Intel Thermal Analysis Tool

[Eastcoasthandle]

For those of you that don't have it here it is called TAT


This program is hard to find as it's not available from Intel as of yet...I checked...

[AndrewZorn]

does not work on my 805D. i take it this is C2D only.

i dont understand, you can change the workload level? like make it run at 50% max or something?

[vengance_01]

It will simulate a work load so you can gauge temps. From what I can tell its the same as core temp, so I use core temp, smaller footprint etc....

It see my CD2 as a Pentium M core, so mabey just newer chips and not the P4 chips??

[Ksu]

I've tried it and TAT shows 2C less temp than Core Temp so I like TAT more

[Kashath]

I'm getting 2-3 degrees C less than CoreTemp.

[ted3]

i dont understand, you can change the workload level? like make it run at 50% max or something?

Set max workload to test your max temp, and for sure the temps will go way higher than Orthos with small FFT's, theres no other way i know of to stress a CPU this much.

And you can also set a throttling level to reduce the processors performance to a preset percentage, up to 87.5% throttling. Apps will run quite slow if you set both cores to 87.5% throttle =)

What makes TAT more useful than CoreTemp is the throttlewatch, i posted a screenie in another thread not long ago, this is what happened 2 minutes after CPU fan was disconnected (3.6GHz E6600 at 1.6volt):

[Ace-a-Rue]

as much as we would like to use this, it was designed for the mobil core...might be close but not exact...

[ted3]

Ace-a-rue, Merom and Conroe have the same cores, i cant see what makes a big difference. Merom has lower FSB, and what else is different?

Edit: In task manager, when running TAT stresstest, it shows up as MeromMaxPowerVer0p3, in the folder its installed there is another file for Core Duo, YonahMaxPowerVer1p0.exe.

[Artmic]

I think you can use this thing for anything, it will report the temps, besides it is close to Everest. so how bad could it be if it is from Intel, and somehow i doubt anyone would be stupid enough not to design this software to handle more than one type of CPU.

[taemun]

It won't run at all with my E6600, P5W-DH etc... the dialog comes up asking me to agree, I hit agree, and the program exits...

t

[Eastcoasthandle]

It won't run at all with my E6600, P5W-DH etc... the dialog comes up asking me to agree, I hit agree, and the program exits...

t

bad OC?

[taemun]

bad OC?

nar its the same at stock or oc'd

t

[eallan]

It won't run at all with my E6600, P5W-DH etc... the dialog comes up asking me to agree, I hit agree, and the program exits...

t

Weird, works fine on mine, same system.

(Hello everyone by the way.)

[Eastcoasthandle]

Weird, works fine on mine, same system.

(Hello everyone by the way.)

[eallan]

Thanks

You guys have quite the selection of smileys haha.

[taemun]

Hmm .... you guys are all running XP? I'm running Windows 2003... (there is some killer pricing on it for educational purposes )

Anyone out there with 2003 x86 (32-bit) and working / not working?

t

[AndrewZorn]

Set max workload to test your max temp, and for sure the temps will go way higher than Orthos with small FFT's, theres no other way i know of to stress a CPU this much.

And you can also set a throttling level to reduce the processors performance to a preset percentage, up to 87.5% throttling. Apps will run quite slow if you set both cores to 87.5% throttle =)

What makes TAT more useful than CoreTemp is the throttlewatch, i posted a screenie in another thread not long ago, this is what happened 2 minutes after CPU fan was disconnected (3.6GHz E6600 at 1.6volt):

hotter than orthos!? wow... this combined with the fact that the 'core' probe is deeper than a normal die sensor... no wonder AMD fans think even the 65nm conroe runs hotter than an AX2 (not that it DOESNT, im not sure, but it doesnt run as hot as they think)

anyway the throttling sounds interesting.
so i could say, cut it to 50% clockspeed if it gets to 70*C? do you have to let this program run in the background? is there a way to reduce voltage (reducing clockspeed only wont help much)?

[Anemone]

Welcome Eallan

[MrBean]

Yeah, it's fine on my E6600, P5W DH, and GSkill HZ Ram.

Chipped clocked to only 3.33gig (370fsb 1:1), watercooled, and when using TAT thermal testing tool, temps stabilize around 47 deg C core 1, and 45 for core 2.

TAT reads 1 deg C lower on both cores vs Coretemp.

Very nifty little app, thanx Intel

Orthos will only take temps up to around 38-40 on water, this is quite a bit more intensive. Good stability checker I would say.

[fhpchris]

Does not work @ server 2003

[ted3]

anyway the throttling sounds interesting.
so i could say, cut it to 50% clockspeed if it gets to 70*C? do you have to let this program run in the background? is there a way to reduce voltage (reducing clockspeed only wont help much)?

The on-demand throttling isnt something most people want to use i think I guess its just there for testing purposes, like the name of the app indicates its an "analysing tool". Temps will drop quite a bit when setting throttle, but as my little test without cpu fan showed, the cpu will throttle itself down when going above 85c as a safety meassure. As long as you can run TAT for 5-10 minutes at dual 100% worklevel without throttling you are safe, if it aint broken dont fix it! High temps are nothing to worry about, throttling is what we dont want. The reduction of voltage will still be the job of C1E, dont know if theres a way to do that on demand, probably not.

[Shintai]

Just on a sidenote, please ask before you paste a link to my home server again
But on the good side, you didn´t link my picture directly.

But else, its the best tool. And it will never be a public download, since its for OEMs internal testing only.

[Fred_Pohl]

no wonder AMD fans think even the 65nm conroe runs hotter than an AX2 (not that it DOESNT, im not sure, but it doesnt run as hot as they think)

AMD fans think Conroe runs hot because they have no clue how hot their 100W+ Athlons really run. It doesn't help that so many Conroe users mistakenly think that TAT and Core Temp readings are somehow comparable to other cpu temps, because they aren't.

[vengance_01]

AMD fans think Conroe runs hot because they have no clue how hot their 100W+ Athlons really run. It doesn't help that so many Conroe users mistakenly think that TAT and Core Temp readings are somehow comparable to other cpu temps, because they aren't.

thats very true. my Scythe Mine's heatpipes do not even get warm under full load with my 6300@2.8GHZ@1.32 volts(bios) I think this is reading at the very heart of the core thus giving very accurate temps at that spot. Its cooler running than my old Opteron 165@2.7GHZ as I can feel the back dumping off much less heat.

[Fred_Pohl]

thats very true. my Scythe Mine's heatpipes do not even get warm under full load with my 6300@2.8GHZ@1.32 volts(bios) I think this is reading at the very heart of the core thus giving very accurate temps at that spot. Its cooler running than my old Opteron 165@2.7GHZ as I can feel the back dumping off much less heat.

Finally! Thank you. I'm so tired of reading opinions about how the bios, speedfan, easytune, etc. are reading some nonexistent mobo socket thermistor and that Core Temp, iTAT and Everest are reporting "true" on-die temps. While the second part may be true, it's important to understand that Core Temp, etc. are reading sensors that don't exist on any other cpu and are therefore not comparable. As for the socket thermistor theory, that's just plain BS. Every Intel CPU since the P2 has had an on-die thermal diode which has been read by the mobo to get the cpu temp.

Nice to meet someone else who "gets it".