Sign in with
Sign up | Sign in
Your question

Core Temp vs. Speedfan vs. TAT

Tags:
  • CPUs
  • Core Temp
  • Speedfan
  • Overclocking
Last response: in Overclocking
Share
October 24, 2006 8:06:04 PM

Core Temp and Speedfan and Thermal Analysis Tool (from Intel) all seemingly give different temps for my cpu (E6300 with DS3 board). Why is that? Which one is most reliable/which one should I only look at?

More about : core temp speedfan tat

a c 85 à CPUs
a b K Overclocking
October 24, 2006 8:40:29 PM

TAT

Although it will indicate throttle at 84.1 (according to some sources) below that all temp readings from all sources are a bit dodgy.

Think of the CPU as having a switch inside it, at a certain temp it will flick into a low power state for throttling, however we can only infer what the temp is by looking at some other characteristics of the switch that are not good for reading temp from. hence the uncertainty in the measure...
October 24, 2006 8:54:42 PM

TAT and core temp should read the same and speedfan just reads the bios temp
Related resources
October 24, 2006 9:07:51 PM

Go with TAT. It works, and you can log temps over days of usage.
October 24, 2006 9:36:19 PM

Try MBM5, noone seems to use it that much but it's a much nicer universal tool I find.
October 24, 2006 10:05:20 PM

I've never heard of motherboard monitor. LOOKS GOOD. Lets get a thread going MESA. I wanna get some accuracy figures on that. GOOD FIND!
October 24, 2006 10:47:28 PM

Quote:
I've never heard of motherboard monitor. LOOKS GOOD. Lets get a thread going MESA. I wanna get some accuracy figures on that. GOOD FIND!


So uh... which one is most accurate? So then I can just not care about the other ones :p 
October 24, 2006 11:02:29 PM

TAT
a c 85 à CPUs
a b K Overclocking
October 24, 2006 11:04:29 PM

Quote:
TAT

Although it will indicate throttle at 84.1 (according to some sources) below that all temp readings from all sources are a bit dodgy.

Think of the CPU as having a switch inside it, at a certain temp it will flick into a low power state for throttling, however we can only infer what the temp is by looking at some other characteristics of the switch that are not good for reading temp from. hence the uncertainty in the measure...


TAT
October 24, 2006 11:24:09 PM

TAT or CoreTemp.
I prefer Coretemp because it is still being devloped..
Plugin supprt as well as minimizing to icon task bar aslo in works.

I like TAT for stressing my CPU though...

my 2¢
October 25, 2006 1:12:53 AM

It would be nice if both Core TEmp and TAT had a taskbar icon like Speedfan does that shows your cpu temp. And would be nice if TAT minimizes without having to close it.
a c 145 à CPUs
a b K Overclocking
October 25, 2006 1:53:12 AM

Quote:
I've never heard of motherboard monitor. LOOKS GOOD. Lets get a thread going MESA. I wanna get some accuracy figures on that. GOOD FIND!


Been using MBM5 for years on other computers...but it is not accurate for core 2's

well at least compaired to CoreTemp

but i still find it hard to think that my E6600 runs hotter than my A643200+(754 with cnps 7000alcu) did. damn i wish i have a thermal diode

Hell acording to core temp...my dad's A64 X2 4200 runs at like 20-25 idle

Ohh well as long as it works....i don't care
October 25, 2006 2:59:12 AM

Quote:
You need to understand that all of these use the same chip on the motherboard to get there information. Typicly refured to as the winbond chip because that's the main manufacture of these chips. All motherboard and CPU temps are reported by this chip and read by verious software. All software uses the same information that the bios recieves. Just ech software has a diffrent way of presenting that information. One thing to keep in mind is that this chip has a margine of error do to resistance in the trases on the motherboard. It's also important to note that each software has it's own margin of error. This is why you see slight diffrenceds in the reporting of temps and voltages. Some software like speedfan needs to be configured for your motherboard. Not all boards support temp reporting of more then just the CPU.

In the next 6 months or so i hope to get all the equipment i need to start testing each reporting softwares margin of error. I will then post articls about it in verious places on the web and her in the Forumz.


Instead of 6 months, how about tomorrow? :D 
October 25, 2006 4:26:03 AM

i have been using MB5 for a while now. i find it great. it always has the same reading as the bios i think which is good.
i have CoreTemp aswell. it usually reads 3-4 degrees hotter than MB5.
i have an A64 3000+ BTW

i consider MB5 more accurate coz u have to put in your MB model etc and coretemp just runs as is. i dont know if that makes a difference though

EDIT: i just got speedfan to try. it reads the same temps as CoreTemp does.
However the "Temp2" reading in speedfan is the same reading as CPU from MB5??????????
October 25, 2006 3:20:27 PM

An analysis of the accuracy of these programs would be great. I ran 3DMark06 and PCMark04 last night on my new build. I got TAT and see that it is similar to CoreTemp. So I have TAT about the same as CoreTemp and SpeedFan about the same as the M/B monitor software (and lower by about 10-15C). While running the above programs the max temp was 59C on TAT and 50C on Speedfan. My understanding is over 60C is bad so now I'm concerned. This is still a stock E6400 configuration with a AC7 Pro. I'm wondering if I'll be able to overclock this after all.

:?
October 25, 2006 10:41:33 PM

Quote:
It's not, don't worry about it. And don't ask me if I'm sure or not like other idiots do.
What you want is get to the stage where 60C reported by CoreTemp/TAT isn't even stable enough for the processor. :wink:


But even if the processor is stable, running it at high temps could still shorten its life. Water cooling it down to room temperature would really be nice :p 
October 26, 2006 1:58:41 AM

Ah my misunderstanding of "good"

I'll see what I get to be stable and see how the O/C changes the temp

Dave
October 27, 2006 3:12:08 PM

Has anyone noticed that the TAT stress test makes the core temps jump 3-5C higher on average than running Prime95/Orthos?

I'm running Orthos right now and my temps hover around 47-49C with a Gigabyte Galaxy 3D II water cooler. I probably could have gotten about the same temps with a Big Typhoon, couldn't I?
October 27, 2006 5:18:07 PM

SpeedFan 4.32 beta 4 now has full support for INTEL CORE internal temperature readings.

http://www.almico.com/sfbetaprogram.php

I have tried all 3 and find that then have roughly the same temperature readings... (+-2or3 deg)
Anonymous
a b à CPUs
a b K Overclocking
January 27, 2009 10:48:32 AM

hi all,

I have an asus mobo with a q6600, i have to tell you that mi reads on speedfan are the same exactly that runing coretemp. I have to test tat too...ah my vcore and other v's are the same too with both programs
a b à CPUs
a b K Overclocking
January 27, 2009 1:40:47 PM

This thread is 27 months old. :sleep:  The information is obsolete, and has been superseded.
Anonymous
a b à CPUs
a b K Overclocking
January 29, 2009 6:59:41 AM

oops...thx!

Anything more to add?
a b à CPUs
a b K Overclocking
January 29, 2009 8:07:46 AM

1) Check post dates first ;) 
2) Click on the link in CompuTronix's signature, read, read again, then ask questions in a new thread :) 
!