computer as long as the computer is turned off from time to time and the interface is allowed to cool to room temperature. Once the break-in is complete, the computer can be left on if desired.

graysky: Intel has never documented absolute temperatures for any of the desktop processors.

Anyone interested in an absolute temperature can use CoreTemp 0.95 and will get a realistic looking load temperature number but it is not documented and may or may not be 100% accurate.

What is fully documented by Intel is the DTS. It provides accurate temperature information that is relative to the thermal throttling point for any Core based processor. If you use CoreTemp 0.95 and set it to "Show Delta to Tjunction", you will be able to read the DTS in real time. If you do this, it doesn't matter what the real TjMax is and it doesn't matter what changes Intel made to the latest G0 steppings. The information returned by the DTS did not change.

Post a few links to any questionable G0 temperature readings and I'll have a look at them and try to come up with a reasonable explanation.

Thanks unclewebb for the clear and thorough explanation again...I hope people do get it this time

[synergy]

INCREDIBLE !! I can't believe I'm on air and having this unreal quad core thrill....am I dreaming?!

Nice overclocks / temperatures.

I notice you have a post about running BIOS 404.
Did your motherboard come with an earlier BIOS version
AND you were able to boot WITH THE Q6600 G0 and
REFLASH even though the G0 wasn't supported before 404?

[XtremeTiramisu]

FSB 470 * 8 @ 3760MHz 8hrs+ Small FFT Prime Stable.


Core temp are 67~70c so this is what I'll have as my max 24/7 OC for this season.
Ambient room temperature is 24.9 to 25.1C measured with a digital thermometre; system temp is where the hottest place in my room. Ambient room temp is always 1~2C below it.

I'm in a closed room so temp wise my room can get quite stuffed. Not much air circulation in my room but I've excellent ventilation/air flow inside my case

Again ambient temp in my room is measured from 24.9c ~ 25.1c.
It's very accurate becouse whenever people comes into my room, they felt it's uneasy to breath and hotter than other places in the house. When I brought the digital thermometre to my bathroom, hallway, livingroom, the temp drops considerable down to any where from 21c to 22c.

So, please don't question about the ambient temp in my room again.

[Bail_w]

your system temperature is 26 degree C, which is only 1 degree hotter than ambient temperature?

[XtremeTiramisu]

How to manually correct Speedfan 4.32 core temp accuracy

SpeedFan 4.32 assumes the maximum Tjunction temperature for the Q6600 is 85C and CoreTemp 0.95 assumes 100C.

Reported Temp = TjMax - Digital Thermal Sensor reading

A 15C difference in the assumed TjMax results in reported temperatures that are different by exactly 15C.

Motherboard bios temps are meaningless. Stick to either one of the above programs and you can set a 15C offset in SpeedFan to correct for the error in assumed TjMax by clicking on Configure -> Advanced:



You lucky Quad guys will have to enter in 4 offsets.

[XtremeTiramisu]

your system temperature is 26 degree C, which is only 1 degree hotter than ambient temperature?

Yes becouse I'm in a closed room so temp wise my room can get quite stuffed. Not much air circulation in my room but I've excellent ventilation/air flow inside my case

Again ambient temp in my room is measured from 24.9c ~ 25.1c.
It's very accurate becouse whenever people comes into my room, they felt it's uneasy to breath and hotter than other places in the house. When I brought the digital thermometre to my bathroom, hallway, livingroom, the temp drops considerable down to any where from 21c to 22c.

So please stop questioning about my ambient temp again.

[XtremeTiramisu]

Thermalright Ultra120 Extreme (Lapped) vs Ambient Room Temperature


Ambient temp in the room when screen was captured = ~22.7C
As you can see the Idle temp is roughly 13~15C above ambient temp in my room.
NOTICE: Given Vcore Idles at 1.4880v, even so the temp is still amazing.
So without a doubt, the Thermalright Ultra120 Extreme is the best air cooled heatsink avaliable.
(For similar results, "lapping" the heatsink base is strongly recommended )

[XtremeTiramisu]

Nice overclocks / temperatures.

I notice you have a post about running BIOS 404.
Did your motherboard come with an earlier BIOS version
AND you were able to boot WITH THE Q6600 G0 and
REFLASH even though the G0 wasn't supported before 404?

Good question:

My motherboard came with 0202 BIOS but I had my mobo for nearly a month now. If you noticed in my first post again, you can see that I had the E6700 OC'ed with the same mobo just before I swapped it out for the Q6600 G0. So the mobo was already flashed to 0404 BIOS before the Q6600 G0 came in.

My guess with the P5K-Dlx is that the P35 chipset should still recongnize the Q6600 G0 stepping regardless of the BIOS version. I'd worry if you've older chipset mobos such as the P965/975 or 680/650I SLI (without their newest BIOS flash).

I cannot guarantee if the Q6600 G0 will boot with older P5K-Dlx BIOS becouse obviously I didn't try it. If you can flash the mobo to the 0404 BIOS or newer before you put in the G0 that'd be great but if not just put it in anways and boot it up. Worst case senerio is the system will not post but you'll not kill it.

[fornowagain]

Thermalright Ultra120 Extreme (Lapped) vs Ambient Room Temperature

Ambient temp in the room when screen was captured = ~23.2C
As you can see the Idle temp is roughly 11~13C above ambient temp in my room.
So without a doubt, the Thermalright Ultra120 Extreme is the best air cooled heatsink avaliable.
(For similar results, "lapping" the heatsink base is strongly recommended )

You worked out the chip is using approximately 210W, right? Having a look around the 'extreme' with a good fan mounted correctly could have thermal resistance of up to 0.1 C/W. Assuming its mounted well, then a fairly accurate way of confirming real world temps is to work out the delta.

So the temperature delta for that HSF and cpu is 0.1 x 210 = 21C at load, 13C at idle seems reasonable. I assume its in a case, or is the room ambient the same as the air moving over the HSF, i.e. case temps?

Just add the case/room temp to the delta for Tc. So 21+23.2 = 43.2C The core temp is higher, opinions vary, maybe +10C. Gives an approximate junction temperature of 53C.

[[XC] Synthetickiller]

Thanks for posting this EPIC OC thread on the q6600. This gives me high hopes that my OC will be possibly 4.0ghz on water if i ever get my cpu from tankguys!

Thanks. Now that you've got 3.8ghz stable, lets go for 3.9! It seems doable.

[AndyM]

Try Prime blend (or whatever gives large fft's)

[XtremeTiramisu]

You worked out the chip is using approximately 210W, right? Having a look around the 'extreme' with a good fan mounted correctly could have thermal resistance of up to 0.1 C/W. Assuming its mounted well, then a fairly accurate way of confirming real world temps is to work out the delta.

So the temperature delta for that HSF and cpu is 0.1 x 210 = 21C at load, 13C at idle seems reasonable. I assume its in a case, or is the room ambient the same as the air moving over the HSF, i.e. case temps?

Just add the case/room temp to the delta for Tc. So 21+23.2 = 43.2C The core temp is higher, opinions vary, maybe +10C. Gives an approximate junction temperature of 53C.

+1 for another informative post.
I've x2 Scythe S-Flex 1600rpm in Push-Pull configuration.
Thanks for working out the detail calculations for me. It seems this area is your experty

Thanks for posting this EPIC OC thread on the q6600. This gives me high hopes that my OC will be possibly 4.0ghz on water if i ever get my cpu from tankguys!

Thanks. Now that you've got 3.8ghz stable, lets go for 3.9! It seems doable.

I can acheive 3.8GHz prime stable on air for x amount of time and relative given voltage at low 1.5v but temp wise would get a little higher than my liking at ~73 to 74c consistent max full load all 4 cores for hours. So I eventually backed down to 3.76GHz and lowered the Vcore a bit. Stable as a rock now. Temp wise, it peaks at 70c so it's still within my limit but it can go beyond that as you can see from my previous screen shots above.

As for 3.9GHz Prime stable goal? I guess I'll leave this to you watercoolers.
I'm already topping out the air cooled Q6600 chart with 8+ hours Prime stable screen shot. Clock speed : Voltage : Temp : Air cooling = I think I'm currently ranked #1.

[XtremeTiramisu]

Try Prime blend (or whatever gives large fft's)

Indeed, my next goal and I'm quite confident

[Icer]

Of course orion23
There's no point to get shy and keep them to myself right

I'm also curious about your mobo settings






JumperFree Configuration Settings
AI Overclocking: Manual
CPU Ratio Control: Manual
- Ratio CMOS Setting:
FSB Frequency:
PCI-E Frequency:
DRAM Frequency:
DRAM Timing Control:
CAS# Latency:
RAS# to CAS# Delay:
RAS# Precharge:
RAS# Activate to Precharge:
TWR:
TRFC:
TWTR:
TRRD:
TRTP:
DRAM Static Read Control:

Transaction Booster:
Clock Over-Charging Mode:

CPU Spread Spectrum:
PCIE Spread Spectrum:

CPU Voltage:
CPU Voltage Reference:
CPU Voltage Damper:
CPU PLL Voltage:
DRAM Voltage:
FSB Termination Voltage:
North Bridge Voltage:
North Bridge Voltage Reference:
South Bridge Voltage:

Advance CPU Settings
CPU Ratio Control:
Ratio CMOS Setting:
C1E Suppport:
Max CPUID Value Limit:
Vanderpool Technology:
CPU TM Function:
Execute Disable Bit:
PECI:

[XtremeTiramisu]

Like I've promised Orion, I'll post my detail P5K-Deluxe BIOS configuration to serve as a reference for all future quad core OC'ers

I won't have time and results by the end of today.
Meanwhile just sit tight and I should have it everything confirmed and ready by around 11pm tonight

[michaLcoughliN]

Tiramisu. Your quad is absoltely remarkable. Congrats once again

[sdumper]

Is this the norm for these Q6600 G0

[XtremeTiramisu]

Tiramisu. Your quad is absoltely remarkable. Congrats once again

Thanks
GL in OC your E6850 too

Is this the norm for these Q6600 G0

My case is definetely something out of the "norm"

It's really a combination of various factors including the choice of your hardwares and your luck in getting a really good batch.

So far from the statistic of all Q6600 G0 I've seen, the norm would fall some where btw 3.5GHz ~ 3.6GHz with reasonable amount of volts.

Nevertheless, the G0 Q6600 has certainly raise the bar for what we used to thought a Q6600 could OC. Honestly, I really don't see a reason of buying a dual core when a Quad core can acheive the same or surpass the clock speed of most C2Ds

[jeffyjaixx]

Very nice.

You're using a 2 GB kit?

and what vcore in bios?

[Icer]

Is this the norm for these Q6600 G0

I think with the right mobo & cooler, 3.6 sounds like a good avg. Even though I'm pretty sure my mobo settings are correct, I can't get a stable 3.8GHz. I am currently using the Thermaltake V1 cooler and am not real happy with it. My Idle temps are good but when I put a Orthos load on it, the temps get too high. It's like the eveready bunny, they keep going and going I'd get the Thermalright Ultra-120 Extreme but it's too tall for my case... I'm going to buy the Thermalright Ultima-90 soon and see how it works.

[XtremeTiramisu]

Very nice.

You're using a 2 GB kit?

and what vcore in bios?

Yes 2GB Kit but will try out 4GB configuration of the same clock/voltage

In order to get 3760MHz 8hrs+ Prime SmallFFT & LargeFFT stable, the Vcore needed to be set to 1.5250v in BIOS; 1.4880v Idle and 1.4800v/1.4720v SmallFFT/LargeFFT Load respectively.

More details in the complete BIOS configuration will be updated shortly.
Currently Prime LargeFFT 7hrs+ and counting towards the 8hrs mark.
LargeFFT puts out the maximum heat, power consumption and stresses Rams.
But nevertheless, it's still smooth sailing @ 3760MHz on air

I think with the right mobo & cooler, 3.6 sounds like a good avg. Even though I'm pretty sure my mobo settings are correct, I can't get a stable 3.8GHz. I am currently using the Thermaltake V1 cooler and am not real happy with it. My Idle temps are good but when I put a Orthos load on it, the temps get too high. It's like the eveready bunny, they keep going and going I'd get the Thermalright Ultra-120 Extreme but it's too tall for my case... I'm going to buy the Thermalright Ultima-90 soon and see how it works.

G0 Q6600 runs very cool and you don't really need the best air cooled heatsink to acheive decent clock
It really lives up to the hype afterall

[atlr]

That's exactly what I found when playing around with my E6400. As the MHz went up I needed to leave more temperature head room to maintain long term, 100% Orthos stability.

Here are my approximate findings:

3200 MHz ~ DTS=5
3400 MHz ~ DTS=15
3600 MHz ~ DTS=25

Cool. Nice observations.

I think you can see this progression of lower temperatures needed for stable higher frequency operation in Intel's specs. The Thermal Specification for the highest frequency processor is lower than that of lower frequency CPUs.

A Q6600 G-0 has
Thermal Design Power: 95W
Thermal Specification: 71C

A QX6850 G-0 has
Thermal Design Power: 130W
Thermal Specification: 64.5C

I think a correct re-wording of these specs for a Q6600 G-0 is that a CPU cooler must dissipate thermal energy at a rate of 95W to keep Tcase temperatures under 71C to prevent throttling at 2.4GHz.

Likewise, the QX6850 G-0 needs a cooler to dissipate 130W to keep Tcase below 64.5C for 3GHz operation.

[Spanki]

Cool. Nice observations.

I think you can see this progression of lower temperatures needed for stable higher frequency operation in Intel's specs. The Thermal Specification for the highest frequency processor is lower than that of lower frequency CPUs.

A Q6600 G-0 has
Thermal Design Power: 95W
Thermal Specification: 71C

A QX6850 G-0 has
Thermal Design Power: 130W
Thermal Specification: 64.5C

I think a correct re-wording of these specs for a Q6600 G-0 is that a CPU cooler must dissipate thermal energy at a rate of 95W to keep Tcase temperatures under 71C to prevent throttling at 2.4GHz.

Likewise, the QX6850 G-0 needs a cooler to dissipate 130W to keep Tcase below 64.5C for 3GHz operation.

Just to add to your info above, they've changed the Xeon parts to some new specs recently...

A X3220 G-0 has (equiv to Q6600)
Thermal Design Power: 100W
Thermal Specification: 85C

A X3210 G-0 has (equiv to non-existent Q6400 @2.13GHz, 8x multiplier)
Thermal Design Power: 100W
Thermal Specification: 85C

...so it looks like the server parts have more headroom.