Real Temp - New temp program for Intel Core processors

[unclewebb]

RealTemp 2.5 has been released.

I've moved the download and documentation over to TechPowerUp.

http://www.techpowerup.com/realtemp/

I warned them that they might need to upgrade their servers. The XS guys take their temps pretty seriously.

Now that RealTemp has a home and a Tray Icon and a separate webpage with documentation, what's next? I need some feedback here.

Some things that I've been thinking about are RealTemp for dual CPU systems. Either a small button to display 4 cores at a time or a super size version to display all 8. Maybe a special Dual Core specific version with a cleaned up GUI so you don't have to look at a bunch of empty space. Maybe a larger temp readout in that version similar to TAT. Maybe the temp number displayed in the System Tray instead of the icon.

Anything you can think of beside Logitech G15 support? I'm still thinking about that one!

[jas420221]

Now that RealTemp has a home and a Tray Icon and a separate webpage with documentation, what's next? I need some feedback here.

How about a link to a donations page for you!!!

I think this is worth a little something personally!

[Uberbob102000]

Sorry if this has been asked but my idle calibration settings won't stick with 2.5.

[unclewebb]

Sorry if this has been asked but my idle calibration settings won't stick with 2.5.

I just downloaded the version that is available at the link I posted and it works for me.

Have a look in your INI file and make sure there isn't a leading semi-colon.

;Idle0=1

is wrong. The ; will block this setting.

Idle0=1

should work fine.

jas420221: How about a link to a donations page for you!!!

Thanks for the thumbs up. I'm very happy with all of the free software on my computer and being able to give something back to the community for a change is great.

[Uberbob102000]

Oh wow, That was stupid on my part, Thanks.

[railmeat]

Niiiice!!
Under Small FFT's test, now real temp shows 10C lower temps than coretemp, is that right?
I am happy now!

me 2 e8400 at 1.12 v stock on my water 29-31c oc time!! now that i can somewhat trust something temp wise

[unclewebb]

Before I wrote a single line of code for RealTemp, I did some initial testing with an E8400. CoreTemp assuming that TjMax=105C for my E8400 is wrong. I plan to re-install my E8400 next week and do a real thorough test. Previously, 35C measured with the Fluke 62 IR thermometer was being reported as 50C by CoreTemp.

RealTemp was only an idea during that previous test. Time for some pictures of RealTemp vs CoreTemp along side the IR thermometer and you can decide which program is giving you values closest to the truth.

[RedAlert-]

Hey guys,

I've been using RealTemp for a while, but I noticed something interesting. When my Intel Q9300 (overclocked to 3,45GHz) is under 100% load (Orthos), RealTemp says that core temperatures are 60 to 64. But, HWMonitor, Everest and Coretemp say that core temperatures are like 70 to 74.

Should I believe Realtemp? It seems to be a better program, but this is 3 vs. 1. I'm little bit confused with this now, so would somebody tell me, which is right.

[unclewebb]

RedAlert: If 3 guys ahead of you jumped off a bridge would you follow them and be #4? There's a reason why no one has been too happy with their core temps since the introduction of these chips. The 10C difference you are seeing is because these programs use a different TjMax. 95C vs 105C means their results will always be 10C different. Read the post above yours about how far off CoreTemp is with my 45nm E8400. Ask those other programmers to show you their research. Read the last 37 pages of this and you will see some of my research.

[gx-x]

@unclewebb:

Thanks for the new version

Two suggestions :
1. Make the app show temps in tray instead of an icon
OR
2. Make an option to disable the tray icon when realtemp shows temps in taskbar (when minimized) which is what I am using. I am just one of those who always likes to have an eye on the temps

PS. In this version (compared to 2.49) I have that problem with idle0/idle1 -2 setting. 2.49 was fine (?!) but 2.5 is like like 2.41(I think it was that one), it showing -3 ~-4C lower instead of -2. (idle0=0 is 40C, idle0 -2 is 36C, strange)

Best regards !

[k4vz0024]

I´d like to know is TjMax= 85 C correct value for Intel Quad Q6600 Revision B3?

Best regards

[Big Lar]

Version 2.5 here working fine, Thank you Very much yet again
Vid;
Coretemp; 0.9875v
Realtemp; 1.1000v

Larry

[Ace-a-Rue]

RealTemp 2.5 has been released.

I've moved the download and documentation over to TechPowerUp.

http://www.techpowerup.com/realtemp/

I warned them that they might need to upgrade their servers. The XS guys take their temps pretty seriously.

something is wrong with the download link after jumping over to techpowerup.com...when you click on the download link it diverts me to a comcast.net advertisement for high speed internet...in order to circumvent that ad, i had to click on the download link at the very top and then click on real temp for downloading...all worked fine going the cirumvention route.

[rge]

RealTemp says that core temperatures are 60 to 64. But, HWMonitor, Everest and Coretemp say that core temperatures are like 70 to 74.

Should I believe Realtemp? It seems to be a better program, but this is 3 vs. 1. I'm little bit confused with this now, so would somebody tell me, which is right.

HWMonitor, Everest, and Coretemp are all using tjmax of 105 for E8400 and all from the same one faulty source of information, and presumably based on 45nm mobile cpus, so it is really 1 vs 1...but...

Mathematically, using intels own formulas and specs, you can show a tjmax of 105 is not plausible for E8400.

known facts
1) (see pic of intel formula)
2) at idle EIST enabled state TDP = 6-8W (intel specs)
3) thus at idle 6x200 mhz, .95vcore, TDP<6-8W
4) casing temps measured by multiple fluke measurements of E8400 is 95C when DTS=0.
5) casing temps are going to be equal or slightly lower than Tcase measurements, they can not be higher.

Suppose 105 tjmax
- Solving for theta (core to case) = 0.5 C/W
- At 0.5C/W, max gradient from casing to tjmax at 6x200 mhz and vcore .95 is 0.5C/W x 6W or 3C, more likely gradient .5x4W or 2C, plus or minus 1-2C error.
- therefore if 105 tjmax is correct, casing temps would have to be 102C to 103C, plus or minus 1 or 2C error, which is not the case.
- Or for 105 tjmax to be correct, at 6x200mhz and vcore .95, TDP would have to equal 20W, which is not even close.
- Without a heat sink, the gradient would be even less, which makes 105 again, highly unlikely.

105 is not plausible unless some of intels specs are way off.

Suppose 95 tjmax
-Solving for theta (core to case) = 0.35C/W (which as an aside is a similar value to whats listed in intel research papers)
- At 0.35C/W, max gradient from casing to tjmax at 6x200 mhz and vcore .95 is 0.35C/W x 6W or 2C, more likely gradient .35x4 = 1.4C.
- For 95 tjmax to be correct, casing temps would have to be 93-94C +/- 1-2C error. Casing temps of 95, as measured, would be within 1-2C error.

95 tjmax is possible based on intel specs and formulas. Also with no heatsink, the gradient should be less than 1.4C, and has been measured by independents to be around .5C, making 95 not only possible, but likely.

I would not argue with anyone who claims tjmax is 94-98, but 105 contradicts known intel specs and formulas and is way outside measurement error.

[zlojack]

Here's a comparison of the VID as read by different monitoring programs.

Looks like CoreTemp and Everest are off on this one.

[Ace-a-Rue]

@Uncle:

is there any way to marry up the"Distance to TjMax" for four cores?...i've tried everything, including the individual adjustment of TjMax in the "ini" file....if i increase or drecrease the individual TjMax's, all that happens is the actual core temp goes up or down, not the distance to TjMax values...

i am happy that i can marry up the actual core temps but i would like to also marry up the "Distance to TjMax" too!...anyway of doing that?

[unclewebb]

is there any way to marry up the"Distance to TjMax" for four cores?

The Distance to TjMax displayed by RealTemp can not be changed, now or ever. It is a direct reading of the sensors within all Core processors. RealTemp then takes that raw data and tries to convert it to some meaningful temperatures as best it can after considering the limitations of the sensors as well as the limited documentation supplied by Intel.

The reason I call it a Distance is because there is not a direct, linear relationship between the data coming out of the digital thermal sensors and how many degrees you are away from TjMax, especially at idle.

Let me know if I didn't understand your question or the reason why you would want this directly read information to be altered.

As for advertising on websites, I use Firefox with Adblock Plus. I rarely see any ads. TechPowerUp wants to make some money to cover their cost of hosting RealTemp. Now that RealTemp is a pretty decent little app, I decided to go with them for some more exposure. If things don't work out, I might go back to hosting it myself.

k4vz0024: I´d like to know is TjMax= 85 C correct value for Intel Quad Q6600 Revision B3?

I'll admit that this is one of the few processors that I am not 100% sure about. I wish I could get my hands on one for proper testing. I have a Q6600 G0 and it showed TjMax=95C when tested with the IR thermometer. The two main reasons I went with TjMax=85C for the B3 Quad is because a B3 Quad consists of two B2 dual processors. I've tested my E6400 - B2 and it showed TjMax=85C.

The next item is that the B2 Dual Core and the B3 Quad core have very similar Tcase temperature ratings published by Intel while the B3 is about 10C lower than the G0. My G0 tested at TjMax=95C so about 10C less is 85C.

http://processorfinder.intel.com/det...px?sSpec=SL9UM
http://processorfinder.intel.com/det...px?sSpec=SLACR

There were also early reports of some B3 Quads easily getting up to or near the thermal throttling point when used with the retail heatsink and fan. Intel might have raised TjMax when they went to the G0 stepping to reduce the chance of this happening.

I have been questioned on this before but so far only one user with a B3 has done any testing and when his results started to show that I might be right, he decided to dismiss his results.

I've found that the typical Dual Core B2 with a top notch air cooler ( Thermalright extreme ) will show idle readings a couple of degrees below the room temperature. That's impossible which is why I introduced the Idle calibration factors in RealTemp. When locked to 1600 MHz and about 1.10 volts, a B3 Quad should idle a couple of degrees higher than that If my pick for TjMax is correct, RealTemp should display a number very close to your room temperature during this test. If I'm wrong, I'm more than willing to swap my G0 for a B3, straight across, to prove this. We need some B3 owners to give this test a try.

[emoners]

@uncle: hope these help... CPU-Z sensor turned OFF

using RealTemp v2.48 & CoreTemp v0.98.1

using RealTemp v2.5 & CoreTemp v0.99

thanks for the program again!

[Leeghoofd]

@uncle: hope these help... CPU-Z sensor turned OFF

thanks for the program again!

why don't you change the ini file for coretemp and set value -10 giving exact same temps then as real temp

[unclewebb]

why don't you change the ini file for coretemp and set value -10 giving exact same temps then as real temp

I remember the early days when people wanted to change RealTemp by 10C so it could look just like CoreTemp.
I'm glad to see the logic on that has changed.

[kpo6969]

Here's what mine look like. CoreTemp and Everest adjusted to TjMax 95c.

[Demo]

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?

From what i know it detects throttling based on the CPU signal w/c is not dependent on the TjMAX

[Demo]

Heres my Q6600 G0 at 15mins into prime v25.6 SmallFFT test

Its a VERY cold evening here thats why max temp is 51c, i usually get 61~63 during scorching afternoons...





Windows vista ultimate 64bit
Q6600 G0
Bios voltage 1.335v
Thermalright Ultra 120 Extreme

[k4vz0024]

[QUOTE=The next item is that the B2 Dual Core and the B3 Quad core have very similar Tcase temperature ratings published by Intel while the B3 is about 10C lower than the G0. My G0 tested at TjMax=95C so about 10C less is 85C.[/QUOTE]

What about fact that B3 Quad core have Thermal Design Power to 105W?

Best regards

[Leeghoofd]

I remember the early days when people wanted to change RealTemp by 10C so it could look just like CoreTemp.
I'm glad to see the logic on that has changed.

Does the VID readout function work properly ?

Reads out this E8400 at 1.2250 VID while Coremtemp says it is 1.1125

Kpo6969 also has an E8400 and his screen gives the same VID readout as I am experiencing... any user getting same readouts via coretemp and Realtemp ?