Q9450 temperature

[Kewldude]

I'm wondering which program to trust regarding the temperature of my cpu/cores. I've ran 4 programs to monitor the temperatures while Everest had the cpu at 100%, and that's what happened:



I wish Real Temp was the correct one as I wouldn't be worrying so much, but considering 3 out of 4 showed it was reaching nearly 70C it made it worry.
Anyone happen to know how hot a Q9450 is supposed to be in 100% load?
Also, why is the difference from the CPU temperature to the cores so incredly high? On 3 programs the temperature of the cores are 33C hotter than the temperature of the CPU according to the readout from the BIOS, and then there's Real Temp which shows the cores' temperatures being 21C higher than the CPU's.
In the end which program should I trust? The ones that say the cores' temperatures are 21C hotter than of the CPU's or the ones that say it's 33C hotter than the CPU's? Or neither of those 4 are correct and I should just go by what the CPU temperature is and maybe add an extra 5-10C to it to figure out the real cores' temperatures?

Thanks in advance.

edit: took a new screenshot using the version of real temp that came out yesterday.

 

[Kewldude]

right now the cpu temperature according to the bios is 29C, and CoreTemp shows the cores at 50-54C.
I'm mainly wondering why is there such a great difference from the cpu temperature to the cores temperatures. Could something be wrong? Or that's normal to have such a great difference of temperatures?

 

[Kewldude]

so I should just knock off 10C of whatever core temp is displaying to get the accurate temperature? I would end up having the temperature which is displayed on Real Temp 2.60, which is 10C cooler than core temp and every other program out there. Maybe Real Temp 2.60 has nailed it after all.

 

[Reynod]

Try using the search box above you ... the little magnifying glass in blue.

It will help you search threads in the forums on the topic of interest to you.

Here are a couple that might help you:

http://www.tomshardware.com/forum/221745-29-core-quad-temperature-guide

http://www.tomshardware.com/forum/248429-28-temperature-monitoring

 

[Granite3]

Real Temp is the only one reading correct on the 45nm quads.

I verified this with a thermocouple on the die side, it was a variance of around 2-3c.

Core Temp, TAT and even speedfan are off right now, but at least you can calibrate Speedfan.

My Q9450, on water, runs 34c idle, and around 52C +- at load on prime 95.

Scared crap out of me, as my e6750 ran 34c at load on the same setup... Tried a Q6700, and it runs 13c hotter than the 9450!

Everest runs way higher temps, very dissapointing to shell out that money last year, and not useable now....

 

[Guest]

Granite3
I was concerned about my temps as my WC system cooled my AMD 5200+ to about 34 C on full load but stock im getting 39-32-39-39 using real temp on the q9450.
only thing i did notice is that moderately oc'd to 3.33GHZ im getting temps of roughly 53-47-49-44 (not too much higher), but the 3rd and 4th cores dont heat up for a good 8 mins? I was wondering if my temp reporting wasnt working properly for the 3rd and 4th cores?

What are your exact temps at stock? is the 32 C on myne accurate? core no. 2? and what system are you running? do your cores do the same thing? sticking

I am using thermochill PA120.2, D-Tek block, MWC30 NB, MCP655 pump, fans and pump flat out! oh and the 4.5mm injector nozzle

 

[Cannon Fodder]

Hey,

I have the same cpu and oc'ing to 3.2 with a Zalman 9700. My temps by Real temp are in the mid 30C at idle and max out in the high 40C when playing Crysis. About your dicussion on which temp program to trust, go to the Real Temp web site and the author gives a good description of how these programs work. I was looking at your screen shot and noticed that your temps are 10 C difference. Real temp Tj max is set default at 95 C and Core Temp Tj max is 105 C. Intel has not released the Tj max for the 9450 yet. The author of Real Temp talks about his rational for the 95 C setting. I am not an expert on this by no means I just read these threads to gauge how my system is doing as I am very green to oc'ing and do not wish to burn up $300. Good luck

 

[Grimmy]

If you use apps that give core readings... they go by a tjmax.

If the tjmax is different, you'll have different readings. (coretemp, realtemp, speedfan, HDmonitor)

Intel doesn't publish what the tjmax are, so the software makers guess/assume what it is.

You can make all the apps, except HDmonitor to have the same tjmax, and wella, they all say the same temp.

 

[mikeynavy1976]

If adding more data helps. I have an air-cooled 9450 at 3.2GHz (and 1.29V) and my core temps at idle, using RealTemp 2.60 are:

C1 - 35C
C2 - 35C
C3 - 38C
C4 - 38C

With an ambient temp in the room right now of around 27C I think these are fine, even with the 3 degree difference between the cores. Any thought? I assume you'll always have some variance in temps amongst the cores, correct?

 

[Srotten]

Funny,

I've been dealing with the same thing, and what everyone is saying here is what I'm hearing from others as well.

My Cores are around 8c difference at load. Using HWMonitor they are around 55c to 61c with core 1 being the hottest.

I'm just going to take 10c off the reading to give me a rough estimate of temps. I'm not sure there is a perfect temp program out for the new chips.

Just wanted to let you know you are not the only one seeing this.

-Mark

 

[Grimmy]

Its usually best to go by the Tcase sensor, which Intel does provide a max thermal spec. All in all I wouldn't worry too much about core temps, since its based on a guess, unless you get in the nitty gritty and do the calbrations. Although the 45nm seem to have cores that stick at certain temps, so again, its better to go with the Tcase sensor, which your bios reads, and speed fan can also read, as well as the MB utility app should read.

Edit:

Only other option is to RMA it, and hope for the best on the replacement chip.

 

[mikeynavy1976]

It's not worth replacing the chip unless your core temp sensors are really stuck. I've looked at two 9450s and the results are about the same. As for what utility to use, my understanding is that Real Temp is the best for 45nm temperatures. In the research I've done, many people say that CoreTemp, HW Monitor, and some of the other ones are more accurate with 65nm processors but are usually 10C off with 45nm processors (Wolfdale and Yorkfield).

 

[Grimmy]

Never said it was worth replacing the chip.. just saying that is the only other option if you are not satisfied enough, or bothered by the temp readings.

I've already explained about RealTemp, CoreTemp, HD Monitor, as well as SpeedFan, they all go by a tjmax, which Intel doesn't publish or specify.

As for myself, I really go by the Tcase sensor, rather then the cores.

 

[mikeynavy1976]

Sorry Grimmy...I wasn't attacking you...just saying that the chances of getting a chip with the same problem seem to be good. Yes, he might try it but might also be disappointed or risk getting some other problem. I agree with you though regarding the Tcase. I tried a few times to reseat my heat sink to try to get all four cores to register the same idle temp and just wasted my time.