Interesting results... I am still researching the topic of heat and what not in regards to my Q6600 B3. I found some interesting results, things that just don't add up over time.
Xbit labs is one of my favorite review sites. I prefer their testing, to each their own I suppose. But I did take notice to something that clearly doesn't make much sense at all (to me). Their reviews are located here.
Lets examine the two tests...
First Test
Ok, now stay with me here...
Second Test
These possibly cannot be the results of a 6c ambient temperature difference. The fact that a Conroe E6400 CPU runs HOTTER than a Core 2 Quad is ridiculous. They are both Conroes.
I wish we had more information in regards to temperature discrepencies.
A few answers I have seen
1) Bad IHS
2) Improperly mounted HSF
3) To much Thermal Grease
4) Convex or Concave CPU
5) Ambient Temperatures
6) Incorrect Sensors
7) Different CPU Revisions
Also, for those who do not know, I also did my OWN test. Let me share the results here...
Both of these were done on non-lapped chips and HSF. Later, I decided to lap my Q6600 and it dropped my temps 9c across the board. I didn't lap the G0 E6750 though, as I don't own (friends chip) it.
Anyway, my results are extremely different than the results that were tested by Xbit labs. In my test system, the QC ran 25c hotter than the higher clocked and higher volted DC with the same ambient temps. Xbit labs had a hotter running C2D, and slightly higher ambient temps. Even dropping their numbers 10c (generous too) to account to the 6c ambient temperature difference, still does not make much sense.
The reason this is an important issue is that you have people running around telling others than they did something wrong and that is the reason their temps are high, when it is plain to see that some respectable review sites are seeing some abnormally high temperatures as well. Is it possible this is more sensor related? It would be nice for an Intel Engineer to comment the discrepencies of their CPUs. Things just don't add up, honestly.
Even read on Extreme Forums that some of the hardcore water rigs are running a load of 65c on their Q6600 B3 lapped, while others claim to have a Q6600 B3 with a Tuniq and are claiming 59c load temps and both report ambients temps in the range of 25-26c.
Also, for those who do not know my background. I work for DuPont on the automotive computer side of things and we have a lot of Intel machines in environments that are just downright terrible. So much dust goes through the auto shops that quite often the Pentium D's would be running without the CPU or case fans spinning and they would run for year(s). This suggests, IMO, that CPU's have a far higher heat tolerance than many people believe both on this forum and on many others.
You have people saying that a CPU running at 65c all the time is going to kill it in two years (I just read that in another forum, I believe extreme systems). Whereas, I have machines out in the field that are running in 55c ambient temperatures! A CPU cannot idle lower than its sorrouding ambient temperature! So, here we have machines that idle, quite litterally am minimum of 60c. These seem to have no ill effects, even years later... I just see a lot of misinformation out there without any hard data to back them up. Even the whole Tcase @ 65c that Intel has in their specifications isn't clear cut. Because as far as I understand, it is not possible to know for certain what the Tcase temperature is and also find it odd that the Tjunction would not throttle itself until after this supposed Tcase temp is exceeded (assuming a 15 degree delta that is most commonly stated). Intel is not in the business to be replacing their CPU's for such an easy heat tolerance to be exceeded. Hell, if aftermarket cooling both lapped still run near their max Tcase (assuming 15c delta here, take from Tjunction) something tells me that the sensors are wrong, or these chips can very easily thrive in their less than ideal environment.
I'd be glad to know everyones thoughts on this, more important, someone who is far more knowledgable than myself in the matters of CPU operating temperatures. But I am getting the impression that not many here (or on any other forum for that matter) design and test CPUs.
Xbit labs is one of my favorite review sites. I prefer their testing, to each their own I suppose. But I did take notice to something that clearly doesn't make much sense at all (to me). Their reviews are located here.
Lets examine the two tests...
First Test
Ok, now stay with me here...
Second Test
These possibly cannot be the results of a 6c ambient temperature difference. The fact that a Conroe E6400 CPU runs HOTTER than a Core 2 Quad is ridiculous. They are both Conroes.
I wish we had more information in regards to temperature discrepencies.
A few answers I have seen
1) Bad IHS
2) Improperly mounted HSF
3) To much Thermal Grease
4) Convex or Concave CPU
5) Ambient Temperatures
6) Incorrect Sensors
7) Different CPU Revisions
Also, for those who do not know, I also did my OWN test. Let me share the results here...
Versus
Both of these were done on non-lapped chips and HSF. Later, I decided to lap my Q6600 and it dropped my temps 9c across the board. I didn't lap the G0 E6750 though, as I don't own (friends chip) it.
Anyway, my results are extremely different than the results that were tested by Xbit labs. In my test system, the QC ran 25c hotter than the higher clocked and higher volted DC with the same ambient temps. Xbit labs had a hotter running C2D, and slightly higher ambient temps. Even dropping their numbers 10c (generous too) to account to the 6c ambient temperature difference, still does not make much sense.
The reason this is an important issue is that you have people running around telling others than they did something wrong and that is the reason their temps are high, when it is plain to see that some respectable review sites are seeing some abnormally high temperatures as well. Is it possible this is more sensor related? It would be nice for an Intel Engineer to comment the discrepencies of their CPUs. Things just don't add up, honestly.
Even read on Extreme Forums that some of the hardcore water rigs are running a load of 65c on their Q6600 B3 lapped, while others claim to have a Q6600 B3 with a Tuniq and are claiming 59c load temps and both report ambients temps in the range of 25-26c.
Also, for those who do not know my background. I work for DuPont on the automotive computer side of things and we have a lot of Intel machines in environments that are just downright terrible. So much dust goes through the auto shops that quite often the Pentium D's would be running without the CPU or case fans spinning and they would run for year(s). This suggests, IMO, that CPU's have a far higher heat tolerance than many people believe both on this forum and on many others.
You have people saying that a CPU running at 65c all the time is going to kill it in two years (I just read that in another forum, I believe extreme systems). Whereas, I have machines out in the field that are running in 55c ambient temperatures! A CPU cannot idle lower than its sorrouding ambient temperature! So, here we have machines that idle, quite litterally am minimum of 60c. These seem to have no ill effects, even years later... I just see a lot of misinformation out there without any hard data to back them up. Even the whole Tcase @ 65c that Intel has in their specifications isn't clear cut. Because as far as I understand, it is not possible to know for certain what the Tcase temperature is and also find it odd that the Tjunction would not throttle itself until after this supposed Tcase temp is exceeded (assuming a 15 degree delta that is most commonly stated). Intel is not in the business to be replacing their CPU's for such an easy heat tolerance to be exceeded. Hell, if aftermarket cooling both lapped still run near their max Tcase (assuming 15c delta here, take from Tjunction) something tells me that the sensors are wrong, or these chips can very easily thrive in their less than ideal environment.
I'd be glad to know everyones thoughts on this, more important, someone who is far more knowledgable than myself in the matters of CPU operating temperatures. But I am getting the impression that not many here (or on any other forum for that matter) design and test CPUs.
Facebook
Twitter