Heat Problem with Q6600

[snowmelt]

My Q6600 is burning HOT! I have the following setup:

Q6600 OC @ 3.3GHz - 8 x 412FSB
Asus P5E-VM HDMI (Bios 0605, latest from last summer)
HSF Zalman CNPS 8000 running full speed
Corsair 4G RAM (5,5,5,15) Running 1:1
Roswell Case with two fans at max speed
RaidMax 730W

I understand thermal compound can make a difference but for now all I have is white silicone compound which I applied thin layer to both CPU and HS. I can't get the core temperature to go down any more than what it shows. The system seems to run stable for now but with Summer approaching, I really got to get this down... Even when it's hot, I can touch the base of the HSF with my bare hand and it just feels warm not hot. So could it be possibly Tjunction (DTS) sensor misread? I can't imagine Core Temp and Speedfan all reading the same Core volts though even if that's the case.

Do you have any suggestions you can throw at this?

 

[irkjab]

Try lowering the core voltage if you can. My Q6600 is stable at 3.2ghz with 1.175v

 

[orangegator]

Get a better heatsink or lower the overclock. Overclocked q6600s put out a lot of heat.

 

[snowmelt]

Try lowering the core voltage if you can. My Q6600 is stable at 3.2ghz with 1.175v

What is rest of your settings? NB, SB, Memory, etc...

 

[snowmelt]

Get a better heatsink or lower the overclock. Overclocked q6600s put out a lot of heat.

I have seen others with stock HSF at this setting that runs a lot cooler than mine...

 

[irkjab]

All my other voltages are at the lowest/stock setting.

Edit: I was getting 70c temps when I had the voltages on auto settings. Figured out that the auto settings were overvolting everything by a fair bit.

 

[orangegator]

irkjab - You're using a much better heatsink and have a lower overclock than the op. That's why your temps are better.

 

[irkjab]

Yes, however, my point about the voltages is still valid. I saw a 15c drop in load temps when I switched from auto voltage settings to default.

 

[snowmelt]

Yes, however, my point about the voltages is still valid. I saw a 15c drop in load temps when I switched from auto voltage settings to default.

That was a great advice. Thank you. It's not perfect but I was able to lower it to 68C by lowering the Vcc volt to 1.225V. So far I have been running Prime for 30 minutes with no problem but we will see. I had to set some settings to a little higher volt cause it wouldn't boot into Windows otherwise.

NB 1.38V
SB 1.2V
DRAM 1.8V
CPU Spread Spectrum Disabled
CPU Volt Reference Auto
CPU Volt Damper Auto
PLL Volt 1.5V (lowest)
FSB Termination Volt 1.35V (lowest was 1.2V but wouldn't work)

I tried Vcc 1.2V but BSOD after 30 seconds of Prime. Increased to 1.225V and it seems stable so far. If you can suggest something else I can try to lower the temp, please do. BTW, I did pencil mod on this MB.

 

[snowmelt]

Perhaps now I just need to get a better HSF...

 

[irkjab]

That would help to lower temps, yes. You could try reapplying the thermal paste.
By the way the NB voltage seems a bit high. Will it not boot on lower volts?

 

[snowmelt]

That would help to lower temps, yes. You could try reapplying the thermal paste.
By the way the NB voltage seems a bit high. Will it not boot on lower volts?

Ok I did just now lower NB to 1.27V and FSB Termination Volt to 1.3V. It did boot into Windows and now under torture test. But I am still confused about this Core temperature and CPU temperature difference. My understanding is there should only be a few degree difference between those but as you can see in SpeedFan, the difference is 17 degree. Curiously, Everest shows the difference of only 7 degree because of the lower core voltage it reports. This is why I tend to believe Everest is more accurate in reporting core volts, but then why would Core Temp and SpeedFan both show the same core temps, different than Everest?

I hope you can follow my reasoning here...

 

[irkjab]

Core temp does not show the core voltage, it shows the VID, which I believe is the default voltage for your particular chip. This varies between processors, my Q6600's VID is 1.2125v for example. Speedfan and Everest both report the same core voltage.
Since Speedfan, Real Temp, and Core Temp all give you the same temperature readings, I would ignore Everest's.

 

[CompuTronix]

snowmelt,

Uncalibrated temperatures are rarely accurate. During Prime95 Small FFT's, mean Core temperature should be 5c higher than CPU temperature, when calibrated, and is documented in the following Intel paper - http://arxiv.org/ftp/arxiv/papers/0709/0709.1861.pdf

If you'd like to learn how to calibrate your temperatures, then check out the Core i7 and Core 2 Temperature Guide - http://www.tomshardware.com/forum/221745-29-core-core-temperature-guide

Comp

 

[snowmelt]

snowmelt,

Uncalibrated temperatures are rarely accurate. During Prime95 Small FFT's, mean Core temperature should be 5c higher than CPU temperature, when calibrated, and is documented in the following Intel paper - http://arxiv.org/ftp/arxiv/papers/0709/0709.1861.pdf

If you'd like to learn how to calibrate your temperatures, then check out the Core i7 and Core 2 Temperature Guide - http://www.tomshardware.com/forum/221745-29-core-core-temperature-guide

Comp

I feel honored the actual GURU of that guide is responding

Are you saying in a nut shell that my core temp readings from Core Temp, SpeedFan and Real Temp are wrong since they are not calibrated? If so, do you think that they are quite a bit off or just a few degrees? I guess I am gasping for some answer first before I delve into this...

 

[CompuTronix]

... I am still confused about this Core temperature and CPU temperature difference. My understanding is there should only be a few degree difference between those but as you can see in SpeedFan, the difference is 17 degree...

The mean Core temperature is probably within 3c of being accurate, however, your CPU temperature is definately way off the mark.

 

[snowmelt]

CompuTronix,


Here is my readings from ASUS AI Suite which shows the same CPU temp of 54 degree which is the same as the one shown in SpeedFan. Does it mean then if caliburated correctly, the core temperature(s) should be around 59 degree? Am I to trust that the CPU reading from ASUS AI Suite is accurate since it is the maker of the motherboard?

 

[CompuTronix]

No. It means that your CPU temperature is too low, as shown in both the Asus utility and SpeedFan. Since your Zalman 8000 ranks as a low mid-range cooler, then at 22c ambient, your Q6600 G0 should Idle at 27c CPU temperature and 31c mean Core temperature, but only under the following conditions:

Computer Case Covers = Removed
Computer Case Fans = Manual 100% RPM
CPU Fan = Manual 100% RPM
CPU Frequency = Auto (See Note 1: below) **
CPU Internal Thermal Control = Enabled
Enhanced C1 Control (C1E) = Enabled
Internet = Disconnected
Memory Frequency = Auto
PECI (If Equipped) = Enabled
Speedstep (EIST) = Enabled
Vcore = Auto (See Note 1: below) **
Vdimm = Auto
Windows Programs = Closed

** Note 1: If BIOS does not respond properly to Auto Vcore, Frequency and Multiplier settings, then use an appropriate combination of manual settings to provide the following in CPU-Z:

Core Voltage = 1.100 V
Core Speed = 1600 Mhz (Core i7 processors)
Core Speed = 1600 Mhz (Core 2 - 65 nm processors)
Core Speed = 2000 Mhz (Core 2 - 45 nm processors)

(A) Measure Ambient near the computer case air intake, clear of warm exhaust. A trusted indoor analog or digital thermometer will suffice. The accuracy of this device and measurement will determine the overall accuracy of the Calibrations.

(B) Boot into Windows. Close all programs, background processes, Screen Savers, SETI, Folding and Tray software. Press Ctrl-Alt-Delete, click on Task Manager, then click on the Performance tab to confirm CPU Usage is less than 1%. Use the Applications and Processes tabs to close programs if necessary.

(C) Open CPU-Z and SpeedFan. Observe CPU-Z for Intel's Speedstep to decrease Core Voltage, Core Speed and Multiplier to minimum values. Observe SpeedFan, allow 10 minute at Idle to ensure that temperatures decrease to minimums, then record Tcase Idle.

These test settings assure that differences in hardware configuations and cooling variables are eliminated, so that apples can be accurately compared to apples.

Comp

 

[daship]

I just built a q6600 with a zigmatek s1283 and oc to 3.6 with 1.38 v-core 32 idle 72 load.

 

[CompuTronix]

daship,

With Prime95 Small FFT's? To which 72c load temperature are you referring? CPU or Core? There's a difference.

Since temperatures and overclocking are all about Intel's specifications, if we're not specific, then temperatures are reduced to nothing more than apples and oranges thermal fruit salad in a blender.

Nevertheless, nice OC. You probably have a VID of about 1.2625.

Comp

 

[snowmelt]

ComputTronix,

Thanks for the detailed explanation. Looks like I have some readings to do. Either way it is bit daunting to go from where I am to that low core temp.



daship,

What are the other settings?

 

[daship]

Whatever real temp goes by, and 2 get to 72 while the other to hover between 68-69.

 

[daship]

Auto except my DDR3 ram is 1.66.

 

[CompuTronix]

Real Temp, SpeedFan, Core Temp and Everest all use 100c Tjunction Max for the Q6600 G0, however, that doesn't gaurantee completely accurate Core temperatures. Tjunction Max is actually between 95c to 100c, and most frequently is 97c to 98c.

The calibration procedures in my Guide will get you closer without using any Tjunction Max value, and closer than using Intel's canned default value of 100c. This is why Tcase Max, which is the Thermal Specification shown in Intel's Processor Spec Finder - http://processorfinder.intel.com/Default.aspx - is shown to the tenth of a degree, while Tjunction Max is a round figure.

 

[snowmelt]

Just a quick status report.

I am running stable at 3.4Ghz (8x425) with load core temp at 72C. Still hot but it's manageable for now.

- CPU Vcc Volt 1.3V
- NB 1.38V
- SB Auto
- DRAM Auto Volt running 1:1 (850 mhz)
- All BIOS Energy saving controls disabled

My 3dmark Vantage score looks like this. Does any of the numbers look little too low to you? I am really not familiar with this score thingy. I saw someone with AMD Phenom 940 (OC 3.8GHZ) and GeForce 8800 GTX getting 38000 3DMarks03 score. I thought that was phenomenal but it is 03 version of 3Dmark. I am not sure if it would be the same with 3DMark Vantage.

My specs: Q6600 OC 3.4GHZ, 4G Kingston DDR2-1066, ATI HD 4850 1GIG GDDR3.....