Sign in with
Sign up | Sign in
Your question

TAT vs. speedfan inconsistencies between cores

Last response: in Overclocking
Share
May 18, 2007 10:43:23 PM

If I run both TAT and speedfan, the Intelcore temps are different. Yeah, yeah, I know, I read the thread that explains they should be 15 °C off. On my system, I can calculate an average difference which differs for each core. Not only that, but the difference changes if the machine is in an idle state vs. a load state.

What I did was simply run 2x orthos, then log the temps in both speedfan and TAT (you can write a log in TAT after the time period is up). After I parsed out the non-temp data from the logs, I simply averaged the numbers per core. If I subtract the TAT temps from the speedfan temps, I get the average differences, which as I said are both different per core, and different for an idle vs. load state.

For my setup (P5B-Del and Q6600 @ 9x333) I get the following:



Anyone else willing to try this experiment? If you don't want to or know how to do the averaging in a spreadsheet, you can email your logs to me and I'll do it for you.
May 18, 2007 10:51:24 PM

LOL, actually those discrepancies are EXACTLY correct. What you don't know if there is supposed to be a difference because they measure differently.

Read this, it will enlighten you....look for the tcase and tjunction comments, and how that translates to approx. 15C difference in TAT and Speedfan:

http://forumz.tomshardware.com/hardware/Core-Duo-Temper...
May 18, 2007 11:58:13 PM

Yeah, I've read that, but my errors are not only >1 °C, they're inconsistent from idle to load. I don't see how there's so much guess work in reading a sensor.
Related resources
a b à CPUs
a b K Overclocking
May 20, 2007 11:20:15 PM

graysky, SpeedFan is incorrect and requires +15c Offsets for all Cores. Your findings will then conform, since slight variations are normal. From the Guide:

Quote:
Interpretation

...Since Intel's Thermal Analysis Tool (TAT) is a Notebook tool, and desktop C2D's have an integrated heat spreader, TAT will typically indicate ~ 2c lower than SpeedFan...


Hope this helps,

Comp 8)
May 22, 2007 8:38:04 PM

I figured out what my problem was. TAT sucks for reading temps on C2D chips. I should have taken a hint from the application itself. If you launch it and look in the upper left under Processor Details, you'll see, "Processor: Pentium M." I just ran RMClock and Speedfan and had both apps log a set of load temps. Then I meticulously analyzed the log files and averaged the exactly (down to the last second) the same time points from each log to ensure a fair comparison. If I take the difference of the average RMClock core temp numbers and the average Speedfan numbers, I arrived at:

[code:1:9be30f3c22]Core 0: 15.002
Core 1: 15.069
Core 2: 15.049
Core 3: 14.979[/code:1:9be30f3c22]

Conclusion: Speedfan can log temps as precisely as RMClock can and the offset is indeed -15 °C. Also, don't use TAT for a C2D or Quad C2D!
May 23, 2007 12:51:54 AM

Ok, I'll venture a noob question........

If TAT sucks for C2D, then what to use to get accurate temp readings? I see people quoting all kinds of different numbers when posting temps on their overclocked C2D's. And there is NO WAY they're running temps lower than my liquid cooling.....so when they say, oh it's 35C....I say BS! Apples to Apples dammit, give the right and proper comparison and temps!!!! :evil: 

So, what's the simple answer for "What program to use?" and "What temps to follow?"

Seriously, there must be a simple answer. I can't tell you how MANY posts I see about temps, C2D, what's right, etc, etc, etc.

From what I understand, 60C in TAT for a C2D is generally the long-term max......65C is pushing it, and 70C is what I'd consider shut-down level. IF that is the case, then simply subtract 15C from each of those to get the same equivalents for SpeedFan/CoreTemp???
May 23, 2007 1:03:46 AM

RMClock = accurate readings. Read this thread paying attention to what uncleweb writes about using CrystalCPUID to read the DTS directly. You can convince yourself that whatever thermal software you're using either is or is not accurate.

I did followed his advice and found that RMClock was dead on for all my cores whether idle or load.

Try it and let us know.
May 23, 2007 1:18:48 AM

hmmmm.....d/l'ing now. I've always used CoreTemp and Speedfan....but sadly I can't get CoreTemp to work with Vista....which is why I switched to TAT.

Unclewebb's comment that "the programming tricks that will work in XP may not work in other versions of Windows", to which you replied "No way would I put that o/s on my machine. At least not until the first major SP is released." So, since you have XP you are in luck and avoid these issues. I similarly had no issues in XP either. Even though I still have XP, I use Vista as my primary OS and don't wish to reboot every time I want to check my temps, so unfortunately I'm stuck. Regardless of what anyone thinks of Vista, the situation is the situation.

Frankly, I'm not concerned about discrepanies <2C...I'm more concerned with the "big picture"....ie--is 62C in TAT too hot and will pose long-term problems or not? That sort of thing.......

Regardless, with my cooling I no longer have to worry about it, but it would help to know to pass along the info to others that have the same questions I suppose........:) 
May 23, 2007 1:23:46 AM

Cool man, just post to this thread if RMClock works under vista for you... you might wanna try that CrystalCPUIP trick and report back that result as well.
May 23, 2007 2:33:34 AM

Ok, here's what I got with the 3 different apps. Can you interpret??

May 23, 2007 5:54:58 AM

Speedfan is reporting erroneous results. If you click on the "CPU 1" button in RMC, you'll get the core temp of your other core. If you check the "Log monitoring data" button, you'll get a nice log file updated every 1 sec of temps and all the other parameters which you can load up in excel or Openoffice's calc or whatever your fav. spreadsheet program might be and manipulate the data later on, graph, average, etc.

The other thing I noticed from your screenshot is that you're using an older version of speedfan. 4.32 is current and might fix the problem, I dunno.
May 23, 2007 2:03:42 PM

Thanks, I'll try taking a look at that tonight, and also update SpeedFan to the current version.

I'll try to get new screenies to post for your review. Thanks!
May 24, 2007 5:16:28 AM

There appears to be a 2C difference between CPU0 and CPU1 in RM. TAT also reports approximately the same between CPU0 and CPU1.

Speedfan 4.32 is reporting a difference between 1C and 2C....it's almost in sync with the other 2 apps, but not quite.

There is a very minor <0.5C difference between what is being reported in RM and TAT. So, all things considered, I would surmise those are consistent and correct.

All 3 apps are very close....RM and TAT being the closest.

So, I guess there may be very minor discrepancies as you reported, but I find them within "acceptable" tolerances for my needs. They're all showing around 42-43C......so for my overclocked 6300, I know I'm fine. 2C +/- is fine for me. But you are correct, RM is accurate. And closer to TAT from what I can see.
May 24, 2007 7:02:06 AM

cool... gotta get the link to rmclock to get added to the temp sticky.
!