Core 2 Temps?

C

chanchan

Limp Gawd
Joined
Sep 14, 2005
Messages
482
Couldn't find a good thread on Core 2 temps regardless of version.

I would like to know what temps people are getting in Idle/Load and their clock speed and cooling method.

I'm using water cooling @ 3.06 GHz

These were old values which are incorrect! Idle 55c load is about 62c.

After new BIOS 1002 for my Striker and reporting using CoreTemp

IDLE 42c Load 55c!
 
tisb0b said:
stock cooler with [email protected] (1.3v) idle: 51-53 load: I'm not sure as I haven't stressed it for awhile
Click to expand...

Idle at 51-53? Wow that's not too good. I'm glad you haven't stressed it in a while because I don't think it would survive o_O. If you're going to clock it that high you should grab an after market heatsink or clock it down until you can afford one.
 
Yeh... regardless of your ambient room temp that's way too high... Tell us what you're using to measure temps.... google core temp and see if it differs...

My new system, overclocked to 3.5Ghz and on air idles at 28c and loads to 55c ... check the seating of your water block...
 
Battle_Rattle said:
Yeh... regardless of your ambient room temp that's way too high... Tell us what you're using to measure temps.... google core temp and see if it differs...

My new system, overclocked to 3.5Ghz and on air idles at 28c and loads to 55c ... check the seating of your water block...
Click to expand...

I find it hard to believe your temps. I got the same temps using CoreTemp but when I use TAT, I get a more realistic reading of 42-44C idle and 68-70C load at 3.3 GHz (1.425v).

Try with TAT and look.
 
My highest load temp was at 1.45 volts, got me up to 70+ Celsius. Having the extra 240 mhz was not worth the 12C change in load temps, so now I run at 1.325 volts and 3.06 ghz and see a normal load temp of 58-62C.
 
Xilikon: Could you try SpeedFan v.4.32 that just came out. It now has the option to report directly from the C2D on chip temperature sensors.

On my P5B Deluxe, SpeedFan and CoreTemp are now in agreement but I've heard of some E4300 users having problems with CoreTemp. It would be interesting to see how CoreTemp and the new SpeedFan compare on your board. Can you post your room temperature as well so I can compare to what I'm getting.
 
unclewebb said:
Xilikon: Could you try SpeedFan v.4.32 that just came out. It now has the option to report directly from the C2D on chip temperature sensors.

On my P5B Deluxe, SpeedFan and CoreTemp are now in agreement but I've heard of some E4300 users having problems with CoreTemp. It would be interesting to see how CoreTemp and the new SpeedFan compare on your board. Can you post your room temperature as well so I can compare to what I'm getting.
Click to expand...

I could do this tonight without problems... I'm a bit afraid of using speedfan after downloading 4.31 and borking XP by kicking some power saving, reducing voltage drastically (dunno how to disable) so I was forced to reinstall XP. If you say 4.32 works fine, i'd give another try after making a new image of my XP setup.
 
SpeedFan when first installed shouldn't do anything to your system. It can be used to monitor CPU volts and temps without having to allow it to automatically adjust anything.

SpeedFan doesn't allow you to adjust voltages, only fan speeds, on my P5B Deluxe so maybe something else screwed up your computer. It's not worth trying again to satisfy my curiosity if it screwed your system up that bad last time.
 
unclewebb said:
SpeedFan when first installed shouldn't do anything to your system. It can be used to monitor CPU volts and temps without having to allow it to automatically adjust anything.

SpeedFan doesn't allow you to adjust voltages, only fan speeds, on my P5B Deluxe so maybe something else screwed up your computer. It's not worth trying again to satisfy my curiosity if it screwed your system up that bad last time.
Click to expand...

On the BX2, there are reports of Everest screwing with the cpu voltage when trying to read sensors. It was a bug with the program who got fixed in a subsequent version. I believe it happened the same way with SpeedHan but I dunno exactly what I did to screw it so i'll do a research on it. However, it may be a moot point since the new 4.32 should work fine.

I just checked the speedfan site and I noticed 4.32 is now final with the change log. Items I'm probably interested is :

- added full support for INTEL CORE DUO internal temperature readings
- improved compatibility with Intel D975XBX2
- improved Intel SMBus routines
 
With the same chip running 3.065 ghz on air...
Idle 33-34C
Load (Orthos @ 5 min.) 55-56C
using Core Temp and room temp 20C.
 
Donkers: Your C2D is running exactly as it should. Those numbers are what a person should be seeing with a room temp of 20C / 68F when air cooling.
 
At what point do temps get uncomfortably high at load for a C2D? From what I've read, it seems like most don't like beyond 65C? Is this a pretty accepted number?
 
unclewebb said:
Donkers: Your C2D is running exactly as it should. Those numbers are what a person should be seeing with a room temp of 20C / 68F when air cooling.
Click to expand...

Precisely, but what chanchan is getting is quite higher than what it should be.

Chanchan, maybe you should re-install your water cooler, you should be running cooler than air.
 
I use the Arctic Cooling Freezer 7 Pro and a e6300 overclocked to 2.8Ghz. My idle with TAT is in the low to mid 40's and my load using TAT and running folding@home on each core is in the low 50's.
 
Thermaltake BigWater 745
Click to expand...

see my sig, we need to talk, something is wrong. Load is acceptable/marginal but idle is at least 10C too high. As I sit here runing Folding at Home in the background temps are 40C 38C on core not running F@H and I am pushing it a lot harder than you are. Have some tips and tricks mentioned in the site below most for my board but there are a lot we can do with yours too) : To start, lap the block and look at my waterblock mounting.

http://mysite.verizon.net/ressdxka/

Whats your Vcore ?

Are you using the newer version with the 1x120 and 2 x 120 rads and 10mm tubing ?
 
unclewebb, I downloaded and ran SpeedFan 4.32 now and it match CoreTemp readings. TAT is consistenly about 12C higher so I believe it's buggy and unreliable since when I touch the heatsink, it felt warm, almost cool. If TAT is accurate, I should feel the heat more than that.

I will retry to push the overclock again and see if it will get over 80C in TAT without throttling. If it throttle at this speed even if the 2 other is under 80C, TAT is accurate and the others is if it is not the case.
 
I would like to exactly what is what myself.

E4300
ABIT IB9
Vapochill Micro w/fan on controller

CoreTemp: 21c-23c idle ; 50c-51c load
Everest: 39c-40c idle ; 55c-58c load
SpeedFan 4.31: 39c-40 idle ; 63c-65c load

The only thing that matches up are idle temps of SpeedFan and Everest. Don't ask me about TAT cuz it won't run on my system, I have tried.
 
I guess my idle temps are slightly higher than normal considering that it's on water, but when loaded it doesn't differ too much. Guess that 8800 dumps quite abit of heat into the loop.

Idle: 43-46 for first 2-3 hours, then 47-50 after.
Load: 55-58C

Temps measured using TAT & Coretemp agrees with it completely. E6400 @ 3.2ghz with 1.285 volts.
 
My temps are 50 idle. 60-62 when gaming (in TAT, CoreTemp does not work for me). Room temp: 21C. Computer in sig.
:( It seems a bit high even though its stock hs.
 
jbmx4life: SpeedFan v.4.32 that just came out is the first version that properly reports temperatures from the temp sensors built into the C2D chip. Upgrade to this version and believe what it tells you.

Too many people have been trusting TAT because it has that nice Intel logo on it but I'm seeing too many different combinations where it's not reporting temperatures correctly.

On the 965 boards like the Asus P5B series and the Gigabyte S3, DS3, DQ6 it works fine.

Speedfan v.4.32 supports XP and Windows Vista 64 so there should be no more problems with people reporting bogus temperatures no matter what OS they're running.
 
I just tried 4.32 and I'm getting:

Speedfan: 33C Idle
TAT: 50 idle

Trust speedfan?
Is there a specific ver of CoreTemp to work w/Vista?
 
E6300 @ 3.24ghz @ 1.40v with Zalman 9500 (using AS5) gets 44 idle and 55 load (orthos). Using CoreTemp. Ambient room temperature is 74 degrees Fahrenheit.
sdzf5.jpg
 
As soon as you start adding voltage to the C2D the heat starts to go up, both idle and load.

The Intel heatsink and fan is more than adequate at default voltage.

At a room temp of 18C, here's how my mildly overclocked E6400 at 1.328 volts looks at idle.
Very cool. :cool:

e64003200mhzsb9.png


I'm using a Gigabyte 3D Aurora 570 case.
http://www.gigabyte.com.tw/Products/Chassis/Products_Spec.aspx?ProductID=2322

I disconnected the front fan because I thought I saw a layer of ice forming on my hard drives. It has two rear exhaust fans which are very quiet and the Seasonic power supply is about as quiet as they come. It's also very efficient and is the coolest power supply I've ever used. It all adds up to some nice core temps.

Speedfan v.4.32 and CoreTemp are very consistent. It's not likely that they are both equally wrong. Everest does not read the Intel on chip sensor and I don't trust TAT on every board and now I don't trust it on Vista either. It was designed for laptop testing and it needs to be updated to properly support the C2D and the new motherboards available.
 
I did look around yesterday evening and I stumbled across a topic on xtremesystems (http://www.xtremesystems.org/forums/archive/index.php/t-106951.html) and it was a interesting read. What I gathered here will help us clarify the facts :

Core 2 processors have 3 sensors in it (or 5 for quad cores). There are 2 Thermal Digital Sensor, one in each core, reported by CoreTemp and SpeedFan 4.32 as core temps. There is also a Thermal Diode on top as well, reported in SpeedFan 4.32 as CPU temp and is also the temperature reported in the BIOS.

Thermal throttling or shutdown occurs when the Thermal Diode get over 80C even if core temps stay below this. This seems accurate since I did test to find the best overclock on my E4300 and at 3414 MHz @ 1.5v, the core temps stay around 65-68C but the CPU temp in SpeedFan hover around 75C. At 3510 Mhz, I ran OCCT for 20 mins then it reboot unexpectedly so I missed the readings. However, I believe it was due to thermal protection since it is hovering at 78-79C, very close to the threshold.

This kind of testing and results tell me that the temperature we should all check and keep in control should be the CPU temp reported by SpeedFan since it's what trigger the BIOS. With CoreTemp, thermal throttling or shutdown would occur at around 70C since the difference between core temp and cpu temp is approximately 10C. CoreTemp match SpeedFan but is a bit more faster at updating the data while speedfan lag for 1-2 seconds. If we want to get a computer temperature in a good range for daily usage, we should try to keep core temperatures below 60C at all costs.

In my case, I officially trashed TAT since it's grossly inaccurate.
 
Speedfan 4.32 doesn't seem to work perfectly for me yet, the Core 2 temps agree with Core Temp Beta, but I have one temp that registers at -2C, and either my PSU is grossly failing, the voltage readings are waaaaay off too. (+12V 0.90V, -12V -16.97V? That normal?)

So Core Temp and Speedfan report the same temperatures, and TAT is about 20C above them at all times. Looks like I should trust Core Temp / Speedfan?
 
Xilikon said:
for daily usage, we should try to keep core temperatures below 60C at all costs.
Click to expand...
I have to disagree with that. I think a maximum core temp of 60C is way too conservative.
The C2D is very reliable even when CoreTemp is reporting numbers up into the low 70C range.



It really depends on what you do with your computer. Many users that are gaming and / or running single threaded apps will never get near 70C even if they're going higher than that when running Orthos.

CoreTemp is a good program because you can log your temps to a file. When the C2D starts getting near its maximum it generally tends to reboot. After it does this you can have a look in the log file and see what core temperature made it crash. I found that when temps got up over 75C to 77C for a prolonged period of time is when stability started to suffer.

skudmunky: I wouldn't trust any program to report the 3.3 volt, 5 volt or 12 volt rails 100% correctly. Whenever I plug in my digital multimeter those numbers are never accurate. I agree that SpeedFan reports some bogus numbers but the important numbers, core temp and core voltage look very accurate to me. Everything else really isn't that important.
 
unclewebb, I based my opinion on the fact that I'm folding 24/7, hence why I would suggest a lower core temp of 60C. If all you do is casual gaming, browsing, ripping and not in continous periods, we can live with 70C for sure.
 
Xilikon said:
Thermal throttling or shutdown occurs when the Thermal Diode get over 80C even if core temps stay below this.
Click to expand...
I have to disagree with this statement as well. On my P5B Deluxe, thermal throttling didn't start to occur until CoreTemp was reporting 80C.
 
Xilikon: How does your core temperature compare when you're just running Orthos to when you're folding 24/7?
 
When folding, I'm always getting 3-5 degrees lower compared to Orthos. OCCT is the one who pushed the hardest by 2-3C over Orthos.

About throttling, I believe it's depending on motherboards. I noticed the behavior on mine when the thermal diode registered 80C, not when core temp reached 80C.
 
I used TAT to check out thermal throttling. It seems to put more heat into my CPU than Orthos or anything out there. It also reports when thermal throttling has started.
 
Okay I'm using CoreTemp to find out my values, where do I find TAT?

Maybe I'm a Google idiot, can someone post the link?

I'm wondering if my high temps are related to my entire water system installed INSIDE my case? My GTX does pump out a lot of hot air, and some of the air will pass the RAD's.

Wonder if I'll get better results with a high performance air cooler? x___x

Here's a picture ... but it's kinda old, i have the ACS3 card now, and on the side panel, there's one 120mm fan sucking air out where the video card is, and one 120mm fan blowing air into the CPU area.

rig_full_cpu.jpg
 
Hold on folks!

Seems like after updating to the newest BIOS 1002, my temps seem lower ...

I will do multiple hours of Orthos before posting any solid values.
 
unclewebb said:
I used TAT to check out thermal throttling. It seems to put more heat into my CPU than Orthos or anything out there. It also reports when thermal throttling has started.
Click to expand...

Ok, I was basing my statements on the feelings, not notifications. I will bring TAT and redo some testing tomorrow along with the new FireStix. I will report more accurate findings for the good of the community :D
 
Xilikon said:
I will report more accurate findings for the good of the community :D
Click to expand...
In the name of science, we thank you! :D

The D9GMH Micron chips, which I believe your new FireStix are using, are incredibly flexible for timings, speed and voltage. My DDR2-800 Team memory has seen DDR2-1067 speeds without having to do too much tweaking. Add voltage and I was done!
 
I noticed on my computer that the CPU diode that previous versions of SpeedFan was using reports about 10C to 12C higher at idle than the core temp readings.

According to Intel, they positioned their new digital core sensors at the hottest spot on each core so it seems impossible for the center of the core to be warmer than the hottest spots!

Trust the core temp readings and doubt everything else.

I just rebooted and noticed now that the diode temp is 8C less than the core temps. That seems right but I still wouldn't trust it.
 
unclewebb said:
In the name of science, we thank you! :D

The D9GMH Micron chips, which I believe your new FireStix are using, are incredibly flexible for timings, speed and voltage. My DDR2-800 Team memory has seen DDR2-1067 speeds without having to do too much tweaking. Add voltage and I was done!
Click to expand...

No problem :)

Yes, it's the D9GMH one I will get. I know it can scale very nicely and my personnal goal is to set it 1:1 at 3-3-3-9 unless benchmarking tell me it's better to run the ram faster. This is the tradeoffs nobody is clear about (bandwidth or timings). This is something for a different thread tho.

However, in the actual subject, I wanted to get rid of all the bottlenecks whenever possible and test only the CPU to avoid any confusion or issues tainting my observations. I got reboots at 3.5 GHz @ 1.5v, core temp seems manageable at 70C so I cannot tell if it is the RAM being pushed (280+ MHz for a set rated at 200 MHz is pushing a bit far).
 
You must log in or register to reply here.
Back
Top