COOL! New version of Core Temp

coo-coo-clocker

Limp Gawd
Joined
Sep 6, 2007
Messages
183
The 0.96 version of core temp is now available!
Check their website for details.
I just tried the new version on my conroe and it shows much lower temps that the 95 beta version was showing...
 
Cool - thx for posting.

Just downloaded it and . . . same temps as before for me :)
 
Same temps here, too. :duh: ;) But I also rigged it to work under Vista x64 on my laptop, so I'm happy with it. Hope v1.x has better x64 support.
 
How accurate is this program?

uGuru says I'm at 57C CPU temp
Core Temp says I'm at 62-68C, keeps fluctuating.
 
I don't know. I'm seeing the same thing. This version has way different temperatures. About 15-16C higher than the old CoreTemp and way higher than every other program I can find, including Linux and the MB's BIOS. I think CoreTemp is wrong unless everything else is measuring Tcase instead of the core (I don't think so though).
 
Still no love for Vista 64, I'll stick with HWMonitor.

Thanks for the heads up OP.
 
On my A64x2 it shows the temps for both cores and they tie in with putting work on each core. I've never seen a program report what appear to be both actual core temps before on an A64 x2 chip.
However one core seems way too high under load.

Asus probe = 54C load, Coretemp = 59C and 71C.
Asus probe = 38C idle, Coretemp = 37C and 39C.

Given that the idle temps are very close, I'm inclined to believe there is some truth to what it is tellling me about the load temps.
Whether the values are exactly right isnt a big concern as it is stable but maybe one core doesnt have a good contact with the heat spreader.

To verify this I used Prime 2004 on the cores one at a time.
With Core 0 at full load, the temp increases and both cores show almost exactly the same temp, 45 and 46C, Asus Probe reports 45C.
With Core 1 at full load, Core 0 is 49C, Core 1 is 67C, Asus Probe reports 47C.
It does seem to show both core temps and it does look like one isnt being cooled very well!

Maybe my chip (4200+ x2 Toledo socket 939) can clock a lot higher if I sort this out.
I'm going to look at removing the heat spreader safely to put some Arctic Cermique under it.
Alternatively, as my heatsink wont clamp tight without the Heat Spreader, I will look to see if an easy mod can be done so I can put the heatsink directly on the core as this for sure will net lower temps and hopefully better clock :)
Then I will be able to tell if Coretemp really is getting it right.
 
On my A64x2 it shows the temps for both cores and they tie in with putting work on each core. I've never seen a program report what appear to be both actual core temps before on an A64 x2 chip.
However one core seems way too high under load.

Asus probe = 54C load, Coretemp = 59C and 71C.
Asus probe = 38C idle, Coretemp = 37C and 39C.

Given that the idle temps are very close, I'm inclined to believe there is some truth to what it is tellling me about the load temps.
Whether the values are exactly right isnt a big concern as it is stable but maybe one core doesnt have a good contact with the heat spreader.

To verify this I used Prime 2004 on the cores one at a time.
With Core 0 at full load, the temp increases and both cores show almost exactly the same temp, 45 and 46C, Asus Probe reports 45C.
With Core 1 at full load, Core 0 is 49C, Core 1 is 67C, Asus Probe reports 47C.
It does seem to show both core temps and it does look like one isnt being cooled very well!

Maybe my chip (4200+ x2 Toledo socket 939) can clock a lot higher if I sort this out.
I'm going to look at removing the heat spreader safely to put some Arctic Cermique under it.
Alternatively, as my heatsink wont clamp tight without the Heat Spreader, I will look to see if an easy mod can be done so I can put the heatsink directly on the core as this for sure will net lower temps and hopefully better clock :)
Then I will be able to tell if Coretemp really is getting it right.

Asus probe reports CPU temps. core temp is well the Core temp's, two deffrent temp's.
 
I have an E4550 @ stock with a stock cooler. With the new version of coretemp I have a tj. max of 100 instead of 85. So my idle temps are now 42-44c. After 10mins of orthos my temps were 65-67c. Are these temps too high? Only thing I did wrong that I know of is I applied a bit of Artic Silver on the cpu not knowing the stock cooler already had some thermal paste on it. I'm such a nub when it comes to these newer chips. I'll probably reseat the cooler on my day off Wednesday. Any tips on how to clean the chip off with out taking the mobo out? My case is a pain in the ass to take everything out.
 
Asus probe reports CPU temps. core temp is well the Core temp's, two deffrent temp's.

Yep I realised that, if indeed there really are sensors for each core on A64 x2 chips.
I'm trying to prove whether or not there really are 2 core sensors on an A64 x2 as prior to this I thought there was only one.
Should this be true, one of my cores needs better heat spreader contact.

If there is only one sensor, it should show the same temp on different programs or at least follow the same curve as temp increases.
That is what makes me think its probable there are 2 sensors.
 
I have an E4550 @ stock with a stock cooler. With the new version of coretemp I have a tj. max of 100 instead of 85. So my idle temps are now 42-44c. After 10mins of orthos my temps were 65-67c. Are these temps too high? Only thing I did wrong that I know of is I applied a bit of Artic Silver on the cpu not knowing the stock cooler already had some thermal paste on it. I'm such a nub when it comes to these newer chips. I'll probably reseat the cooler on my day off Wednesday. Any tips on how to clean the chip off with out taking the mobo out? My case is a pain in the ass to take everything out.

Just pull the heatsink off and use some alcohol on a Qtip to clean the chip, then reapply and reattach the heatsink. Should be no need to remove the MB. And my E4300 loads up at 65C so you're not out of the ballpark with your temps on the stock cooler.
 
Just to weigh in on this coretemp being off the mark. I run both Everest and CoreTemp and both are very close to each other if not dead on. And if I use SpeedFan program and add the required 15c to each temp listed I get about the same readings as well.

Unknown why others having issues.
 
Exactly the same readings here, 4 cores with a 3 degree difference as it always has been.
 
IMO the temps in coretemp 0.96 for my cpu (e2180) and probably other MO stepping intel cpus are incorrect, I put mine at 1.2 v and 1.6ghz and it read a idle temp of 38-39c and load of 51 which to me seems very high. When read with all other temp programs it is about 15 degrees cooler. I understand that it was misreading the Tj max before at 85c and now it reads 100c, but adding 15 degrees to the overall temps doesn't add up..

temps16ghz.jpg
 
Just pull the heatsink off and use some alcohol on a Qtip to clean the chip, then reapply and reattach the heatsink. Should be no need to remove the MB. And my E4300 loads up at 65C so you're not out of the ballpark with your temps on the stock cooler.

This is a perfect example of the misreading, your cpu is overclocked alot and on a little higher voltages, yet it is about the same temp as his at stock speed and voltage. I have a feeling that his cpu is not that hot and the new coretemp is just misreading MO stepping cpus. But if he is going to re-apply his HSF then I'll be interested in the results.
 
i like coretemps minimize feature. Seeing temps in the lower corner is just so cool. If HWmonitor can do this, I'd soley use this program. Maybe it can be done and I haven't figured it out. If it can, someone show me please.
 
When read with all other temp programs it is about 15 degrees cooler.

temps16ghz.jpg

TAT is not made for the current line of Intel desktop cpu's and thus, reports incorrect temps. Coretemp is the most accurate and you would be best to believe it. That easy tune looking software is also notoriously incorrect.

The best way to use coretemp is to go into the settings and tell it to report the remaining temp before the max is reached. That temp is where the cpu will start to throttle, if you keep the chip 10c or more below that, your cpu will be happy.
 
Yup, the new M0 100C Tjunction reference shows just much the E21xx die to IHS thermal compound blows...or maybe it isnt 100C and really is 85C spec. Intel isnt telling how to read temps, that's for sure. I kinda wonder though, I have a E2180 lapped with a lapped Tuniq & Coolabotory LM running 3.0Ghz at 1.2875vcore and Orthos load is ~70C, before the 100C reference it was high 50's very low 60's under Orthos. From experience with this cooler and my E6600 running various speeds and vcores it seems fishy. But keeping in mind that the M0 stepping CPU's no longer have a solder based die to IHS interface one has to wonder how thermal transfer is working out. Anyone planning on yanking their E21xx IHS off and getting a feel for what the temps are like running that way? :p
 
Same temps here, too. :duh: ;) But I also rigged it to work under Vista x64 on my laptop, so I'm happy with it. Hope v1.x has better x64 support.

http://www.alcpu.com/forums/viewtopic.php?t=211

In Vista x64 Microsoft decided that it was time to make developers pay. If your application uses a Kernel mode driver, you must certify it using a special tool. This certification costs you around $400 per year and unless you get your driver certified, Vista x64 will refuse to run it.



So, I don't think the developer will ever pay the fee. However, it seems the SpeedFan guys either got around this issue, or bought the "special tool."
 
Yup, the new M0 100C Tjunction reference shows just much the E21xx die to IHS thermal compound blows...or maybe it isnt 100C and really is 85C spec. Intel isnt telling how to read temps, that's for sure. I kinda wonder though, I have a E2180 lapped with a lapped Tuniq & Coolabotory LM running 3.0Ghz at 1.2875vcore and Orthos load is ~70C, before the 100C reference it was high 50's very low 60's under Orthos. From experience with this cooler and my E6600 running various speeds and vcores it seems fishy. But keeping in mind that the M0 stepping CPU's no longer have a solder based die to IHS interface one has to wonder how thermal transfer is working out. Anyone planning on yanking their E21xx IHS off and getting a feel for what the temps are like running that way? :p

Yea, the new readings can't be right, I had my [email protected] and only 1.2v core it should not be showing 38-39deg at idle, it should only be a few degrees above ambient. Same with the guy that had the e4500, his temps at stock were to high. Unless we all installed our HSFs incorrectly.

Just going by reviews for my HSF, my temperatures seem inline at 30c idle and 51-53c load with a 33% overclock.

look at this review he has a quadcore @ 3600 and 1.39vcore and its at 69c. My cpu is hitting 67c at only 2.66ghz and 1.3vcore with the new coretemp, same HSF.
http://www.guru3d.com/article/cooling/483/5
 
Y
look at this review he has a quadcore @ 3600 and 1.39vcore and its at 69c. My cpu is hitting 67c at only 2.66ghz and 1.3vcore with the new coretemp, same HSF.
http://www.guru3d.com/article/cooling/483/5

Keep in mind YOUR ambient temps, case, etc. Reviews are typically done in very cool environments, with no case.

Again, use coretemp, set to show the DELTA to Tjunction Max.....it's way more accurate that way.
 
I have done some more research on this and the M0 stepping CPU's like the E2xxx series do in fact have a 100C Tjunction Max (looking at the actual Intel datasheet). This may mean the new CoreTemp is actually accurate and everything else is wrong (sort of*).

* Keep in mind that the Tcase temp is going to be lower than the core value. It may be 10C-15C lower and this is what the Intel thermal specification for your processor is based on.
 
Well I cleaned off my E4500 and stock cooler. Reapplied some Artic Silver and popped the cooler back on. Ran Orthos for 10mins and my temps remained the same under load. 65-67c. Granted I didn't give it time for the "break in" period so temps may drop a few degrees. Looks like this chip is running warmer than I thought it was if these new temps from Core Temp are correct.
 
Well I cleaned off my E4500 and stock cooler. Reapplied some Artic Silver and popped the cooler back on. Ran Orthos for 10mins and my temps remained the same under load. 65-67c. Granted I didn't give it time for the "break in" period so temps may drop a few degrees. Looks like this chip is running warmer than I thought it was if these new temps from Core Temp are correct.

IMO, it's more likely that the temps are being read wrong in the new coretemp and your actual load is 10-15 degrees below that, but I'm not positive.
 
IMO, it's more likely that the temps are being read wrong in the new coretemp and your actual load is 10-15 degrees below that, but I'm not positive.

Easiest way to find out is change coretemp to the "delta from max temp" setting, you might not know the actual temp, but you will know how close it is to the max. That reading is right off the chip, so you know it is correct.
 
I'm getting odd readings with CoreTemp on my E4500 as well - 45C while idle, up to 60C under max load.
This is all at stock speeds right now, 1.245V.

SpeedFan is reporting between 10C & 15C below that, and it's temperatures actually seem to make sense.

BIOS reads around 29C after a restart, which seems to line up with SpeedFan's 30C at idle.
 
The E4500 has a Tjunction Max temp of 100C (temperature at which the CPU shuts down). I think most software is incorrectly assuming some of these chips have a 85C Tjunction Max and that's why it shows a temp 15C lower. The old CoreTemp had this problem.

See, the temperature is reported from the Digital Thermal Sensors as a delta value from Tjunction Max and if you use the wrong max value and try to calculate the temperature it will also be wrong.

I think SpeedFan is not using DTS/Tjunction Max, it uses a Thermal Diode which is measuring Tcase. The problem is that particular Tcase value is adjusted by the motherboard BIOS and may be incorrect. The DTS/Tjunction measurement is calibrated by the factory for that specific CPU.

The E4500 has a 10C (+/- 3C) delta between Tjunction Max and Tcase on the DTS according to Intel. So if CoreTemp is showing 60C then the Tcase is about 50C. The max Tcase for the E4500 is 73C or so. The E4500 will start to throttle as it approaches the 100C Tjunction Max.

Confusing but I highly recommend reading this:
http://www.tomshardware.com/forum/221745-29-core-temperature-guide
 
I have a quick question. Does anyone know the specs for Intel's stock cooler? In the bios it says the fan is running at 1700rpm. I would like to have it running at 100% but my mobo (ASRock 4CoreDual-SATA2) doesn't have much control over these kinds of settings. I just have an option to disable or enable quiet fan. As of right now I have it disabled. Is 1700rpm running at 100%?
 
I have a quick question. Does anyone know the specs for Intel's stock cooler? In the bios it says the fan is running at 1700rpm. I would like to have it running at 100% but my mobo (ASRock 4CoreDual-SATA2) doesn't have much control over these kinds of settings. I just have an option to disable or enable quiet fan. As of right now I have it disabled. Is 1700rpm running at 100%?

If you turn off fan control in the CMOS it will run at 100%.
Make sure you havent got a util in windows also trying to control the fans speed.
 
If you turn off fan control in the CMOS it will run at 100%.
Make sure you havent got a util in windows also trying to control the fans speed.

That's what I thought. 1700rpm just seems a bit slow for 100%. I have no utilities in windows controlling fan speeds so thats not an issue. 1700rpm it is I guess until I get a better cooler.
 
Easiest way to find out is change coretemp to the "delta from max temp" setting, you might not know the actual temp, but you will know how close it is to the max. That reading is right off the chip, so you know it is correct.


I have been trying to tell these guys that......if you use coretemp that way, it reports the SAME across all revisions of cpu, since it is taking the digital data direct from the in die sensor, and reporting how close you are to the tmax temp, no matter if your tmax is 85c or 100c, the delta to it is the SAME.........that way you know for sure how hot your cpu is....mine at idle...

6ym0kk1.jpg
 
Back
Top