I think core temp.... is officially buggered.

Printable View

Show 100 post(s) from this thread on one page

08-24-2006, 04:34 PM
G H Z

Quote:

Originally Posted by Yoxxy

But at 85 there is no throttling whatsoever.

85 as read by what ;)

Quote:

Originally Posted by ted3

i suggest to use a permanent marker and write 28 Celsius beside your screen and sit back and enjoy your super-cool system.

That would be as/more accurate than the numbers we get from some boards.
08-24-2006, 04:37 PM
G H Z

Quote:

Originally Posted by dinos22

i just don't understand how everything the same and two different motherboards (both DS3s) could have such major variation in reported temperature in Coretemp i mentioned earlier

Let's see these screen shots.
08-24-2006, 04:42 PM
dinos22

Quote:

Originally Posted by G H Z

Let's see these screen shots.

i've sold off both DS3s and that CPU and another CPU after that

i'm starting from scratch today

hence no screenshots but i wouldn't lie about this
08-24-2006, 05:10 PM
nikhsub1

1 Attachment(s)

OK guys, I downloaded the INTEL Thermal Analysis Tool... Granted it is designed to be used with MOBILE CPU's it seems to be in close agreeance with coretemp... It has a cool logging feature and you can load one or both cores to check temps - not a stability tester. You can even on demand throttle each core, pretty cool! As I watch both monitors, it seems coretemp sometimes reads 4-6C hotter than the Intel software... I like them both but check it out, I have uploaded it to my server as it is a PITA to find...

http://www.anonforums.com/builds/TAT.zip
08-24-2006, 05:27 PM
Absolute_0

3.65 Ghz, and 1.55v, you get 51c load by CoreTemp.

Guess what i get?

http://img.photobucket.com/albums/v7...temp-sucks.jpg

If CoreTemp is right, i want to know what's causing the 19c difference, if my room temp is 72F right now, while yours is higher...
08-24-2006, 05:30 PM
nikhsub1

Absolute, what is your cooling system, specs please. Also, I have my WB CRANKED tight to the CPU, no springs, bolts to the block, probably 100lbs of force.
08-24-2006, 05:35 PM
Absolute_0

Swiftech apogee lapped
MCP600
120.2 heatercore shrouded with 2 120mm U1As

Contact is perfect

I also don't use the mounting gear, i have 4 screws with those brass thumb nuts tightened down as tight as possible.

If my contact was bad, i wouldn't be dual prime stable at 3.73 Ghz...
08-24-2006, 05:40 PM
Pjoeloe

yup, see my last reply in http://www.xtremesystems.org/forums/...103639&page=13 :

"coretemp alway`s does like 10 degrees above the speedfan temps "

I think the prog is crazy :rolleyes:

Quote:

Originally Posted by dinos22

how do you know CoreTemp is reporting accurate temps

Quote:

Originally Posted by vintage_guitar

because he's pissed and triple posted

:D
08-24-2006, 06:16 PM
nikhsub1

Quote:

Originally Posted by Pjoeloe

yup, see my last reply in http://www.xtremesystems.org/forums/...103639&page=13 :

"coretemp alway`s does like 10 degrees above the speedfan temps "

I think the prog is crazy :rolleyes:

:D

So then the Intel utility is crazy too? Keep yer head in the sand.
08-24-2006, 06:18 PM
nikhsub1

Quote:

Originally Posted by Absolute_0

Swiftech apogee lapped
MCP600
120.2 heatercore shrouded with 2 120mm U1As

Contact is perfect

I also don't use the mounting gear, i have 4 screws with those brass thumb nuts tightened down as tight as possible.

If my contact was bad, i wouldn't be dual prime stable at 3.73 Ghz...

ERr, 31 mins of prime is not proven stable, but W/E. With your gear compared to my gear, there would be a sizeable difference in temps... I stand by ~7-10C.
08-24-2006, 06:25 PM
SturmoV

My CoreTemp results are pretty much in agreement with what TAT is showing.

Thus my 6600 @ 3.6 GHz ~ 1.48 Vcore tops out at 66-68 Celsius with both cores loaded, on Scythe Ninja +. Dual prime unstable (puter runs, but one of the tests shuts down with "hardware failure" error)
Asus Probe II shows 14 degrees less, so does SpeedFan. So I'd say they're full of :banana::banana::banana::banana:.
This proc is one hot puppy, that's for sure! Gonna keep it 3GHz on stock voltage until I put together a WCing system...
08-24-2006, 06:27 PM
Absolute_0

Yeah, up to 10c difference between good watercooling and superb watercooling... and the other 9c?!?!

That program is crap, my head certainly isn't in the sand for questioning temps when they make no sense and lack consistency.

Head in the sand is blindly believing the program.
08-24-2006, 06:31 PM
aashburns

CoreTemp simply reads a value via a serial interface from the Conroe core. Intel embeds digital diodes into the hottest parts of the die - if you have a more accurate way of measuring the temperature then speak up...otherwise please stop making ignorant comments.
08-24-2006, 06:34 PM
nikhsub1

Quote:

Originally Posted by Absolute_0

That program is crap, my head certainly isn't in the sand for questioning temps when they make no sense and lack consistency.

Head in the sand is blindly believing the program.

That comment was not directed at you :cool:
08-24-2006, 06:41 PM
pcdoc1

I don't have a horse in the race and really don't give a shyt: Asus probe, AI Booster, and Core Temp report exactly the same on a P5B Deluxe...
08-24-2006, 06:59 PM
fhpchris

Quote:

Originally Posted by aashburns

CoreTemp simply reads a value via a serial interface from the Conroe core. Intel embeds digital diodes into the hottest parts of the die - if you have a more accurate way of measuring the temperature then speak up...otherwise please stop making ignorant comments.

Yes, measure the difference by measuring power consumption.

Heat that is given off by the cpu must actually be electricity pulled from the wall socket, so measure the ammount of electricity it uses :)

You could even compare this to another LGA775 cpu in the same system :)

Once you know the delta between idle and load, it will be much easier to understand how much power the system is consuming...
08-24-2006, 07:00 PM
Pjoeloe

there is http://www.skyaid.org/images/heads_in_sand.jpg
08-24-2006, 07:01 PM
Pjoeloe

a button http://www.sheffieldgreenparty.org.u...ates/sand.jpeg
08-24-2006, 07:01 PM
Pjoeloe

they call it http://www.affordablehousinginstitut...sand_small.jpg
08-24-2006, 07:02 PM
Pjoeloe

the edit http://www.dealingtime.com/upload/fi...0in%20sand.jpg
08-24-2006, 07:02 PM
Pjoeloe

button http://guy-sports.com/fun_pictures/stand_on_head.jpg

sorry :D

conclusion : core temp works correct with some asus boards and some not , gigabyte same story
08-24-2006, 07:09 PM
ted3

Core Temp and TAT, gotta love them, seems to be close to perfect for my system. Just thinking about the crazy readings i got from various MBs over the years makes me sick. I never forget the Abit 945 board that insisted on reporting 28 celsius idle on a D830. LoL, what a joke, the only way to get a D830 to that temp on air cooling was to shut down and wait at least an hour. I can understand that ppl who actually used to believe those MB reported temps has a hard time adjusting to this "new" temps.

However i must agree that some ppl here makes valid points to question the accuracy of CT, but it could just as well be a problem with the temp-circuits in the CPU.
08-24-2006, 07:16 PM
Btrice

Quote:

Originally Posted by fhpchris

Yes, measure the difference by measuring power consumption.

Heat that is given off by the cpu must actually be electricity pulled from the wall socket, so measure the ammount of electricity it uses :)

You could even compare this to another LGA775 cpu in the same system :)

Once you know the delta between idle and load, it will be much easier to understand how much power the system is consuming...

That only works if the conversion of consumed power into heat is linear (or you know the shape of the trend, which is unlikely unless you designed the bugger).
08-24-2006, 07:26 PM
Ad1tya

Hmm...... comparision between 3 different progs measusing the temparatures:

http://img174.imageshack.us/img174/2397/tempsry5.jpg

this is on a T2300 yonah in a laptop.

Core Temp is reading it using the Intel Prescribed method. End of Story....
08-24-2006, 07:39 PM
new-b

@Absolute_0..... for me Core Temp seems to have been affected by what bios I use.

IIRC Before flashing to 1334, Core Temp was around 10-15c higher than IDCC.
After flashing the temps are exactly the same Idle or load. Core Temps readings changed not IDCC's....... strange. :stick:

http://img.photobucket.com/albums/v1.../coretemp1.jpg http://img.photobucket.com/albums/v1.../coretemp2.jpg

Show 100 post(s) from this thread on one page

All times are GMT -8. The time now is 06:10 PM.

XtremeSystems