CPU or Core Temps

[CranialBlaze]

This is something that really confuses me, i have tried google but as i am not sure exactly what to search for i have not yet found an actual answer.

What temperature do i need to actually worry about, the temp in the BIOS matches that of the "CPU" temp, but now the "Core 1" temp is almost double that and i know danger zone is around 75 for the Q8300 and in testing the Core temps easily hit 70, but then the CPU is only around 40/45.

I just put on a much better cooler it dropped the CPU temp from 31 idle to now 20 idle but the core temps have remained the exact same as before.

I would like to OC this CPU, i know it can handle up to 3.6, i have tested it that far but the temps made me rather not, now with a better cooler i only see 1 reading change and just what to know which is the reading i really need to concern myself with when pushing the living hell out of this thing.

Thanks guys

 

[nakedpeanut]

I would stand to be corrected. But the Core temps are the temps actually being generated by each core. The CPU temp is the temp of the CPU casing (which is attached to your cooler) that is why it's less, as the heat is being dissipated from it.

Makes sense in my little mind anyways

 

[Archer]

Its the individual cores you need to worry about. What your room temperature by the way? I have not yet seen a CPU cooled that low (21 degrees) on air cooling in a normal room, even more reason not to trust that number. What program you using to monitor temps?

 

[nakedpeanut]

Found this Thread might help u..

Intel provides separate thermal specifications for 2 different sensor types; a CPU Case (not computer case) Thermal Diode located within the CPU die between the Cores, and Digital Thermal Sensors located within each Core. The Case Thermal Diode measures Tcase (Temperature Case), which is CPU temp, and the Digital Thermal Sensors measure Tjunction (Temperature Junction), which is Core temp. Since these sensors measure 2 distinct thermal levels, there is a constant temperature gradient between them, which is Tcase to Junction Delta. C2Q`s have 1 Tcase and 4 Junction sensors, while C2D`s have 1 Tcase and 2 Junction sensors.

 

[CranialBlaze]

Its the individual cores you need to worry about. What your room temperature by the way? I have not yet seen a CPU cooled that low (21 degrees) on air cooling in a normal room, even more reason not to trust that number. What program you using to monitor temps?

I have gotten temps much lower than that, mind you not on a qaud core, im using a Zalman 9700-NT. Using i think it was an 8000, the flat copper 1 i got an E8400 to 9 degrees. When u put i think its 700g of copper on it its quite possible. Remember intels run quite a bit cooler than AMD's, especially when you have a very well ventilated chassis. For your money the big ass Zalmans are the best fan coolers i have ever seen.

No idea what the room temp is, i don't keep thermometers in my house.

Monitors, i'm using BIOS, but it only shows the 21, as well as everest and CoreTemp. The 21 used to be 31-33 with my old cooler

 

[nakedpeanut]

Well using RealTemp my min core temp is 23.. An yes it's freezing 2 nyt so room temp must be low... But i got a nice big cooler on mine as well..

 

[Archer]

I have gotten temps much lower than that, mind you not on a qaud core, im using a Zalman 9700-NT. Using i think it was an 8000, the flat copper 1 i got an E8400 to 9 degrees. When u put i think its 700g of copper on it its quite possible. Remember intels run quite a bit cooler than AMD's, especially when you have a very well ventilated chassis. For your money the big ass Zalmans are the best fan coolers i have ever seen.

No idea what the room temp is, i don't keep thermometers in my house.

Monitors, i'm using BIOS, but it only shows the 21, as well as everest and CoreTemp. The 21 used to be 31-33 with my old cooler

Not possible. If the air in your house is more than 9 degrees (which I assume it is since 9 degrees is bloody cold) then you simply can not cool something with that same air to a lower temp. Basic physics. Absolute best case scenario you can cool the CPU to the same temp as the surrounding air, more likely you get to within 10 degrees. Your temp monitoring software, or the thermometers inside the CPU are faulty.

Also AMD chips run cooler than Intel -> i7 and i5 chips run way hotter than Phenom II chips. Not quite sure what point you were trying to make there anyway.

 

[CranialBlaze]

Not possible. If the air in your house is more than 9 degrees (which I assume it is since 9 degrees is bloody cold) then you simply can not cool something with that same air to a lower temp. Basic physics. Absolute best case scenario you can cool the CPU to the same temp as the surrounding air, more likely you get to within 10 degrees. Your temp monitoring software, or the thermometers inside the CPU are faulty.

Also AMD chips run cooler than Intel -> i7 and i5 chips run way hotter than Phenom II chips. Not quite sure what point you were trying to make there anyway.

Then there have to be lots of faulty sensors out there as i have done it 3 times, i got an E8300 and an E8400 to under 10 on idle 1 had a Zalman CNPS7500-Cu Led, the other a zalman CNPS9700-NT. I also got an AMD Althlon 2.7 to idle at 13 with the CNPS7500-Cu Led. Remember the sensors read the cpu, these heatsink are essentially massive blocks of copper, if i stuck a thermometer on the sink itself it would probably be much hotter, butt he whole point is to draw heat away from the cpu and into the sink. Also, i fitted all these machines during summer, not winter, never actually checked their temps in winter.

My current temp is 16 and according to my weather app its supposed to be about 10 outside, should be slightly higher than that inside.

Here are the temps from someone who replied to my thread on another forum using the Coolermaster Sphere, and i had absolutely nothing to do with his machine

Weather or not is "scientifically possibly" it's happening, the copper is probably just removing all the heat and dissipating it quickly like it should.

 

[Archer]

Weather or not is "scientifically possibly" it's happening, the copper is probably just removing all the heat and dissipating it quickly like it should.

It is not possible, you simply can not cool the CPU to a temp lower than the air temp you use to cool the CPU with, even if you had 5 tons of copper sitting on it. Like I said, either the temp monitoring software is buggy, or the temp probes are faulty. Faulty temp probes btw are not that uncommon. IIRC a whole line of Phenom CPUs had temp sensors that under read the temperature. Download HWMonitor, see what temps that reports. You'll probably find it reports close to the core temps values you are getting which look way more normal.

 

[CranialBlaze]

It is not possible, you simply can not cool the CPU to a temp lower than the air temp you use to cool the CPU with, even if you had 5 tons of copper sitting on it. Like I said, either the temp monitoring software is buggy, or the temp probes are faulty. Faulty temp probes btw are not that uncommon. IIRC a whole line of Phenom CPUs had temp sensors that under read the temperature. Download HWMonitor, see what temps that reports. You'll probably find it reports close to the core temps values you are getting which look way more normal.

The likelihood of 4 different processors on 4 different occasion with 4 different hardware configuration is extremely unlikely no matter how common it is, occam's razor dude.

You problem is the assumption that the fan cools the CPU, which is does not, it cools the heatsink which draws heat away from the CPU. The sensors do not check the temp of the heatsink, they check the temp of the steel plate making contact with the heatsink, if all the heat gets pulled away from that metal plate then logically its temperature will drop.

We already confirmed that the core temps are the proper temps to actually monitor, not that of the casing covering them and the 21 degrees is that of the casing covering them.

 

[Archer]

The likelihood of 4 different processors on 4 different occasion with 4 different hardware configuration is extremely unlikely no matter how common it is, occam's razor dude.

You problem is the assumption that the fan cools the CPU, which is does not, it cools the heatsink which draws heat away from the CPU. The sensors do not check the temp of the heatsink, they check the temp of the steel plate making contact with the heatsink, if all the heat gets pulled away from that metal plate then logically its temperature will drop.

We already confirmed that the core temps are the proper temps to actually monitor, not that of the casing covering them and the 21 degrees is that of the casing covering them.

Even the integrated heatsink on top of the 4 cores cant be cooled to less than air temp. Physics! It will likely be a lower temp than the individual cores, but still cant be lower than air temp. Pity your hard drive does not have an additional temp sensor for incoming airflow, am curious to see what it is.

And I made no assumptions that the fan cools the CPU?! I do know how a heatsink works, dont assume that I do not.

Edit: how can something common be unlikely?

 

[CranialBlaze]

I have 2 thermal sensor cables in my cupboard that i never bother using.

Sensor 2 is just behind the intake fan and sensor 3 is between the fins of the heatsink. The heatsink itself is 4 degrees above intake temp and 10 above the cpu plate.
The only logical explanation would be that its pulling the heat off the plate and dissipating it the way copper was scientifically proven to do.

Here are pics of the case and the sensor itself, just to prove the results are not fabricated.

 

[Archer]

I have 2 thermal sensor cables in my cupboard that i never bother using.

Sensor 2 is just behind the intake fan and sensor 3 is between the fins of the heatsink. The heatsink itself is 4 degrees above intake temp and 10 above the cpu plate.
The only logical explanation would be that its pulling the heat off the plate and dissipating it the way copper was scientifically proven to do.

Here are pics of the case and the sensor itself, just to prove the results are not fabricated.

Argh!!!
I'm going to give the air intake temp an arbitrary value -> 20 degrees. So the heatsink (zalman copper block) is 24 degrees then. Thats fine assuming CPU is idle. Now you say that the heatsink is 10 degrees above the CPU IHS (integrated heatsink), making the IHS 14 degrees. Simply not possible! Heat goes from CPU core 1/2/3/4 to the IHS to the Zalman to the air. So the absolute coldest that anything in that process can be is whatever the air temp is. The air can only cool the Zalman heatsink to 20, the heatsink can then only cool the IHS to 20, the IHS can then only cool the cores to 20. Cant go lower. Just because the zalman heatsink is made of copper does not mean the IHS can get colder than the zalman heatsink.

What you are proposing is that the IHS on the CPU is cooling the zalman heatsink on top of it. If I read the part in bold correctly.

 

[CranialBlaze]

Argue all you wan't, the proof is in front of you. Believe what you want. Just because u cannot achieve such results does not make them false.
The first batch of Q8400's had extremely faulty sensors, i have checked with Intel a while ago and mine is from the second batch, the safe batch.

Its my ability to replicate these results in numerous conditions that has prevented me from waisting money on Water cooling.

I have done this many times and seen these kind of temps many times and none of the machines i have configured like that have air conditioned houses.

 

[Archer]

Argue all you wan't, the proof is in front of you. Believe what you want. Just because u cannot achieve such results does not make them false.
The first batch of Q8400's had extremely faulty sensors, i have checked with Intel a while ago and mine is from the second batch, the safe batch.

Its my ability to replicate these results in numerous conditions that has prevented me from waisting money on Water cooling.

I have done this many times and seen these kind of temps many times and none of the machines i have configured like that have air conditioned houses.

Its physically impossible, and hence I wont believe the results. The proof provided is not sufficient. In no way do you ever rule out the possibility of a faulty, or incorrectly calibrated IHS temp sensor. Try push it all you want, but you have not broken the rules of physics/thermodynamics.
Having said all that, +-40 degrees on idle for the cores is what is it should be on stock. But the load temps are what you should worry about if wanting to OC. Run prime95 for a few min, check the max temp. If its under 71 degrees (max for your chip afaik) you got some OC headroom.

 

[Budza]

Is there no 'wind chill' type effect of air cooling? Whereby the HS can reach a temperature lower than ambient? I'm inclined to side with Archer on this one- just wondering if there is a legit case for CB's temps...

 

[CranialBlaze]

Its physically impossible, and hence I wont believe the results. The proof provided is not sufficient. In no way do you ever rule out the possibility of a faulty, or incorrectly calibrated IHS temp sensor. Try push it all you want, but you have not broken the rules of physics/thermodynamics.
Having said all that, +-40 degrees on idle for the cores is what is it should be on stock. But the load temps are what you should worry about if wanting to OC. Run prime95 for a few min, check the max temp. If its under 71 degrees (max for your chip afaik) you got some OC headroom.

Having been able to achieve these types of results in a number of machines with different CPU's i have no reason to doubt the possibility and the logic behind it. i ran prime95 last night and my max load on core was 49 for core 0, the rest were all lower than that. I have a friend running an i5 750 with stock cooling OC'd to 3.6Ghz and his idle temps are 29 - 31 and load temps of 55 - 62.

The Zalman i got in now is his old cooler and when he had his 775 his temps were pretty similar to what i am getting now, we swapped because i had the 1156 adapter and the greater need for his cooler, if his stock gives him those temps than my old 9500 will be more than enough for his system.

 

[Monkey101]

OMG is this guy a retard? lol. A CPU cannot be cooler than the surrounding air unless there is an object of lower temperature causing it to drop. ie, your CPU will NEVER be cooler than the surrounding air unless you are running it with dry ice, ln2 or so on. Its even impossible with water cooler as the water temperatur will eventually get to room temperature and wont go any lower.

 

[nakedpeanut]

LOL funniest thread i've read in ages..
budza's therory might hold some credibility..
But i have to agree with Archer on this..
Youe air particles have a specific temperature eg 20. By moving over your heatzinc which is say 24 they are absorb the extra heat energy (thus dissipating the heat). The air is now warmer having left the heatzinc and is expelled from the case. Thats why a well ventilated case is needed so that constant "cool" air can make contact with the heatzinc and then be expelled.
If the air eg 20 is higher than your heatzinc say 15, the air will transfer this energy to the haetzinc making it hotter and the air cooler. But as more air of 20 gets pushed past your heatzinc the hotter the heatzinc will get.. We not even taking into account the temps being put into the heatzinc by the CPU...

So no way do i see it possible for the heatzinc to be cooler than the ambient temp unless budza's idea is correct...

 

[Monkey101]

Read up here CranialBlaze:
Wind Chill

For inanimate objects, the effect of wind chill is to reduce any warmer objects to the ambient temperature more quickly. It cannot, however, reduce the temperature of these objects below the ambient temperature, no matter how great the wind velocity.