which temp is correct?

Intel TAT, it is their processor after all. :p

TAT is reading (directly I beleive) the thermal diode inside the chip and converting to temp.

Speedfan is getting its information from the SMS buss monitoring chip so there are several "layers" if you will, of code and conversions going on.

run Coretemp, almost all of us use TAT or Coretemp and so far (new 680i boards excepted) they agree very well.

Also based on my experience, beleive me when I say you aint gonna get 43c on a 6300 at 3GHz at full load
 
BillParrish said:
Intel TAT, it is their processor after all. :p

TAT is reading (directly I beleive) the thermal diode inside the chip and converting to temp.

Speedfan is getting its information from the SMS buss monitoring chip so there are several "layers" if you will, of code and conversions going on.

run Coretemp, almost all of us use TAT or Coretemp and so far (new 680i boards excepted) they agree very well.

Also based on my experience, beleive me when I say you aint gonna get 43c on a 6300 at 3GHz at full load
Click to expand...

so based on this don't use coretemp? use tat? why? explaination needed. thanks
 
no, Coretemp and TAT good !!! use either one you want. they should agree and typically do. Both read chip temp directly.

TAT is nice because it has the loading software included. Coretemp is just temps and we typically use Orthos to load the cpu for testing when using Coretemp.


It really does not matter, both programs have shown themselves to be very accurate on Intel chipset boards.
 
BillParrish said:
no, Coretemp and TAT good !!! use either one you want. they should agree and typically do. Both read chip temp directly.

TAT is nice because it has the loading software included. Coretemp is just temps and we typically use Orthos to load the cpu for testing when using Coretemp.


It really does not matter, both programs have shown themselves to be very accurate on Intel chipset boards.
Click to expand...

but i just don't see why coretemp is lower then TAT? by a big difference too. so i should beleive the lower load temp or the higher load temp?
 
TAT gives me insanely high temps with my E4300 too ... like 80-85C on full load ... which is impossible ... so I just use Core Temp which gives me ~65C on full load.

I think TAT has some issues with newer motherboards/chipset.
 
TAT won't run at all on my Abit AB9 QuadGT. I just use Coretemp, and it seems to report the same temperatures I was getting on my 680i board.
 
OH JEEZ I AM AN IDIOT So Sorry no wonder you are confused. My bad !!!! :(

I did a quick glance at the pics and saw Speedfan was running and missed you had coretemp up too, thought it was cpu-id and didnt even look at it. dumb dumb dumb bad Bill BAD !


I hate to say it but speedfan and coretemp are in close agreement and I dont trust speedfan and the speedfan and coretemps temps seem a little too good frankly. Soooooooo I dont farking have any idea but if you put a gun to my head...

actully scratch that, I dont have a clue, and gonna stf up, you need to find another owner of that board.
 
BillParrish said:
OH JEEZ I AM AN IDIOT So Sorry no wonder you are confused. My bad !!!! :(

I did a quick glance at the pics and saw Speedfan was running and missed you had coretemp up too, thought it was cpu-id and didnt even look at it. dumb dumb dumb bad Bill BAD !


I hate to say it but speedfan and coretemp are in close agreement and I dont trust speedfan and the speedfan and coretemps temps seem a little too good frankly. Soooooooo I dont farking have any idea but if you put a gun to my head...

actully scratch that, I dont have a clue, and gonna stf up, you need to find another owner of that board.
Click to expand...

i thot something is wrong!
 
judging your idle temps i'd say TAT is the acurate one. 19c is 66f, and at the same time your cpu is at 66, your motherboard is registering 89f. average room temperature is about 71f, so 66 at idle couldn't happen, slightly over 100 seems a bit more like it, and thats where my 6600 idles.

i have no proof as to what i say, just and educated guess.
 
spint3 said:
judging your idle temps i'd say TAT is the acurate one. 19c is 66f, and at the same time your cpu is at 66, your motherboard is registering 89f. average room temperature is about 71f, so 66 at idle couldn't happen, slightly over 100 seems a bit more like it, and thats where my 6600 idles.

i have no proof as to what i say, just and educated guess.
Click to expand...
my room temp is actually 60F without heater on.
 
I agree with spint3 and tell that TAT is the most accurate. No way the cpu is idling at the same or below room temperature. When in doubt, check the idle temps in the BIOS then compare with which one to find out who are the closest.

When I booted my E4300 for the first time, it is idling at 31C in the BIOS, jumping to 33C when I overclocked to 2.4 GHz (damn crappy ram holding me).
 
On most boards TAT and CoreTemp are usually pretty close since they are reading from the same sensors but in your case, TAT appears to be right and CoreTemp wrong.

SpeedFan v4.32b17 is available which now includes reading of the sensors within the Intel core. Previous versions were reading from a motherboard sensor which isn't the same.

On my P5B Deluxe, CoreTemp and SpeedFan 4.32 are exactly the same from low temp to high temp. It would be interesting to see what this new version reports on your board. Remember to let all of the temps stabilize before snapping a screen shot.

http://www.almico.com/sfbetaprogram.php
 
unclewebb said:
On most boards TAT and CoreTemp are usually pretty close since they are reading from the same sensors but in your case, TAT appears to be right and CoreTemp wrong.

SpeedFan v4.32b17 is available which now includes reading of the sensors within the Intel core. Previous versions were reading from a motherboard sensor which isn't the same.

On my P5B Deluxe, CoreTemp and SpeedFan 4.32 are exactly the same from low temp to high temp. It would be interesting to see what this new version reports on your board. Remember to let all of the temps stabilize before snapping a screen shot.

http://www.almico.com/sfbetaprogram.php
Click to expand...

coretemp and new SpeedFan 4.32 beta 17 (09 feb 2007) is exactly the same on load. whereas TAT is a good 12C higher. so load on coretemp is 46-48C TAT would be about 58-60C
 
I'm starting to believe that your CoreTemp and new Speedfan readings are correct on your board.

The C2D starts to put out heat when the voltage goes up but you're running 2950 MHz at 1.29 volts which is less than the Intel default of 1.325 volts. Your 60F room temperature is also much lower than most users and your Zalman 9700 can also keep things very cool.

I'll drop my volts and my MHz to match yours and move my computer to my sub 60F basement later today to get you some real numbers to compare to.

Everyone assumes that TAT is correct since it is made by Intel but both CoreTemp and the new Speedfan are showing identical temps which means their programmers followed the Intel guidelines exactly the same.

TAT was not specifically designed for C2D and there does seem to be some problems with it on some of the newer motherboards.
 
Here's some interesting numbers.

I have an E6400 but I dropped the multi down to 7 and set the FSB to 422 MHz to simulate your setup. I also set it so CPU-z was reporting 1.312 volts at idle and 1.288 volts while under load running Orthos.

Room temperature when I started was 13C in my basement when I started this test and went up to 14C when I ran the idle test. This is just slightly cooler than your room temperature of 60F (~15C ).

Using CoreTemp & Speedfan the temps under load went as high as 52C and then settled out at about 50C. Idle temp was reported as 20C.

The biggest difference between our setups is the heatsink and fan. Your Zalman 9700 significantly outperforms my Intel OEM heatsink and fan so your temps should be cooler than this and they are. If my computer was operated at your room temperature then my load temp would probably be about 53C. Your load temp of about 47C with a much better cooler makes perfect sense.

Anandtech did a review of the Zalman 9700. On an X6800 at 2.93 GHz their load temp dropped by 20C from 56C to 36C and their Idle temp went from 41C to 28C. Knowing this, it's possible that after you properly apply some AS5 and give it time to setup, your CoreTemp readings might even decrease a few degrees.
http://www.anandtech.com/casecooling/showdoc.aspx?i=2932&p=1

While experimenting I've seen CoreTemp readings of 80C at which point the Intel C2D thermal throttling kicked in to keep the temps from hitting the Intel maximum of 85C. I have no doubt that CoreTemp is reporting the correct temperature on my P5B Deluxe.

I'm also convinced that on your board CoreTemp and the new version of SpeedFan are reporting your real core temperature. The Intel TAT program is reporting about 3C less at full load on my board but seems to be out to lunch on your board. If I was you, I'd be trusting CoreTemp.





NOTE: Orthos, CoreTemp and TAT are all reporting 3376 MHz because they're incorrectly reading the multiplier. It knows it is an E6400 so these programs assume that I am running at 8 X 422 MHz = 3376 MHz when I'm actually running at 7 X 422 MHz = 2954 MHz.
 
unclewebb said:
Here's some interesting numbers.

I have an E6400 but I dropped the multi down to 7 and set the FSB to 422 MHz to simulate your setup. I also set it so CPU-z was reporting 1.312 volts at idle and 1.288 volts while under load running Orthos.

Room temperature when I started was 13C in my basement when I started this test and went up to 14C when I ran the idle test. This is just slightly cooler than your room temperature of 60F (~15C ).

Using CoreTemp & Speedfan the temps under load went as high as 52C and then settled out at about 50C. Idle temp was reported as 20C.

The biggest difference between our setups is the heatsink and fan. Your Zalman 9700 significantly outperforms my Intel OEM heatsink and fan so your temps should be cooler than this and they are. If my computer was operated at your room temperature then my load temp would probably be about 53C. Your load temp of about 47C with a much better cooler makes perfect sense.

Anandtech did a review of the Zalman 9700. On an X6800 at 2.93 GHz their load temp dropped by 20C from 56C to 36C and their Idle temp went from 41C to 28C. Knowing this, it's possible that after you properly apply some AS5 and give it time to setup, your CoreTemp readings might even decrease a few degrees.
http://www.anandtech.com/casecooling/showdoc.aspx?i=2932&p=1

While experimenting I've seen CoreTemp readings of 80C at which point the Intel C2D thermal throttling kicked in to keep the temps from hitting the Intel maximum of 85C. I have no doubt that CoreTemp is reporting the correct temperature on my P5B Deluxe.

I'm also convinced that on your board CoreTemp and the new version of SpeedFan are reporting your real core temperature. The Intel TAT program is reporting about 3C less at full load on my board but seems to be out to lunch on your board. If I was you, I'd be trusting CoreTemp.





NOTE: Orthos and CoreTemp report 3376 MHz because they're incorrectly reading the multiplier. It knows it is an E6400 so it assumes I am running it at 8 X 422 MHz = 3376 MHz when it's actually running at 7 X 422 MHz = 2954 MHz.
Click to expand...

hmm! if more would confirm wut you did i would be even mroe happy... majority tells me to use TAT since they don't beelieve my idle temps
i read coretemp forum and even the owner (thecoolest) maker of coretemp is confused by this too
 
If my testing doesn't convince you then nothing will.

Too many people have been brainwashed into believing that TAT is perfect because it was made by the big Intel corporation but obviously on some boards, like yours, it is not correct. It needs to be updated for proper C2D operation but it hasn't been updated for quite a while.

CoreTemp and Speedfan are simply reading a register within the C2D processor and displaying the result. On my P5B Deluxe you could argue whether CoreTemp or TAT is displaying the true temperature but on your board there is nothing to argue about. The cool room you use your computer in leads to some cool CoreTemp numbers but they certainly make perfect sense compared to what TAT is showing on your board.

CoreTemp has reported my E6400 at every temperature from 20C to 80C without any problems. Your numbers and my testing in a similar environment confirm that CoreTemp is reporting the correct temperatures. I'll have to go and post on his forum and tell him that his program is working as it should.

By the way, did you use AS5 on your processor when you installed it and how did you apply it? I found that the AS5 guide that shows that you should draw a line down the centre of your C2D didn't work very well for me. Covering the centre of the heat spreader where the core is located with a thin even layer was more effective.

OT: Another thing that most users don't want to hear is that I gained absolutely nothing by replacing the Intel thermal paste with AS5. So far after a few days of break-in, the before and after temps are exactly the same. Can anyone find me a link to a reputable review of AS5 in a controlled environment vs the 3 strips of goo that Intel puts on their heatsinks?
 
unclewebb said:
If my testing doesn't convince you then nothing will.

Too many people have been brainwashed into believing that TAT is perfect because it was made by the big Intel corporation but obviously on some boards, like yours, it is not correct. It needs to be updated for proper C2D operation but it hasn't been updated for quite a while.

CoreTemp and Speedfan are simply reading a register within the C2D processor and displaying the result. On my P5B Deluxe you could argue whether CoreTemp or TAT is displaying the true temperature but on your board there is nothing to argue about. The cool room you use your computer in leads to some cool CoreTemp numbers but they certainly make perfect sense compared to what TAT is showing on your board.

CoreTemp has reported my E6400 at every temperature from 20C to 80C without any problems. Your numbers and my testing in a similar environment confirm that CoreTemp is reporting the correct temperatures. I'll have to go and post on his forum and tell him that his program is working as it should.

By the way, did you use AS5 on your processor when you installed it and how did you apply it? I found that the AS5 guide that shows that you should draw a line down the centre of your C2D didn't work very well for me. Covering the centre of the heat spreader where the core is located with a thin even layer was more effective.

OT: Another thing that most users don't want to hear is that I gained absolutely nothing by replacing the Intel thermal paste with AS5. So far after a few days of break-in, the before and after temps are exactly the same. Can anyone find me a link to a reputable review of AS5 in a controlled environment vs the 3 strips of goo that Intel puts on their heatsinks?
Click to expand...

i'm not saying i don't believe your testing on the temperature. but i'm just saying why do most people believe it is right to use TAT? haven't people tested for themselves? or everyone like you said is truely brainwashed. i did see how as5 is supposed to be applied. but what i did when i applied it was put a small dot amount on the center and then put my heatsink on it and then turned the heatsink right and left to smudge it then screw it on. i didn't draw a line like it said i should.

another thing is that i live in SF. it's cold these few days. i have my heater set to 68F. so i dunno if my idle temp is still correct since it could range 18-20c at 68F room temp

posted the test of yours in XS
http://www.xtremesystems.org/forums/showthread.php?p=2034158#post2034158 hope that's okay with you
 
Mine displayed the same discrepancies as yours. At first, I thought TAT is right since the idle temp is more correct at 33C but since I forgot to disable C1E or EIST on my board, it is throttling at 6x and CoreTemp is displaying 19-20C idle, 56-57C load at 3.2 GHz while TAT is displaying 68-69C load. I then thought that TAT is probably a bit old and not optimized for C2D while CoreTemp is. Even SpeedFan 4.32 agree with CoreTemp...

I touched the heatsink during orthos testing after 12 hours and it seems pretty cool so CoreTemp might be right actually.
 
chinked|out said:
i'm not saying i don't believe your testing on the temperature. but i'm just saying why do most people believe it is right to use TAT? haven't people tested for themselves? or everyone like you said is truely brainwashed. i did see how as5 is supposed to be applied. but what i did when i applied it was put a small dot amount on the center and then put my heatsink on it and then turned the heatsink right and left to smudge it then screw it on. i didn't draw a line like it said i should.

another thing is that i live in SF. it's cold these few days. i have my heater set to 68F. so i dunno if my idle temp is still correct since it could range 18-20c at 68F room temp

posted the test of yours in XS
http://www.xtremesystems.org/forums/showthread.php?p=2034158#post2034158 hope that's okay with you
Click to expand...

Something wrong because I just checked both of these on my rig. The stock E6600 speeds Core Temp shows temps are;
Core 0 = 34C
Core 1 = 35C
TAT temps are:
Core 0 = 35C
Core 0 = 36C
I went it at stock to compare to others so overclocking wasn't affecting it. I'm using a Ninja Cooler with the 120mm fan set to low speed for the test. TAT monitors more often than Core Temp, I had speed-ed it up yet. The difference I see is about 1 to 2C.
 
Donnie27 said:
Something wrong because I just checked both of these on my rig. The stock E6600 speeds Core Temp shows temps are;
Core 0 = 34C
Core 1 = 35C
TAT temps are:
Core 0 = 35C
Core 0 = 36C
I went it at stock to compare to others so overclocking wasn't affecting it. I'm using a Ninja Cooler with the 120mm fan set to low speed for the test. TAT monitors more often than Core Temp, I had speed-ed it up yet. The difference I see is about 1 to 2C.
Click to expand...

yes because i'm on allendale core. which ameks a huge difference in temp. the 6600 and above aren't allendales.
 
I am having the same issues with my E4300 and GA965P-DQ6 Mobo.
My TAT temps are typically 12C higher than Coretemp readings.
Full load reading 67C with TAT while Coretemp reads 54C running at 3150Mhz with Water cooling.
I ran the new Speedfan and this ties up with Coretemp.
Here is a link to the coretemp developer forum discussing this very issue.
http://www.alcpu.com/forums/viewtopic.php?p=836&sid=6ebdb09ff910e20e9ce29295572afcb9

Mac.
 
Donnie27 said:
Something wrong because I just checked both of these on my rig.
Click to expand...
On a board with the 965 chipset like a DS3 or P5B, CoreTemp and TAT report very similar temperatures.

The problem is that on other boards TAT is not accurate. Too many people are relying on this program even though it hasn't been updated for the C2D processors and the collection of different chipsets available for the C2D.

People have been doubting CoreTemp because they assumed that TAT was correct but testing shows that TAT is definitely not correct on some motherboards.
 
Quick Q. specs in sig.

Is it 50C high when just having aim & IE open? I'm using TAT. Core temp didn't work. All stock.
 
That's the problem with forums. If the same lie is told enough times, people start to believe it, even if it isn't true.

Now that the real E4300 Allendale is available, some programmers are updating their software to report the Intel C2D properly.
CPU-z, v.1.38 reports the E6300 / E6400 as Conroes, which they really are.

conroeallendalevl6.png


http://techreport.com/reviews/2006q3/e6300-vs-sff/index.x?pg=1
The first of our two contestants is the Core 2 Duo E6300, the humblest of Intel's new Core 2 processors. Unlike its fancier big brothers, the E6300 has only 2MB of L2 cache to share between its two execution cores. You'll find plenty of sources that will tell you the code name for these 2MB Core 2 Duo processors is "Allendale," but Intel says otherwise. These CPUs are still code-named "Conroe," which makes sense since they're the same physical chips with half of their L2 cache disabled. Intel may well be cooking up a chip code-named Allendale with 2MB of L2 cache natively, but this is not that chip.
 
unclewebb said:
That's the problem with forums. If the same lie is told enough times, people start to believe it, even if it isn't true.

Now that the real E4300 Allendale is available, some programmers are updating their software to report the Intel C2D properly.
CPU-z, v.1.38 reports the E6300 / E6400 as Conroes, which they really are.

conroeallendalevl6.png


http://techreport.com/reviews/2006q3/e6300-vs-sff/index.x?pg=1
The first of our two contestants is the Core 2 Duo E6300, the humblest of Intel's new Core 2 processors. Unlike its fancier big brothers, the E6300 has only 2MB of L2 cache to share between its two execution cores. You'll find plenty of sources that will tell you the code name for these 2MB Core 2 Duo processors is "Allendale," but Intel says otherwise. These CPUs are still code-named "Conroe," which makes sense since they're the same physical chips with half of their L2 cache disabled. Intel may well be cooking up a chip code-named Allendale with 2MB of L2 cache natively, but this is not that chip.
Click to expand...


Another proof it's a Conroe is the presence of VT on these processors. Allendales doesn't have any VT support ;)
 
chinked|out: I moved my computer back upstairs today where the room temperature is 19C (66F) and ran another quick full load vs idle test.

Yesterday:
Room Temp 13C
Idle 20C
Load 50C

Today:
Room Temp 19C
Idle 27C
Load 56C

temps19cjb2.png


Room temps went up and the idle and load temps went up by an equal amount.

I ran it at the same 2954 MHz ( 422 X 7 ) and 1.288 volts under load like previously.

At a normal room temperature of 21C (70F) people should be seeing idle temps in the low 30C to 35C range and under Orthos load they shouldn't be seeing much more than 60C to 65C as long as their case is properly ventilated and their heatsink is properly installed.

Those numbers are for an E6300/E6400 with the Intel heatsink and fan, at default volts and 2954 MHz. The extra cache in the E6600 and E6700 will likely make them run a little warmer.

If you have two 8800 cards in SLI mode in a small case then obviously your temps will be higher.
 
chinked|out said:
yes because i'm on allendale core. which ameks a huge difference in temp. the 6600 and above aren't allendales.
Click to expand...

But as far as I know, they use the same sensors and I wasn't expecting you to have the same temps, but readings differences.
 
unclewebb said:
On a board with the 965 chipset like a DS3 or P5B, CoreTemp and TAT report very similar temperatures.

The problem is that on other boards TAT is not accurate. Too many people are relying on this program even though it hasn't been updated for the C2D processors and the collection of different chipsets available for the C2D.

People have been doubting CoreTemp because they assumed that TAT was correct but testing shows that TAT is definitely not correct on some motherboards.
Click to expand...

Thanks for the info!
 
chinked|out said:
just got confirmation regarding conroe and allendales.
Click to expand...
Do you have a date code on your L2 E6300? I believe the original E6300 chips were all Conroe stepping B2 and it's only recently that the L2 E6300 chips have been in production. Thanks for finally clearing things up.
 
unclewebb said:
Do you have a date code on your L2 E6300? I believe the original E6300 chips were all Conroe stepping B2 and it's only recently that the L2 E6300 chips have been in production. Thanks for finally clearing things up.
Click to expand...

batch # is Q648A289
SL9TA

from readin the below. i think i'm confused again. it's a allendale.
here we got some problem again
Cracker said:
Actually, an E6300 can be either a Conroe or an Allendale depending on the stepping with the earlier chips being Conroe and some of the newer ones being Allendale. Here is a good thread on this subject LINK

And for a listing of the s-spec's and steppings to further explain here is an Intel link. LINK L2's are Allendale and B2's are Conroe.
Click to expand...
 
You must log in or register to reply here.
Back
Top