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

If Yoxxy's right there is no throttling at 85C if coretemp is to be believed. What a way to prove a point killing a 6300, but throttling started at coretemp = 100C and the IHS measured 83C. Its know from IHS removal attemps, that the solder on the core melts anywhere from 77C to 85C. Even if the temperatures are taken from the core, throttling has to be lower than at most 85C.

The other thing that makes me think is the idle temps. I've two idles to choose from:

http://img169.imageshack.us/img169/3953/tempshp0.th.jpg

My HSF-Tuniq has an approx thermal resistance of TR=0.18 C/W including some losses. Just using the simplified dissipation equations:

If it idles at 31C (20C ambient) according to ProbeII/Speedfan = 61W
If it idles at 46C (20C ambient) according to coretemp = 144W

Which one is more likely for an idle C2D at 3.6GHz/1.48V?



REF TDP is 65W.

TR = RJC + RCS + RSA

Tj = Ta + Pt*TR

where:
TR = Total Thermal Resistance
Tj = junction temperature (or cpu temp)
Ta = ambient temperature
Pt = total power dissipation
RJC = thermal resistance, junction-to-case, = 0 for this.
RCS = thermal resistance, case-to-sink, = 0.0045 for AS5
RSA = thermal resistance, sink-to-ambient = 0.1775

Quote:

---

Originally Posted by 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:

---

If it's affected by BIOS flashes, then it's definitely messed up. You're not the first person i've heard of who said their CoreTemp readings changed a ton because of a bios flash too.

Quote:

---

Originally Posted by Absolute_0

If it's affected by BIOS flashes, then it's definitely messed up. You're not the first person i've heard of who said their CoreTemp readings changed a ton because of a bios flash too.

---

yeah for sure man

Quote:

---

Originally Posted by Absolute_0

If it's affected by BIOS flashes, then it's definitely messed up. You're not the first person i've heard of who said their CoreTemp readings changed a ton because of a bios flash too.

---

Coretemp says that my T2600es cooled by a TT BT loads at like 50c.

Yea, right....:stick:

HOLY CRAP, most active thread ive ever started :). well, i dont think the IHS is poor, as the heatsink is heating up, and i doubt its even close to that temp in reality, because if it was, i doubt it would run orthos for a few hours ;). SO, i will conclude, at least in my personal case, core temp doesn't work correct. It may be down to specific motherboard/cpu combo's, i dont know, ill leave that up to the coders.

Well to comment on this if Bios flashes are changing what is read that seems to make me believe it is buggered. I have never disagreed that this worked right for Yonah and Merom. TCaseMax is 100 for them meaning 100-Delta = Your Proc temp. Having read the Intel whitesheet for Conroe nowhere does it mention the "TCaseMax" or anything alike to this for Conroe. This would seem to be a mystery number. I think from all the tests with Core Temp it may be a motherboard problem and not so much the programs fault.

http://www.xtremesystems.org/forums/...3&postcount=18

regards

unfortuantly, i think i believe the high temps that core temp gives me. @ 3.44ghz and 1.45v, i was seeing around 63-65c load temps, and mid 40's idle. I noticed during a dual prime, that my 1900xt would go up about 4c in idle. From 40c to 44c. Also my zues 650w wuld start to spin up and fairly warm air would be exhausting out. With my x2 3800 @ 2.7ghz and 1.5v, it would not get this hot, so it sounds like the cpu really is generating alot of heat.

Here's my experience:

I have been priming for 40min at the moment at 2.7Ghz and 1.35V v-core and I get the following temps:

-Speedfan: 55C
-Intel TAT: 66C, 66C
-Core Temp: 67C, 67C

High huh? But keep in mind that room temp has been above 30C for a couple of months now (33C at the moment) and ambient temp in the case is around 38C right now.

EDIT: What's the max temp for a E6600? In this thread I saw that "Max Cover Temp" is 60C. What does this mean?

Quote:

---

Originally Posted by Mr.X

But keep in mind that room temp has been above 30C for a couple of months now (33C at the moment) and ambient temp in the case is around 38C right now.

What does this mean?

---

Means you need to look into getting an air conditioner:p:

Sry, couldn't resist:)

Quote:

---

Originally Posted by Icer

Means you need to look into getting an air conditioner:p:

Sry, couldn't resist:)

---

Hey, you altered my post! :p:

Quote:

---

Originally Posted by Mr.X

"Max Cover Temp" is 60C. What does this mean?

---

Quote:

---

Please note that the 'Max. Case Temp.' is the maximum temperature allowable as measured on the exterior of the processor package in the middle directly above the chip die (under the heatsink). It is not the maximum internal ambient temperature of the system enclosure. The term 'case' refers to the processor's package, not the system case. For that, AMD suggests:

---

Processor Electrical Specifications

It's funny that most of you guy's that keep on saying that Core Temp is screwed didn't already posted screenshots showing Intel's Thermal Analysis Tool readings hosted by nikhsub1. I guess you would trust how Intel is handling the matter.
C'mon show us what you get.

Cover Temp and Case Temps = the same?

60C can't be right though.

Absolute. Did you actually take the CPU off and verify thermal compound contact was good? Please read my post about verifying temps. Don't say it wouldn't 32MB prime 3.73 if it were that high because as I stated in that post I had good pi/3d tests all the way to 3.8Ghz. Even though a major problem existed.

Eric

I'm sick of looking at these high temps no matter where they are coming from. I'm getting a vapochill. ;)

Seriously, where do all these programs get their readings from? Are there many different temp monitoring diodes? Isn't there only one MB/cpu source of cpu temp reads that the software reads? It's not like every different program has it'as own HW diode? So why is the same source of temperature reading so much differently from one progy to the next?

Quote:

---

Originally Posted by LowRun

It's funny that most of you guy's that keep on saying that Core Temp is screwed didn't already posted screenshots showing Intel's Thermal Analysis Tool readings hosted by nikhsub1. I guess you would trust how Intel is handling the matter.
C'mon show us what you get.

---

Look at post #51

Its been around awhile, 2002? Lawrywild posted a Rapidshare a few days back.

I'm not entirely sure its reading the sensor correctly, these interrupts can't be disabled. I'm not placing to much faith in it. After all. it's designed for Mobile Pentium M processors.

http://img140.imageshack.us/img140/6982/tatdp6.jpg

Quote:

---

Originally Posted by xgman

I'm sick of looking at these high temps no matter where they are coming from. I'm getting a vapochill. ;)

Seriously, where do all these programs get their readings from? Are there many different temp monitoring diodes? Isn't there only one MB/cpu source of cpu temp reads that the software reads? It's not like every different program has it'as own HW diode? So why is the same source of temperature reading so much differently from one progy to the next?

---

My guess would be that the data the software gets is in some other form than temperature and the software calculates the temp. Each might calculate it differently/incorrectly.

Quote:

---

Originally Posted by ewitte

Absolute. Did you actually take the CPU off and verify thermal compound contact was good? Please read my post about verifying temps. Don't say it wouldn't 32MB prime 3.73 if it were that high because as I stated in that post I had good pi/3d tests all the way to 3.8Ghz. Even though a major problem existed.

Eric

---

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...

---

Quote:

---

Originally Posted by Absolute_0

readings in the 80s can't be right though. Throttling and motherboard protection turn-off occurs there.

I'm on watercooling, for 24/7 i use 1.4v actual, coretemp gives me load temps up to 58c. I checked contact and it's perfect, i got a nice flat IHS.

Pentium D (HOTTER THAN CONROE) with this many volts loaded in the low 40s. You know something is wrong...

---

yep, as i've said before, GOOD contact, better than what i anticipated because of everyone's IHS problem. I don't have the IHS problem.

One question about CoreTemp:

If the readings it provides are so accurate, then surely CPU throttling temperature (which is controlled by the BIOS isn't it) would actually be completely useless, because the CPU temperature shown in the BIOS is usually at least 20 degrees lower than CoreTemp seems to say - so when the BIOS actually thinks your CPU has reached 85C its actually 100C+ ?

Quote:

---

Originally Posted by Durzel

One question about CoreTemp:

If the readings it provides are so accurate, then surely CPU throttling temperature (which is controlled by the BIOS isn't it) would actually be completely useless, because the CPU temperature shown in the BIOS is usually at least 20 degrees lower than CoreTemp seems to say - so when the BIOS actually thinks your CPU has reached 85C its actually 100C+ ?

---

The cpu has it own internal throttle, THERMTRIP# or one of the others, haven't looked to closely. But if you look at what Yoxxy is saying, it does throttle, but at a coretemp reading of a 100C. The motherboard should have its own overtemp based on its thermistor, looking at the values the motherboard doesn't read the core temps in bios.

Quote:

---

Originally Posted by Durzel

One question about CoreTemp:

If the readings it provides are so accurate, then surely CPU throttling temperature (which is controlled by the BIOS isn't it) would actually be completely useless, because the CPU temperature shown in the BIOS is usually at least 20 degrees lower than CoreTemp seems to say - so when the BIOS actually thinks your CPU has reached 85C its actually 100C+ ?

---

Yes... IF core temps readings were accurate. But from everything I've seen using it and everyone I've talked to (100c at idle, temps changing with different BIOSes, etc) its a giant load of BS and the worst temp monitoring app I have ever used.

Not sure if this helps or just confuses.But I'll keep reading the thread for more input before I say which one to trust.Maybe I'll just put that ol rectal thermometor on the CPU and see if shows the true temp.
:scope:

As far as I can see, some of you believe, that the temperatures shown by coretemp are OK while the others say absolutely not. Well, it's very easy to find it out. I wasn't taken aback by the results. I'll show you two screens - the first one taken while using Gigabyte DQ6 and the second one while using Asus P5B deluxe (both P965 mobos). In both cases I used the same CPU, same cooling (Ninja + Fander 120mm 1180RPM), OCZ Ultra 5+, room temperature 24 celsius (while testing P5B) and 23 celsius (while testing DQ6).

GIGABYTE DQ6
http://www.xtremesystems.org/forums/...9&d=1156537216

ASUS P5B DELUXE
http://www.xtremesystems.org/forums/...0&d=1156537216

As far as the results are concerned, while using my Asus P5B, the CPU teperature was 6-7 celsius lower, than using the DQ6. So the question is: is coretemp accurate? No way.

Hmm room temp +1c on P5B, yet CoreTemp shows -6c...

I agree with you.