Real Temp - New temp program for Intel Core processors

[jas420221]

I can help with that also.. I will report back this evening.

[phill]

I've tried this on my E6600 rig and it works fine, I'll try it on the QX and see what happens

[Nevrsadie]

Thanks for the reply Uncle. The data I sent you was idle temps overclocked, so I got home today and did actual idle temps per your original post.

Heres my idle temps 266*6 @ 1.02v room temp 22C

29 26 22 29


cores 0 and 3 are 7c over room temp, core 1 is right in line and core 2 is low.

I adjusted my idletemps in the realtemp.ini as follows

-1 0 1 -1

This put all four cores within 3-5c of my room temp of 22c.


I then rebooted and overclocked my Q6600 to 9*378(3400Mhz) at a vCore of 1.381. Then started Prime95, small fft's, to achieve full load.



All four cores reading the same temp, Sweet!
I did not have to adjust the tjmax.

[HDCHOPPER]

[QUOTE=unclewebb;2906894]RealTemp 2.4 with less cryptic features is now available.
http://www.fileden.com/files/2008/3/...7/RealTemp.zip



How's that for listening and responding to user input?

get em unclewebb !

[IanB]

"As a programmer.." ??
You are talking about GUI/application design here. Not programming.
Programming is a skill at different levels/different technologies (tools/languages..).
Has really little to do with the layout/design of eg an app like RealTemp.

Nonsense, I'm afraid. You are suggesting that the "programmer" just writes the algorithm, the raw code that does the work, then sends it to someone else to wrap a user interface around it. No, the programmer does both in building a GUI app. The algorithm and interface are totally intertwined, not just the "press this and it does that" paradigm but the way the result is reported back to the user. The user interface is more than half the work (and often more than half the code, actually) - if you'd ever built a Windows app from scratch you'd understand that.

Why do you think Microsoft spends millions on user trials to test how people interact with their software and feed that back into the design? Not because they have money to burn... Almost anyone with sufficient programming knowledge can write an app that just "works", the real trick is to take it from something merely functional to something that's easily and intuitively understood by the end user.

Windows is the success it is because of over 15 years of GUI design and improvement, where we all know how to use any new app in minutes because they all work in substantially the same way. An app that breaks those conventions is not going to get the same usage, user satisfaction or recommendations simply because the learning curve to use it is too high, and people may give up before they get to grips with it. Making it "geek-friendly" only would be a suicidal strategy, and why do you believe that only geeks should have access to the important information this app provides?

@ unclewebb. Thanks for listening. Since you aren''t using menus, another way of handling this would be to have a single button on the main interface that takes you to a second "advanced options" dialog. That way you get a perfect separation of action functions (setting calibration, clearing logs etc.) from your current clean display-only UI.

The only other thing I'd suggest is that you could add comments into the INI file directly, so that the number ranges for the various parameters are explained in the file itself for easy reference, rather than having to refer to separate documentation. If you are using GetPrivateProfileString/WritePrivateProfileString then the comments are simply ignored, else they are easily parsed over. The convention for INI comments is a semicolon at the start of the line.

[unclewebb]

IanB: Thanks for your input. RealTemp is a better program because of it. This is my first use of INI files and I'll try to include comments within the INI file for the next release.

Thanks HDCHOPPER. You're my biggest fan!

Nevrsadie: I had a brief look at your data and I can't come up with a reasonable explanation for why your temps sometimes, instantly jump by 5C or 6C without any change in load. It could be another one of those undocumented sensor issues. I got a little side tracked today but hopefully tomorrow I can investigate my Q6600 and its uneven temperatures at full load further.

[Richard Dower]

I don't about egos...all i care about is that the program is accurate, style and design is scondary.

[IanB]

I don't about egos...all i care about is that the program is accurate, style and design is scondary.

This has NOTHING to do with egos.

If an app is poorly designed and forces you to go through non-intuitive hoops to do simple things, you'd soon get frustrated with it. And the more difficult it is to use, the more effort the author has to expend in supporting it, which reduces the time available for further R&D, which is definitely a bad thing.

Style is one thing, but a good design is critical. You generally don't notice good information design and usability until it ISN'T there.

[ChrisZ]

This is my first use of INI files and I'll try

May I suggest three usability changes?

1. Editing INI file should be possible even when RT is running. As it is now, RT will nullify all changes on exit by rewriting INI. Any particular reason for this?

2. Calibration can be either +- 0, 1 or 2. This is not enough. On attached screenshot, even with +2 on last two cores, they are still BELOW ambient (which is 21 C) despite the maximum correction. Maybe +3 would help?

3. On my box, RT does not remembers its last position on a screen. It's a minor issue though.

BTW, The GUI is now very clear
(the one in 2.4, not the one in screenshot)

Best!

[Captn]

unclewebb, thank you very much for this wonderful program.

[Ace-a-Rue]

This has NOTHING to do with egos.

If an app is poorly designed and forces you to go through non-intuitive hoops to do simple things, you'd soon get frustrated with it. And the more difficult it is to use, the more effort the author has to expend in supporting it, which reduces the time available for further R&D, which is definitely a bad thing.

Style is one thing, but a good design is critical. You generally don't notice good information design and usability until it ISN'T there.

maybe you need to evaluate how you deliver the message about your objections or recommendations...

putting it to pen is a whole lot different than speaking it out loud to the individual(s)

i'm not trying to flame you, just give you a nudge in the right direction

PEACE to you my friend......

[unclewebb]

1. Editing INI file should be possible even when RT is running. As it is now, RT will nullify all changes on exit by rewriting INI. Any particular reason for this?

The reason behind this is because it made my life simple. I told you I was new to using INI files. In the next major release I'll make some changes so RealTemp is more logical and doesn't create an all new INI file on each exit.

2. Calibration can be either +- 0, 1 or 2. This is not enough.

The factors I've chosen were based on what I've observed to date. I can certainly add a 3 to the list. I thought that if a user had to go beyond 2 then it was possible that RealTemp had guessed wrong at TjMax.

Are you air or water cooled? My observed idle temperatures during my low MHz / low voltage test were with a Tuniq Tower with the fan at its highest speed. More data from other users running this test with a variety of processors would be helpful.

I'm 100% confident that TjMax=95C is correct for my 45nm E8400 but I have not had my hands on a 45nm Quad for testing. This choice looks appropriate for core0 and core1 of your Quad but not so appropriate for core2 and core3. I'm still trying to sort this behaviour out on my Q6600 and what is the real cause of it. Can you PM me a RealTempLog.txt file for your Quad with the calibration factors all set to 0 and run it from idle to full load to idle using Prime small FFTs on each core? Do all 4 cores report more or less the same at full load?

3. On my box, RT does not remember its last position on a screen.

Neither does CPU-Z! I'll see if I can save some X,Y co-ordinates in the INI file.

BTW, The GUI is now very clear

Even SpongeBob is happy now!

[Ramaistro]

Hi programs looks nice and simple

One question is it also aqurate for a q6600 ? I get these different temps and that with the quad at 3375 and 100% cpu usage it seems kinda wrong..

[ChrisZ]

Are you air or water cooled?

I'm on water.

Can you PM me a RealTempLog.txt file for your Quad with the calibration factors all set to 0 and run it from idle to full load to idle using Prime small FFTs on each core? Do all 4 cores report more or less the same at full load?

I'll do it for you, no problem.

But this raises an important issue, namely what *ambient* temp do you prefere?
See, on air cooling it's no brainer -- just let the stuff go for a few minutes on idle and you can assume the temps are actually *deltas* from the ambient room temperature.
Not so on water, once when it has got heated to say, 50 C, this is now your new "ambient", at least for a good while, isn't it correct?

Just let me know what is your point of view on that issue.

Cheers!

[unclewebb]

I'm on water.

That's sort of what I thought. Room temp is what you compare to when you are air cooled but if you are water cooled then you would need to idle for a while until your core temp and water temp are stable. You would then compare RealTemp to your water temp.

Ramaistro: You need to be a lot more specific about your set up, method of cooling, room temperature, MHz, core voltage, etc., etc.

If the TjMax I picked for your B3 is wrong, then your temperatures would be reported too low. The best test for this is to lower your MHz and core voltage as outlined in post#1 and compare your core temperature to your room or water temperature. In your screen shot your Minimum reported temps look reasonable. If you can make them go 10C lower than your room temp during testing then I probably screwed up and you should be using a higher TjMax. The B3 Quad uses two B2 dual cores which are both TjMax=85C so that's one of the reasons why I went with that number.

[ZenEffect]

will this program work for negative temps? say... phase change?

[unclewebb]

will this program work for negative temps?

It depends on your DTS sensors and how far below zero you are. In theory, the temperature data is being stored internally in the CPU in 7 bits of data so at most you can get down to 95C-127 = -32C.

I don't know if the previous 65nm sensors have any sticking issues when things get cold but I know you'd probably have to go through quite a few E8400 processors before you got lucky and found one that could go that low without at least one of the sensors, and probably both, sticking at a fixed temperature.

[steve30x]

TJ Max = 95 is right for my Q6600 G0 right?

[ZenEffect]

It depends on your DTS sensors and how far below zero you are. In theory, the temperature data is being stored internally in the CPU in 7 bits of data so at most you can get down to 95C-127 = -32C.

I don't know if the previous 65nm sensors have any sticking issues when things get cold but I know you'd probably have to go through quite a few E8400 processors before you got lucky and found one that could go that low without at least one of the sensors, and probably both, sticking at a fixed temperature.

sounds good to me
thanks!

[C-N]

It depends on your DTS sensors and how far below zero you are. In theory, the temperature data is being stored internally in the CPU in 7 bits of data so at most you can get down to 95C-127 = -32C.

I don't know if the previous 65nm sensors have any sticking issues when things get cold but I know you'd probably have to go through quite a few E8400 processors before you got lucky and found one that could go that low without at least one of the sensors, and probably both, sticking at a fixed temperature.

Uncle, I did give Realtemp a quick go last time I had my Q6600 on phase but I found my temp sensors were stuck at 4 & 10oC on each core.

CN

[NaeKuh]

Did you try calibrating idle temps as per the 1st post?

yes, it showed all my diode movement. But my cores still idle at 16C when the air intake on my radiator is reading 20C. The coolant levels in my system on idle stay a constant 21-22C.

Trust me, my system is spec'd out to be fully monitored. I cant get real temp to display temps above ambient at idle unless i do the obvious, put load.

If you need something let me know i'll be more then helpful. We do need a good monitoring program.

That's sort of what I thought. Room temp is what you compare to when you are air cooled but if you are water cooled then you would need to idle for a while until your core temp and water temp are stable. You would then compare RealTemp to your water temp.

No, im serious there is something wrong with the middle cores in your program. Its not how long the system is up. Mine has been up for Weeks and it still says 16C.

Here is how i monitor my loop. I have the new scythe fan controler which has 4 temp display.

Probe 1: Cpu loop res Temp probe <gets coolant temps>
Probe 2: is taped to the front of my radiator intake to get accurate ambients
Probe 3: GPU Loop reservoir
Probe 4. Gpu Radiator intake.

Probe 5 which is on my matrix orbital using dalas double confirms the water reading with my scythe.
Probe 6 which i have on my gpu loop, another dalas probe also confirms my water temps.

the H2o system on my unit is probably ranked in the top 1&#37; too if you want to know.

So no, my messurements are pretty much on the dot, and im telling you, i cant get my middle cores to NOT be lower then ambients on idle unless i put load on it.

Dayam i wish you had a yorkfield, you'd probably understand what i mean.

[bryanw1995]

ramaistro, you need to upgrade to the latest coretemp. I don't think that tjunction is 100c on a Q6600. I don't know if realtemp has it right at 85c, either, unfortunately...

[steve30x]

TJunction is showing as 95 celcius on real temp for my Q6600

[i43]

@ unclewebb
Apreciate Your hard work and effort you have put into this nice program RealTemp.

Here is my full air cooled system with 45n QX9650 at minimum CPU voltage 1.100v. / what Maximus Ext board allows
@ 1600 mhz (6x266) everything other on Auto / exept had to disable C1E for this low speed run
(other ways was not able to boot into winXp normally ... it does'nt load nvidia drivers and etc..)

- At Low speed screen, those ambient readings are marked with red box on Everest- and cpuid HwMon.panel
So, ambient was 21*C / double checked / with external probe (asus ME board allows this luxury also, i mounted a simple digital thermometer close to the CPU fan,
side panel is open and system was on idle, running 15min before screen shot.
(*Btw i have photos/screens of most Bios settings on that LowSpeed run, i can post or send if needed)



If compare temperature readings on [Stock speed] and [Low (downclock) speed], - at idle state the core sensors are never moving -
below 35 ; 35 : 31 ; 37
but they will go step by step upwards by rising case internal temp
( if i close panels and put all fans to minimum rpm, also at my case, there is much ''help'' from my video card what can do the ''best'' heat-up job ever )

here some Idle/Load runs @ Stock speed 2 screens (3000mhz) All auto C1E Enabled
Idle
http://img409.imageshack.us/img409/1...tockautke0.png
Load (everest stability test)
http://img255.imageshack.us/img255/2...tockautci8.png
----
don't know does this help or not .. anyways
Thank You

[jason4207]

yes, it showed all my diode movement. But my cores still idle at 16C when the air intake on my radiator is reading 20C. The coolant levels in my system on idle stay a constant 21-22C.

Trust me, my system is spec'd out to be fully monitored. I cant get real temp to display temps above ambient at idle unless i do the obvious, put load.

If you need something let me know i'll be more then helpful. We do need a good monitoring program.




No, im serious there is something wrong with the middle cores in your program. Its not how long the system is up. Mine has been up for Weeks and it still says 16C.

Here is how i monitor my loop. I have the new scythe fan controler which has 4 temp display.

Probe 1: Cpu loop res Temp probe <gets coolant temps>
Probe 2: is taped to the front of my radiator intake to get accurate ambients
Probe 3: GPU Loop reservoir
Probe 4. Gpu Radiator intake.

Probe 5 which is on my matrix orbital using dalas double confirms the water reading with my scythe.
Probe 6 which i have on my gpu loop, another dalas probe also confirms my water temps.

the H2o system on my unit is probably ranked in the top 1&#37; too if you want to know.

So no, my messurements are pretty much on the dot, and im telling you, i cant get my middle cores to NOT be lower then ambients on idle unless i put load on it.

Dayam i wish you had a yorkfield, you'd probably understand what i mean.

Did you use the (+) or (++) adjustment? That is what I meant, not the diode movement test.

From the 1st post:

Idle Temperature Calibration:

If you are interested in this feature then you need a way to find out if your processor reads too high or too low at idle. The method I use is I set my processor to run at a front side bus speed of 266 MHz and then I set the multiplier to 6.0 which results in approximately 1600 MHz. I also lower the core voltage as low as it will go. My Asus board goes as low as 1.08 volts which is fine for this test. You could also use SpeedStep and set C1E to on in the bios to accomplish the same thing. The goal is to reduce the heat output of your processor to a minimum.

By reducing your processor down to a common fixed value, I've found that a well air cooled processor's cores will idle at approximately 3C to 5C above the ambient temperature. If your reported idle temperatures are going below the ambient temperature then you need to use a correction factor of (+) or (++) to get them up into a more believable range. If your CPU is like my E8400 and you notice reported temperatures that are significantly higher than the 3C to 5C range then you'll need to use a (-) or (--) Idle Temperature Calibration factor to bring them down to a reasonable range. That's all there is to it. As long as TjMax is chosen correctly and you do a simple calibration, your reported Core Temperatures will be very accurate and comparable to other users who are using the same software and have also individually calibrated their processor.