As many of us know, trying to maintain the lowest possible temperature for our CPU's is something that we absess about! It consumes us when we hear of some of our fellow brethren claiming some ridiculous low temps with his Thermalright Ultra 120 Xtreme! So we begin our quest to figure out why the hell are my temps so high? Well, in a nutshell, it is ambient temperature that determines lowest possible idle temperature when using air cooling. So much of it has to do with ambient temperature.
This guide borrows quite a bit from Tom's Hardware Forum in this thread. So, I really take no credit, but rather am pointing to his work in a possibly easier to understand post. Essentially, my goal is to condense that version into something quick and easy. Because, to be honest, I am not sure many people have really read that thread in detail nor have they put it into practise. Because if people did, we wouldn't see people claiming idle temperatures lower than their ambients (I have seen a few people claim it) or claiming idle temps so low that their ambients would had to have been extremely low (Antarctica anyone?).
Here are the values in a nutshell that apply to the Q6600 B3/G0
Tcase = Geometric Center of IHS. Now, Tcase cannot be lower than your ambient case temperatures. Blowing 100F air over an object is not going to cool it below 100F, the best it can hope to do is cool it down to 100F. Thus the principle applies to Tcase. If your computer case temperatures are 35c, then your Tcase cannot be anything but 35c+. If your CPU & HSF are not lapped (reduced metal to metal contact) then you are realisticly looking at a minimum of 2c+ computer case temperature for Tcase.
Tjunction = Diode inside the center of each core themselves to find the 'true' temperature. Tjunction for Q6600 B3/G0s is anywhere from 7-13c+ Tcase. My particular Q6600 B3 sees 13c+ at the hottest core, and 9c+ at the coolest core.
Idle Versus Load = The temperature of the chip when you are doing nothing except sitting within windows at 0% utilization (difficult to get this at absolute zero for obvious reasons). The load temperature is usually measured by a program called P95/Orthos loading all four cores, or TAT (if it works). TAT is said to put a 114% workload on the CPU. Of course, I am not sure how you can get another 14% beyond 100%, but hey, those are the numbers claimed. Also, the delta between idle and load is typically 15c - 30c depending on your cooler/fan, environment and program used to stress the CPU.
Putting it all into a formula using room temperature in celcius as our starting point.
RT + 5 = CT
CT + 2 = TCase
Tcase + 10 = Tjunction
Lowest possible idle temperature cannot be exceeded. That means that if you can hit your lowest possible idle with the stock cooler, then nothing will be gained by switching to your Tuniq or Thermalright for idle temperatures. Load, on the other hand is a different story. The better the cooler, the smaller the delta you have between your idle and load temperatures.
Breaking it down under average conditions
Breaking it down for the best possible scenario
Anyway, I hope this helps some people to better understand all the different variables that can affect temperature. But in the end, it really is quite simple to figure out if your temperatures are believable or not.
If I made a mistake, or if I have a typo, please let me know so that I can correct it. The last thing I want to do is spread false information. The goal of this thread to give ATers something they can expect from the C2Q family.
Edit ** I also wanted to add my experience when I went to the store today at circuit city. I was shopping for a new monitor 22" or 24" and stumbled across an HP Q6600 machine. I was curious as to how well the case was ventelated and so I decided to stick my hand out the only exhaust fan it had. The temperature in my guestimation was 110F (43c). Applying the same rules here, we know that Tjunction cannot idle less than 53c... So, the question one really has to ask is why would HP allow a machine to run so close to max specification of the CPU? Unless the max specification is extremely conservative? That seems most likely.
Keep in mind as well that this case was baking in CC with an ambient temp of around 22-23c. A home user buys that computer, sticks in their desk and and ambient temps are raised a further 7c. That means the thing will idle at 60c... Food for thought. Consider that maybe, just maybe the Tcase max temperature stated on Intel's website specification is probably extremely conservative because I cannot imagine HP wanting warranty calls on the CPU's burning up or overheating.
I would love to get my hands on an OEM Q6600 computer and test the temperatures. We could have a far better guideline for what is deemed 'safe' for longevity.
This guide borrows quite a bit from Tom's Hardware Forum in this thread. So, I really take no credit, but rather am pointing to his work in a possibly easier to understand post. Essentially, my goal is to condense that version into something quick and easy. Because, to be honest, I am not sure many people have really read that thread in detail nor have they put it into practise. Because if people did, we wouldn't see people claiming idle temperatures lower than their ambients (I have seen a few people claim it) or claiming idle temps so low that their ambients would had to have been extremely low (Antarctica anyone?).
Here are the values in a nutshell that apply to the Q6600 B3/G0
Tcase = Geometric Center of IHS. Now, Tcase cannot be lower than your ambient case temperatures. Blowing 100F air over an object is not going to cool it below 100F, the best it can hope to do is cool it down to 100F. Thus the principle applies to Tcase. If your computer case temperatures are 35c, then your Tcase cannot be anything but 35c+. If your CPU & HSF are not lapped (reduced metal to metal contact) then you are realisticly looking at a minimum of 2c+ computer case temperature for Tcase.
Tjunction = Diode inside the center of each core themselves to find the 'true' temperature. Tjunction for Q6600 B3/G0s is anywhere from 7-13c+ Tcase. My particular Q6600 B3 sees 13c+ at the hottest core, and 9c+ at the coolest core.
Idle Versus Load = The temperature of the chip when you are doing nothing except sitting within windows at 0% utilization (difficult to get this at absolute zero for obvious reasons). The load temperature is usually measured by a program called P95/Orthos loading all four cores, or TAT (if it works). TAT is said to put a 114% workload on the CPU. Of course, I am not sure how you can get another 14% beyond 100%, but hey, those are the numbers claimed. Also, the delta between idle and load is typically 15c - 30c depending on your cooler/fan, environment and program used to stress the CPU.
Putting it all into a formula using room temperature in celcius as our starting point.
RT + 5 = CT
CT + 2 = TCase
Tcase + 10 = Tjunction
Lowest possible idle temperature cannot be exceeded. That means that if you can hit your lowest possible idle with the stock cooler, then nothing will be gained by switching to your Tuniq or Thermalright for idle temperatures. Load, on the other hand is a different story. The better the cooler, the smaller the delta you have between your idle and load temperatures.
Breaking it down under average conditions
Breaking it down for the best possible scenario
Anyway, I hope this helps some people to better understand all the different variables that can affect temperature. But in the end, it really is quite simple to figure out if your temperatures are believable or not.
If I made a mistake, or if I have a typo, please let me know so that I can correct it. The last thing I want to do is spread false information. The goal of this thread to give ATers something they can expect from the C2Q family.
Edit ** I also wanted to add my experience when I went to the store today at circuit city. I was shopping for a new monitor 22" or 24" and stumbled across an HP Q6600 machine. I was curious as to how well the case was ventelated and so I decided to stick my hand out the only exhaust fan it had. The temperature in my guestimation was 110F (43c). Applying the same rules here, we know that Tjunction cannot idle less than 53c... So, the question one really has to ask is why would HP allow a machine to run so close to max specification of the CPU? Unless the max specification is extremely conservative? That seems most likely.
Keep in mind as well that this case was baking in CC with an ambient temp of around 22-23c. A home user buys that computer, sticks in their desk and and ambient temps are raised a further 7c. That means the thing will idle at 60c... Food for thought. Consider that maybe, just maybe the Tcase max temperature stated on Intel's website specification is probably extremely conservative because I cannot imagine HP wanting warranty calls on the CPU's burning up or overheating.
I would love to get my hands on an OEM Q6600 computer and test the temperatures. We could have a far better guideline for what is deemed 'safe' for longevity.
Facebook
Twitter