Real Temp - New temp program for Intel Core processors

[Richard Dower]

Uncle...here is the info you wanted:



[URL=http://g.imageshack.us/g.php?h=521&i=capturepd5

[Richard Dower]

As per Uncle's advise...right clicking on the program fixed the 0.5x multi issue, it's now displaying 3400MHz...cheers.

[sirheck]

If you have a DTS sensor that is sticking will restarting the compy (unstick)
it?

I have core 0 sticking on my brothers x3110 while F@H.
I dont want to restart it just yet, (want to finish the WU first).
(Stanfords servers are all messed up now and its hard to get a WU
from them).

I guess i will find out tomorrow.

[unclewebb]

Most DTS sensors that get stuck get stuck as the sensor cools down when you switch from full load to idle. One or more sensors will start reporting the exact same number even though your CPU is cooling down. This can effect all sensors, 65nm and 45nm, but it seems to be a much bigger problem with the new 45nm chips.

Unfortunately, there is no solution and based on user feedback, Intel does not consider this to be a big enough problem to justify replacing your CPU for you. As long as your processor can crunch some numbers without over heating or melting then it is doing what it was designed to do.

[DHP]

I tried to find an answer in the thread, but is Realtemp good to use for the QX9650 C0? It gives me about 55-60C stressed @ Quad PRIME. Coretemp gives me 10C more.... Which is not very good imho.

[jason4207]

didn't mean to offend you UncleWeb I read the 1st 10 pages & understood most of your work but couldn't find out if it worked on Vista 64 ,I like your program as it is running fine on my 64 sys

i'm getting 33,33,32,35 in core temp right now & with your program 28,28,27,30 sooo maybe just idle about 2 pages of IE open right now & both programs open

your programstest >
test went core 0 = 10
core 1 = 9
core 2 = 9
core 3 = 9

all this on rig#1 in my signature

agin sorry if i offended you ,

i have alot of PC projects going all at once here and the weekend is my only time to do it .As I drive all week 60 hrs plus & was just asking if someone could save a guy some time with a quick post about vista64 & your program ..

I tried to find an answer in the thread, but is Realtemp good to use for the QX9650 C0? It gives me about 55-60C stressed @ Quad PRIME. Coretemp gives me 10C more.... Which is not very good imho.

I'm pretty sure the very first post clearly outlines the discrepancy you guys are seeing. It has to do w/ the chosen Tjmax. CoreTemp has chosen different values than RealTemp. If you're not happy w/ the chosen Tjmax you can change the value via the ini file.

Or you can just look at the 'distance to Tjmax' being reported which comes directly from the DTS sensor in the CPU, and is the same across every temp program.

[gx-x]

Hello UncleWebb

First - Thank you for this great app !

second: when I set idle to -2 in .ini file, temp drops by 4 instead of 2 -1 works ok. To clarify: with no calibration my CPU idles @ 40C with -1 it says 39, with -2 it says 36C

PS. If understand correctly, for my e8200 I need to put -2 into ini file to calibrate it properly and compensate for DTS ? (I read your first post but somehow I am still not sure :P )

Thank You once more !

[DHP]

This really bugs me!



Which sensors are right? Which ones are wrong?

This really sucks!

I just remounted the cooler to see if I had bad contact: it was mounted very good and gave me similar temps as shown...

Good job Intel, 800 euro CPU's with defective sensors!

Maybe I should RMA it.

Edit: before I remounted it was 2Hrs PRIME stable @ 3,8GHz 1,3V, now its giving me an error after 8 minutes...

[xtclocker]

This really bugs me!



Which sensors are right? Which ones are wrong?

This really sucks!

I just remounted the cooler to see if I had bad contact: it was mounted very good and gave me similar temps as shown...

Good job Intel, 800 euro CPU's with defective sensors!

Maybe I should RMA it.

Edit: before I remounted it was 2Hrs PRIME stable @ 3,8GHz 1,3V, now its giving me an error after 8 minutes...

Should RMA it, and check the temps again.

[DHP]

Remounted again... Pulled the CPU out of the socket and put it in again... Load temps in Realtemp seem to be a bit lower... 52C to 56C stressed on 2 cores, stressed 38C on the other 2 @ 3800MHz.

Now using Thermalright paste instead of AS5. Maybe the AS5 was too old.

Lets see what happens now. I really want to hit 4GHz... This is my second QX9650. The previous one only did 3,6GHz, which was extremely bad.

But the sensors on that one worked fine. It idle'd @ 30C in Coretemp @ all 4 cores.

If this one fails PRIME @ 1.3V again i'll RMA it and get a E8500 for the time being.

LOL. Isn't this a great hobby? One has to be more patient overclocking and building PC's than you have to be @ fishing.

[phill]

I've started to ignore the temps as I was told from the company I bought my QX9650 from that as long as its reporting temps in the bios it will throttle down if it gets too hot and so will be fine (oh they added as long as its run at stock settings... )

£630 on a CPU and they want me to run it at stock settings???? They must be mad!

[phill]

I dont understand my QX9650 as I can run it at Stock speed at around 1.08vcore full load no problems, yet when I start to overclock it, the damn thing needs stupid volts to get to go anywhere..

I'm running at 4Ghz and it appears stable at the moment but I have to set the vcore in the bios at 1.5v and even then its only drawing off 1.38vcore in speedfan (its the only program that reads the temps the same as the bios does) when under load!!

If I try to run anything further it does require some big voltage.. I managed a run at 4.4Ghz which required me to set it at 1.6750vcore (or so) in the bios, but even then under a load, it was pulling just 1.53vcore.. This board gives me the most applauing Vdroop and trying to overclock on it (its a 780i XFX board) isnt the nicest thing in the world...

One of the guys mentioned in a thread I'd posted said that he could get his QX to 4Ghz with just 1.26vcore!! I was drooling when I heard this and when its come to mine, it just sucks!!
Mind you I dont think the temp probes help a heck of a lot...

[kadir_slayer]

I would just like to know when the cpu throttles does it mean that it really gets that hot or it throttles just because the sensors tells it to throttle even they are wrong?? ı mean can the cpu throttle because of broken sensors even though it is not that hot in real??

[phill]

I've no idea about it throttling as I dont know what the CPU is limited too for temps... The maximum I've seen with my limited readings and cooling is about 70c... I decided to give up at that point...

[ChrisZ]

I dont understand my QX9650 (...) required me to set it at 1.6750vcore (or so) in the bios, but even then under a load, it was pulling just 1.53vcore.

I think you are on thin ice now...
http://processorfinder.intel.com/det...px?sSpec=SLAN3

[Gir92]

I'm liking my Q6600 temps a lot better now. Confirmed working for Vista x64

This SS was taken at idle with ambient temp at 19-20C. D5 vario and 480GTX rad, apogee GT block.

[phill]

I think you are on thin ice now...
http://processorfinder.intel.com/det...px?sSpec=SLAN3

I'm not sure I follow you? Are you meaning this:

VID Voltage Range: 0.85V – 1.3625V

Mines not running anything more than that 24.7 regardless.. I only used the 1.53vcore for a very short period

[Demo]

Where did you guys get version 2.49, the latest posted is 2.41

[loonym]

Where did you guys get version 2.49, the latest posted is 2.41

http://www.fileden.com/files/2008/3/...alTempBeta.zip
Grab that then simply replace the exe in realtemp folder.

[unclewebb]

I only used the 1.53vcore for a very short period

That sounds a lot like the guy that only forgot to use a condom once.

With 45nm chips, it may only take once to permanently damage one.

[phill]

Whats one of them!?

I dont know how the other guys run 2v through their's then! A guy I was talking to on another forum said he was running his at 4.2Ghz 24.7 and I think he was using around 1.4 to 1.5vcore perminately.. I'm under 1.4 under a load but idle seems to higher it up for some reason??

[Richard Dower]

Is the program going to final yet?

[Cy4n1d3]

Don't know if this has been asked before..
Does the program detect throttling due to signals sent by the cpu or simply by comparing temps to tjmax?

[unclewebb]

Does the program detect throttling due to signals sent by the cpu or simply by comparing temps to tjmax?

Good question. The PROCHOT# flags that will be displayed in RealTemp are based on reading that specific data from the CPU. For the techy types, this info comes from reading MSR 0x19C. Bits [22:16] contain the 7 bits of DTS data and bits 0 and 1 in that same MSR contain the flags for PROCHOT#.

You can read all about what they mean by downloading volume 3A of the Intel manuals and reading this chapter:

13.4.2.5 Thermal Status Information

This feature was tested and it definitely works. The interesting thing is that the History flag gets set just before the Status flag gets set as shown in this picture and throttling gets flagged just before DTS reaches zero.



That's a good thing since if there is any throttling at all, this bit will remain set until you reboot. You don't even need to have RealTemp running since this bit is set within the processor and will remain set even after the processor cools down.

When you get more carried away then they'll all light up.



But if your brother thinks he can turn off RealTemp to cover up the evidence then you can still catch him by starting up RealTemp.



The evidence is there even though RealTemp doesn't show any unusual temperatures in the Maximum columns.

Richard Dower: This round of beta testing is pretty much done so RealTemp 2.5 should be available by tomorrow.

[Cy4n1d3]

Ah ok, I see.
Was just curious how my cpu was able to reach throttling-temp with ~1.35 volts on air cooling