Real Temp - New temp program for Intel Core processors

[ReverendMaynard]

sweet, runs without a hitch....I think.

[rusty]

how's that ? winxp x64. for some reason temps almost the same as everest.

[unclewebb]

That is about 8~10c rise above ambient when my processor is idle, is that the normal range for the Q6600 GO's ?

I'm not sure about that. You could send me your Q and I'll promise to give it a thorough evaluation!

WoZZeR999: Thanks for your patience. I guess I'm still in disbelief to find that it really does work in Vista x64, especially without any extra programming on my part. That's one hell of a driver I switched to.

I'll go wrestle with FileDen some more and update the main download.

[jaredpace]

Unclewebb, is it possible to include CPU temp as well. (tcase measurement) into your program? Or is this unnecessary because Realtemp eliminates the need for this by assuming the "real" core temps and tcase are relatively equal?

[Ricasete]

Great work.

Thanks to your effort!

Works like a charm on Vista64(as it is).

Idle and load:





Thanks again.

[scottc19]

Good job unclewebb my vistax64 is working great... temps are now~ -5C underload than core temp and about~ -5C compared to everest also

[unclewebb]

At idle with low volts and MHz I believe that my IHS measurements accurately relate to the true core temperature but when a processor is under a normal variable load there are instantaneous hot spots developing in various regions of the core. That is the whole reason behind Intel going to multiple on chip DTS sensors. To keep better track of this activity. I've never been a big fan of CPU diode temps. Some are correct but it's impossible to tell which ones are and which ones are not. Core temperatures are what's needed to be measured. Maybe this will help encourage Intel to release proper documentation for these sensors.

[jaredpace]

okay, gotcha. nobody has exact documentation telling the Tjmax of G0 q6600s and E8400's? This is where we see discrepancies in CT & RT.

[unclewebb]

Correct. Intel does not document TjMax. Both programs are left guessing but I'd like to think one program's guesses are closer to the truth than the other.

The main download as listed in post #1 has been updated with the x64 drivers so that should work now for whatever OS you're using.

http://www.fileden.com/files/2008/3/...7/RealTemp.zip

[Kryckter]

Very nice program. Thanks!

[lasershock]

Working fine here on Vista x64. No more stuck sensors. Good job.

[Gen]

are Core #2 and 3 bunk? they almost always report exactly the same temp.

approximate 20-22c ambient

[unclewebb]

are Core #2 and 3 bunk? they almost always report exactly the same temp.

The two cores in my E8400 track each other almost identically from idle to full load temps and back again. The difference between cores is rarely 1C. I don't know if that is normal or how that compares to the average processor. If the DTS sensors are working properly, the IHS is square and well attached to the cores and a heatsink has been installed properly, then that is what you should expect to see. Quad cores are different and it seems a lot rarer to see all 4 cores that are balanced like that.

Where's that smart guy with Excel when you need him to examine the temp logs to see what the average delta is during normal use?

It's a good idea for all users to properly apply thermal paste in the orientation that matches the layout of your cpu. (Dual vs Quad). I think some mistakes here might be responsible for some of the uneven temperatures seen. The AS5 site provides good documentation for this. It's also easy enough to make a mistake when you're trying to carefully manipulate a large air cooler into position.

[Ron 61]

Thanks again unclewebb for a well researched and well thought out program. Your program is IMHO the best core temp monitoring utility on the net bar none, keep up the good work.

[WoZZeR999]

I'm not sure about that. You could send me your Q and I'll promise to give it a thorough evaluation!

WoZZeR999: Thanks for your patience. I guess I'm still in disbelief to find that it really does work in Vista x64, especially without any extra programming on my part. That's one hell of a driver I switched to.

I'll go wrestle with FileDen some more and update the main download.

Heh, no problem. I'm glad that my 'off chance' of using the x64 drivers actually worked. I actually got the idea when it said 'WinRing0.dll' wasn't there, and I was looking at the file. I'm assuming that winring just dynamically renames itself within memory or somthing.

I figured that the actualy x64 programming wouldn't really be any different because I assume that you are just pulling the numbers from some basic call from the drivers. The only thing that the x64 could bring is a new 'optimized' x64 bench. You score xxx with x86 instructions, and xxxx with x64 instructions.

Thanks for the program.

[Demo]

What is the official TjMAX for the E8400 ?

[karateo]

Ok...the program runs fine in vista x64. the thing now, is how reliable the readings are.

I shows the same temp as my dfi mobo (without the +12C I added to agree with coretemp) so I believe this is a verification. Apart from that and your experiments, how can we be sure that it reads the actual temp of our E8xxx cpu?

[jas420221]

What is the official TjMAX for the E8400 ?

Correct. Intel does not document TjMax. Both programs are left guessing but I'd like to think one program's guesses are closer to the truth than the other.

[rge]

Thermal gradient within the die can achieve 20~40+ °C between the hottest and the coldest point, depending of the load.

I believe that is the story of the fish that grew from 10inches to 40 inches, unless you can produce a document showing otherwise. I had read the paper when it first came out, but took a while to find it again. The paper published by senior intel engineers and architects of the core duo processor titled "Temperature measurement in the Intel® CoreTM Duo Processor" shows the die map, and a chart revealing the temp gradient of roughly 150 different load programs they tested.

Quote from paper.
"In order to evaluate the DTS temperature reading, we performed a study to identify the impact of different workloads on the difference between diode and the hot spot, as measured by the DTS. A set of workloads including all SPEC-2K components and other popular benchmarks and applications, at single thread and multithread were executed on the CPU. Several iterations were done to reach a thermal steady state and then the diode and DTS temperatures were measured. Before taking the measurement, a calibration process has been performed, leaving only the temperature offset. Figure 5 shows the offset between the analog diode and the hot spot, as measured by the DTS... It also can be noted that some workloads display high temperature gradients (largest ~5.2C) while others have no offset."

Highlights:
Gradient from Die hot spot to central diode is 0C at idle and was measured 0C at roughly 15% of the "popular" load programs/benchmarks tested after steady state load was reached. Nearly a third of load benchmark/programs had a gradient less than 1C, and the maximum gradient between hot core and cooler central diode was ~5.2C, likely on programs like TAT, described with such a gradient in another article. The hottest transient gradient measured from hot spot on one core to an inactive core was 10C, using TAT like program or hotter, and this was on a larger cpu than 45nm.

What intel describes as a large hot spot is again stated in their article describing the reasoning between putting DTS on die versus junction..."The better accuracy translates either into 3%-7% higher performance or into improved ergonomics" or "In desk-top computers the impact is even higher due to the lower thermal resistance and 1C accuracy translates into 2 Watt of CPU power." Intel considers this 5C max gradient seen between hot spots and junction to be worth the move to die sensors.

The only possibility imo that tjmax was still 105, was the potential hot spots. But at idle, and even at load on a large number of the tested "popular benchmarks/programs the gradient was measured 0C from core to between cores, even at high temps 93% of tjmax.

The only other gradient, from sitting on core to Tcase was measured at 0.4C in previous link.

45nm seems to have less gradient, illustrated by Unclewebb showing no more than 1C difference between loaded core an idle core.

I think the burden of proof is clearly on those who want to claim a gradient larger than 1C exists from core to tcase at either idle or steady state load for moderate load programs.

http://eda-publishing.imag.fr/spip/IMG/pdf/TMI23.pdf
good read

[Vipeax]

Thanks for this very nice program .

Now, is it normal to have this amount of sensor movement? (10/11/12/13)


PS: Finally no more Vista 64bit issues, like CoreTemp had .

[unclewebb]

rge: Many thanks for your continued effort in killing the old "gradient myths." Last summer, after plenty of research, I brought this topic up on XS but it was an uphill battle trying to convince some of the old school types about gradients in a modern Core cpu. Where are you joebuffalo?

On the 45 nm dual cores, Intel doesn't even need multiple DTS sensors. They should almost concentrate on creating one high quality sensor with a wide dynamic range rather than multiple sensors with sticking issues when trying to report idle temps.

WoZZeR999: The quick Bench feature in RealTemp was provided more as a convenience feature for users wanting to make sure that their processor is running OK and the temps are moving around under load without having to reach for another program to confirm this. One interesting thing about this bench is it can't be tricked by booting up at one frequency and then using SetFSB to get a higher frequency. Your bench score is based only on what you're capable of booting up at. Just trying to keep people honest!

Vipeax: The sensor test was designed with my dual core E8400. It quickly brings all cores up to full load and does a quick compare of before and after temperatures. Quad core processors are going to generate more overall heat during this test so it's normal that they will report higher movement numbers during this test compared to a dual core CPU at the same core voltage and MHz. Depending on all of those variables, there's a wide range of what a good reading in this test is. Sensors that move more or less equally is what you want to see. Sensors that don't move (0) during this test is a sign that they are getting stuck and not properly responding to the change in temperature that your processor just experienced. Too many 45nm chips have this issue.

I use SpeedFan and like its graphing capabilities. It uses a 100C TjMax compared to 95C that RealTemp uses. Anyone that wants SpeedFan to read the exact same as RealTemp needs to click on the Configure button and go into the Advanced section and set an offset for each of your cores to -5. That will equalize TjMax between the two programs, and will equalize your readings.

[Dostoyevsky77]

Very nicely done. Will download ASAP & spread the word on other forums.

Vipeax, latest CoreTemp is Vista 64 compatible.

***EDIT: Interesting phenomenon, but I tried this out at work on our P4 3.0 & the temperatures just get stuck at TjMAX (85 degrees). The time stamp does not advance. Hopefully this isn't the case on my systems at home. Someone else with a P4 is reporting the same thing. LINKBACK

[Darth Vader]

unclewebb
Please, can you write log of changes in the program under versions on the first page?

[unclewebb]

Vipeax, latest CoreTemp is Vista 64 compatible.

***EDIT: Interesting phenomenon, but I tried this out at work on our P4 3.0 & the temperatures just get stuck at TjMAX (85 degrees). The time stamp does not advance. Hopefully this isn't the case on my systems at home. Someone else with a P4 is reporting the same thing.

This program is designed for all of the Core series processors. I included the ability to run it on some P4 processors for testing purposes. If it doesn't run on your P4 properly then there is nothing I can do. The P4 doesn't have the sensor needed for core temp readings so it will read 0 for DTS and show temp at max. I tried!

[DOM.]

unclewebb, so was mine like yours ?